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
|
|---|---|---|---|---|---|
Systemic sclerosis ( SSc ) is an orphan , complex , inflammatory disease affecting the immune system and connective tissue . SSc stands out as a severely incapacitating and life-threatening inflammatory rheumatic disease , with a largely unknown pathogenesis . We have designed a two-stage genome-wide association study of SSc using case-control samples from France , Italy , Germany , and Northern Europe . The initial genome-wide scan was conducted in a French post quality-control sample of 564 cases and 1 , 776 controls , using almost 500 K SNPs . Two SNPs from the MHC region , together with the 6 loci outside MHC having at least one SNP with a P<10−5 were selected for follow-up analysis . These markers were genotyped in a post-QC replication sample of 1 , 682 SSc cases and 3 , 926 controls . The three top SNPs are in strong linkage disequilibrium and located on 6p21 , in the HLA-DQB1 gene: rs9275224 , P = 9 . 18×10−8 , OR = 0 . 69 , 95% CI [0 . 60–0 . 79]; rs6457617 , P = 1 . 14×10−7 and rs9275245 , P = 1 . 39×10−7 . Within the MHC region , the next most associated SNP ( rs3130573 , P = 1 . 86×10−5 , OR = 1 . 36 [1 . 18–1 . 56] ) is located in the PSORS1C1 gene . Outside the MHC region , our GWAS analysis revealed 7 top SNPs ( P<10−5 ) that spanned 6 independent genomic regions . Follow-up of the 17 top SNPs in an independent sample of 1 , 682 SSc and 3 , 926 controls showed associations at PSORS1C1 ( overall P = 5 . 70×10−10 , OR:1 . 25 ) , TNIP1 ( P = 4 . 68×10−9 , OR:1 . 31 ) , and RHOB loci ( P = 3 . 17×10−6 , OR:1 . 21 ) . Because of its biological relevance , and previous reports of genetic association at this locus with connective tissue disorders , we investigated TNIP1 expression . A markedly reduced expression of the TNIP1 gene and also its protein product were observed both in lesional skin tissue and in cultured dermal fibroblasts from SSc patients . Furthermore , TNIP1 showed in vitro inhibitory effects on inflammatory cytokine-induced collagen production . The genetic signal of association with TNIP1 variants , together with tissular and cellular investigations , suggests that this pathway has a critical role in regulating autoimmunity and SSc pathogenesis .
Systemic sclerosis ( MIM181750 ) is a connective tissue disease characterized by generalized microangiopathy , severe immunologic alterations and massive deposits of matrix components in the connective tissue . Being an orphan disease , SSc presents a major medical challenge and is recognized as the most severe connective tissue disorder with high risk of premature deaths [1] . Epidemiological data on SSc vary in different parts of the world and depend on selection criteria for the study population . Inasmuch , the prevalence of the disease fluctuates across global regions and population-based studies result in higher prevalence than do hospital records-based studies . In North America , the prevalence of SSc has been reported as 0 . 7–2 . 8 per 10 , 000 in a Canadian study , whereas in the U . S . figures of 2 . 6 per 10 , 000 versus 7 . 5 per 10 , 000 were reported by medical records - versus population-based studies , respectively . In Europe , a prevalence of 1 . 6 per 10 , 000 was reported in Denmark , 3 . 5 per 10 , 000 in Estonia , 1 . 58 per 10 , 000 adults ( 95% confidence interval , 129–187 ) in Seine-Saint-Denis in France [2]–[4] . The risk of SSc is increased among first-degree relatives of patients , compared to the general population . In a study of 703 families in the US , including 11 multiplex SSc families , the familial relative risk in first-degree relatives was about 13 , with a 1 . 6% recurrence rate , compared to 0 . 026% in the general population [5] . The sibling risk ratio was about 15 ( ranging from 10 to 27 across cohorts ) . The only twin study reported to date included 42 twin pairs [6] . The data showed a similar concordance rate in monozygotic twins ( 4 . 2% , n = 24 ) and dizygotic twins ( 5 . 6% ) ( NS ) and an overall cross-sectional concordance rate of 4 . 7% . However , concordance for the presence of antinuclear antibodies was significantly higher in the monozygotic twins ( 90% ) than in the dizygotic twins ( 40% ) suggesting that genetics may be important for the auto-immune part of the disease . The aetiology of SSc is still unclear but some key steps have been described , in particular early endothelial damage and dysregulation of the immune system with abnormal autoantibody production [7] . At the cellular level , early events include endothelial injury and perivascular inflammation with the release of a large array of inflammatory mediators [8] , [9] . In the advanced stage , a progressive activation of fibroblasts in the skin and in internal organs leads to hyperproduction of collagen and irreversible tissue fibrosis [9] . Epidemiological investigations indicate that SSc follows a pattern of multifactorial inheritance [10] . Previous candidate-gene association studies have only identified a handful of SSc risk loci , most contributing to the genetic susceptibility of other autoimmune diseases [9]–[16] . So far , two genome-wide association studies of SSc have been conducted [17] , [18] . The studies differ according to the ancestry of the studied population ( Korean vs US/European ) and the genome-wide association data: map density ( ∼440 K vs 280 K SNPs ) and sample size ( ∼700 vs ∼7300 subjects ) . They provided evidence of association with known MHC loci , but only one ‘new’ locus was identified at CD247 in the US/European dataset , variants at CD247 being known to contribute to the susceptibility of systemic lupus erythematosus [18] . The diagnosis of SSc is based on recognized clinical criteria established decades ago however , these do not include specific autoantibodies or recent tools for assessment of the disease [19] , [20] . Therefore , phenotypic heterogeneity is a concern for SSc and genetic heterogeneity is also highly probable with regards to data obtained in other connective tissue disorders . Given these considerations , and previous findings in other autoimmune diseases , it is apparent that additional risk variants for SSc remain to be discovered . Therefore , to identify further common variants that contribute to SSc risk in the European population , we conducted a two-stage GWAS , in two case-control samples ( total >8 , 800 subjects ) .
We established a collaborative consortium including groups from 4 European countries ( France , Italy , Germany and Eastern-Europe ) from which we were able to draw upon a combined sample of over 8 , 800 subjects ( before quality control ) and conducted a two-stage genome-wide association study . In stage 1 , we genotyped 1 , 185 samples on Illumina Human610-Quad BeadChip and genotypes obtained using the same chip from 2 , 003 control subjects were made available to us from the 3C study [21] , [22] . After stringent quality control , we finally tested for association in stage-1 , 489 , 814 autosomal SNPs in 2 , 340 subjects ( 564 cases and 1 , 776 controls ) ( Table 1 ) . We tested for association between each SNP and SSc using the logistic regression association test , assuming additive genetic effects . The quantile-quantile plot and estimation of the genomic inflation factor ( λ = 1 . 035 ) indicated minimal overall inflation ( Figure 1A ) . The genome-wide logistic association results are presented in Figure 1B . Table S1 provides details for all SNPs with P<10−4 , including one SNP exceeding P<10−7 , the Bonferroni threshold for genome-wide significance . The three top SNPs were located on 6p21 , in the HLA-DQB1 gene: rs9275224 , P = 9 . 18×10−8 , OR = 0 . 69 , 95%CI[0 . 60–0 . 79]; rs6457617 , P = 1 . 14×10−7 and rs9275245 , P = 1 . 39×10−7 ( Figure 1B and Table 2 ) . Several associated SNPs in HLA-DQB1 have already been reported but rs6457617 was also identified as the most associated SNP in the previous US/European GWAS study [18] . Of note , the three SNPs in HLA-DQB1 are in strong LD ( r2>0 . 97 ) . Within the MHC region , the next most associated SNP ( rs3130573 , P = 1 . 86×10−5 , OR = 1 . 36[1 . 18–1 . 56] ) is located in the psoriasis susceptibility 1 candidate 1 ( PSORS1C1 ) gene ( Table 2 ) , a candidate gene for psoriasis [23] . Conditional analyses of susceptibility variants within MHC showed that there were two independent association signals at rs6457617 ( HLA-DQB1 ) and at rs3130573 ( PSORS1C1 ) . Indeed , the association at PSORS1C1 remained significant ( P<2 . 1×10−5 ) after controlling for the association at HLA-DQB1 and the association at HLA-DQB1 remained also significant ( P<1 . 5×10−7 ) after controlling for the association at PSORS1C1 ( Table S2 ) . Outside the MHC region , our GWAS analysis revealed 7 top SNPs ( P<10−5 ) that spanned 6 independent genomic regions ( Figure 1B and Table S1 ) . Conditional analyses of each of them on HLA-DQB1 showed no significant drop in the association signals ( Table S3 ) . The 6 loci having at least one SNP with a P<10−5 were selected for follow-up analysis . Within each locus we selected the SNPs with the strongest ( P<10−4 ) association signals to be genotyped in a post-QC replication sample of 1 , 682 SSc cases and 3 , 926 controls ( Table 1 ) . To this list we added two top SNPs in HLA-DQB1 and the SNP in PSORS1C1 . Finally , we further included 4 SNPs at the two known loci ( STAT4 and TNPO3-IRF5 ) and at the newly identified locus ( CD247 ) by Radstake et al [18] . Out of a total set of 21 SNPs submitted for replication , 20 passed the quality-control analyses . Stratified association analyses in stage 2 data ( Table 2 ) , confirmed the strong association for HLA-DQB1 ( rs6457617 , P = 1 . 35×10−28 ) at 6p21 . 3 and also with the PSORS1C1 variant ( rs3130573 , P = 4 . 98×10−3 ) at 6p21 . 1 . Of the 6 remaining loci selected in stage 1 , only 2 were replicated with nominal P<5% and with same direction of effect . They mapped at 2p24 ( rs342070 , P = 0 . 026; rs13021401 , P = 0 . 024 ) and 5q33 ( rs3792783 , P = 4 . 14×10−5; rs2233287 , P = 4 . 38×10−3; rs4958881 , P = 2 . 09×10−3 ) . None of the replicated SNPs showed evidence for heterogeneity of effects among the 4 geographical origins ( Breslow-day P>0 . 10 ) . As expected , ORs estimated in the discovery tended to be higher than those obtained in the replication stage data . Afterwards , association signals from joint analyses of the 2 datasets ( Table 2 ) consistently showed highly significant association for HLA-DQB1 ( P = 2 . 33×10−37 ) , PSORS1C1 ( P = 5 . 70×10−10 ) and TNIP1 ( P = 4 . 68×10−9 ) , and also showed some evidence of association for RHOB ( P = 3 . 17×10−6 ) . All populations showed same direction of effects ( Figure 2 ) . Finally , we also replicated association signals at IRF5 ( P = 3 . 49×10−5; combined-P = 4 . 13×10−7 ) , at STAT4 ( P = 1 . 9×10−10; combined-P = 2 . 26×10−13 ) and at the recently identified new SSc risk locus , CD247 ( P = 2 . 90×10−5; combined-P = 1 . 30×10−6 ) ( Table 2 ) . In our combined data , the locus-specific PAR estimates were 24% for HLA-DQB1 , 4% for TNIP1 , 8% for PSORS1C1 , 7% for CD247 , 8% for STAT4 and 3% with IFR5/TNPO3 . The combined PAR estimate was 47 . 4% . As secondary analyses , we assessed homogeneity of SNP's effect between sub-categories of SSc ( cutaneous sub-types and auto-antibodies ) . Case-only analyses revealed no significant evidence for heterogeneous ORs between cutaneous sub-types of SSc patients for any of the 5 replicated SNPs at 2p24 or 5q33 loci ( Table S4A ) . Indeed , similar association signals were obtained from case-category association analyses ( Table S4B ) . Altogether , the results did not suggest that the association signals in the newly identified 5q33 locus were driven by a specific sub-type of SSc . Conversely , for HLA-DQB1 and PSORS1C1 we found evidence of heterogeneity in OR estimates in positive vs negative ACA or TOPO auto-antibody SSc patients ( Table S4A ) . Yet , the association signals in each of these sub-types of patients remained strong ( Table S4B ) . These results support the previously reported hypothesis that the magnitude of the HLA-DQB1 effect on SSc susceptibility may depend on auto-antibody status [11] . The GWAS stage had 78% power to detect loci of the effect sizes observed in the discovery sample for TNIP1 variants ( OR = 1 . 50 ) at a significance of P<10−5 . However , it is widely acknowledged that effect sizes of significant GWAS loci are overestimates of true effects and other genes of lower effect sizes are unlikely to reach stringent significant thresholds . Our GWAS analysis revealed strong association with PSORS1C1 , which is ∼1 Mb of HLA-DQB1 . Notably , PSORS1C1 is known to be involved in autoimmune response [23] . In the combined data , association with PSORS1C1 was highly significant ( P = 5 . 70×10−10 ) and remained significant after controlling for the association at HLA-DQB1 . Altogether , our results suggest that this region is likely to contain more than one gene playing a role in the pathogenesis of autoimmune disorders [23] , [24] . Fine mapping at this locus is warranted to identify causal variants . The three strongly associated SNPs at the 5q33 locus are located within the TNFAIP3 interacting protein 1 ( TNIP1 ) gene . TNIP1 is a very interesting new candidate gene for SSc . The protein encoded by this gene exerts a negative regulation of NF-kappaB via two sequential activities: deubiquitination of Lys63-based chains and synthesis of Lys48-based chains on the TNF receptor-interacting protein and also inhibition of NF-KappaB processing [25] . TNIP1 interacts with A20 ( TNFAIP3 ) to negatively regulate NF-kappaB . Several recent studies have suggested that the activation of some inflammatory factors may upregulate fibrotic mediators through Toll-like receptors ( TLRs ) , thereby contributing to SSc pathogenesis [8] . It has been shown that TLR engagement leads to A20 induction in macrophages and that TNIP1/A20 is essential for the termination of TLR-induced NF-kappaB activity and proinflammatory cytokine production [26] . Although interactions between TNIP1 and A20 are not well known , A20 also acts as a deubiquitinating enzyme , suggesting a molecular link between deubiquitinating activity and the regulation of TLR signals [26] . Therefore , TNIP1 and A20 may play a critical role in the regulation of downstream TLR signals , and this issue will have to be addressed in SSc . Interestingly , variants at TNIP1 have been shown to be implicated in systemic lupus erythematosus susceptibility [27] , [28] and in psoriasis [29] . Furthermore , we have recently reported an association of one TNFAIP3 variant with SSc [30] . In our stage-1 data , evidence of association at TNFAIP3 was nominal ( lowest P = 0 . 047 ) and no pairwise interaction was found ( P>0 . 06 ) between TNFAIP3 and TNIP1 variants . Analysis of the LD structure across the TNIP1 gene revealed that the 3 strongly associated SNPs belong to the same LD-block ( Figure 3 ) . No residual association signals were observed when rs3792783 and each of the other 2 SNPs were paired in conditional analyses . Therefore , any of them , or other variants yet to be identified , could be the causal variant ( s ) . Interestingly , rs3792783 is located upstream from the transcription start site in exon 2 ( Figure 3 ) . It is noteworthy that previously reported lupus TNIP1 variants were located in the same LD-block [27] , [28] . Because of the compelling evidence of the potential role of NF-kappaB in autoimmune diseases and our raised new signal association for SSc at TNIP1 ( a negative regulator of this pathway ) we performed ex vivo and in vitro investigations to assess TNIP1 expression in SSc patients and healthy controls . For SSc patients , the results showed a strikingly reduced expression of TNIP1 in skin tissue ( Figure 4A ) , and of both mRNA ( Figure 4B ) and protein ( Figure 4C ) synthesis by cultured dermal fibroblasts . Addressing the question of the potential link between the NF-kappaB pathway and the fibrotic propensity that characterizes SSc , we next assessed the influence of pro-inflammatory cytokines and TNIP1 on the synthesis of extra-cellular matrix by dermal fibroblasts in culture . Using cells from the skin of healthy controls ( Figure 5 ) and SSc patients ( Figure 6 ) , we showed that recombinant TNIP1 abrogated collagen synthesis induced by inflammatory cytokines both at the mRNA and protein levels . It must be acknowledged that TNIP1 is described as an intra-cellular protein whereas we used recombinant protein added to cell supernatant in these experiments . The observed effects may be related to different hypotheses . TNIP1 has been described as a nuclear shuttling protein and it could have a chaperon-like activity , highly interacting with other protein that could result in engulfment of TNIP1 through interaction with a cell surface protein . Such intra-cellular effects of extra-cellular proteins has been shown also for the S100 family of proteins that have no leader sequence and for clusterin for which it is postulated that the protein could be taken up by interacting with either a yet unidentified receptor or by a mechanism related to their chaperon-like activity [31] . More work is needed to determine which of these hypotheses has to be retained and to investigate more in depth soluble TNIP1 In this first attempt to explore TNIP1 functional disturbances , we could not investigate a relationship between specific TNIP variants and in vitro or in vivo changes; this will need to be addressed ideally after the identification of the causal variant and using a much larger sample size . Nevertheless , our results raise a potential relationship between inflammation and fibrosis and open a new and highly relevant field of investigation in SSc pathogenesis and in fibrotic disorders . Our next most associated SNPs at 2p24 are in strong LD ( r2 = 0 . 98 ) and map ∼30 kb from RHOB . RHOB is the Ras homolog gene family member B that regulates protein signalling and intracellular protein trafficking . RhoB is essential for activity of farnesyltransferase inhibitors and also statins that are two strong potential future drugs in SSc [9] , [32] . To our knowledge , association to RHOB has never been reported so far . The signal for association was weaker at this locus and therefore will need to be confirmed in other samples and more investigations are warranted to assess RHOB implication in this disease . In conclusion , we have conducted a large genome-wide association study of SSc and identified two new SSc-risk loci , PSORS1C1 and TNIP1 . We also confirmed the association of SSc with variants at STAT4 , IRF5 and CD247 , in the European population . We also found compelling evidence of association to a putative new SSc risk locus on 2p24 , close to the RHOB gene . None of the newly identified 3 loci have been previously reported associated to SSc . The TNIP1 variants identified do not have precise functional implications; however , their localization within a regulatory region strongly suggests an impact on transcription of the gene . This is supported by our ex vivo and in vitro investigations . Altogether , our results are consistent with a reduced inhibition of NF-kappaB , therefore favoring inflammatory/immune responses and potentially contributing to the overproduction of extra-cellular matrix . This raises a new clue for a link between inflammation and SSc that could also be of importance in other fibrotic disorders .
Stage-1 included 654 SSc patients and 531 controls recruited through the French GENESYS project [11] , [13] , [30] and 2 , 003 controls from the French Three-City ( 3C ) cohort [21] , [22] . The stage-2 data included an independent collection of 4 , 492 samples ( pre quality controls ) from several University Hospitals in France , Italy , Germany and Eastern Europe . It also included 721 Italian controls recruited through nationwide efforts by HYPERGENE consortium and 481 Illumina HumanHap550 for the KORA S4 study [33] , recruited in the city of Augsburg , Southern Germany . In both stage 1 and stage 2 samples , SSc patients fulfilled ACR criteria [34] and were classified in cutaneous subsets according to LeRoy's criteria [19] . Table 1 shows the main characteristics of the post-QC SSc patients and controls . All participants gave written informed consent , and approval was obtained from the relevant local ethical committees . Association analysis of the genotype data was conducted with PLINK ( v1 . 07 ) software [35] . All reported P values are two sided . In stage 1 , we applied logistic regression assuming an additive genetic model . The quantile-quantile plot was used to evaluate overall significance of the genome-wide association results and the potential impact of residual population substructure . A conservative genome-wide significance threshold of 0 . 05/489 , 918 = 1 . 02×10−7 was used . Stage 2 association and combined analyses were carried out with the Mantel-Haenszel test to control for differences between geographical groups . A Breslow-Day test was performed to assess the heterogeneity of effects in different populations . In the replication analysis , P values<0 . 05 and direction of effect as observed in the stage-1 data , were considered to indicate statistical significance . Secondary statistical analyses were conducted to assess independency of multiple association signals within and between loci and homogeneity of effects between subgroups of SSc patients . Case-only association analyses were conducted using the three main clinical variables ( Table 1 ) . The LD structure of the identified loci was analyzed using Haploview 4 . 1 [37] and LD blocks delimited using the D′-based confidence interval method [38] . The locus-specific Population attributable risk ( PAR ) was calculated for each of the 6 replicated loci ( HLA-DQB1 , TNIP1 , PSORS1C1 , STAT4 , IFR5/TNPO3 and CD247 ) according to the following formula: PAR = RAF× ( OR-1 ) / ( RAF× ( OR-1 ) +1 ) , where RAF is the frequency of the associated allele in the controls , and OR is the odds ratio associated with the risk allele . The combined PAR was computed as 1−Pj ( 1−PARj ) . Fibroblast cultures were prepared by outgrowth cultures from lesional skin biopsy specimens of eleven SSc patients and from twelve healthy controls matched for age and sex . The median age of SSc patients was 49 years old ( range: 22–67 years ) and their median disease duration was 7 years ( range: 1–17 years ) ; seven had the limited cutaneous subset and four the diffuse . Immunohistochemistry was performed on paraffin-embedded skin sections from 5 SSc patients and 5 controls using mouse anti-human TNIP1 antibodies ( eBioscience , Frankfurt , Germany ) . Total RNA , issued from cultured dermal fibroblasts , isolation and reverse transcription into complementary DNA were performed as previously described [39] . Gene expression was quantified by SYBR Green real-time PCR , with a specific primer pair available upon request . Protein assessment was performed on western blots , as previously described [40] using mouse anti-human TNIP1 antibodies ( eBioscience , CA , USA ) . In selected experiments , dermal fibroblasts from healthy control subjects and patients with SSc were treated for 24 hours with recombinant TNIP1 ( 2 µg/ml , Abnova , Tapei City , Taiwan ) in the presence or not of the following proinflammatory cytokines: TNFα ( 20 ng/ml , R&D systems , Abingdon , UK ) , IL1β ( 1 µg/ml , Immunotools , Friesoythe , Germany ) or IL6 ( 1 µg/ml , Immunotools ) . mRNA levels of human α1 ( I ) and α2 ( I ) procollagen were quantified by quantitative real-time PCR , specific primers are available upon request . The collagen content in cell culture supernatants was analyzed with the SirCol collagen assay ( Biocolor , Belfast , UK ) [41] . Comparisons were performed using Student's T test .
|
Systemic sclerosis ( SSc ) is a connective tissue disease characterized by generalized microangiopathy , severe immunologic alterations , and massive deposits of matrix components in the connective tissue . Epidemiological investigations indicate that SSc follows a pattern of multifactorial inheritance; however , only a few loci have been replicated in multiple studies . We undertook a two-stage genome-wide association study of SSc involving over 8 , 800 individuals of European ancestry . Combined analyses showed independent association at the known HLA-DQB1 region and revealed associations at PSORS1C1 , TNIP1 , and RHOB loci , in agreement with a strong immune genetic component . Because of its biological relevance , and previous reports of genetic association at this locus with other connective tissue disorders , we investigated TNIP1 expression . We observed a markedly reduced expression of the gene and its protein product in SSc , as well as its potential implication in control of extra-cellular matrix synthesis , providing a new clue for a link between inflammation/immunity and fibrosis .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Material",
"and",
"Methods"
] |
[
"medicine",
"rheumatology",
"public",
"health",
"and",
"epidemiology",
"cytokines",
"immunology",
"rheumatoid",
"arthritis",
"genetic",
"epidemiology",
"systemic",
"lupus",
"erythematosus",
"inflammatory",
"and",
"psoriatic",
"arthritis",
"disease",
"mapping",
"scleroderma",
"connective",
"tissue",
"diseases",
"major",
"histocompatibility",
"complex",
"epidemiology",
"immune",
"system",
"psoriasis",
"genetics",
"of",
"the",
"immune",
"system",
"clinical",
"immunology",
"autoimmune",
"diseases"
] |
2011
|
Genome-Wide Scan Identifies TNIP1, PSORS1C1, and RHOB as Novel Risk Loci for Systemic Sclerosis
|
Deep brain stimulation ( DBS ) of the subthalamic nucleus ( STN ) is modeled to explore the mechanisms of this effective , but poorly understood , treatment for motor symptoms of drug-refractory Parkinson’s disease and dystonia . First , a neural field model of the corticothalamic-basal ganglia ( CTBG ) system is developed that reproduces key clinical features of Parkinson’s disease , including its characteristic 4–8 Hz and 13–30 Hz electrophysiological signatures . Deep brain stimulation of the STN is then modeled and shown to suppress the pathological 13–30 Hz ( beta ) activity for physiologically realistic and optimized stimulus parameters . This supports the idea that suppression of abnormally coherent activity in the CTBG system is a major factor in DBS therapy for Parkinson’s disease , by permitting normal dynamics to resume . At high stimulus intensities , nonlinear effects in the target population mediate wave-wave interactions between resonant beta activity and the stimulus pulse train , leading to complex spectral structure that shows remarkable similarity to that seen in steady-state evoked potential experiments .
Deep brain stimulation ( DBS ) has become an effective treatment for a number of neurological disorders such as Parkinson’s disease ( PD ) and essential tremor [1 , 2] . In Parkinson’s disease DBS treatments , a macroelectrode is chronically implanted in a target nucleus , typically either the globus pallidus internus ( GPi ) , subthalamic nucleus ( STN ) , or the ventral intermediate nucleus of the thalamus; this electrode delivers high frequency ( >100 Hz ) electrical stimulation as a series of pulses . More broadly , studies have also shown the efficacy of DBS treatments in dystonia [3] , epilepsy [4] , and obsessive-compulsive disorder [5] . Significant progress has been made exploring the influence of DBS on neural activity [6] . However , the efficacy of DBS treatments could be improved with a greater understanding of the underlying therapeutic mechanisms . Furthermore , it is unclear what stimulation parameters , electrode geometries , and electrode locations are most effective for the present and future uses of DBS technologies . Electrical stimulation of the brain influences a variety of mechanisms involved in the function and signaling of neurons . The sensitivity of different contributing elements depends upon the amplitude and temporal properties of the stimulation [7] , geometry of the stimulus field [8] , target cell physiology and geometry [9] , and the possible pathophysiology of different disease states [10] . It is known that distinct neuron types possess different types of ion channels and that these may have different voltage-sensitive activation and inactivation properties [11] . Thus , the effect of DBS on a single neuron’s dynamics may vary significantly between brain regions . However , by averaging over millimeter to whole-brain scales , a generalized description of DBS at a population level may provide insights into the net effect of these different mechanisms and allow prediction of effective stimulation protocols . It was initially thought that deep brain stimulation had a predominantly inhibitory effect on the stimulated population due to similar therapeutic effects to lesioning [12 , 13] . This inhibition hypothesis is supported by several experimental findings in STN-DBS of rats [14] and monkeys [15] and GPi-DBS and STN-DBS in humans [16 , 17 , 18] . Furthermore , [19] demonstrated an inhibitory response to GPi-DBS that was mediated by the GABA receptors and that this inhibition could be blocked via a GABA antagonist . Seemingly in contradiction with the inhibition hypothesis , several experiments with recordings in efferent nuclei of the DBS target population indicate an increase in the stimulated populations activity [20 , 21 , 22] , and an entrainment of local neural firing during DBS [23] . Several modeling approaches have been used to elucidate DBS effects across multiple scales . Finite element methods used in conjunction with multi-compartment neuron models have explored DBS responses in small neural assemblies [24 , 25 , 8] , and the distribution of the applied electric field [26] . These single cell models have demonstrated a disassociation of activity at the soma relative to the axon during extracellular stimulation [24] , and a systemic activation of axons both efferent and afferent to the stimulation site [27] . The variable nature of response to DBS between brain regions might then be understood by approximating DBS as an activation of intrinsic axons within a certain effective range of the electrode , and thus could explain observations supporting the contradictory excitation and inhibition hypotheses . Parkinson’s disease ( PD ) is a neurodegenerative disorder characterized by motor dysfunction including akinesia , bradykinesia , tremor , and rigidity [28] . These clinical manifestations have been linked to dopaminergic denervation in the substantia nigra pars compacta ( SNc ) and synucleinopathy leading to Lewy bodies , and neurites in the SNc and other brain regions [29] . The firing pattern hypothesis regarding PD proposes that pathological oscillations and/or synchronization play a primary role in motor symptoms of the disorder . Single unit and local field potential recordings have shown enhanced activity within and between the basal ganglia ( BG ) , thalamus and motor cortex at about 4–8 Hz and 13–30 Hz [30 , 31 , 32 , 33 , 34] , which seems to correlate with significant coherence at these frequencies [35 , 36 , 37 , 38 , 39 , 40] . It has thus been suggested that these pathological rhythms cause a disturbance of motor-related information processing in the BG [41] , which could explain some PD symptoms . The enhanced beta band oscillations ( 13-30 Hz ) found in the STN of PD patients are thought to be related to symptom severity , based on direct correlation results [42] , as well as observations of a reduction in beta power following treatments that ameliorate PD symptoms , such as dopaminergic supplementation [43] and deep brain stimulation [44] . Several experimental and modelling studies have suggested that the circuit formed between the STN and GPe may be responsible for beta activity generation [45 , 46] , and that cortical excitatory inputs to the STN amplify them [47] . However , the origin of parkinsonian beta activity is still a matter of debate . Physiologically based mean-field models of the brain provide a tractable framework for the analysis of large-scale neuronal dynamics by averaging microscopic structure and activity [48 , 49 , 50 , 51 , 52] . Neural field theory incorporates realistic anatomy of neural populations , nonlinear neural response , interpopulation connections and dendritic , synaptic , cell-body , and axonal dynamics [48 , 50 , 52 , 53 , 54 , 55 , 56 , 57 , 58] . Neural field models have been successful in accounting for many characteristic states of brain activity including sleep stages , eyes-open , and eyes-closed in waking , nonlinear seizure dynamics , anesthesia and many other phenomena [52 , 53 , 55 , 51 , 59 , 60 , 61 , 62] . In the particular case of Parkinson’s disease , neural field models of the corticothalamic-basal ganglia system have been able to account for several electrophysiological correlates of the disease including changes in population average activity , and ∼4-8 Hz and ∼13–30 Hz oscillations characteristic of EEG and LFP spectra [63 , 64 , 65] . However , the generative mechanisms of these characteristic PD rhythms is still a matter of debate . The core aims of this work are to develop a population level description of DBS of the corticothalamic-basal ganglia ( CTBG ) system that can account for experimental observations and the results of other modeling studies . The work will explore parkinsonian states of the CTBG system and determine whether subthalamo-pallidal circuits can sustain characteristic beta oscillations in this framework . Finally , the effects of DBS on these parkinsonian states will be analyzed and provide insights into the efficacy of DBS treatments .
Fig 1 shows a schematic of the CTBG model . The system contains nine distinct neural populations across three brain regions . The cerebral cortex contains populations of excitatory pyramidal neurons , e , and inhibitory interneurons , i . The thalamus is divided into an excitatory population for the specific relay nuclei ( SRN ) , s , and an inhibitory population for the thalamic reticular nucleus ( TRN ) , r . The basal ganglia ( BG ) contains two inhibitory populations within the striatum , one expressing the D1 dopamine receptor , d1 , and one expressing the D2 dopamine receptor , d2 . The striatum projects to two inhibitory populations , the globus pallidus pars externa , p2 , and a population representing the globus pallidus pars interna and substantia nigra pars reticulata , p1 . The subthalamic nucleus ( STN ) is represented by an excitatory population , ζ . Finally , deep brain stimulation is defined as an input source , x , which is coupled to STN as well as to its projection sites . This is discussed in detail in a later section . The substantia nigra pars compacta ( SNc ) and ventral tegmental area ( VTA ) are not explicitly defined as a population within the model , however , they are included in Fig 1 as an indication of the neural pathways affected by dopamine . The mean firing rate , Qa ( r , t ) , of a neural population can be approximately related to its mean membrane potential , Va ( r , t ) , by [66 , 67] Q a ( r , t ) = S a [ V a ( r , t ) ] , ( 1 ) = Q a max 1 + exp [ - { V a ( r , t ) - θ a } / σ ′ ] . ( 2 ) where Eqs ( 1 ) & ( 2 ) define the sigmoidal mapping function Sa , Q a max is the maximal firing rate , Va is the average membrane potential relative to resting , θa is the mean neural firing threshold , and σ ′ π / 3 is the standard deviation of this threshold . A number of experimental studies have revealed waves of neural activity spreading across the cortex [68 , 49 , 69 , 70] , which have been analyzed theoretically [71 , 72 , 73 , 51 , 74 , 75 , 52 , 58] . This propagating activity is represented as a field of mean spike rates in axons , ϕa . A population a , with a mean firing rate Qa , is related to ϕa by the damped wave equation D a ( r , t ) ϕ a ( r , t ) = Q a ( r , t ) , ( 3 ) where D a ( r , t ) = 1 γ a 2 ∂ 2 ∂ t 2 + 2 γ a ∂ ∂ t + 1 - r a 2 ∇ 2 . ( 4 ) Here , γa = va/ra represents the damping rate , where va is the propagation velocity in axons , and ra is the characteristic axonal length for the population . The propagation of these waves is facilitated primarily by the relatively long-range white matter axons of excitatory cortical pyramidal neurons . Later in our model the simplifying local interaction approximation rb ≈ 0 is made for b = i , r , s , d1 , d2 , p1 , p2 , ζ due to the short ranges of the corresponding axons which implies ϕb ( r , t ) = Qb ( r , t ) for these populations [52 , 57 , 54 , 55 , 76 , 77] . The mean soma potential Va of a population a at position r and time t is given by sum of the postsynaptic potentials ( PSPs ) : V a ( r , t ) = ∑ b V a b ( r , t ) , ( 5 ) where Vab ( r , t ) is the postsynaptic potential generated by projections arriving at population a from population b . The influence of incoming spikes to population a from population b is weighted by a connection strength parameter , νab = Nabsab , where Nab is the mean number of connections between the two populations and sab is the mean strength of response in neuron a to a single spike from neuron b . The postsynaptic potential response in the dendrite is given by D α β V a b ( r , t ) = ν a b ( r , t ) ϕ a b ( r , t - τ a b ) , ( 6 ) where τab is the average axonal delay for the transmission of signals to population a from population b . The operator Dαβ describes the time evolution of Vab in response to synaptic input , and is given by D α β = 1 α β d 2 d t 2 + ( 1 α + 1 β ) d d t + 1 . ( 7 ) where β and α are the overall rise and decay response rates to the synaptodendritic and soma dynamics . It has been shown that nominal brain activity is well characterized by perturbations about a mean value [55] . Hence , we first find the time independent states of the CTBG system . Following the approach of previous neural field models , excitatory and inhibitory synapses in the cortex are assumed proportional to the numbers of neurons [50 , 78] . This random connectivity approximation results in νee = νie , νei = νii , and νes = νis , which implies Ve = Vi and Qe = Qi . Inhibitory population variables can then be expressed in terms of excitatory quantities and are thus not neglected even though they do not appear explicitly below . The steady states are obtained by setting all time derivatives to zero in Eqs ( 3 ) , ( 4 ) and ( 6 ) . Using the connectivity shown in Fig 1 , and excluding DBS , Eqs ( 5 ) and ( 6 ) give V e ( 0 ) = ( ν e e + ν e i ) ϕ e ( 0 ) + ν e s ϕ s ( 0 ) , ( 8 ) V r ( 0 ) = ν r e ϕ e ( 0 ) + ν r s ϕ s ( 0 ) , ( 9 ) V s ( 0 ) = ν s e ϕ e ( 0 ) + ν s r ϕ r ( 0 ) + ν s p 1 ϕ p 1 ( 0 ) + ν s n ϕ n ( 0 ) , ( 10 ) V d 1 ( 0 ) = ν d 1 e ϕ e ( 0 ) + ν d 1 s ϕ s ( 0 ) + ν d 1 d 1 ϕ d 1 ( 0 ) , ( 11 ) V d 2 ( 0 ) = ν d 2 e ϕ e ( 0 ) + ν d 2 s ϕ s ( 0 ) + ν d 2 d 2 ϕ d 2 ( 0 ) , ( 12 ) V p 1 ( 0 ) = ν p 1 ζ ϕ ζ ( 0 ) + ν p 1 d 1 ϕ d 1 ( 0 ) + ν p 1 p 2 ϕ p 2 ( 0 ) , ( 13 ) V p 2 ( 0 ) = ν p 2 ζ ϕ ζ ( 0 ) + ν p 2 d 2 ϕ d 2 ( 0 ) + ν p 2 p 2 ϕ p 2 ( 0 ) , ( 14 ) V ζ ( 0 ) = ν ζ e ϕ e ( 0 ) + ν ζ p 2 ϕ p 2 ( 0 ) . ( 15 ) The system’s steady states then can be determined by considering the simultaneous zeros of the five functions F ( ϕ e ) = ϕ e - S e [ ( ν e e + ν e i ) ϕ e + ν e s ϕ s ] , ( 16 ) F ( ϕ s ) = ϕ s - S s [ ν s e ϕ e + ν s r ϕ r + ν s p 1 ϕ p 1 + ν s n ϕ n ] , ( 17 ) F ( ϕ d 1 ) = ϕ d 1 - S d 1 [ ν d 1 e ϕ e + ν d 1 s ϕ s + ν d 1 d 1 ϕ d 1 ] , ( 18 ) F ( ϕ d 2 ) = ϕ d 2 - S d 2 [ ν d 2 e ϕ e + ν d 2 s ϕ s + ν d 2 d 2 ϕ d 2 ] , ( 19 ) F ( ϕ p 2 ) = ϕ p 2 - S p 2 [ ν p 2 d 2 ϕ d 2 + ν p 2 p 2 ϕ p 2 + ν p 2 ζ ϕ ζ ] . ( 20 ) where ϕr , ϕp1 , and ϕζ can be determined from Eqs ( 9 ) , ( 13 ) and ( 15 ) , respectively , in conjunction with Eq ( 1 ) . The roots of Eqs ( 16 ) – ( 20 ) are computed numerically using the MATLAB function fsolve ( ) with a tolerance of 10−15 V . A linearized form of the CTBG model can be used to derive the transfer function of the system [63 , 64 , 65] . This is a function of the internal gains of the system , which represent the additional activity generated in postsynaptic nuclei per additional unit input activity from presynaptic nuclei , and are [53 , 55] G a b = ρ a ν a b ( 21 ) where ρ a = d Q a d V a | V a ( 0 ) = ϕ a ( 0 ) σ ′ [ 1 − ϕ a ( 0 ) Q a max ] . ( 22 ) All numerical simulations of the CTBG neural field model in this work are performed using the NFTsim code package detailed by [79] . This package is used to solve Eqs ( 1 ) – ( 7 ) numerically for the spatially uniform case where the ∇2 in ( 4 ) is zero . The solutions to these delay differential equations are found using a standard forth-order Runge-Kutta integration method with a time step of 10−4 s . Nominal brain states have been found to exist near stable fixed points [55] . Hence , all simulations in this work are performed with the system initialized to the low firing steady state found in the previous section using the parameters given in Table 1 , unless otherwise specified . Many different stimulus protocols have been used in clinical DBS—with different pulse geometries ( i . e . sinusiodal or square-wave ) , signal amplitudes , stimulation frequencies , and/or transient stimulation phases , followed by varied quiescent periods . In this work we seek a general formulation of a neural populations response to fluctuations in an applied electric field that will allow for the effects of various stimulus protocols to be determined . The minimum current necessary to stimulate a given neural element with a long stimulus duration is called the rheobase [80] . The minimum length of time required to activate a given neural element using a stimulus amplitude twice as large as the rheobase is called the chronaxie . Extracellular stimulation experiments have demonstrated a chronaxie time for the myelinated axons which is substantially smaller than the chronaxies of the cell body and dendrites [7 , 81 , 82] . Hence , our key assumption is that the net effect of fluctuations of an applied electric field is a stimulation of voltage-gated ion channels that induces transmembrane current flow predominantly in both afferent and efferent axons of a subset of neurons within the stimulated population . A mean-field model has recently been used to describe population effects of transcranial magnetic stimulation [83 , 84] . A modification of this approach is used by defining an external pulse rate ϕx ( t ) that consists of a train of pulses with a width twidth similar to time series used in DBS treatments . The applied stimulation is then given by ϕ x ( t ) = ϕ x max ∑ j R ( t - t j p ) , ( 23 ) where ϕ x max is the pulse amplitude and R ( t ) is a top-hat function of width twidth , R ( t ) = { 1 , 0 < t < t width , 0 , otherwise . ( 24 ) The time-integral of ϕx ( t ) , Eq ( 23 ) , over the pulse width twidth is the average number of additional spikes generated in the target axon by the applied stimulation . The external stimulus is then coupled to a target population a via a connection parameter νax with a pulse frequency fstim . In the case of STN-DBS , ϕx ( t ) is coupled to the STN , but also to the GPi and GPe populations as an approximation of the activation of axons terminals near the stimulation site .
An afferent spike rate to any population in the CTBG system induces a change in the dendritic membrane potential of that population with a time evolution described by Eq ( 6 ) . Depending on the connection type , this change may be positive ( excitatory ) or negative ( inhibitory ) . Each inter-population connection then produces a change in voltage which is integrated at the soma , as described by Eq ( 5 ) . In the case of DBS , ϕx ( t ) , the mean voltage perturbation observed at the soma of neurons , can be shown by numerically convolving the stimulus time series with the normalized impulse response function given in differential form in Eq ( 7 ) . Fig 2 shows the evoked response potential generated by a stimulus pulse train , which resembles typical 130 Hz clinical stimulation , and the resulting perturbation to the target population firing rate . The temporal parameters used for the stimulus in Fig 2 ( a ) prescribes an inter-pulse quiescent period of about 7 ms . It can be seen in Fig 2 ( b ) that during this period the impulse response function only decays to about 80% of its maximum value . Fig 2 ( c ) can then be understood as showing small oscillations in the evoked response potential about a constant mean perturbation that results from stimulus time scales which are shorter than population response time scales . The evoked response potential is integrated at the soma of the target population along with intrinsic afferents from other populations within the network . A constant perturbation applied to the soma potential of a population changes its mean firing rate by moving the population along its corresponding sigmoidal response function , ( 1 ) . Fig 2 ( d ) demonstrates that in the case of inhibition mean soma potentials correspond to lower mean firing rates when compared with unperturbed values . In the case of excitation , the effect is reversed with mean soma potentials corresponding to higher mean firing rates when compared with the unperturbed values . Enhanced activity at ∼13-30 Hz is a common feature of Parkinson’s disease patient LFP recordings in the GPi and STN which has been correlated with symptom severity [43 , 44 , 42] . Recent works have suggested that the neural circuit formed between the GPe and STN can generate these beta oscillations [45 , 46] and that excitatory inputs from the cortex may facilitate their amplification [47] . Table 1 contains parameter estimates for parkinsonian states of the CTBG model adapted from [63] . Changes to the [63] connection strength estimates were made in order to explore the effects of a dominant GPe-STN-GPe pathway . In Fig 3 ( a ) and 3 ( b ) , power spectra of the STN firing rate demonstrate enhanced activity at ∼26 Hz , as well as at ∼6 Hz . By increasing STN-GPe coupling vp2ζ , damping of the GPe-STN-GPe loop is weakened and results in a strengthened hyperdirect pathway . Together these loops drive 26 Hz oscillations in the STN firing rate which project to the GPi population through STN efferents and then on to thalamic and cortical populations . The 6 Hz STN oscillations observed in Fig 3 ( b ) are weaker than the 26 Hz beta oscillations . However , the power spectrum for the cortical population shows an opposite relationship with stronger 6 Hz activity than the 26 Hz beta oscillations . This is an interesting result because tremor oscillations in PD patients measured via electromyography ( EMG ) are typically about 6 Hz and these correlate well with motor cortical activity measured via electroencephalography ( EEG ) [85] . However , STN LFP recordings about the 4–6 Hz tremor frequency have yet to be demonstrated as a reliable source for tremor detection [86] . Furthermore , other studies have suggested the thalamo-basal ganglia circuit as the origin of tremor oscillations [87] . The model configuration required to produce a dominant GPe-STN-GPe loop resonance involves both an increase in STN-GPe coupling vp2ζ , with respect to previous parameter estimates [63] , as well as an increase in cortico-STN coupling νζe . This is consistent the findings of a recent conductance-based modelling study where cortical inputs amplified parkinsonian oscillations generated by the subthalamo-pallidal circuit [47] , although the frequencies observed in that study were lower , 8–14 Hz , and represent dopamine depleted states of primates [45] . In this section the CTBG system is numerically simulated using parkinsonian parameters defined in Table 1 . These parameters yield strong GPe-STN and hyperdirect loop resonances , which results in large amplitude ∼26 Hz oscillations in STN activity . In Fig 4 ( a ) parkinsonian ∼26 Hz STN activity is simulated for 30 s and then 150 Hz DBS is applied . Following the application of this stimulation , a damping of the ∼26 Hz oscillation is observed . A comparison of STN power spectrums pre-stimulus and during stimulation is given in Fig 4 ( b ) and this shows peak power is reduced as a result of DBS . Fig 5 ( a ) demonstrates that increasing DBS pulse frequency strengthens the suppression of 13-30 Hz STN activity . As discussed in previous sections , coupling a DBS input to any population in the model results in an effective constant perturbation to the membrane potential Va of that population . Because DBS is coupled to the STN , GPe , and GPi populations via νζx = −1 . 2 mVs and v p 1 x , v p 1 x = 1 . 2 mVs , each corresponding mean membrane potential is perturbed by |ΔV| . This perturbation is −ΔV ( inhibitory ) for the STN population , and +ΔV ( excitatory ) for the GPi and GPe populations . In Fig 5 ( b ) we compare power suppression at 13–30 Hz for two cases: In the first case , a direct constant perturbation is made to the membrane potentials of the STN , GPi , and GPe populations . In the second case , the stimulus input used to produce Fig 5 ( a ) is convolved with an impulse response function , as discussed in a previous section . This allows the DBS input to be approximately represented as a constant perturbation to the membrane potential of a given population . Fig 5 ( b ) shows how peak power between 13-30 Hz is effected by directly perturbing the mean membrane potential for the STN , GPi , and GPe populations relative to indirectly perturbing them with an oscillating DBS input . The suppression of pathological beta activity by DBS in our model is then largely attributable to this effective perturbation to the mean membrane potential . Fig 5 ( a ) and 5 ( b ) also show constructive wave interactions for stimulus pulse frequencies equal to the 26 Hz beta oscillation and destructive interactions near the beta peak and its harmonic ( 52 Hz ) and subharmonic ( 13 Hz ) . Studies have shown low-frequency stimulation may worsen PD motor symptoms [88 , 89] as well as improve them [90] . The dependence of key network gains on the DBS pulse frequency is shown in Fig 6 ( a ) and 6 ( b ) . The parkinsonian parameters define a pathologically strong STN-GPe-STN loop gain as well as a strong hyperdirect pathway . As the pulse frequency of the DBS inputs increases , so too does the net inhibition in the system . This is due to DBS inputs activating the inhibitory pallidal populations ( GPe and GPi ) more strongly . In contrast , the remaining DBS input inhibits the STN population , which is critical to the generation of a ∼ 26 Hz resonance . It is important to note that the same suppressive effect can be achieved for a lower stimulus frequency if the stimulus amplitude is correspondingly increased . In DBS treatments , using larger signal amplitudes has the potential to increase the area directly affected by the applied stimulation , possibly incorporating non-motor projecting segments of the STN or even adjacent populations . Our results demonstrate that a high stimulus pulse frequency ( fstim > 100 Hz ) is necessary for beta suppression when the signal amplitude is constrained to be small relative to other STN inputs , e . g . , for the Table 1 parameters DBS constitutes ∼6% of the connection weighted activity arriving at the STN over a time interval greater than several stimulus pulse widths . Fig 7 shows the power spectrum of the STN firing rate time series as a function of DBS pulse frequency . Strong oscillations are seen at ∼26 Hz and its second harmonic ∼52 Hz . When the stimulus pulse frequency reaches about 140 Hz the ∼26 Hz power decreases by several orders of magnitude . Overall power within the 0–60 Hz frequency band also decreases and is redistributed to higher frequencies . Additionally , Fig 7 shows peaks at the stimulus frequency fstim ( 1:1 ) and also at its harmonics fstim ( N:1 ) , however , the harmonics are much weaker and not clearly discernible in this plot . Nonlinear wave interactions have previously been demonstrated in a neural field model of the corticothalamic system [91] which shows good agreement with human EEG studies [92] . In [91] , a periodic nonlinear input was used to drive the CT system and resulted in nonlinear interactions between the drive frequency and an intrinsic alpha oscillation . Spectral peaks were found at frequencies equal to the sum and difference of the drive frequency and the intrinsic alpha frequency , f± = |fstim ± fα| , as well their respective harmonics . Our model also demonstrates these nonlinear interactions . In Fig 7 , spectral peaks are seen at the sum and difference of the stimulus pulse frequency and the beta frequency f± = |fstim ± fβ| where fβ = 26 Hz . These nonlinear interactions are much more distinct in Fig 8 with peaks seen at f± = −fstim + 2fβ , 2fstim − fβ , and −2fstim + 3fβ . Additionally , Fig 8 demonstrates an entrainment of STN activity as a result of DBS inputs . The intrinsic parkinsonian beta peak is shifted to match the stimulus pulse frequency within the 25 . 5-26 . 2 Hz range . This result is consistent with experimental findings in human PD studies where a local entrainment of neural activity was observed during GPi-DBS [23] .
In this work we have developed a novel description of deep brain stimulation and incorporated it into a neural field model of the corticothalamic-basal ganglia system . The model has enabled us to explore generative mechanisms for the pathological beta band activity observed in Parkinson’s disease and the influences of DBS on these oscillations . The main results of the paper are as follows: Overall , our work provides insights into the generative mechanisms of pathological oscillations in human Parkinson’s disease and the population level effects of deep brain stimulation upon these oscillations . Furthermore , the model provides a framework for predicting effective stimulus protocols systematically rather than by trail and error , as has been the case to date . Closed-loop adaptive DBS systems use feedback from local field potential measurements made via the implanted simulation electrode to modulate stimulus protocols [94] . Our model could be used in conjunction with an adaptive DBS system to increase the efficacy of clinical treatments . Cortical and subthalamic firing rate spectra in this model could be fitted to EEG and LFP spectra during an on-off DBS treatment cycle . The change in spectra corresponds to specific variable changes in the model and the trajectory of these changes could then be used as a detection method for parkinsonian states that are specific to the patient . Several studies have observed antidromic activation as a result of deep brain stimulation [95] , and activation of pallido-thalamic fibers during STN-DBS [96] , which could be included in future generations of the model .
|
Pathological 13-30 Hz ( beta ) oscillations within the basal ganglia are a characteristic feature of human Parkinson’s disease which seem to correlate with symptom severity . The origin of these oscillations and the suppressive mechanism of effective deep brain stimulation treatments remains to be shown . We formulate a physiologically based population model of the corticothalamic-basal ganglia system that produces 13-30 Hz oscillations in the neural circuit formed between the globus pallidus pars externa and the subthalamic nucleus and the hyperdirect corticothalamic-basal ganglia pathway . We then develop a model of deep brain stimulation applied to the corticothalamic-basal ganglia system that permits systematic determination of effective stimulus protocols , which have been estimated by trial and error to date . Our results demonstrate that high pulse frequency ( >140 Hz ) stimulation is required to effectively suppress the pathological oscillations , which agrees with clinically used values . Interactions between these oscillations and the applied stimulus also lead to complex spectral structure that shows remarkable similarity to that seen in steady-state evoked potential experiments .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"myoclonus",
"pathology",
"and",
"laboratory",
"medicine",
"neurodegenerative",
"diseases",
"membrane",
"potential",
"brain",
"electrophysiology",
"electrophysiology",
"neuroscience",
"surgical",
"and",
"invasive",
"medical",
"procedures",
"deep-brain",
"stimulation",
"functional",
"electrical",
"stimulation",
"signs",
"and",
"symptoms",
"brain",
"mapping",
"nerve",
"fibers",
"bioassays",
"and",
"physiological",
"analysis",
"research",
"and",
"analysis",
"methods",
"animal",
"cells",
"axons",
"biological",
"tissue",
"electrophysiological",
"techniques",
"movement",
"disorders",
"cellular",
"neuroscience",
"diagnostic",
"medicine",
"cell",
"biology",
"anatomy",
"neurology",
"parkinson",
"disease",
"physiology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"ganglia",
"neurophysiology"
] |
2018
|
Quantitative theory of deep brain stimulation of the subthalamic nucleus for the suppression of pathological rhythms in Parkinson’s disease
|
In malaria-naïve individuals , Plasmodium falciparum infection results in high levels of parasite-infected red blood cells ( iRBCs ) that trigger systemic inflammation and fever . Conversely , individuals in endemic areas who are repeatedly infected are often asymptomatic and have low levels of iRBCs , even young children . We hypothesized that febrile malaria alters the immune system such that P . falciparum re-exposure results in reduced production of pro-inflammatory cytokines/chemokines and enhanced anti-parasite effector responses compared to responses induced before malaria . To test this hypothesis we used a systems biology approach to analyze PBMCs sampled from healthy children before the six-month malaria season and the same children seven days after treatment of their first febrile malaria episode of the ensuing season . PBMCs were stimulated with iRBC in vitro and various immune parameters were measured . Before the malaria season , children's immune cells responded to iRBCs by producing pro-inflammatory mediators such as IL-1β , IL-6 and IL-8 . Following malaria there was a marked shift in the response to iRBCs with the same children's immune cells producing lower levels of pro-inflammatory cytokines and higher levels of anti-inflammatory cytokines ( IL-10 , TGF-β ) . In addition , molecules involved in phagocytosis and activation of adaptive immunity were upregulated after malaria as compared to before . This shift was accompanied by an increase in P . falciparum-specific CD4+Foxp3− T cells that co-produce IL-10 , IFN-γ and TNF; however , after the subsequent six-month dry season , a period of markedly reduced malaria transmission , P . falciparum–inducible IL-10 production remained partially upregulated only in children with persistent asymptomatic infections . These findings suggest that in the face of P . falciparum re-exposure , children acquire exposure-dependent P . falciparum–specific immunoregulatory responses that dampen pathogenic inflammation while enhancing anti-parasite effector mechanisms . These data provide mechanistic insight into the observation that P . falciparum–infected children in endemic areas are often afebrile and tend to control parasite replication .
In previously unexposed individuals , blood-stage Plasmodium falciparum parasites rapidly replicate and almost invariably induce fever and other symptoms of malaria [1] through the production of pro-inflammatory cytokines and chemokines [2]–[4] . Although the initial systemic inflammatory response is crucial for setting in motion the innate and adaptive immune effector mechanisms that control blood-stage parasites [5] , [6] , dysregulated inflammation has been linked to severe malaria [7] , [8] which only occurs in a minority of individuals with infrequent or no prior malaria exposure [9] . Conversely , in malaria endemic areas where individuals are repeatedly exposed , P . falciparum infections more commonly cause a mild febrile illness or no symptoms at all , and parasite numbers in the blood are generally kept in check , even in young children [10]–[12] who have yet to acquire a fully protective antibody repertoire [13] . The nature of the immune response that enables most children to restrain P . falciparum-induced inflammation while maintaining control of parasite replication remains elusive [6] , [14] . The notion of malaria ‘tolerance’ has long been invoked to explain the common finding of low-level , asymptomatic blood-stage infection in endemic areas [15] , particularly among children , as antibodies that reliably protect against febrile malaria are only acquired after many years of exposure to genetically diverse and clonally variant P . falciparum antigens [13] . Several mechanisms have been proposed to explain malaria tolerance or ‘anti-disease’ immunity [14] , [16] including antibody-mediated neutralization of P . falciparum pathogen-associated molecular pattern ( PAMP ) molecules such as GPI anchors [14] , [17] , [18]; desensitization of pattern-recognition receptor ( PRR ) -mediated signaling as a result of repeated stimulation [16]; and the production of anti-inflammatory mediators such as IL-10 [2] , [19]–[21] and TGF-β [21]–[23] that suppress inflammation-driven anti-parasite effector mechanisms once parasite replication has been controlled [14] . Interestingly , it has long been speculated that parallels exist between malarial tolerance and bacterial endotoxin tolerance ( reviewed in [16] ) . This speculation is based in part on early studies in humans that showed that malaria induces cross-tolerance to the febrile response normally induced by bacterial endotoxin [24] , [25] . The traditional view of endotoxin tolerance holds that immune cells that are exposed to endotoxin have an altered response when re-challenged with endotoxin , such that the production of pro-inflammatory cytokines and chemokines is attenuated relative to the response induced at homeostasis [26] . Whether P . falciparum infection in humans ‘tolerizes’ immune cells in an analogous manner is unknown . More specifically , there is no direct evidence that febrile malaria in humans induces regulatory responses that limit the production of pro-inflammatory cytokines and chemokines upon re-exposure to P . falciparum parasites relative to responses induced at homeostasis before malaria in the same individual . Also , it is unclear how current views of malaria-induced tolerance/regulatory responses account for the observation that most children in endemic areas are able to restrain P . falciparum-induced inflammation and simultaneously control parasite replication , since generalized suppression of P . falciparum-triggered immune responses predicts that parasite replication would proceed unhindered and cause severe disease . A potential solution to this problem is that febrile malaria temporarily alters immune cells such that the host responds to P . falciparum re-exposure by downregulating the production of acute phase pro-inflammatory mediators that contribute to fever and other malaria symptoms while enhancing anti-parasite effector mechanisms that control parasite replication , such as phagocytosis-mediated clearance of blood-stage parasites . This hypothesis is supported by a recent study by Foster et al . who used an in vitro model of lipopolysaccharide ( LPS ) tolerance in murine macrophages to show that the regulation of LPS-triggered inflammation is component-specific , such that pro-inflammatory mediators are transiently silenced or ‘tolerized’ while antimicrobial effectors are primed or enhanced upon re-challenge with LPS [27] . A similar phenotype was observed in a human model of LPS tolerance [28] , but whether these observations reflect events induced by natural infections in humans is unknown . Here we tested the hypothesis that febrile malaria alters children's immune cells such that pro-inflammatory mediators are downregulated and anti-parasite effector responses are upregulated upon re-exposure to P . falciparum parasites compared to that induced at homeostasis before malaria in the same children . In addition , given the pivotal role of IL-10 in regulating Plasmodium-induced inflammation in murine models [19] , [29] , we sought to elucidate the identity , function and kinetics of P . falciparum-specific IL-10-producing cells . We further asked if P . falciparum-inducible regulatory responses that limit inflammation are maintained in untreated children who harbor chronic asymptomatic infections , and whether these responses can be recalled in children who have not been exposed to P . falciparum for extended periods of time . To address these questions we applied a systems biology approach [30] to a longitudinal analysis of peripheral blood mononuclear cells ( PBMCs ) sampled from Malian children over a 12-month period—starting at their healthy baseline before the six-month malaria season; seven days after treatment of their first febrile malaria episode of the ensuing malaria season ( when malaria symptoms had resolved ) ; and after the subsequent six-month dry season , a period of little to no P . falciparum transmission . We found that febrile malaria induces a marked shift in the response to P . falciparum re-exposure with cells producing lower levels of pro-inflammatory cytokines and chemokines and higher levels of anti-inflammatory cytokines compared to responses induced at homeostasis before malaria . Re-exposure was also associated with enhanced expression of pathways involved in phagocytosis and activation of adaptive immunity . This shift was accompanied by a marked increase in P . falciparum-specific CD4+Foxp3− T cells that co-produce IL-10 , IFN-γ and TNF . IL-10 remained partially inducible in untreated children with chronic asymptomatic infections whereas IL-10 was no longer inducible in children whose infections had been cleared by treatment .
To obtain a global view of transcriptional changes that persist in children's PBMCs after the clinical resolution of febrile malaria , compared to each child's own healthy baseline , we profiled RNA expression of PBMCs collected from 34 healthy Malian children before the six-month malaria season , when blood smears were negative for P . falciparum parasites , and 7 days after treatment of their first febrile malaria episode of the ensuing malaria season , when malaria symptoms had resolved . The average age of these children was 8 . 5 years and 32% were female . At their first febrile malaria episode of the season , children had an axillary temperature of >37 . 5°C ( or their parents reported fever within 24 hours ) , were infected with P . falciparum ( geometric mean density 17 , 817 asexual parasites/µl of blood ) and had no other cause of fever discernible on physical examination . The average incidence of febrile malaria during the six-month malaria season was similar for the 34 children in this study compared to children in this age group in the larger cohort ( 1 . 6 and 1 . 5 episodes , respectively ) [31] . By definition , these malaria-susceptible children had yet to acquire P . falciparum-specific antibodies that reliably protect from febrile malaria . Individual demographic and clinical data are shown in Table S1 . Malaria was effectively treated in all subjects with a standard 3-day course of artemether/lumefantrine . PBMCs were first analyzed directly ex vivo ( not re-stimulated ) . Principal components analysis of the microarray data showed segregation of transcription profiles based on time-point ( healthy baseline vs . day 7 after malaria ) , but not age , gender or batch effects ( Figure S1A ) . Within-subject gene-expression changes were computed and resulted in 1497 differentially expressed genes ( DEGs ) —1 , 351 increased and 146 decreased after the resolution of febrile malaria relative to baseline ( Table S2 ) . Ingenuity Pathway Analysis ( IPA ) identified “Infectious Disease” , “Immunological Disease” and “Inflammatory Disease” among the top five functional categories enriched with DEGs . Notably , all differentially expressed genes encoding cytokines and chemokines that have been associated with P . falciparum-induced fever and inflammation [2]–[4] , [32] , including the canonical pyrogenic cytokines IL1B and TNF as well as the pro-inflammatory chemokines IL8 , CCL3 ( MIP-1α ) and CXCL2 ( MIP-2α ) , were suppressed after the resolution of malaria to lower levels than observed at baseline ( Figure 1A ) . Conversely , genes expressing molecules directly involved in microbial killing and activation of adaptive immunity were significantly upregulated after the resolution of febrile malaria relative to the healthy baseline ( Figure 1A ) . Specifically , IPA identified the following canonical pathways as significantly upregulated: “Toll-like receptor signaling” ( P = 2 . 15e-6 ) , “Fcγ receptor-mediated phagocytosis in macrophages and monocytes” ( P = 1 . 8e-4 ) , “Production of nitric oxide and reactive species in macrophages” ( P = 1 . 09e-7 ) , “Antigen presentation pathway” ( P = 0 . 0463 ) , “T cell receptor signaling” ( P = 1 . 19e-5 ) and “Interferon signaling” ( P = 5 . 08e-6 ) ( Figure S1B ) . The expression of several PRRs including Toll-like receptor ( TLR ) 2 and TLR4 was increased after malaria relative to baseline ( Figure 1A ) , consistent with recent in vivo exposure to P . falciparum PAMPs such as GPI anchors [33] , hemozoin [34] , CpG-containing DNA motifs bound to hemozoin [35] and AT-rich DNA motifs [36] , but of note , NLRP3 , a putative receptor for P . falciparum hemozoin-induced IL-1β production [37] , was the only PRR to be downregulated after malaria relative to homeostasis ( Figure 1A ) . Because changes in mRNA levels in PBMCs can reflect an altered cell composition of the PBMC compartment or changes in gene expression in discrete cell populations , we analyzed the PBMCs used for gene-expression profiling by FACS and found no significant differences in the percentage of CD4+ or CD8+ T cells , B cells or monocytes after the resolution of febrile malaria relative to the healthy baseline ( Figure 1B ) . Moreover , at the individual subject level we found that genes encoding myeloid-expressed pro-inflammatory mediators were downregulated after resolution of malaria relative to baseline irrespective of changes in the percentage of monocytes ( Figure 1C and Figure S1C ) , and even as mRNA levels of other genes identified as myeloid-specific [38] were unchanged or increased after malaria relative to baseline ( Figure S1D ) . Taken together , these data indicate that the changes in mRNA levels that persist after the resolution of febrile malaria relative to the healthy baseline reflect differential regulation of gene expression rather than gross alterations in the composition of the PBMC compartment . These data indicate that febrile malaria induces transcriptional changes in PBMCs that persist after the resolution of febrile malaria; namely , we observed that the expression of genes encoding pro-inflammatory mediators is suppressed while the expression of molecules involved in microbial killing and activation of adaptive immunity is enhanced after malaria relative to baseline . On the basis of these findings we hypothesized that re-exposure to P . falciparum parasites soon after the resolution of febrile malaria would induce a qualitatively different immune response relative to that induced at the healthy baseline . To test this hypothesis and to investigate the molecular and cellular basis of regulatory responses induced upon P . falciparum re-exposure , we analyzed the same PBMCs from the same 34 children ( collected at the healthy baseline before the malaria season and seven days after treatment of the first febrile malaria episode ) following in vitro stimulation with P . falciparum-infected red-blood cell ( iRBC ) lysate . iRBC-inducible gene expression as well as secreted and intracellular cytokine production were examined with each child serving as his or her own healthy baseline control . Within-subject gene expression changes induced by iRBC stimulation at both time points were computed and resulted in 456 DEGs—148 decreased and 308 increased after the resolution of febrile malaria relative to that induced at the healthy baseline ( Table S3 ) . IPA identified “Inflammatory Response” as the functional category with the highest enrichment score ( P = 3 . 36e-34 ) . Within this functional category , all differentially expressed pro-inflammatory cytokines and chemokines ( IL1B , IL6 , IL8 , IL19 , IL24 , CCL3 , CCL3L1 , CCL19 , CCL20 , CCL22 , CXCL1 , CXCL2 , CXCL3 , and CXCL6 ) were downregulated in response to iRBC stimulation after the resolution of malaria relative to the iRBC-induced response at baseline ( Figure 1D ) . In line with this result , IPA identified the following canonical pathways as significantly downregulated: “Acute phase response signaling” ( P = 5 . 94e-4 ) , “Role of hypercytokinemia/hyperchemokinemia in pathogenesis of Influenza ( P = 1 . 19e-3 ) , “IL-6 signaling” ( P = 1 . 18e-3 ) , “Agranulocyte adhesion and diapedesis” ( P = 8 . 82e-6 ) , “Granulocyte adhesion and diapedesis” ( P = 6 . 55e-9 ) and “Role of cytokines in mediating communication between immune cells” ( P = 0 . 017 ) ( Figure S1E ) . Consistent with the downregulation of pro-inflammatory responses , NFKB1 and TREM-1 , a positive regulator of inflammation [39] , were downregulated , while several negative regulators of inflammation were upregulated including IL18BP , IL1R2 , BTLA and SAMHD1 ( Figure 1D ) . In contrast to the downregulation of pro-inflammatory cytokine and chemokine responses , genes encoding molecules involved in microbial killing and activation of adaptive immunity were upregulated by iRBC stimulation after malaria relative to responses induced by iRBC stimulation at baseline . These included molecules involved in opsonic and non-opsonic phagocytosis , phagolysosome maturation , antigen presentation and co-stimulation ( Figure 1D ) . IPA identified the following canonical pathways as significantly upregulated: “iNOS signaling” ( P = 0 . 0012 ) , “Antigen presentation pathway” ( P = 6 . 87e-5 ) , “T helper cell differentiation” ( P = 1 . 88e-3 ) and “T cell receptor signaling” ( P = 0 . 035 ) ( Figure S1E ) . Together these data suggest that re-exposure to P . falciparum parasites after a recent episode of febrile malaria induces the differential expression of functionally distinct components of the immune response , whereby acute phase pro-inflammatory cytokines and chemokines that drive the initial systemic inflammatory response are restrained , while pathways involved in microbial killing and activation of adaptive immunity are upregulated . To validate the expression of selected immune-related genes , we used quantitative real time ( qRT ) -PCR to analyze PBMCs from the 17 children who had microarray data from the ex vivo and iRBC-stimulated experiments at both time points ( before and after malaria ) . We found a positive correlation ( r = 0 . 8653; P<0 . 0001 ) for gene expression as detected by microarray and qRT-PCR ( Figure 1E and Table S4 ) . The qRT-PCR data confirmed decreased expression of the canonical fever-inducing cytokines IL1B and IL6 and increased expression of the anti-inflammatory cytokine TGFB after resolution of febrile malaria relative to the healthy pre-malaria baseline , in both the unstimulated and iRBC-stimulated experiments ( Figure 1F ) . Therefore , the qRT-PCR data confirmed a molecular pattern of restrained P . falciparum-inducible inflammation in children who had recently recovered from febrile malaria . IL-10 plays a critical role in controlling and resolving inflammation by limiting the production of pro-inflammatory cytokines and chemokines [40] , yet we did not observe differential expression of IL10 in the microarray data . Given the known temporal dissociation between IL-10 transcription and translation [41] , we assayed supernatants of iRBC-stimulated PBMCs for secreted IL-10 using a multiplex assay that also measured IL-1β , IL-6 , IL-8 and TNF . We observed that P . falciparum-inducible IL-10 production was higher after the resolution of febrile malaria relative to the healthy baseline of the same children before the malaria season ( P<0 . 0001; Figure 2A ) , while P . falciparum-inducible production of the pro-inflammatory chemokine IL-8 was lower after the resolution of febrile malaria relative to baseline ( P<0 . 0001; Figure 2A ) , in agreement with the microarray data . P . falciparum-inducible production of IL-1β and IL-6 trended toward lower levels after the resolution of febrile malaria relative to baseline , but the decrease was not statistically significant ( Figure 2A ) . Because IL-1β and IL-6 are primarily produced by monocytes/macrophages , we isolated monocytes/macrophages ( Figure S2A ) from 9 additional children who had PBMCs available at their healthy baseline before the malaria season and 14 days after their first febrile malaria episode of the ensuing malaria season ( Table S1 ) . We stimulated these monocytes/macrophages with iRBCs for 6 hours and measured IL-1β and IL-6 in the supernatants . We found that P . falciparum-inducible production of IL-1β and IL-6 by monocytes/macrophages was lower after the resolution of febrile malaria relative to that induced at baseline ( P = 0 . 0066 and P = 0 . 0003 for IL-1β and IL-6 respectively; Figure 2B ) , consistent with a reduced risk of fever in children who are exposed to ongoing P . falciparum transmission during the malaria season . In an independent experiment , we sought to determine if the upregulation of P . falciparum-inducible IL-10 production after malaria influences the production of pro-inflammatory cytokines . We observed that blocking IL-10 activity with antibodies specific for IL-10 and the IL-10 receptor ( Figure 2C ) enhanced iRBC-inducible TNF and IL-6 production in some but not all children after the resolution of febrile malaria compared to baseline ( Figures 2D and E ) . We next asked if P . falciparum-inducible IL-10 responses could be recalled in children who had not been exposed to P . falciparum transmission for an extended period of time . We performed iRBC stimulation of PBMCs collected from 18 additional children ( Table S1 ) at their healthy baseline before the malaria season , 7 days after treatment of their first malaria episode of the ensuing 6-month malaria season , and after the following 6-month dry season , a period of little to no P . falciparum transmission . This independent experiment confirmed that P . falciparum-inducible IL-10 is upregulated after the resolution of febrile malaria relative to baseline ( P = 0 . 0082; Figure 2F ) . However , in the absence of ongoing malaria exposure , children reverted to an apparent homeostatic baseline in which IL-10 production was no longer inducible ( Figure 2F ) , suggesting that ongoing malaria exposure is required to maintain P . falciparum-inducible IL-10 production capacity . To test this hypothesis we identified 16 untreated children ( Table S1 ) whose asymptomatic P . falciparum infections persisted through the six-month dry season , and compared their P . falciparum-inducible IL-10 response to age-matched children who were uninfected at the same time point at the end of the dry season . We observed that P . falciparum-inducible IL-10 responses of persistently infected asymptomatic children were higher than responses of age-matched uninfected children ( P = 0 . 0166; Figure 2G ) , suggesting that P . falciparum-inducible IL-10 upregulation is partially maintained by ongoing P . falciparum exposure and that IL-10 upregulation may contribute to protection from febrile malaria in the context of ongoing P . falciparum exposure . Despite evidence that IL-10 plays a critical role in regulating Plasmodium-induced inflammation in murine models , the cellular sources of IL-10 and the functionality and kinetics of IL-10-producing cells in the context of human malaria remain unclear [42] . To identify the predominant cellular source of P . falciparum-inducible IL-10 and to investigate longitudinally the functionality and kinetics of IL-10-producing cells in children exposed to intense seasonal malaria , we analyzed PBMCs by FACS with intracellular staining for IL-10 , IFN-γ and TNF after in vitro iRBC stimulation at the healthy baseline before the malaria season , 7 days after malaria treatment ( when symptoms had resolved ) , and after the 6-month dry season . Consistent with the kinetics of P . falciparum-inducible secreted IL-10 production described above ( Figure 2F ) , we found that P . falciparum-inducible IL-10 production by CD4+ T cells increased significantly after the resolution of febrile malaria relative to the healthy baseline ( P = 0 . 0095; Figure 3A ) , and then reverted to a state in which IL-10 production was no longer inducible by the end of the following six-month dry season ( Figure 3A ) . Interestingly , the majority of P . falciparum-inducible IL-10-producing PBMCs following febrile malaria were CD3+CD4+ T cells ( Figure 3B; mean 53 . 1% , 95%CI: 44 . 8–61 . 5 ) and most of these were CD25+FOXP3− ( Figure 3B; mean 78 . 8% , 95%CI: 72 . 2–85 . 5 ) , while FOXP3+CD4+ T cells ( regulatory T cells ) represented only a small percentage of IL-10-producing T cells . P . falciparum-inducible IFN-γ- and TNF-producing CD4+ T cells also increased after the resolution of malaria compared to baseline , and like IL-10 , reverted to homeostasis after the dry season ( Figure 3A ) . At the single-cell level the majority of IL-10-producing P . falciparum-inducible CD4+ T cells also produced IFN-γ , or IFN-γ plus TNF ( Figure 3C–E ) , thus identifying these cells as ‘self-regulating’ Th1 effector cells [43] . From these results and the microarray data emerges a consistent theme whereby re-exposure to P . falciparum parasites after recent febrile malaria induces exposure-dependent regulatory mechanisms that limit the production of pro-inflammatory mediators that drive systemic inflammation while enhancing effector mechanisms that control parasite replication . Because whole microbe stimulation with P . falciparum iRBCs involves a complex mixture of antigens and stimuli for innate receptors , we asked whether P . falciparum-inducible IL-10 production by CD4+ T cells requires antigen presenting cells ( APCs ) and T cell receptor engagement . We magnetically isolated CD4+ T cells that had been collected after the resolution of febrile malaria and found that they failed to produce IL-10 in response to iRBC stimulation in the absence of antigen-presenting cells ( Figure 3F ) . Moreover , iRBC-induced IL-10 production by CD4+ T cells was abrogated in PBMC cultures in the presence of antibodies that block major histocompatibility complex ( MHC ) class II molecules ( Figure 3G ) . Together these data demonstrate that P . falciparum-inducible IL-10 production by CD4+ T cells is T cell receptor-dependent . Having established that iRBC-inducible IL-10 production by CD4+ T cells requires APCs and T cell receptor engagement , we sought to understand the role that in vivo conditioning of APCs plays in modulating IL-10 production by P . falciparum-specific CD4+ T cells . We magnetically isolated CD4+ T cells collected after the resolution of febrile malaria and cultured these cells with autologous APCs collected at the healthy baseline before the malaria season or after the resolution of febrile malaria . Under both conditions iRBC-inducible IL-10 production by CD4+ T cells was restored to similar levels ( Figures 3F ) , suggesting that the in vivo conditions during acute febrile malaria shape the functional response of CD4+ T cells in a manner that is independent of the in vivo conditioning of APCs .
In our previous investigations at this study site we observed that the risk of febrile malaria slowly decreases over years as individuals are exposed to intense seasonal P . falciparum transmission such that adults rarely experience febrile malaria when infected with blood-stage parasites [44] . The gradual acquisition of blood-stage immunity that reliably protects from the onset of febrile malaria likely reflects the need for repeated infections over years to achieve levels of broadly reactive antibodies that exceed a protective threshold [13] , [45] . However , even malaria-susceptible children at this study site ( who by definition have yet to acquire reliably protective antibodies ) experience only 1 to 2 febrile malaria episodes per six-month malaria season despite ≥100 infective mosquito bites per person each season , and generally these children manage to keep parasite numbers in the blood in check [44] . These observations prompted us to investigate immune mechanisms beyond antibody responses that might contribute to protection from febrile malaria and parasite replication in children who are exposed to repeated P . falciparum infections , and also to investigate how children become susceptible again to febrile malaria after a period of decreased P . falciparum exposure . We found that acute febrile malaria alters children's PBMCs such that P . falciparum re-exposure results in downregulation of acute phase pro-inflammatory cytokines that drive fever and systemic inflammation ( e . g . IL-1β and IL-6 from monocytes/macrophages ) , and upregulation of immune mechanisms involved in control of inflammation ( e . g . IL-10-producing CD4+ T cells ) and parasite clearance ( e . g . IFN-γ-producing CD4+ T cells , phagocytosis and phagolysosome maturation ) ( Figure 4 ) . The maintenance of this regulatory state appears to depend on recent or ongoing P . falciparum exposure as children revert to a homeostatic baseline in the absence of ongoing P . falciparum exposure during the six-month dry season . The short-lived , exposure-dependent nature of this response mirrors the kinetics of P . falciparum-specific antibody responses in children [44] , [45] , suggesting that these responses work in concert to protect children as long as P . falciparum exposure is ongoing . These data offer mechanistic insights into how children who are repeatedly infected with P . falciparum commonly manage to remain afebrile and control parasite replication , and how they become susceptible again to febrile malaria after a period of reduced P . falciparum exposure . The possibility that treatment with artemether/lumefantrine contributed to these findings cannot be excluded . These data shed light on the long-standing and enigmatic clinical notion of ‘premunition’—a partially effective , exposure-dependent immune response that protects against illness and high numbers of parasites in the blood without completely eliminating the infection [12] , [46] , [47] . Although premunition is often viewed as a state of immune dysregulation or suppression [12] , [46] , [47] , we speculate that it evolved as an appropriate immune response in the face of unrelenting exposure to genetically and antigenically diverse parasites such that young children are at least partially protected from potentially life-threatening inflammation and unchecked parasite replication before they acquire durable , broadly reactive antibodies that reliably protect against the onset of malaria symptoms . Although we did not study severe malaria per se—an overlapping set of syndromes [48] which have been linked to excessive inflammation [49]—it is conceivable that the ability to rapidly downregulate P . falciparum-inducible inflammation in early life contributes to the rapid acquisition of strain-transcendent immunity to severe malaria which may occur after only one or two symptomatic infections [50] , and conversely , that the small percentage of children who develop severe malaria are those whose genetic background , environment ( e . g . co-infection history , microbiota , nutritional status ) or specific interaction with parasite virulence factors [51]–[54] tips them toward dysregulated pathologic inflammatory responses . The prospective design of this study , in which each subject served as their own healthy control , provides a rare view of the regulation and functional plasticity of innate and adaptive immune cells in response to a natural infection in humans . In general , innate immune cells such as monocytes/macrophages first detect pathogens through PRRs such as TLRs and NOD-like receptors ( NLRs ) which recognize highly conserved PAMPs [55] . Through these initial host-pathogen interactions , innate immune cells provide the first line of defense against pathogen invasion and also direct the quality of antigen-specific B and T cell responses . To date , only a handful of P . falciparum PAMPs and their respective PRRs have been identified . These include GPI anchors ( TLR2>TLR4 ) [33] , hemozoin ( NLRP3 ) [34] , CpG-containing DNA motifs bound to hemozoin ( TLR9 ) [35] and AT-rich DNA motifs ( unknown cytosolic receptor ) [36] . Studies in vitro and in animal models show that these PAMPs drive monocytes/macrophages to produce pro-inflammatory cytokines and chemokines such as IL-1β , IL-6 , IL-8 and TNF [33]–[35] . These observations are consistent with studies in humans that show these cytokines and chemokines rise and fall in the serum of individuals treated for febrile malaria [2]–[4] . However , prior to this study , the nature of the inflammatory response induced by P . falciparum re-exposure relative to that induced at the healthy baseline of the same individuals was unknown . Here we show that the capacity of monocytes/macrophages to produce the canonical pyrogenic cytokines IL-1β and IL-6 is reduced upon re-exposure to P . falciparum parasites relative to that induced at the healthy baseline of the same individuals—a finding we observed at the mRNA level by microarray and qRT-PCR in PBMCs , and at the protein level in isolated monocytes/macrophages . Although the precise molecular mechanisms that regulate P . falciparum-inducible inflammation remain to be fully elucidated , this study offers insight into the multiple levels at which this regulation might occur . For example , we observed that NFKB1 expression was downregulated after malaria relative to baseline . NFKB1 encodes the p50 component of the canonical p65/p50 NF-κB heterodimeric transcription factor that stimulates the expression of pro-inflammatory cytokines such as IL-1β and IL-6 [56] . We also observed decreased expression of NLRP3 after malaria relative to baseline . NLRP3 encodes a component of the NALP3 inflammasome which is expressed in myeloid cells and activates caspase-1 , thereby promoting the maturation and secretion of IL-1β [57] . Other differentially expressed PRRs such as TLR2 and TRL4 were upregulated after malaria relative to baseline , suggesting that the regulation of P . falciparum-inducible inflammation does not occur at the level of TLR expression . Interestingly , TLR expression is also upregulated in the context of tolerance induced by gram positive bacteria [58] , whereas tolerance induced by gram negative bacteria is associated with reduced expression of TLR2 and TLR4 [59] , underscoring the microbe-specific nature of immune-regulation . After the resolution of febrile malaria we also observed increased expression of genes encoding proteins that limit the inflammatory response including IL18BP , IL1R2 , CTLA4 , BTLA , SAMHD1 and TNFSF10 , as well as decreased expression of genes encoding proteins that promote inflammation including PTGS2 and TREM1 . Further studies are needed to more clearly elucidate the signaling networks involved in regulating the immune response to P . falciparum infection and to fully understand the relationships between perturbations to these networks and the variability in malaria clinical outcomes . The gene expression microarray data also shed light on the regulation of chemotactic responses [60] in malaria . Relative to the response induced at the healthy baseline , re-exposure to P . falciparum parasites was associated with downregulated expression of chemokines that recruit macrophages ( CCL3 , CCL3L1; Figures 1A , 1D and 4 ) and neutrophils ( CXCL1 , CXCL2 , CXCL3 , CXCL6; Figures 1A , 1D and 4 ) , but upregulated expression of monocyte/macrophage-specific chemokine receptors ( CCR1 , CCR5; Figure 4 ) . We postulate that this pattern reflects the fine-tuning of chemotactic responses in the face of ongoing or repeated P . falciparum exposure , whereby systemic chemokine release is restrained to decrease the potential for tissue damage caused by aberrant trafficking and accumulation of effector cells such as neutrophils , whereas the reciprocal regulation of monocyte/macrophage-specific chemokines ( repressed ) and chemokine receptors ( increased ) enhances the sensitivity of monocyte/macrophages to detect decreased concentrations of chemokines . Murine models clearly demonstrate that IL-10 and TGF-β play critical roles in regulating Plasmodium-induced inflammation [19] , [22] . In humans , IL-10 and TGF-β levels increase in serum during acute febrile malaria and then fall after treatment [2] , [20] , consistent with a role for these cytokines in restraining and resolving P . falciparum-induced inflammation during a single malaria episode . However , whether febrile malaria conditions the immune system to modify the production of IL-10 and TGF-β upon subsequent exposure to P . falciparum within the same individual remained an open question . Here we show that P . falciparum-inducible IL-10 and TGF-β production/expression is upregulated after the resolution of febrile malaria relative to that which is inducible at the healthy baseline of the same individuals . In addition , we observed that IL-10 blockade in vitro enhanced IL-6 and TNF production in some but not all children , consistent with a role for IL-10 in controlling inflammation in the setting of P . falciparum re-exposure , but also highlighting the complexity of regulatory responses that restrain P . falciparum-induced inflammation . Given the role of IL-10 in regulating Plasmodium-induced inflammation , we sought to illuminate the identity , function and kinetics of P . falciparum-specific IL-10-producing cells . Previous studies in humans have shown that total FOXP3+ T regulatory cells ( Tregs ) increase in response to experimental [23] and natural [61] P . falciparum infection ( reviewed in [42] ) , which suggested that Treg-generated anti-inflammatory cytokines play an important role in controlling P . falciparum-inducible inflammation . However , subsequent cross-sectional studies failed to conclusively show significant differences in Treg responses between individuals with mild and severe malaria [62] , [63] . Here we show that CD4+CD25+Foxp3− T cells are the predominant source of P . falciparum-inducible IL-10 , whereas Tregs contributed minimally to the overall IL-10 response . We demonstrate that IL-10 producing CD4+CD25+Foxp3− T cells are P . falciparum-specific , in that they require APCs and T cell receptor engagement to produce IL-10 . We also found that P . falciparum-inducible IL-10 production by CD4+ T cells isolated after malaria did not change significantly when these cells were co-cultured with homologous APCs collected before malaria , although there was a trend toward enhanced IL-10 production by these CD4+ T cells when cultured with APCs collected after malaria . Interestingly , we observed that a significant proportion of P . falciparum-specific IL-10 producing CD4+CD25+Foxp3− T cells co-produced the Th1 cytokines IFN-γ and/or TNF . Similar ‘self-regulating’ Th1 cells that co-produce IL-10 and IFN-γ were first identified in the lungs of patients with active pulmonary tuberculosis [64] and have since been observed in mice infected with Toxoplasma gondii [65] and Leishmania major [66] as well as in humans with visceral leishmaniasis [67] . Intriguingly , we observed that P . falciparum-specific IL-10 was only inducible in activated Th1 cells after recent febrile malaria , and that after the dry season IL-10 and IFN-γ were no longer inducible through P . falciparum stimulation . This is consistent with a recent study of Ugandan children by Jagannathan et al . in which the frequencies of P . falciparum-specific CD4+ T cells co-producing IFN-γ and IL-10 were inversely associated with days since last malaria episode [68] . Together these data support the hypothesis that IL-10 production by antigen-specific Th1 cells represents a normal phase of their differentiation program which is reached after full activation in order to restrain the inflammatory response while still allowing an efficacious immune response [43] , [65] , namely , IFN-γ production that promotes phagocytosis-mediated clearance of blood-stage parasites . Importantly , we show that P . falciparum-specific IL-10 production remains inducible in some but not all untreated children whose low-level asymptomatic P . falciparum infections persisted through the six-month dry season , suggesting that the production of IL-10 and IFN-γ is finely tuned such that parasitemia is controlled without inducing clinically overt inflammation—potentially explaining the long-standing clinical observation that most individuals , if left untreated after their initial bout of febrile malaria , become afebrile and maintain control of parasitemia for months before the infection is finally cleared . The exposure-dependent inducibility of IL-10 production by Th1 cells may also explain our previous observation at the same study site that children with asymptomatic P . falciparum infection at the end of the dry season are at lower risk of febrile malaria during the ensuing malaria season [31] , whereas uninfected children at the end of the dry season are at increased risk—corresponding temporally with their return to a homeostatic baseline in which P . falciparum exposure induces a pro-inflammatory phenotype . The exposure-dependent inducibility of IL-10 is also consistent with anecdotal reports of rapidly waning clinical immunity to febrile malaria in those who emigrate from malaria endemic areas [69] . Taken together these data point toward a protective effect of P . falciparum-specific IL-10 producing Th1 cells in malaria , a hypothesis supported by a cross-sectional study in The Gambia which showed a higher frequency of total IL-10 producing Th1 cells in children with mild versus severe malaria [62] . In contrast , a study in Uganda recently reported that frequencies of CD4+ T cells co-producing IFN-γ and IL-10 were not associated with protection from future malaria , although imprecise measures of malaria exposure may have led to spurious associations with protection [68] . More studies are needed to define the potential role of these cells in protection from malaria and to elucidate the molecular basis of their remarkable functional plasticity [70]—information that could define ways in which these cells could be safely induced and maintained through vaccination . Further studies are also needed to disentangle the relative contributions of IL-10 upregulation and antibodies to protection from malaria . Malaria-induced regulatory responses that control inflammation are often viewed as globally immunosuppressive , which predicts that parasites would grow unimpeded in individuals residing in areas of ongoing P . falciparum transmission . However , this model is at odds with the common finding of low-level , asymptomatic infection among children in endemic areas . Therefore a key finding of this study is that despite the downregulation of P . falciparum-inducible inflammation after the resolution of malaria , pathways involved in clearance of blood-stage parasites and activation of adaptive immunity were upregulated . Specifically , we observed P . falciparum-inducible upregulation of genes encoding proteins that mediate opsonic ( e . g . FCGR1 ) and non-opsonic ( e . g . CR1 , CD36 ) phagocytosis , phagolysosome maturation , antigen processing and presentation , and T cell co-stimulation ( Figure 4 ) . The limited blood volume available from children enrolled in this study precluded concomitant functional confirmation of these observations; however , the observed gene expression pattern of enhanced anti-microbial activity after the resolution of febrile malaria is consistent with the results of a study in which malaria-naïve adults , who were experimentally infected with P . falciparum , showed enhanced macrophage phagocytic activity after treatment relative to baseline [71] . Non-opsonic phagocytosis of iRBCs is considered to be an important first line of defense in non-immune or partially immune hosts who have yet to acquire P . falciparum-specific opsonizing antibodies [72] . Indeed , others have shown that the scavenger receptor CD36 mediates phagocytosis of non-opsonized iRBCs [73] , and interestingly , does so without inducing pro-inflammatory cytokines [73] , [74] . This study reveals several intriguing parallels between the regulation of P . falciparum-triggered inflammation and endotoxin tolerance [75] , a link that is particularly germane in light of earlier studies in humans that showed that malaria induces cross-tolerance to the febrile response normally induced by bacterial endotoxin [24] , [25] , suggesting at least partial overlap of regulatory pathways induced by Plasmodium and gram negative bacteria . Indeed , similar to what has been described in an in vitro model of LPS tolerance in murine macrophages [27] , we observed that the regulation of P . falciparum-triggered responses is component-specific , such that acute phase pro-inflammatory mediators such as IL-1β and IL-6 are transiently downregulated or ‘tolerized’ , while anti-parasitic effector pathways are primed or enhanced upon re-challenge with P . falciparum parasites . Further reductionist studies are needed to define the molecular mechanisms by which this regulation occurs including the potential role of chromatin modification [76] and microRNAs [77] . It will be of interest to understand how malaria-induced epigenetic reprogramming of innate immune cells—or “trained immunity”—differs from that induced by other pathogens [78] , [79] . To our knowledge , no human study has evaluated the genome-wide transcriptional response to a natural infection in which each subject serves as his or her own healthy control . Nearly all individuals at the study site become infected with P . falciparum within a predictable window of time each year [80] , which enabled us to compare intra-individual changes in PBMC gene expression at the healthy baseline before the malaria season and after the resolution of febrile malaria—both directly ex vivo and after re-exposing PBMCs to P . falciparum parasites in vitro . An important limitation of blood transcriptome analysis is that changes in mRNA levels can be driven by de novo transcriptional regulation or changes in the composition of PBMCs in peripheral blood [81] . Three lines of evidence indicate that the observed changes in mRNA levels in this study are driven by de novo transcriptional regulation . First , by flow cytometry we did not observe gross changes in the composition of the study subjects' PBMCs from before to after malaria . This is consistent with the observation that immune cells traffic out of the peripheral circulation during acute malaria but then return to the peripheral circulation after the infection has resolved [82] . Second , at the individual subject level we found that genes encoding myeloid-expressed pro-inflammatory mediators were downregulated after malaria relative to baseline , irrespective of changes in the percentage of monocytes , even as mRNA levels of other genes identified as myeloid-specific [38] were unchanged or increased relative to baseline . And finally , we observed that in vitro stimulation of fixed populations of cells ( PBMCs and isolated monocytes/macrophages ) induces de novo expression of immune-related genes . In summary , this longitudinal study of Malian children shows that febrile malaria induces exposure-dependent P . falciparum-specific regulatory responses that limit pathogenic inflammation and enhance anti-parasite effector responses upon P . falciparum re-exposure . These findings offer mechanistic insights into several long-standing clinical observations in malaria including the high incidence of asymptomatic P . falciparum infection in endemic areas [69] , reduced fever with repeated experimental Plasmodium infections in humans [83] , the rapid acquisition of immunity to severe malaria [50] , the rapid loss of clinical immunity to febrile malaria in the absence of ongoing P . falciparum exposure [6] and Plasmodium-induced hetero-tolerance to endotoxin challenge [25] . Longitudinal studies of symptomatic and asymptomatic individuals who are repeatedly exposed to P . falciparum will refine our understanding of the mechanisms underlying the regulation and dysregulation of Plasmodium-induced inflammation and may help define the potential for interventions that safely prevent or mitigate Plasmodium-induced immunopathology without compromising control of parasite replication .
The Ethics Committee of the Faculty of Medicine , Pharmacy , and Dentistry at the University of Sciences , Techniques , and Technologies of Bamako , and the Institutional Review Board of the National Institute of Allergy and Infectious Diseases , National Institutes of Health approved this study . Written informed consent was obtained from the parents or guardians of participating children . Study subjects were enrolled in an observational cohort study conducted in Kambila , Mali , a rural village of ∼1500 inhabitants where intense seasonal P . falciparum transmission occurs from July through December . The cohort is an age-stratified random sample of the entire village population . A detailed description of the study site and design of the cohort study has been published elsewhere [31] . The present study focused on children aged 5–13 years who had PBMCs collected at their healthy baseline before the malaria season , and 7 or 14 days after treatment of their first malaria episode of the ensuing malaria season , as well as a subset of children who also had PBMCs collected after the following six-month dry season , a period of little to no P . falciparum transmission . Individual demographic and clinical data are given in Table S1 . Febrile malaria episodes were detected prospectively by self-referral to the study clinic , which was staffed by a physician 24 hours/day . Malaria episodes were treated with a standard 3-day course of artemether/lumefantrine . Thick blood smears were stained with Giemsa and counted against 300 leukocytes , and P . falciparum densities were recorded as the number of asexual parasites/µl of whole blood based on an average leukocyte count of 7500/µl . Each smear was evaluated separately by at least two expert microscopists . P . falciparum was detected by PCR from dried blood spots preserved on 903 Protein Saver filter paper ( Whatman ) as previously described [80] . Blood samples ( 8 ml ) were drawn by venipuncture into sodium citrate-containing cell preparation tubes ( BD , Vacutainer CPT Tubes ) and transported 20 km to the laboratory where PBMCs were isolated and frozen within three hours according to the manufacturer's instructions . PBMCs were frozen in fetal bovine serum ( FBS ) ( Gibco , Grand Island , NY ) containing 7 . 5% dimethyl sulfoxide ( DMSO; Sigma-Aldrich , St . Louis , MO ) , kept at −80°C for 24 hours , and then stored at −196°C in liquid nitrogen . For each individual , PBMCs from all time points were thawed and assayed at the same time . The trypan blue dye exclusion assay consistently demonstrated >80% viability of PBMCs after thawing . RNA extraction , cDNA amplification , synthesis and labeling was performed as previously described [84] . Hybridization , fluidics and scanning were performed according to standard Affymetrix protocols . GeneChip Operating Software GCOS v1 . 4 was used to convert the image files to cell intensity data ( cel files ) . All cel files , representing individual samples were normalized using the Robust Multiarray Average ( RMA ) method from the affy package library in the R project for Statistical Computing ( R Core Team 2013 ) . Nine outlier chips were identified among the unstimulated samples using quality control plots from Partek Genomics Suite software ( Partek , inc . St . Louis , Mo . , v6 . 5 6 . 11 . 310 ) and principal components analyses ( PCA ) computed using R . An empirical Bayes moderated paired T-test was computed using the limma package library in R to obtain false discovery rate ( FDR ) adjusted p-values and fold changes . Probes were considered statistically significant if their FDR-adjusted P values were <0 . 05 and their absolute fold change was >1 . 25 . Heatmaps were generated with the gplots package library in R . Log fold change ratios , p-values and false discovery rates from the empirical Bayes T-tests were imported into Ingenuity Pathways Analysis to examine enrichment of pathways and functional groups . Human yeast Sfi1 homolog spindle assembly associated gene ( Sfi1 ) was selected as a reference gene based on its low coefficient of variation ( CV ) across DNA microarray analysis . Seven mRNAs were analyzed by q-RT-PCR to validate DNA microarray findings: chemokine ( C-X-C motif ) ligand 5 ( CXCL5 ) , interleukin-1 beta ( IL1B ) , interleukin-6 ( IL6 ) , interleukin-10 ( IL10 ) , toll-like receptor 2 ( TLR2 ) , and transforming growth factor , beta 1 ( TGFB1 ) . All six probe and primer sets were designed using Primer Express version 3 . 0 ( ThermoFisher Scientific , Waltham , MA ) and are listed inTable S5 . Seventeen out of 34 patients were selected for q-RT-PCR validation . Four RNAs were analyzed from each patient representing the two time points HB and d7 and two experimental conditions ( ‘ex vivo unstimulated’ and ‘in vitro stimulated with iRBC’ ) . Template preparation and q-RT-PCR analysis was performed as described previously [84] . 3D7 P . falciparum parasites were maintained in fresh human ORh+erythrocytes at 3% hematocrit in RPMI 1640 medium ( KD Medical ) supplemented with 10% heat-inactivated ORh+ human serum ( Interstate Blood Bank , Memphis , Tennessee ) , 7 . 4% Sodium Bicarbonate ( GIBCO , Invitrogen ) and 25 µg/ml of gentamycin ( GIBCO , invitrogen ) , at 37°C in the presence of a gas mixture containing 5% O2 , 5% CO2 and 90% N2 . Parasite cultures were shown to be free of mycoplasma and acholeplasma using an ELISA-based Mycoplasma Detection Kit ( Roche ) which contains polyclonal antibodies specific for M . arginini , M . hyorhinis , A . laidlawii and M . orale . P . falciparum schizont iRBCs were isolated in RPMI 1640 medium supplemented with 0 . 25% Albumax ( GIBCO , Invitrogen ) and 7 . 4% Sodium Bicarbonate ( GIBCO , Invitrogen ) using magnetic columns ( LD MACS Separation Columns , Miltenyi Biotec ) . Control preparations of uninfected red blood cells ( uRBC ) from the same blood donor were obtained and tested in all experiments . Lysates of P . falciparum-infected and uninfected RBCs were obtained by three freeze-thaw cycles in liquid nitrogen and 37°C water bath . PBMCs were cultured in complete RPMI ( RPMI 1640 plus 10% fetal calf serum , 1% penicillin/streptomycin , 2-mercaptoethanol ) in flat-bottom 96 well plates , at 37°C in a 5% CO2 atmosphere . 500 , 000 PBMCs were stimulated with lysate of infected red blood cells ( iRBCs ) or uninfected RBCs ( uRBCs ) in a ratio of 3 RBCs per PBMC for 18 h , with or without 1 . 25 µg/ml Brefeldin A ( BFA ) ( Sigma-Aldrich ) for the last 15 h of stimulation . A 3∶1 ratio of RBC to PBMC was used on the basis of titration experiments ( from 5∶1 to 1∶1 ) and is consistent with previous reports [85] . PBMCs stimulated with 1 . 18 µg/ml Staphylococcal enterotoxin B from Staphylococcus aureus ( SEB ) ( Sigma-Aldrich ) was used as a positive control for cytokine production in supernatants and within cells . Following stimulation , cells were centrifuged and supernatants were recovered and frozen at −80°C for cytokine analysis . Cells stimulated in the presence of BFA were centrifuged , washed and recovered for intracellular staining and flow cytometry analysis . Monocyte/macrophages were isolated from PBMCs of Malian children by negative selection using the MACS Pan Monocyte Cell Negative Isolation kit II ( Miltenyi Biotec ) , an indirect magnetic labeling system for the isolation of untouched monocytes/macrophages . Non-monocyte/macrophage cells were directly depleted by using a cocktail of biotin-conjugated antibodies followed by magnetic removal of labeled cells . Monocyte/macrophage purity was verified by flow cytometry using fluorescently labeled antibodies specific for CD3 PE ( UCHT1 ) , CD4 APC ( RPA-T4 ) , CD8 APC-Cy7 ( SK1 ) , CD14 FITC ( M5E ) , CD16 Pacific blue ( 3G8 ) ( BD Biosciences ) , CD19 PerCP-Cy5 . 5 ( SJ25C1 ) ( eBioscience ) , and 7-Aminoactinomycin D ( 7-AAD ) viability staining ( BD Biosciences ) . FACS analysis was performed on a BD LSR II Table flow cytometer ( BD Bioscience ) and analyzed using FlowJo software ( Tree Star ) . Purified monocytes/macrophages were then stimulated in a ratio of 30 RBCs per monocyte for 6 h with lysate of P . falciparum-infected RBCs and cytokines were measured in supernatants . PBMCs were washed in PBS with 4% heat-inactivated FCS and cells were incubated for 30 min at 4°C with fluorescently labeled antibodies specific for CD3 PE ( UCHT1 ) , CD4 APC ( RPA-T4 ) , CD8 APC-Cy7 ( SK1 ) , CD14 FITC ( M5E ) and CD16 Pacific blue ( 3G8 ) purchased from BD Biosciences; and CD19 PerCP-Cy5 . 5 ( SJ25C1 ) purchased from eBioscience . All phenotypic analyses were performed using mouse mAbs specific for human markers conjugated to fluorophores . FACS analyses were performed on a BD LSR II Table flow cytometer ( BD Biosciences ) and analyzed using FlowJo software ( Tree Star , Inc ) . Supernatants were thawed and immediately analyzed with Bio-plex human cytokine assays ( Bio-Rad Laboratories , Inc . ) as recommended by the manufacturer . The following cytokines were measured: IL-1β , IL-6 , IL-8 , IL-10 and TNF . Briefly , 25 µL of supernatant was diluted 1∶2 in medium and incubated with anti-cytokine antibody-coupled magnetic beads for 30 min at room temperature shaking at 300 RPM in the dark . Between each step the complexes were washed three times in wash buffer , using a vacuum manifold . The beads were then incubated with a biotinylated detector antibody for 30 min before incubation with streptavidin-phycoerythrin for 30 minutes . Finally , the complexes were resuspended in 125 µL of detection buffer and 100 beads were counted with a Luminex 200 device ( Bio-Rad Laboratories , Inc . ) . Final concentrations were calculated from the mean fluorescence intensity and expressed in pg/mL using standard curves with known concentrations of each cytokine . PBMCs were cultured in complete RPMI in flat-bottom 96 well plates , at 37°C in a 5% CO2 atmosphere . 500 , 000 PBMCs were stimulated for 18 h with lysate of infected ( iRBCs ) or uninfected RBCs ( uRBCs ) in a ratio of 3 RBCs per PBMC in the presence of anti-IL-10 ( BD Pharmigen , USA ) and anti-IL-10R ( R&D Systems , Inc . ) or in the presence of the respective isotype controls . Following stimulation cells were centrifuged and supernatants were recovered for cytokine analysis . After stimulation a total of 1×106 PBMCs were sequentially stained for surface and intracellular markers in round-bottom 96-well plates at room temperature . To exclude dead cells , PBMCs were stained for 30 min using the LIVE/DEAD Fixable Violet Dead Cell Stain Kit ( Invitrogen ) followed by a surface staining with PerCP-Cy5 . 5 anti-human CD27 ( M-T271 ) and APC-H7 anti-human CD45RO ( UCHL1 ) for 20 min . After fixing and permeabilizing the cells according to the manufacturer's protocol using the FoxP3 Staining Buffer Set ( eBioscience ) , the cells were stained with BD Horizon V500 anti-human CD3 ( UCHT1 ) , PerCP anti-human CD4 ( SK3 ) , Alexa Fluor 700 anti-human IFNγ ( B27 ) , FITC anti-human TNF ( MAb11 ) , APC anti-human IL-10 ( JES3-19F1 ) , PE-Cy7 anti-human CD25 ( BC96 ) and PE anti-human FoxP3 ( 236A/E7 ) for 30 min . Fluorescently labeled antibodies against TNF , CD25 and FoxP3 were purchased from eBioscience , the remaining antibodies were purchased form BD Biosciences . Cells were acquired using a BD LSR II Table flow cytometer ( BD ) and analyzed using FlowJo software ( Tree Star ) and SPICE software [86] . CD4+ T cells were isolated from PBMCs of Malian children by negative selection using the MACS CD4+ T Cell Negative Isolation kit II ( Miltenyi Biotec ) , an indirect magnetic labeling system for the isolation of untouched CD4+ T helper cells . Non-CD4+ T cells were directly depleted by using a cocktail of biotin-conjugated antibodies against CD8 , CD14 , CD16 , CD19 , CD36 , CD56 , CD123 , TCR g/d and Glycophorin A and anti-Biotin microbeads , followed by magnetic removal of labeled cells . CD4+ T cells purity was verified by flow cytometry , using fluorescently labeled antibodies specific for CD3 ( PE ) and CD4 ( APC ) ( BD Bioscience ) and a BD LSR II Table flow cytometer ( BD Bioscience , USA ) and analyzed using FlowJo software ( Tree Star , Inc ) . Purified CD4+ T cells were then incubated either in the presence or absence of non-CD4+ cells and stimulated for 18 h with lysate of infected ( iRBCs ) or uninfected RBCs ( uRBCs ) , for cytokine analysis of supernatants . PBMCs were cultured in complete RPMI in flat-bottom 96 well plates , at 37°C in a 5% CO2 atmosphere . 500 , 000 PBMCs were stimulated for 18 h with lysate of infected ( iRBCs ) or uninfected RBCs ( uRBCs ) in a ratio of 3 RBCs per PBMC , in the presence of anti-human leukocyte antigen HLA-DQ ( SPVL-3; Beckmann Coulter , USA ) , anti-HLA-DP ( B7/21; abcam , USA ) and anti-HLA-DR ( L243; Biolegend , USA ) , or in the presence of respective isotype controls . Following stimulation , cells were centrifuged and supernatants were recovered for cytokine analysis . Continuous data were compared using the paired or unpaired Student's T-test , paired Wilcoxon rank sum test or permutation tests of mean paired differences as appropriate . Bonferroni adjustments were applied to correct for multiple comparisons when appropriate . A linear mixed model for repeated measures ANOVA with Tukey HSD post hoc tests was also used to compare continuous variables . Pearson correlation coefficients and linear regressions with 95% confidence bands were used to examine the correlation between continuous variables . Fisher's exact test was used for contingency table analyses . The statistical test used is specified in the figure legends . Statistical significance was defined as a 2-tailed P value of ≤ . 05 . Statistical tests were computed using R version 2 . 13 . 2 ( http://www . R-project . org ) , GraphPad Prism version 5 . 0d ( http://www . graphpad . com/scientific-software/prism/ ) or JMP 10 . 0 ( www . jmp . com ) .
|
Malaria remains a major cause of disease and death worldwide . When mosquitoes infect people with malaria parasites for the first time , the parasite rapidly multiplies in the blood and the body responds by producing molecules that cause inflammation and fever , and sometimes the infection progresses to life-threatening disease . However , in regions where people are repeatedly infected with malaria parasites , most infections do not cause fever and parasites often do not multiply uncontrollably . For example , in Mali where this study was conducted , children are infected with malaria parasites ≥100 times/year but only get malaria fever ∼2 times/year and often manage to control parasite numbers in the blood . To understand these observations we collected immune cells from the blood of healthy children before the malaria season and 7 days after malaria fever . We simulated malaria infection at these time points by exposing the immune cells to malaria parasites in a test-tube . We found that re-exposing immune cells to parasites after malaria fever results in reduced expression of molecules that cause fever and enhanced expression of molecules involved in parasite killing . These findings help explain how the immune system prevents fever and controls malaria parasite growth in children who are repeatedly infected with malaria parasites .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"infectious",
"diseases",
"medicine",
"and",
"health",
"sciences",
"protozoans",
"malarial",
"parasites",
"clinical",
"immunology",
"biology",
"and",
"life",
"sciences",
"immunology",
"microbiology",
"protozoology",
"plasmodium",
"falciparum",
"parasitic",
"diseases",
"immunomodulation",
"parasitic",
"protozoans",
"malaria",
"organisms",
"tropical",
"diseases"
] |
2014
|
Exposure-Dependent Control of Malaria-Induced Inflammation in Children
|
Insulators are DNA sequences that control the interactions among genomic regulatory elements and act as chromatin boundaries . A thorough understanding of their location and function is necessary to address the complexities of metazoan gene regulation . We studied by ChIP–chip the genome-wide binding sites of 6 insulator-associated proteins—dCTCF , CP190 , BEAF-32 , Su ( Hw ) , Mod ( mdg4 ) , and GAF—to obtain the first comprehensive map of insulator elements in Drosophila embryos . We identify over 14 , 000 putative insulators , including all classically defined insulators . We find two major classes of insulators defined by dCTCF/CP190/BEAF-32 and Su ( Hw ) , respectively . Distributional analyses of insulators revealed that particular sub-classes of insulator elements are excluded between cis-regulatory elements and their target promoters; divide differentially expressed , alternative , and divergent promoters; act as chromatin boundaries; are associated with chromosomal breakpoints among species; and are embedded within active chromatin domains . Together , these results provide a map demarcating the boundaries of gene regulatory units and a framework for understanding insulator function during the development and evolution of Drosophila .
The spatiotemporal regulation of transcription is controlled by the binding of transcription factors to their target cis-regulatory modules ( CRM ) and their resulting interactions with promoters . Such regulatory interactions between CRMs and promoters can occur over short distances when regulatory sequences are immediately proximal to their target promoter or , in many cases , over longer distances involving many thousands of base pairs . Because of the variability in the distances over which CRMs can act , delineating the molecular boundaries of genes can be challenging . Mechanisms by which a CRM targets the appropriate promoter among a collection of adjacent promoters are poorly defined . However , one such mechanism involves the partitioning of the genome into regulatory domains by genome features known as insulators , or boundary DNA elements . Since their initial characterization twenty years ago [1]–[4] , insulator elements have been thought to create distinct regulatory domains , and thus allow enhancers to find their proper target promoter [5] . Insulators have been identified in Drosophila as well as in vertebrate genomes [6] based on their ability to disrupt the communication between an enhancer and a promoter when inserted between them . This enhancer-blocking activity is dependent upon the binding of insulators by several proteins . The CCCTC-binding Factor ( CTCF ) was first identified in vertebrates [7]; its Drosophila homolog , dCTCF , is known to bind several insulators and is necessary for their function [8]–[11] . CTCF is currently the only vertebrate protein known to be associated with insulator elements . In Drosophila however , several other proteins have been identified for their insulator function . Su ( Hw ) is associated with the gypsy retrotransposon insulator and other endogenous binding sites [4] , [12]–[14] . The insulator activity of gypsy is dependent on the recruitment by Su ( Hw ) of two other proteins: Modifier of mdg4 [Mod ( mdg4 ) ] [15] , [16] and CP190 [17] . Three additional proteins have been linked to insulator function in Drosophila . The binding of Zw5 and BEAF-32 on the scs/scs' elements of the hsp70 locus is required for their enhancer-blocking activity [18] , [19] . Similarly , the ubiquitous transcription factor GAF ( GAGA Associated Factor ) is necessary for the enhancer-blocking activity of particular insulators [20]–[22] . Apart from their enhancer-blocking activity , insulators act as chromatin boundary elements . Such boundaries block the spreading of epigenetic marks or chromatin proteins such as repressive heterochromatin proteins or Polycomb Group-dependent ( PcG ) silencing [23]–[27] . While genetic and molecular studies of insulator function suggest that insulators play a major role in the regulatory organization of the genome , functional data have been collected on only a dozen insulator sequences in Drosophila and mammals . The identification of new insulators in flies and mammals by genome-wide approaches has only recently been initiated in different biological sources [10] , [14] , [27]–[30] . Here we provide a uniformly collected dataset and comprehensive analysis from developing embryos for six different insulator proteins .
We mapped the genome-wide binding sites of 6 insulator-associated proteins: CTCF , CP190 , BEAF-32 , Su ( Hw ) , Mod ( mdg4 ) and GAF by Chromatin ImmunoPrecipitation coupled with microarrays ( ChIP-chip ) in Drosophila embryos ( 0–12 h of development ) . For CTCF and Su ( Hw ) , 2 different antibodies for each factor were used as controls to demonstrate the reproducibility of our experiments . At a 1% False Discovery Rate ( FDR ) , we identified between 2 , 500 and 6 , 600 binding sites for each factor ( Figure 1 and Table S1 ) , which included all functionally verified Drosophila melanogaster insulator sequences ( Figure 1 , Table S2 , and Figure S1 ) . The reproducibility of different ChIP-chip experiments for 2 different antibodies for CTCF and Su ( Hw ) is very high , with 94% of CTCF and 87% of Su ( Hw ) binding sites overlapping ( Figure S2 and Figure S3 ) . Moreover , we were able to recapitulate the profiles for CTCF and Su ( Hw ) generated in Drosophila embryos for the homeotic complexes and 3 Mb of the Adh region [10] , [14] with an overlap of 94% ( 31/33 ) for CTCF and 70% ( 27/41 ) for Su ( Hw ) between the published dataset and our mapping in the same genomic region . To characterize the sequence specificity of each insulator-associated protein , we identified significantly enriched sequence motifs for each set of target sites ( Figure S4 ) . The most enriched motif identified for CTCF strongly resembles the CTCF motif identified in vertebrates [31] and Drosophila [10] . Likewise , the motif for Su ( Hw ) that was discovered in this study is similar to a motif previously identified in Drosophila from a limited number of Su ( Hw ) sites [14] , [32] , [33] . The discovered motifs are present in 75 . 6% of CTCF , 86 . 8% of BEAF-32 , 84% of Su ( Hw ) and 88 . 6% of GAF binding sites ( Table S3 ) . Additionally , the motifs identified for one insulator-factor were often also enriched at the binding sites of other insulator-factors ( Figure 2A ) . This cross-enrichment was not observed however , when only binding sites associated with a single factor were considered ( Figure 2B ) , suggesting that each factor retains unique DNA-level binding specificity but associates with other insulator proteins via clustered binding sites and/or protein-protein interactions . Previous analyses have suggested that , in human cells , insulator binding sites are remarkably conserved across cell types [27] , [31] , [34] . Given the large overlap between binding sites here identified in whole embryos and data previously produced in Drosophila S2 cells for CTCF and CP190 [30] , we investigated this trend further . We performed ChIP-chip experiments for CTCF in S2 and Kc cells . Approximately 74–81% of CTCF binding sites identified independently in each cell type overlap ( specifically , have a midpoint to midpoint distance less than 250 bases ) ( Figure S5 and Figure S6A ) . This observation is consistent with a recently published analysis of CTCF binding sites in S2 and Mbn2 cells [35] , in which , by the same criteria , 77–86% of binding sites overlap . However , given the technical differences in protocols for embryos and suspension cell culture and the loss of information inherent in a comparison of independently thresholded binding site calls , we regard this as a conservative estimate . Qualitative observation of binding profiles suggests that many putatively differential binding sites may result from the threshold applied and normalization issues ( Figure S5 and Figure S6B ) . Indeed , we note that the IP signals at non-overlapping binding sites are , on average , four-fold greater than input background , while overlapping binding sites are six-fold greater . In an attempt to avoid such biases , we used a linear mixed model framework to build a binding site detection model that jointly analyzes the data from multiple cell types ( see Text S1 ) . This model identifies 2 , 784 CTCF binding sites , only 166 of which show significant cell type specificity ( Figure S7 ) . In summary , while most insulator sites identified in this study appear to be conserved across cell types , a small fraction appear to function in a regulated fashion . While the six insulator associated proteins mapped in this study often bind independently , we find clusters of overlapping binding sites far more often than would be expected by chance , indicating insulator-associated proteins often bind jointly to the same sequence . Indeed , 45% of the 14 , 145 binding sites identified in this study are occupied by more than one insulator associated protein . For example , 77% of CTCF binding sites cluster with at least another factor ( Figure 2 and Table S1 ) . Analysis of binding site cluster types revealed several notable trends ( Figure 2C and Figure S3 ) . CP190 is frequently ( 5690 out of 6651 total sites ) found to bind with additional factors , BEAF-32 being its most common partner ( 3329/6651 ) . BEAF-32 , CTCF , and CP190 cluster together ( 1378/8872 ) , as do Mod ( mdg4 ) and Su ( Hw ) ( 1101/5381 ) , while GAF displays a significant lack of clustering with other insulator proteins ( 2973 single sites out of 3905 total sites ) . This binding site clustering and the functional data presented below suggest a previously underappreciated compositional complexity of insulator sequences but also clearly identifies two major classes of insulators: Class I principally representing binding sites for BEAF-32/CP190/CTCF and Class II representing Su ( Hw ) -associated binding sites . The distribution of insulator binding sites relative to different classes of functional genomic elements further supports the existence of several distinct functional classes of insulators . BEAF-32 , CP190 , CTCF , GAF , and Mod ( mdg4 ) are clearly enriched at promoters ( Figure 3A , Figure S8 , and Table S4 ) , while Su ( Hw ) is depleted . BEAF-32 , CP190 , CTCF and Mod ( mdg4 ) binding sites are also strongly enriched within 5′UTRs as well as in intergenic regions ( Figure S8 ) and at transcription end sites ( Figure 3B ) . In contrast , they are largely excluded from transposable elements and coding exons ( Figure S8 and Figure S9 ) , suggesting a role of Class I insulator proteins , but not Class II , in regulating the transcription of genes . We reasoned that if insulators act as gene boundaries , they should partition genes into distinct regulatory environments . Indeed , we find that four of the six insulator-associated proteins binding sites are significantly enriched between adjacent consecutive promoters ( Figure 3C ) with a stronger enrichment of BEAF-32 , CP190 , CTCF and Mod ( mdg4 ) between adjacent divergently oriented promoters ( Figure 3D ) . Additionally , as suggested previously in vertebrates for CTCF [31] , Class I and Class II insulator proteins are significantly enriched between alternative promoters , providing a potential mechanism for their independent regulation ( Figure 3E ) . The distribution of insulators relative to a variety of genomic functional element classes suggests a pervasive role in controlling gene regulatory environments . To further address this hypothesis we mapped active promoters in embryos of the same developmental stage that we used for insulator mapping . To identify active promoters , we performed ChIP-chip with antibodies directed against the trimethylated lysine 4 of Histone H3 ( H3K4me3 ) , which is a clear mark of activation [36]–[39] , and against the largest subunit of the RNA Polymerase II ( PolII ) . We combined these two mappings with hybridization on tiling arrays of total RNA extracted from the same material . In Drosophila embryos , H3K4me3 is associated with gene Transcription Start Sites ( TSS ) and colocalizes with PolII immediately downstream of the TSS of active genes ( Figure S10A and S10B ) . We extracted from this dataset a set of high confidence actively transcribed promoters , which overlap with H3K4me3 and PolII signals and whose exons overlap significant RNA signal ( Figure S10C ) . We hypothesized that if insulators do indeed demarcate regulatory units , insulators would separate promoters with differing expression status . We repeated the positional analysis of insulator proteins between divergent , adjacent , and alternative promoters while taking into account the transcriptional status of the promoters ( Figure 4A–4C ) . We observed that the enrichment of BEAF-32 , CP190 , CTCF , GAF , and Mod ( mdg4 ) is greater between promoter pairs when they are differentially expressed ( Figure 4 , Figure S11 , and Figure S12 ) . It is possible however that this result comes from an averaging of promoter activity across all the cell types present in the embryo at this developmental stage . We then repeated H3K4me3 ChIP-chip as a marker of active promoters in 2 embryonic Drosophila cell types: S2 and Kc cells . The overlap of H3K4me3 between embryos and Kc and S2 cell lines is between 71 and 75% respectively , while it is 85% between S2 and Kc cells ( Figure S13A ) . Using H3K4me3 binding sites as a guide , we identified active promoters in each cell type . As in whole embryos , genes flanking CTCF binding sites identified in S2 and Kc cells show a significant enrichment of differentially expressed divergent and alternative promoters ( Figure S13B and S13C ) further demonstrating that Class I insulators delimit the boundaries of gene regulatory units . Consistent with the limited previous functional data demonstrating the enhancer-blocking activity of insulators , we find binding sites for BEAF-32 , CP190 , and Su ( Hw ) are significantly depleted between annotated CRMs and their target promoters across the entire genome ( Figure 4D , Figure S14 ) , while CP190 , CTCF , GAF , and Mod ( mdg4 ) are enriched between cis-regulatory elements and their nearest non-target promoter , distributions that strongly support their proposed enhancer blocking function . Interestingly , we note that binding sites for GAF are significantly enriched between CRMs and their target promoters . Similarly , we find that BEAF-32 , CP190 , and Su ( Hw ) binding sites are depleted between distinct CRMs of the same gene , while GAF is found more frequently than expected ( Figure S14 ) . We note that the enrichment of insulators within such genomic features may , in part , be driven by the effects of differential promoter density or biases in chromatin accessibility . In order to understand how such factors could affect any interpretation of our data , we reanalyzed binding site data for 36 recently published datasets corresponding to 21 transcription factors , from the Berkeley Drosophila Transcription Network Project ( BDTNP ) [40] . We first observed that none of our insulator binding sites preferentially localize with this transcription factor set ( Figure S15 ) . Despite several transcription factors that preferentially bind promoter-proximal sequences ( Figure S16 ) , the enrichment of insulators between promoter pairs is greater than for any of the published transcription factors ( Figure S17A , S17B , S17C ) . In contrast to these findings , and as expected , the published BDTNP transcription factors are not as strongly biased towards CRM , non-target promoter separation ( Figure S17D ) . Previous studies have demonstrated that insulators delimit distinct organizational domains of a genome [27] , [30] . One such chromatin domain is marked by the trimethylated Lysine 27 of Histone H3 ( H3K27me3 ) , a histone modification deposited and recognized by the repressive Polycomb protein complexes [41] . We mapped by ChIP-chip the H3K27me3 mark in Drosophila embryos . We observed in whole embryos , as described previously [42] , [43] , that H3K27me3 is distributed throughout the genome in large domains ( Figure S18 ) . To better define the boundaries of these large genomic regions , we used a hidden Markov model based segmentation algorithm . We confirm that the genes affected by this silencing mark correspond to the previously described Polycomb target genes [42]–[45] . We identified 140 regions of substantial H3K27me3 density and quantified the distribution of each insulator binding site type with respect to the domain boundaries . Interestingly we find that all 6 factors are significantly depleted within and enriched outside these regions ( Figure 5A ) . In addition , CTCF , GAF , and Mod ( mdg4 ) are enriched at the boundaries of regions of high H3K27me3 density , with this enrichment significantly decreasing at increasing distances , further supporting the insulators' role in chromatin domain boundary determination ( Figure 5A and Figure S18 ) . It is possible that this result is confounded by the fact that insulators are enriched at TSSs . We performed Pearson's chi-squared contingency table tests to assess if the frequency of insulator-H3K27me3 boundary overlaps are independent of ( and greater than ) the frequency of TSS- H3K27me3 boundary overlaps . Indeed , CP190 ( p<9 . 8e-6 ) , BEAF-32 ( p<1 . 8e-5 ) , CTCF ( p<0 . 00013 ) , GAF ( p<0 . 0022 ) , Mod ( mdg4 ) ( p<0 . 00035 ) , and Su ( Hw ) ( p<0 . 0088 ) are independently associated with H3K27me3 breakpoints . Given their apparently pervasive role in the establishment of gene regulatory units , we examined the role insulator sequences have played in shaping the evolution of the Drosophila genome . First , insulators show evidence of local sequence constraint . Based on either 15-way insect multiple sequence alignments or pair-wise alignments between the closely related Drosophila melanogaster and Drosophila simulans , insulator binding sites evolve significantly slower than fast evolving introns , although more swiftly than either coding exons or most transcription factor binding sites [46] ( Figure S19 ) . Second , we find that BEAF-32 , CP190 , CTCF , and Mod ( mdg4 ) are significantly enriched near the 12 Drosophila species syntenic breakpoints ( Figure 5B ) [47] . Chi-squared tests demonstrate that for CP190 ( p<0 . 0031 ) , BEAF-32 ( p<0 . 0086 ) , GAF ( p<0 . 027 ) , and Mod ( mdg4 ) ( p<0 . 034 ) , this result is independent of the association of TSSs and syntenic breaks . This finding provides evidence to support the hypothesis [48] that selective pressure has maintained gene regulatory units established by flanking insulators . We find that binding sites for 5 of the 6 insulator-associated proteins ( Su ( Hw ) is the exception ) are regions of reduced nucleosome density relative to surrounding regions ( Figure 6A ) . Reduced nucleosome density often corresponds to sites of high histone replacement or displacement [49] , [50] and classical “active” chromatin as defined by salt solubility properties [51] . We also find that the same 5 of the 6 insulator proteins are preferentially bound in regions characterized by low-salt soluble nucleosomes ( Figure 6B and 6C ) , depleted in the remaining high-salt-soluble fraction ( Figure 6D ) and highly enriched in the salt-washed insoluble pellet ( Figure 6E ) . Similar analyses of only non-promoter proximal insulators reveal the same trends , indicating that the shared solubility properties of insulators and promoters are indeed independent ( Figure S20 ) . Given the correspondence between these results and the regulatory boundary analyses presented above , we hypothesize that this difference in chromatin properties may explain why Su ( Hw ) , defining ClassII insulators , does not act as a gene boundary in the genome .
Insulator identification has been the source of much recent interest . Indeed , in the last 6 months CTCF was mapped in S2 cells [52]; BEAF-32 in embryos ( 6–16 h of development ) [53] , CTCF and CP190 in S2 cells [54] and more recently CTCF , Su ( Hw ) , CP190 and BEAF-32 in Kc cells and Mbn2 cells [35] . Interestingly , the latter paper describes three subclasses of insulators , with CP190/BEAF association being distinct from CP190/CTCF and CP190/Su ( Hw ) . We present in this study the embryonic binding profile of six factors previously known to be associated with insulator function in Drosophila . Our analysis of insulator binding site distributions and protein composition suggest there exist 2 principal categories of insulator elements ( Class I and Class II ) . In particular , we have shown that Class I insulators , identified by the binding of CTCF , CP190 or BEAF-32 , segregate differentially expressed genes and delimit the boundaries of chromatin silencing , while they are depleted between known CRMs and their target genes . We do not find evidence supporting a significant distinction between CP190/BEAF and CP190/CTCF or CTCF/BEAF . In contrast , our analyses suggest that BEAF-32 , CP190 , and CTCF are distributed and function quite similarly , while Su ( Hw ) appears distinct . The Class II insulators , bound by Su ( Hw ) , are often exceptional in our analyses . We note that the analysis of genome-wide mapping data , expression data , and genome annotation provides an endogenous boundary assay that demonstrates that , while Su ( Hw ) has been described as an insulator before , it is not systematically associated with the boundaries of the gene units . By helping to delimit the regulatory boundaries of genes , the Class I insulator map presented here will aid in the identification of transcription factor target genes and the construction of transcriptional regulatory networks . As an example of this concept , we illustrate the distribution of known regulatory elements and insulators across the Antennapedia Complex ( ANT-C ) of homeotic genes ( Figure 7 ) . This region quite strikingly demonstrates the potential utility of insulator binding data for cis-regulatory annotation . Across approximately 500 kb , cis-regulatory elements and their target promoters are found between insulator pairs . For example , a single insulator separates the lab and Edg84A genes , with their respective cis-regulatory elements narrowly partitioned on either side . The adjacent regulatory elements and promoters of zen and bcd are similarly insulator segregated . The presence of an insulator 3′ of ftz was previously hypothesized [55] to explain the ability of distal Scr regulatory elements to bypass ftz by pairing with the proximal SF1 insulator , located between Scr and ftz . Lastly , at Antp , as we observe genome wide , two alternative promoters and their proximal regulatory elements are segregated by a single insulator . We are currently developing analysis methods to systematically partition the entire genome into such regulatory domains . Consistent with their observed regulatory boundary functions , Class I insulators are embedded within local regions of active chromatin and are frequently associated with syntenic breakpoints between species . Previous work has demonstrated that active promoters in yeast and Drosophila are associated with reduced nucleosome occupancy and low-salt soluble and high-salt insoluble chromatin [50] , [56] ( Figure S20 ) . Therefore , surprisingly , dynamic chromatin is a shared feature between promoters and most classes of insulators . It is notable however that some studies have revealed functional similarities between insulators and promoters in transgenic assays [57] . These results have been described as paradoxical , as insulators can negatively affect promoters by blocking communication between enhancers and promoters . One proposed model for insulator function is that they act as promoter “decoys” by recruiting away factors necessary for transcriptional initiation [57] . Alternatively , insulators and promoters might require common chromatin features to function by mechanisms that are still unknown . One potential interpretation is that the dynamic chromatin at insulators forms a flexible chromatin joint that would affect the probability of productive contact between separated regulatory elements . In this way , the similarity between promoters and insulators would be a consequence of their common requirement for dynamic chromatin , although with very different consequences . This model may explain why promoters are so frequently scored as insulators in the classical insulator assay , when an element is placed between an enhancer and a promoter [1] , [58] .
Chromatin immunoprecipitations have been performed as described previously [59] . Briefly , the biological material is homogenized in the presence of 1 . 8% formaldehyde . The cross-linked chromatin is sonicated using a Bioruptor ( Diagenode ) to an average size of 500 bp . Pre-cleared chromatin extract is incubated overnight at 4C with the specific antibody and immunoprecipitated with protein-A Sepharose beads . After purification of the DNA and amplification of the libraries by linker-mediated PCR , the samples are labeled according to Affymetrix protocols and hybridized in parallel with an input sample onto the Affymetrix Drosophila Tiling Array , v2 . 0 R . CTCF-C and CTCF-N antibodies are described in [8] , CP190 antibody is described in [60] , BEAF-32 antibody is described in [18] , Mod ( mdg4 ) antibody is directed against the 67 . 2 isoform and is described in [61] , Su ( Hw ) -1 antibody is described in [62] , Su ( Hw ) -2 is described in [63] , GAF antibody is described in [64] , H3K27me3 antibody is from Upstate ( 07-449 lot DAM1387952 ) , H3K4me3 antibodies is from Abcam ( ab8580 lot 411277 ) and PolII antibody is from Covance ( 8wG16 lot 14861301 ) . Insulator binding data was processed with Model based Analysis Tiling-arrays ( MAT ) software [65] . We ran paired MAT analysis with MaxGap of 500 , MinProbe of 10 , and a Bandwidth of 250 . H3K4me3 , PolII and RNA data were analyzed with TAS ( Tiling Array Software ) and a threshold of 5% of the highest pValues was applied to identify the high intensity signals . The same parameters as for the MAT analysis have been applied to then call the peaks with TAS . We developed a new HMM-based segmentation algorithm to identify H3K27me3 domains , as well as a novel mixed model framework for the joint analysis of ChIP-chip data from more complicated experimental designs , here applied to CTCF binding data from multiple cell types ( see details in Text S1 ) . Motif discovery was performed separately for each insulator . Peak centers that were at least 1 kb away from the peak center of any other insulator were taken ( “uniquely bound peaks” ) and +/− 100 bp windows were generated excluding coding exons , repeats , transposons , 3′ untranslated regions and non-coding RNAs ( “excluded regions” ) . For each insulator up to 500 of the regions were randomly selected and enriched motifs were identified using MEME [66] , AlignACE [67] , and MDscan [68] . All programs were run with default parameters except for MEME , which was restricted to a maximum of 3 iterations and a maximum motif width of 25 . Instances of each of the motifs at conservation levels from 0 . 0 to 1 . 0 confidence ( in steps of 0 . 1 ) were identified in all Intergenic regions ( defined as genomic regions excluding those noted above ) using the motif instance pipeline described in [69] with a PWM threshold corresponding to a p-value of 4−8 as determined by TFM-Pvalue [70] . The motifs were ranked using the fraction of instances found in the uniquely bound regions divided by the fraction for instances of shuffled control motifs at the same conservation cutoff ( Wilson's confidence interval at Z = 1 . 5 was used on the ratios to give a conservative enrichment ) . This procedure is designed to reduce biases due to composition or conservation level . The motif with the highest enrichment at any confidence level was selected . This procedure was repeated using the MAT peak regions ( rather than +/− 100 bp ) to produce the comparison in Figure S4 ( otherwise the +/− 100 bp motifs are used throughout ) . Genomic distribution analyses only used insulators mapped to chromosomes 2L , 2R , 3L , 3R , 4 , and X . All gene annotations , including transcription start site locations and alternative promoter presence were defined according to RefSeq annotations . Transposable element locations were based on Flybase annotations . Divergently transcribed genes were identified as all adjacent transcription start sites , on opposite strands , between 500 and 2500 bases apart . Alternative promoters were identified as all RefSeq annotated genes with more than one distinct transcription start site . The ‘all adjacent’ gene set included all adjacent gene pairs whose transcription start sites were between 1500 and 20000 bases apart , regardless of strand . Cis-regulatory elements and their target genes were defined according to the RedFly database [71] . Breakpoints of regions of conserved synteny across the 12 sequenced Drosophilids were identified in [47] . All genomic distributional analyses were first conducted by mapping protein binding sites relative to the genomic feature of interest . This mapping was performed in one of two ways; First , for genomic features that can be faithfully represented as a single base ( e . g . , a transcription start site ) , the distance from each insulator to its nearest feature was tabulated , second , for paired genomic features ( e . g . , divergent promoters ) , the number of intervening insulators for each feature pair was tabulated . To quantify if the distribution of mapped insulators relative to the genomic feature of interest is significantly different than would be expected by chance ( given the number of insulators and the distribution of the particular feature of interest ) , we performed simulations as follows . First , permuted insulator binding sites were generated by sampling n sites from a random , uniform distribution , the length of each chromosome , where n is the number of observed insulator binding sites , by chromosome . In other words , a simulated insulator is equally likely to be placed at any location across a chromosome . Second , the simulated binding sites were mapped relative to the genomic feature of interest , as with each real dataset . This procedure was repeated 10 , 000 times for each insulator , target element combination . The median simulated values were used to normalize the real data counts to produce enrichment estimates . The 2 . 5 and 97 . 5 percentiles of the simulated distributions were used to produce confidence intervals for display purposes and significance estimates . Empirical p-values were calculated as the fraction of simulations that produced a number of mapped features as extreme as observed in the real data . The position of binding sites have been compared to data of nucleosome density and salt fractionation of the chromatin extraction as described in [51] . Binding sites are defined by their midpoint and nucleosome density and salt fractionation data from S2 cells are plotted as a log ratio of enrichment in a 3 kb interval around the midpoint of the binding site . GSE16245
|
The spatiotemporal specificity of gene expression is controlled by interactions among regulatory proteins , cis-regulatory elements , chromatin modifications , and genes . These interactions can occur over large distances , and the mechanisms by which they are controlled are poorly understood . Insulators are DNA sequences that can both block the interaction between regulatory elements and genes , as well as block the spread of regions of modified chromatin . To date , relatively few insulators have been identified in developing Drosophila embryos . We here present the genome wide identification of over 14 , 000 binding sites for 6 insulator-associated proteins . We demonstrate the existence of two broad classes of insulators . Insulators of both classes are enriched at the boundaries of a particular chromatin modification . However , only insulators bound by BEAF-32 , CP190 , and dCTCF are enriched in regions of open chromatin or demarcate gene boundaries , with a particular enrichment between differentially expressed promoters . Furthermore , insulators of this class are enriched at points of chromosomal rearrangement among the 12 species of sequenced Drosophila , suggesting that insulator defined regulatory boundaries are evolutionarily conserved .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"molecular",
"biology/histone",
"modification",
"genetics",
"and",
"genomics/gene",
"expression",
"genetics",
"and",
"genomics/functional",
"genomics",
"genetics",
"and",
"genomics/epigenetics",
"genetics",
"and",
"genomics"
] |
2010
|
A Comprehensive Map of Insulator Elements for the Drosophila Genome
|
Mammalian Argonaute proteins ( EIF2C1−4 ) play an essential role in RNA-induced silencing . Here , we show that the loss of eIF2C2 ( Argonaute2 or Ago2 ) results in gastrulation arrest , ectopic expression of Brachyury ( T ) , and mesoderm expansion . We identify a genetic interaction between Ago2 and T , as Ago2 haploinsufficiency partially rescues the classic T/+ short-tail phenotype . Finally , we demonstrate that the ectopic T expression and concomitant mesoderm expansion result from disrupted fibroblast growth factor signaling , likely due to aberrant expression of Eomesodermin . Together , these data indicate that a factor best known as a key component of the RNA-induced silencing complex is required for proper fibroblast growth factor signaling during gastrulation , suggesting a possible micro-RNA function in the formation of a mammalian germ layer .
Argonaute proteins comprise a highly conserved gene family necessary for a range of physiological and developmental processes . These proteins are defined by the presence of PAZ and PIWI domains , which modulate protein−protein interactions , nucleic acid binding , and , in some cases , mRNA cleavage [1–5] . Argonaute proteins serve as scaffolds for target-mRNA recognition by short regulatory guide RNAs during the process of RNA interference ( RNAi ) [6] . The Argonaute family was initially linked to RNAi-related phenomena through genetic studies in Caenorhabditis elegans [7] and has since been shown to play a gene-silencing role in plants , yeast , and flies [8–10] . Members of the mammalian Argonaute family associate with micro-RNAs in the RNA-induced silencing complex ( RISC ) , indicating a post-transcriptional gene regulation role in mammals [6] . In the mouse , loss of a single Argonaute family member , eIF2C2 ( Argonaute2 or Ago2 ) , disrupts RISC activity and gives rise to several midgestational developmental abnormalities , including failed neural tube closure , mispatterning of anterior structures , and cardiac malformations [11] . These studies demonstrated that AGO2 has a unique function distinct from its paralogs in the RISC , which indicates the absence of full paralog redundance . However , the specific role played by AGO2 during mammalian development remains unclear . To characterize this role , we investigated Ago2-null embryos during gastrulation and found that Ago2 is required for proper fibroblast growth factor ( FGF ) signaling and mesoderm formation . We further determine that Ago2 haploinsufficiency partially rescues the classic T/+ short-tail phenotype [12] , which is consistent with Ago2 residing in a previously mapped interval shown to modify T [13] . Together , these data reveal a genetic interaction between Ago2 and T and indicate that AGO2 is essential to the formation of a mammalian germ layer .
We explored the role of AGO2 in early mammalian development using gene-trapped embryonic stem cells to generate a mouse line that transmits an interrupted Ago2 allele without an obvious heterozygous phenotype . The interrupted Ago2 allele was characterized , and primers were designed to distinguish wild-type from mutants by genotype ( Figure 1A and 1B ) . This disruption deletes most of the PIWI domain and results in an apparent functional null allele [11] ( Figure 1A and 1C ) . Full-term litters from heterozygous intercrosses did not yield homozygous ( Ago2–/– ) offspring . At embryonic day 9 . 5 ( e9 . 5 ) , we observed two classes of null embryo: intact embryos with assorted morphological phenotypes , such as the neural tube and cardiac malformations that are consistent with the earlier findings of Liu and colleagues ( [11]; unpublished data ) , and embryonic remnants ( Figure 1D ) . Unexpectedly , however , intact e9 . 5 null embryos were observed in numbers significantly lower than predicted based on genetic ratios ( 12/134; p < 0 . 0001; Table 1 ) . Because intact null embryos were recovered in the appropriate genetic ratios during gastrulation ( i . e . , at e7 . 5; Table 1 ) , Ago2 plays an important role at an earlier stage of development than previously reported [11] . Vertebrate gastrulation initiates at e6 . 5 and establishes the three germ layers of the developing embryo ( reviewed in [14] ) . During gastrulation , embryonic ectoderm ( epiblast ) cells are recruited to a transient embryonic structure known as the primitive streak , located on the posterior side of the embryo . At the primitive streak , the epiblast cells undergo an epithelial-to-mesenchymal transition ( EMT ) , before migrating away from the streak and being specified as either the mesoderm or the definitive endoderm germ layers [15 , 16] . By e7 . 5 , a complete mesoderm layer is formed . Brachyury ( T ) , a T-box transcription factor , is expressed in the primitive streak and in the epiblast cells near the primitive streak [17 , 18] . To determine whether a proper primitive streak is formed in the Ago2 mutants , we examined the expression of T in Ago2 null embryos by whole-mount in situ hybridization . We found that homozygous disruption of Ago2 results in expanded expression of T compared to its expression in wild-type e7 . 5 embryos , indicating an abnormal primitive streak in Ago2 mutants ( Figure 2A and 2B and insets ) . Notably , the Ago2 mutants exhibit a variability in the expansion of T expression ( Figure S1B and S1C ) , which may account for the ability of some Ago2 mutants to escape gastrulation arrest and develop until midgestation [11] . Also consistent with previous studies is the reduced extraembryonic region in the e7 . 5 Ago2 mutant embryos; this finding further suggests embryos that survive to later stages have generalized nutritional deficiencies caused by yolk sac and placental defects [11] . Previous experiments have shown that ectopic expression of T is sufficient to induce mesoderm formation [19] , leading us to hypothesize that Ago2 plays a role in mesoderm development . To explore this possibility , we assessed the expression pattern of another known mesoderm marker , Tbx6 [20] , and found that homozygous disruption of Ago2 also results in an expansion of Tbx6 expression compared with its expression in wild-type e7 . 5 embryos ( Figure S2A and S2B ) . These findings , paired with the expanded T expression , argue for an Ago2 function in mesoderm development . To determine the spatial localization of Ago2 during gastrulation , we examined its wild-type expression pattern in sectioned heterozygous Ago2 e7 . 5 embryos by using antibodies against β-galactosidase ( from the gene trap's lacZ insertion driven by the endogenous Ago2 promoter ) and BRACHYURY . We found that wild-type Ago2 expression is restricted to the apical side of the epithelial cell layer and does not overlap with T in the mesenchymal cells of the primitive streak ( Figure 2C , 2E , and 2G ) . Coupled with the fact that homozygous loss of Ago2 results in expanded T expression into the epithelial cell layer ( Figure 2F and 2H ) , these data suggest that Ago2 could play a role in defining the primitive streak . The attenuation of Ago2 expression as cells enter the primitive streak also raises the possibility that AGO2 plays a role in EMT . Indeed , failure to undergo proper EMT is a phenotype observed in embryos with defects in mesoderm development [21] . By contrast , because T is expressed throughout the epiblast of the Ago2 mutants ( Figure 2B [inset] , 2F , and 2H ) , these mutants likely exhibit aberrant EMT because an excess of epithelial cells are being fated to become mesoderm , which ultimately could result in expanded mesoderm at the expense of the epithelial cell layer . Among the mesoderm cell types induced by T expression are the axial and paraxial mesoderms , both of which derive the skeletal tissues that contribute to tail development in vertebrates ( reviewed in [22] ) . In fact , the level of T expression correlates directly with tail length , as evidenced by the short-tail phenotype long recognized in heterozygous T ( T/+ ) mice [12] . Remarkably , previous mapping of T modifier loci defined a small interval on chromosome 15 that includes the Ago2 locus [13] . In order to genetically test whether Ago2 could be the gene responsible for modifying the tail length in T/+ mice , we crossed mice heterozygous for the T deletion with mice heterozygous for the Ago2 disruption ( Ago2+/– ) . We plotted the ratio of tail length to body length for a quantitative comparison of heterozygous mice with double heterozygotes ( Figure 3A ) . While the average tail-to-body ratio in both wild-type and Ago2+/– mice is approximately 0 . 85 , the average ratio in T/+ mice is 0 . 35 ( Figure 3B ) . By contrast , the average tail-to-body ratio in double heterozygote mice is 0 . 58; the double heterozygotes have significantly longer tails than the T/+ mice ( p < 0 . 01 ) . Thus , haploinsufficiency of Ago2 results in a partial rescue of the short-tail T/+ phenotype , demonstrating that Ago2 is a genetic modifier of T expression . As an initial investigation to determine whether Ago2 is one of the previously mapped modifiers of T expression [13] , we searched the entire Ago2 genomic locus ( approximately 80 kb ) for single nucleotide polymorphisms ( SNPs ) [23] and analyzed Ago2 expression between the previously reported background strains . Remarkably , we found only one intronic SNP and that the Ago2 expression levels are indistinguishable between the strains ( unpublished data ) . While this might be interpreted to rule out Ago2 as one of the previously mapped modifiers , this is a gross analysis of Ago2 expression in whole embryos and at only a single stage of development . Indeed , our genetic data clearly show that Ago2 is a modifier of T expression . These studies reveal a genetic interaction between Ago2 and T and demonstrate that AGO2 mediates mesoderm development . The loss of AGO2 is known to disrupt RISC activity [11] , suggesting AGO2 influences T expression via the micro-RNA pathway . Because the homozygous loss of Ago2 results in expanded T expression into the epithelial cell layer ( Figure 2F and 2H ) , AGO2 may utilize its “slicer” activity within the micro-RNA pathway [11] to cleave and degrade T transcripts expressed in the epithelial cell layer . However , in Dicer–/– mutants , RISC activity is disrupted upstream of Ago2 , and these mice do not express T at all [24] , indicating that either AGO2 is more restricted than DICER for RISC activity or the other Argonaute protein family members might retain a low level of functional redundancy to partially compensate for the loss of AGO2 . Alternatively , AGO2 might regulate upstream inducers of T , such as Bmp4 , Eomesodermin , Fgfr1 , or Wnt3a [25–28] . Studies conducted in Xenopus laevis have demonstrated that both transforming growth factor α and FGF signaling are required to initiate T expression as gastrulation commences [18 , 29 , 30] . In mice , mutational analysis of the known FGF genes established that only Fgf4 and Fgf8 are required during gastrulation [31 , 32] . Fgf4 and Fgf8 are coexpressed throughout the primitive streak in an opposing gradient , with Fgf8 expression highest at the posterior end of the streak and barely detectable at the anterior end . Subsequent genetic studies determined that FGF receptor 1 ( Fgfr1 ) is required for the initiation of T expression in the posterior end of the primitive streak , suggesting that Fgf8 is the likely ligand in this region [33] . We examined the expression of Fgf8 in Ago2 null embryos by whole-mount in situ hybridization and found that homozygous disruption of Ago2 results in expanded expression of Fgf8 compared to its expression in wild-type e7 . 5 embryos ( Figure 4A and 4B ) , reminiscent of the expanded T expression pattern ( Figure 2A and 2B ) . These data suggest abnormal FGF signaling causes the expanded T expression in Ago2–/– embryos . In the mouse , direct upstream inducers of Fgf8 are not precisely characterized , but the homozygous loss of either Bmp4 or Eomesodermin ( Eomes ) results in failure to express both Fgf8 and T [27 , 28] . We therefore examined the expression of Bmp4 and Eomes in Ago2-null embryos by whole-mount in situ hybridization and found that homozygous disruption of Ago2 results in expanded expression of Eomes compared to its expression in wild-type e7 . 5 embryos ( Figure 4C and 4D ) , which is consistent with previous data suggesting that Eomes and Fgf8 function similarly during gastrulation [28 , 34] . By contrast , despite the morphological differences , the localization of Bmp4 expression is indistinguishable between Ago2 mutants and their wild-type littermates , in that Bmp4 expression in Ago2 mutants remains restricted to the extraembryonic ectoderm and the proximal embryonic tissue ( Figure 4E and 4F ) . Taken together , these data suggest that Eomes is an upstream inducer of Fgf8 and that Bmp4 is either upstream of Eomes or in a parallel pathway to induce Fgf8 and T gene expression . Finally , as with T , the expansion of both Fgf8 and Eomes expression in the Ago2 mutants is varied , which again suggests a plausible explanation for those Ago2 mutants that escape gastrulation arrest and develop until midgestation ( [11]; unpublished data ) . The induction of T expression has been studied extensively in the 15 years since the gene was cloned . These studies attribute the restricted initiation of T expression to morphogenic movements and cell signaling cascades by showing that disruption of these processes ultimately results in aberrant T expression and mesoderm development [27 , 28] . Coupled with earlier work in X . laevis demonstrating that Bmp4 induces Eomes transcription [35] , our data suggest a T induction working model in which Bmp4 is also an upstream inducer of Eomes in mouse ( Figure 5 ) . At the commencement of gastrulation in wild-type embryos , Ago2 may regulate the proper level of Eomes gene expression , which ultimately induces the downstream expression of Fgf8 and T . In the absence of Ago2 , Eomes may not be regulated properly , leading to its overexpression and a resultant downstream overinduction of Fgf8 and T . Alternatively , Ago2 may regulate an as-yet-unknown upstream inducer of Eomes , or Ago2 may simultaneously have a direct influence on Fgf8 and T gene expression . Because AGO2 is best known to associate with micro-RNA , it might be notable that we find computational algorithms have predicted micro-RNA binding sites in Eomesodermin , Fgf8 , and T ( http://microrna . sanger . ac . uk/targets/v3/ ) , suggesting the modifying influence of Ago2 is mediated by the micro-RNA pathway , although experimental validation of these micro-RNA binding sites awaits further study . In this case , AGO2 may utilize its “slicer” activity within the micro-RNA pathway [11] to cleave and degrade Eomesodermin , Fgf8 , and/or T transcripts expressed outside the primitive streak . Distinguishing among these models will require further analysis of Ago2-null mice that are also null for potential upstream inducers of T . These possibilities notwithstanding , our findings demonstrate that AGO2 is a key factor both in the regulation of T expression and in mesoderm formation , placing a known component of the RNAi machinery in mammalian germ layer development .
Genomic DNA from tail or ear tissue was isolated according to standard procedures . Embryonic and full-term litters were genotyped for the Ago2 disruption via a standard PCR procedure and the following primers: ( a ) 5′-CAGTGCGTCCAGATGAAGAACG-3′; ( b ) 5′-CCCAGGAAGATGACAGGTTG-3′; and ( c ) 5′-GTTTTCCCAGTCACGACGTTG-3′ . The heterozygous T mice ( B10;TFLe-a/a T tf/+ tf/J ) were purchased from The Jackson Laboratory . The Ago2+/– mice are on a congenic C57Bl/6 background , as all the mice used have been backcrossed at least ten generations onto a C57Bl/6 background . While it has been demonstrated that the background strain can affect the heterozygous T tail phenotype , this phenotype is not affected in strains on C57 backgrounds ( e . g . , C57Bl/6 and C57Bl/10; [13] ) . Heterozygote crosses ( T+/– × Ago2+/– ) were set up , and the offspring ( on a mix of C57Bl/6 and C57Bl/10 backgrounds ) were aged 6 to 8 wk to allow for the completion of tail development . At this time , ear tissue was taken , to provide a DNA source , and each animal was subjected to a measurement of tail length as a fraction of body length . Offspring were genotyped for the T deletion using SYBR Green in a standard quantitative PCR procedure and the following primers: ( d ) 5′-CCGGTGCTGAAGGTAAATGT-3′ and ( e ) 5′-CCTCCATTGAGCTTGTTGGT-3′ . The resultant PCR products were quantified using the iQ5 software package and normalized against a known biallelic locus . Embryos were first dissected free from the yolk sac , which was reserved for DNA extraction , then individually boiled in 30 μl of 2× Laemmli buffer before undergoing SDS 10%−PAGE . After transfer to nitrocellulose membrane , the membranes were blocked with 1% milk in PBS−0 . 1% Tween 20 ( Blotto ) and incubated with antibodies against AGO2 ( Abnova ) and EIF4E ( BD Biosciences ) for 1 h at room temperature in Blotto . Membranes were washed in Blotto and incubated with horseradish peroxidase−conjugated anti-mouse antibodies ( Sigma ) for 1 h at room temperature in Blotto . Membranes were washed three times in Blotto and visualized by chemiluminescence in accordance with the manufacturer's ( New England Nuclear ) protocol . Immediately following dissection , embryos were fixed overnight in 4% paraformaldehyde ( Electron Microscopy Sciences ) at 4 °C . Fixed embryos were washed three times in PBS , dehydrated through a methanol series ( 25% , 50% , 75% , 2× 100% ) , and stored at −20 °C . In situ hybridizations were performed on whole-mount embryos , as described [36 , 37] . Antisense riboprobes were synthesized from Brachyury , Fgf8 , Eomesodermin , and Bmp4 cDNA-containing plasmids using a digoxigenin-UTP labeling kit ( Roche ) . Digoxigenin-labeled compounds were detected using alkaline phosphatase−conjugated antidigoxigenin ( Roche ) . Whole-embryo images were captured using a dissection scope ( Zeiss Stemi ) with attached camera ( Zeiss AxioCam MRc ) . Following in situ hybridization , embryos were paraffin embedded using a standard protocol . Then 10-μm sections were dried to positively charged slides ( Surgipath ) . Dried sections were deparaffinized and hydrated by standard procedures . Sections were imaged using a Zeiss Axioskop with attached camera ( SPOT , Diagnostic Instruments , Inc . ) . Immediately following dissection , embryos were fixed for 2 to 3 h in a 6:3:1 ratio of 100% EtOH/37% formaldehyde ( Fisher ) /100% acetic acid ( Fisher ) at 4 °C . Fixed embryos were washed 3× in PBS and were paraffin embedded using a standard protocol . Then 10-μm sections were dried to positively charged slides ( Surgipath ) . Dried sections were deparaffinized and hydrated by standard procedures , before blocking endogenous peroxidases in 100% methanol/3% hydrogen peroxide for 10 min at room temperature . Sections were rinsed with water and PBS prior to antigen retrieval using a standard procedure ( Dako ) . Following PBS washes , sections were blocked in 5% donkey serum/2% BSA for 1 h at room temperature . Blocked sections were incubated overnight with primary antibodies against T ( Santa Cruz ) and β-galactosidase ( Cappel ) at 4 °C . Sections were then rinsed in PBS and incubated for 1 h with the corresponding secondary antibodies ( Invitrogen ) at room temperature . Sections were rinsed in PBS and coverslips were mounted with n-propyl gallate ( Sigma ) . Confocal imaging was performed using the ×20 objective lens and a Zeiss LSM 510 confocal microscope system ( Figure 2 ) . We initially applied a Shapiro-Wilks test to our data to determine whether tail-to-body length ratios followed a normal distribution . When results indicated that the distribution was not normally distributed ( p = 0 . 0043 ) , we applied a nonparametric test of independent samples ( Wilcoxon rank-sum ) to assess differences in the tail-to-body length ratios between single ( T/+; n = 29 ) and double ( T/+ Ago2+/–; n = 33 ) heterozygotes . The double heterozygotes were found to have significantly greater tail-to-body length ratios compared with single heterozygotes ( p = 0 . 007 ) . The endpoints of notches on the notched-box plot are located at the median ± 1 . 58 ( IQR/square root of n ) , where IQR represents the interquartile range and n is the subgroup sample size [38] .
|
Gastrulation is a developmental phase that delineates the three embryonic germ layers: ectoderm , endoderm , and mesoderm . The gene Brachyury is essential for mesoderm development , and short-tail mice , which were later found to be carrying a Brachyury mutation , have been known since 1927 . In this study , we found a genetic interaction between Brachyury and another gene in mouse , Argonaute2 . We show that the loss of Argonauate2 , a necessary component of a recently appreciated pathway of gene regulation called RNA interference , results in embryonic death during gastrulation , abnormal expression of Brachyury , and expansion of the mesoderm layer . This suggests that Argonaute2 is important in early development and in regulating Brachyury function . Consistent with this conclusion , we found that mice simultaneously carrying mutations in both Argonaute2 and Brachyury have significantly longer tails than mice with only a Brachyury mutation . A closer look at other genes involved in mesoderm development revealed that a disruption in fibroblast growth factor signaling may explain the mesoderm expansion in mice carrying the Argonaute2 mutation . Together this work demonstrates that a factor best known as a key component of RNA interference is required for the formation of a mammalian germ layer .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry",
"developmental",
"biology",
"none",
"cell",
"biology",
"mammals",
"eukaryotes",
"vertebrates",
"mus",
"(mouse)",
"genetics",
"and",
"genomics"
] |
2007
|
Argonaute2 Is Essential for Mammalian Gastrulation and Proper Mesoderm Formation
|
There is growing interest in understanding the nature and consequences of interactions among infectious agents . Pathogen interactions can be operational at different scales , either within a co-infected host or in host populations where they co-circulate , and can be either cooperative or competitive . The detection of interactions among pathogens has typically involved the study of synchrony in the oscillations of the protagonists , but as we show here , phase association provides an unreliable dynamical fingerprint for this task . We assess the capacity of a likelihood-based inference framework to accurately detect and quantify the presence and nature of pathogen interactions on the basis of realistic amounts and kinds of simulated data . We show that when epidemiological and demographic processes are well understood , noisy time series data can contain sufficient information to allow correct inference of interactions in multi-pathogen systems . The inference power is dependent on the strength and time-course of the underlying mechanism: stronger and longer-lasting interactions are more easily and more precisely quantified . We examine the limitations of our approach to stochastic temporal variation , under-reporting , and over-aggregation of data . We propose that likelihood shows promise as a basis for detection and quantification of the effects of pathogen interactions and the determination of their ( competitive or cooperative ) nature on the basis of population-level time-series data .
Studies of infectious disease systems typically focus solely on the interaction between the host and the causative agent . This approach has served epidemiologists well , especially for antigenically stable pathogens , such as measles or chickenpox [1]–[4] . It is becoming increasingly evident , however , that a broader perspective may be necessary to take interactions among infectious agents into account [5] . The mechanisms responsible for these interactions may be either immune-mediated or ecological , and their effects may be either competitive or cooperative [6] , [7] . Consider immune-mediated interactions , which are well-studied in the context of , for example , influenza virus infections in humans and other species . Exposure to a particular strain of influenza virus induces a humoral response from the host's immune system that subsequently clears the infection and is thought to confer long-lasting protection against that strain of the virus . This protective immunity may extend to other strains , depending on the similarity between the strains as measured by their “antigenic distance” [8] , or the number of amino acid differences in haemagglutinin epitopes [9] . Under the resulting selective pressure , the influenza virus accumulates amino acid differences in haemagglutinin epitopes to successfully evade immunity present in the population [10] , an evolutionary consequence of competition [11] . Immune-system mediated interactions may also be cooperative . An example of such an effect is the so-called doctrine of original antigenic sin , whereby “the antibody-forming mechanisms appear to be oriented by the initial infections of childhood so that exposures later in life to antigenically related strains result in a progressive reinforcement of the primary antibody” [12] . It is worth emphasizing that immune-mediated interactions are not necessarily restricted to genetically related pathogens . A number of recent studies speculate about the mutually beneficial interaction between HIV and malaria in Sub-Saharan Africa [13] . This interaction is thought to arise because the risk of clinical malaria is increased in HIV-1-positive individuals ( due to immune-suppression ) [14] , [15] , while malarial antigens can stimulate HIV-1 transcription and enhance replication [16] , promoting HIV transmission [17] . In contrast to the competitive dynamics resulting from cross-immunity in influenza , here the interaction is facilitative . Interactions can also arise at the population scale , driven by ecological processes . Previously , Rohani et . al . [18] proposed an interaction between measles and pertussis , two predominantly childhood diseases , via a shared susceptible pool [19] . Because infection with such acute diseases is typically followed by a period of convalescence and perhaps eventual death , an epidemic of one will lead to the temporary ( due to recovery ) or permanent ( due to fatality ) removal of susceptibles for the competing pathogen . This type of interaction –termed ecological interference– was shown to affect the phase relation between disease outbreaks [18] , as well as the inter-epidemic periodicity [20] . An infectious disease system in which many of these facets are thought to be simultaneously at play is the four dengue virus serotypes [21] . After an infection , individuals are immune to subsequent homologous viruses and are thought to be protected for 2–9 months against infection with a heterologous serotype [22] . These cross-reactive antibodies , however , wane over time leading to a scenario where heterotypic viral infections are in fact enhanced through the process known as antibody-dependent enhancement ( ADE ) [23] , [24] . Hence , with dengue , immune-mediated serotype interactions may be both competitive and cooperative , depending on the time since previous infection . Understanding the mechanisms that drive the transmission dynamics of dengue – and other strain polymorphic pathogens – is crucial because they affect interpretation of epidemiological data [25] , [26] , clinical case management [27] and the design , selection and implementation of alternative control programs [26] , [28] . It is thought that ADE is a major determinant of clinical pathogenesis and may explain why prior infection with a heterologous serotype is a significant risk factor for the development of the potentially fatal Dengue Hemorrhagic Fever ( DHF ) . It is not known , however , whether ADE results in increased transmission success of dengue . Attempts to explore this question have focused on comparing the dynamics of various mathematical models with serotype-specific longitudinal data . Reported data for dengue serotypes in hyper-endemic areas such as Mexico [29] , Thailand [30] and Vietnam [31] show the co-circulation of all four serotypes . A pattern of sequential serotype dominance is observed , with outbreaks of serotypes typically out of phase . This phase association has been viewed as a key dynamical signature which a variety of mathematical models of dengue have been challenged to reproduce . However , no consensus has yet emerged as to the primary mechanisms responsible for the observed oscillations in dengue serotypes . A number of studies suggest that cross-immunity may play a central role [25] , [28] , [32]; others have argued for ADE as the primary driving mechanism [31] , [33] , [34] . In an ideal world , the epidemiological impact of pathogen interactions would be quantified from individual-level infection histories observed in hosts in their natural habitats . Apart from exceptional settings [35] , [36] , however , such an undertaking is not feasible . It is inherently a difficult problem to scale biological processes at the level of individual organisms up to their population level consequences . Yet it is precisely the understanding of these population level implications of potential interactions that are fundamental for both public health strategies , and inferring longer-term ecological and evolutionary consequences . A step toward understanding the impact of such interactions might reside in our ability to infer their traces directly from the population level data . However , as we will show , approaches to this based on key dynamical signatures ( such as phase relationships ) can be unreliable as guides . Formal confrontation of mathematical models that include putative mechanisms for pathogen interactions directly with data may enable us to more effectively harvest the data's information and thus to more effectively and reliably distinguish among competing hypotheses and to quantify their relative transmission impact . In this paper , we assess the feasibility of using a likelihood-based inference framework to detect interactions from epidemiological data . Because models with pathogen interactions can generate a rich variety of dynamics [4] , [20] , [28] , [32] , [33] , [37] and may exhibit sensitivity to noise [38] , it is a priori unclear whether such inference is feasible . The question we ask is , if several mechanisms induce dynamics that are qualitatively indistinguishable , might it nevertheless be possible to quantitatively ascertain which are most likely to be operative ? Answering this question is complicated by the ubiquitous presence of stochasticity , which may very well be responsible for “patterns” that appear in data . We use a recently developed set of inference tools [39] , [40] and a flexible and freely available software package [41] , to formulate , estimate , and compare mechanistic models with different mechanisms of pathogen interactions . We seek to understand whether such interactions can be correctly inferred from the data . The techniques we use have been successfully applied in the context of understanding key features of cholera [42] , measles [43] , and malaria [44] , and are amenable to testing mechanistic models of stochastic dynamics with a continuous treatment of time and noisy , incomplete observations . In this proof-of-principle study , we show that , despite the complexity of multi-pathogen models , it is feasible to rigorously compare and distinguish among models having a realistic degree of complexity . We find that when inference is focused on pathogen interactions ( ie , when host demography and epidemiology are known ) , likelihood-based inference leads generally to correct and precise conclusions . Critically , inferential power in these circumstances is determined by the strength of the underlying interaction mechanism . This conclusion is robust even where temporal dynamics are highly variable . That is , we find that accurate inference is reliably possible despite stochasticity- and initial-condition driven phase drift . We conclude that likelihood shows great promise as a basis for the detection of the effects of pathogen interactions and the determination of their ( competitive or cooperative ) nature on the basis of population-level time-series data .
The model we focus upon is designed to be the simplest that admits [ ( i ) ] multiple competing interaction mechanisms , both permanent and temporary effects , and demographic stochasticity . It is a slightly simplified version of the two-pathogen compartmental model proposed by Vasco et al . [37] . It tracks hosts according to their pathogen-specific infection status . We bear in mind that the two pathogens in the model may represent two different strains of the same species or genetically unrelated infectious agents . In general , scaling the model up to deal with more than two interacting pathogens will be a straightforward matter , though the resulting model's complexity , eg , in terms of its state-space dimension , will increase geometrically with the number of interacting pathogens . As shown in Fig . 1 , we assume that individuals are born susceptible to both pathogens . For each pathogen , infection dynamics follow the progression , where , and are the familiar susceptible , infectious and recovered classes , respectively . Compartment has , in previous analyses [18] , [37] , been used to incorporate a period of convalescence , but here might also represent , either a temporary period of immuno-suppression or strain-transcending cross-immunity or a temporary period of enhanced transmissibility associated with ADE , for example . In this model , pathogens interact when an individual currently or previously infected with strain is exposed to pathogen . The consequence of this exposure for individuals previously exposed to strain is determined by positive parameters , and , which modulate the force of infection of strain , experienced by individuals in each of the , , and classes , respectively . Hence , if all , we have the null model in which the dynamics of the two pathogens are mutually independent . A value smaller than reflects either temporary ( as when or ) or permanent ( as when ) cross-immunity . Similarly , when these multipliers are greater than , current or previous infection with one pathogen increases susceptibility to the other , either in a temporary ( or ) or permanent ( ) fashion . This model assumes that all pathogen interactions are via modulation of host susceptibility . In reality , interactions may also operate via transmissibility . Here we ignore effects of heterotypic infections on transmissibility , as explored by , for example , [28] . The model also accounts for host demography in that births replenish the susceptible pool , and natural deaths remove hosts from each compartment . These rates are assumed independent of disease status and are both fixed at . Thus the host population size is held constant . To infer the nature of pathogen interactions in systems with variability in phase relationships , we utilize the framework of partially observed Markov processes [41]–[43] . This consists of three major components: ( i ) the data; ( ii ) the ‘process’ model , proposed to describe the underlying epidemiological and demographic processes ( described in section “Stochastic Model” ) ; and finally ( iii ) the observation model , proposed to describe the process by which the data are generated and linking the process model with the reported data . We assume the data consist of monthly pathogen/serotype-specific case notifications . Since we consider two pathogens , the data comprise parallel time series data and ( ) are related to the true number of infections via a Poisson distribution . Specifically , if is the total number of new recoveries in month for pathogen , and is the reporting probability , then monthly case notifications is assumed to have been drawn from a Poisson distribution with mean . The data sets we use to challenge our inference technique are realizations of this model . We use 40 years of simulated data , unless otherwise stated . For each simulated data set , we compute profile log-likelihoods over the parameters of interest , . The log-likelihood function may be expressed , ie , as the sum of conditional log-likelihoods of given and parameters . These quantities are computed using a Sequential Monte Carlo ( SMC ) algorithm [40] , [46] . Each SMC calculation uses 30 , 000 particles . To estimate Monte Carlo error , we repeat each SMC calculation 5 times at each parameter . The number of SMC particles and the resolution of the grid over are the only algorithmic parameters: it would be straightforward to further reduce Monte Carlo noise in our estimates by using more particles in the SMC calculations and/or a finer grid over . The cost of doing so is purely the greater computational effort required . For further details , refer to the supplementary information ( Text S1 ) .
In the absence of pathogen interaction , the dynamics of our unforced deterministic system are characterized by damped oscillations . However , interaction between pathogens can lead to sustained oscillations depending on parameter values [4] , [37] . When oscillations exist , cooperative interactions tend to generate in-phase cycles while competitive interactions tend to lead to out-of-phase oscillations . However , as Kamo & Sasaki [38] showed in a somewhat similar , but seasonally forced , system , the phase relationship between strains can be sensitive to stochasticity . Specifically , they demonstrated that noise can destabilize the in-phase solution , leading to asynchronous fluctuations . Similarly , in our stochastic system , phase relationships are variable . For all parameter values we examined , stochastic trajectories drift in and out of phase . To assess the reliability of between-strain phase relationship as an indicator of the nature of pathogen interactions , we performed a simulation study . We varied the interaction parameters ( , , ) across broad ranges , simulating realizations of yr duration at each point in parameter space . For each combination of parameters , the phase difference in strain-specific incidence generally varies with time . Fig . 2 shows the fraction of time during which oscillations are in-phase ( left ) and anti-phase ( right ) , as a function of the strength and sense ( cooperative versus competitive ) of both short- and long-term interaction . Even in multiply replicated time series of 5000 yr duration , no consistent association between the cooperative or competitive nature of epidemiological interactions and phase relationship emerges . While permanent cross-immunity ( ) , for example , frequently leads to out-of-phase dynamics , it is also associated with in-phase solutions 10%–20% of the time . More generally , for any combination of parameters , there is a moderate chance ( between 10% to 50% ) of observing either in-phase or anti-phase trajectories . The important practical implication is that phase relationship may be a poor predictor of the mechanism of pathogen interaction . Indeed , phase relationship alone appears to be of little use in indicating even the cooperative or competitive sense of pathogen interactions . To establish whether likelihood offers an improved basis for inferring the nature of pathogen interactions from epidemiological data , we performed another simulation study . We focused on parameters intended to be typical of closely related pathogen strains . In particular , we assumed symmetry between interacting pathogens , ie , , , , , , , and . To keep the complexity manageable in this proof-of-principle study , we focused on the interactions parameters by assuming the strictly epidemiological parameters ( contact rates , infectious periods , immigration rates , and durations of the temporary stage ) to be known . Moreover , we assumed the short-term interaction parameters to be identical , ie , . Table 1 gives the values to which these parameters were set . With these parameters , the net reproductive number , , is 2 . 7 ( in the absence of pathogen interaction ) . We examined the identifiability of the interaction parameters in three distinct scenarios ( Fig . 2 ) : Scenario I: No pathogen interactions , . Since each pathogen is independent of the other , this serves as a null model . Scenario II: Perfect short-term cross-protection , no long-term interaction , , . The ecological interference proposed to explain measles-pertussis interactions ( eg , [18] ) is an example . Scenario III: Moderate short-term cross-protection , permanent enhancement , , . These effects have been posited for the 4 dengue serotypes in hyper-endemic regions . For each scenario , we present log-likelihood profiles for the two interaction parameters of interest: short-term ( ) and long-term ( ) . We plot differences of log-likelihoods , , and compute confidence regions using likelihood ratio tests . We scale log-likelihoods such that the 95% confidence region corresponds to . Further details of the profile likelihood construction are provided in the supplementary information ( Text S1 ) . Having established that likelihood-based inference is computationally feasible in this system and can yield accurate estimates of interaction parameters , we now push further in search of the approach's limitations .
In recent decades , much work has focused on identifying the immunological consequences of infection with one pathogen for subsequent infection with other co-circulating pathogens . This is most obviously applicable to strain-polymorphic pathogens , such as influenza and dengue viruses , but is increasingly thought to affect unrelated infections , including the mutual enhancement of HIV and malaria [13] , TB and macroparasitic infections [47] , competitive/mutualistic dynamics of intestinal worms [35] , [48] , and an entire community of parasites co-circulating in wildlife populations [36] . There are two practical problems . First , it is imperative to establish the extent to which processes of enhancement or exclusion occurring at the level of an individual impact large scale transmission dynamics . Second , in instances where multiple competing mechanisms for pathogen interaction have been mooted , it is important to know whether analyses of incidence data can facilitate hypothesis testing . We have approached these problems by attempting to infer potential interactions from simulated case-notification data . Our choice of a two-pathogen model is deliberately simple . In general , concurrent or prior infection with heterotypic pathogens may modify host susceptibility , transmissibility , virulence , and infection duration , with concomitant impacts on epidemiology . Here , we focus strictly on interactions that affect host susceptibility . Our transmission model is sufficiently flexible and likelihood sufficiently powerful as a basis for inference that investigations of other mechanisms can be straightforwardly accommodated . We are encouraged to find that , in the optimistic case where the epidemiological traits of each pathogen ( and infectious period ) and seroepidemiology ( initial conditions ) are known , it is clearly possible to correctly infer the strength and nature of interactions from longitudinal data . Moreover , even when initial conditions are not known , it is possible to estimate them , with little loss of precision . Examining three distinct scenarios , we have described accurate inference of the presence/absence of an interaction , its strength and , promisingly , the confident identification of multiple co-occurring modes of interactions . Not surprisingly , our ability to infer an interaction is determined by its dynamical impact , with permanent effects better identified than interactions that operate over short time scales . It is important to point out that much of the work on serotype dynamics of dengue [28] , [32] , [33] or interference effects between measles and pertussis [18] , [19] has focused on phase association as a key dynamical signature . In our stochastic unforced model , we find that phase relation is highly variable ( Fig . 2 ) and is , by itself , an unreliable indicator of the type and the intensity of pathogen interaction . To quantify the impact of phase structure in our inference , for each scenario , we deliberately picked two data sets that showed qualitatively different relative phases . Somewhat surprisingly , our ability to make inference about underlying interactions is not driven by the correlation dynamics of the data . This suggests that phase relation–visually suggestive though it may be–is not a characteristic feature that provides reliable information about pathogen interaction . We need to place the relatively encouraging results of this proof-of-principle study within the context of our central–and at times rather optimistic–assumptions . As we have shown , the length of the time-series data directly affects the power of the inference . Within the confines of this project , we find that 40 yr of monthly data appear sufficient for robust inference . However , restricting the data to 20 yr substantially weakens the inference , particularly concerning short-term interactions . Relaxing the assumption of 100% reporting fidelity is an important reality check . We find that under-notification does not significantly impact the identifiability of key model parameters , assuming that the reporting precision is known . Even when the extent of under-reporting is not known and must be estimated along with other parameters , however , our results indicate that pathogen interactions remain identifiable ( see Figs . S3 in Text S1 , & Fig . 8 ) . Another possible data limitation is aggregation , particularly when for disease systems with multiple genetically related strains or serotypes . Our explorations of this problem suggest , although the information on the short term interaction is diluted , it need not substantially impair the correct identification of the interaction . The inference tests carried out here assumed a relatively low basic reproduction ratio ( ) , a short mean duration of temporary interaction ( 7 weeks ) in a large population ( ) . The power of these inferential tests is likely to change with disease epidemiology and population demographics . As we show in Fig . S1 in Text S1 , interactions are more clearly identifiable with a higher . This is to be expected since more interactions occur when infections are more contagious . Similarly , we expect temporary interactions to be more clearly identifiable when they operate over longer periods . In contrast , strain interactions in smaller populations–with their noisier dynamics and lower-amplitude outbreaks–should be less precisely identifiable . All of our inference tests have relied on an important assumption–perfect knowledge of disease epidemiology . In any real world application of our methodology , the transmission rate , infectious period and initial conditions will likely need to be estimated alongside the interaction parameters . It remains to be seen whether likelihood-based inference is up to such a challenge .
|
It is becoming increasingly evident that pathogens associated with infectious diseases interact amongst themselves . Pathogen interactions can occur in a co-infected host , or in host populations where they co-circulate , and they can be cooperative or competitive . Four serotypes of dengue virus , for example , can exhibit both forms of interactions – cross protection for a temporary period and followed by long-lasting enhancement . This bears important consequences for understanding the ecology and developing control and prevention measures . Detecting such interactions in a natural host population , though , can be tricky . We show that studying the phase relation of epidemic cycles , as it has been typically done , is unreliable . We assess the ability of a likelihood based method in detecting such interactions , and find that they are accurate and robust . We propose that this framework shows promise of serving as a basis for detecting and quantifying pathogen interactions .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"population",
"ecology",
"mathematics",
"ecology",
"population",
"modeling",
"statistics",
"theoretical",
"ecology",
"biology",
"computational",
"biology",
"infectious",
"disease",
"modeling",
"statistical",
"methods"
] |
2011
|
Statistical Inference for Multi-Pathogen Systems
|
Residue interaction networks and loop motions are important for catalysis in dihydrofolate reductase ( DHFR ) . Here , we investigate the effects of ligand binding and chain connectivity on network communication in DHFR . We carry out systematic network analysis and molecular dynamics simulations of the native DHFR and 19 of its circularly permuted variants by breaking the chain connections in ten folding element regions and in nine nonfolding element regions as observed by experiment . Our studies suggest that chain cleavage in folding element areas may deactivate DHFR due to large perturbations in the network properties near the active site . The protein active site is near or coincides with residues through which the shortest paths in the residue interaction network tend to go . Further , our network analysis reveals that ligand binding has “network-bridging effects” on the DHFR structure . Our results suggest that ligand binding leads to a modification , with most of the interaction networks now passing through the cofactor , shortening the average shortest path . Ligand binding at the active site has profound effects on the network centrality , especially the closeness .
Extensive experimental studies of dihydrofolate reductase ( DHFR ) have provided rich data toward the structure–function relationship in proteins . Escherichia coli DHFR is a 159–amino acid , monomeric , two-domain protein that is well characterized in terms of structure and function . DHFR catalyzes the reduction of 7 , 8-dihydrofolate ( DHF ) to 5 , 6 , 7 , 8-tetrahydro-folate ( THF ) using the reducing cofactor nicotinamide adenine dinucleotide phosphate ( NADPH ) . DHFR is a clinically important enzyme and is the target of a number of antifolate drugs . Experimental kinetic analysis of various DHFR permutants has identified several loop regions important for catalysis [1 , 2] . Figure 1 illustrates the loop locations . The Met-20 loop ( residues 10 to 23 ) directly controls the ligand binding to DHFR . The FG loop ( residues 116 to 121 ) is behind the Met-20 loop . There is a network of hydrogen bonds connecting the Met-20 loop and the FG loop . The third GH loop ( residues 142 to 149 ) is in contact with both the Met-20 and the FG loops . Based largely on the conformation of the Met-20 loop [3] , the three states of the enzymatic reaction process ( binding and release of cofactor , substrate , and product ) can be defined using available crystal structures ( Figure 1 ) . In the open state ( Figure 1 , green ribbon ) the Met-20 loop is flipped away from the binding site . In the closed state ( Figure 1 , blue ) , the Met-20 loop packs against the cofactor and seals the active site . In the occluded state , the Met-20 loop blocks the binding of the cofactor in the pocket . Simulations of the closed state also indicate changes in the other side of the binding pocket , in the helix region ( residues 44 to 50 ) , which binds the cofactor . Loop region 64–71 , which contacts the helix , also presents large fluctuations . The cooperative movements of these loops couple with the overall dynamics of the protein concerning the binding and dissociation of the cofactor , substrate , and product . The communications among the various parts of the DHFR can be achieved by residue interaction networks and through peptide backbone chain connections and nonbonded residue interactions . Agarwal et al . carried out genomic analysis of sequence conservation , kinetic measurements of multiple mutations , and theoretical calculations , observing that nonbonded residue interactions in DHFR form a network of coupled motions that are important for enzyme catalysis [4] . The effect of the peptide backbone chain connection on the protein dynamics is more complex , since chain connection is coupled with protein folding . Circular permutations of DHFR provide insight into chain connectivity , stability of the fold , and function . Circular permutation of a protein consists of connecting the native N- and C-termini covalently with a peptide linker and cleaving the peptide backbone at another specific site . Iwakura et al . have performed systematic circular permutation of the entire DHFR protein to investigate essential folding elements [5–8] . Other groups [9–11] have also circularly permuted the protein , selectively focusing on several permutations or cutting the backbone connection [11] to test fragment complementation . It was found that the peptide bonds in the protein could be grouped into two categories based on the effects of breaking the backbone connectivity . While cleavage of some peptide bonds results in less active variants or affects the enzyme function only slightly , suggesting that these make only minor contributions to the ability of the protein to fold , cleavage at certain other positions leads to a complete loss of the ability of the protein to fold . When such cleavage sites occurred sequentially in the primary sequence and formed a contiguous peptide segment , the region was named a “folding element , ” which is crucial for a protein to be foldable [5–8] . Folding elements distribute throughout the sequence . It was proposed that a complete set of folding elements is necessary for a protein to fold [5] . By conducting a systematic circular permutation analysis in which the original N- and C-termini of a protein are connected by an appropriate linker and new termini are created sequentially , ten folding elements have been identified in E . coli DHFR , each of which contains two to 14 residues [5–8] ( Figure 2 ) . It was also found that although the positions of the folding elements do not appear to correspond to the secondary structure motifs or to binding sites , almost all of the amino acid residues known to be involved in early folding events of DHFR are located within the folding elements , suggesting a close relationship between the folding elements of a protein and early folding events . In order to delineate the complex relationship among chain connectivity , protein folding , coupled networks , and catalysis by DHFR , we have carried out a systematic network analysis and molecular dynamics ( MD ) simulations of the native ( closed state ) DHFR and 19 of its circularly permuted variants . We first obtained average protein structures from MD simulations of the native DHFR and of the circularly permuted mutants . We investigated the relationship between small-world network behavior and chain connectivity using the crystal and MD average protein structures . Small-world network analysis of the protein structure uses graph theory to explore the bonded and nonbonded amino acid residue network . It was first used to identify key residues in protein folding , as these residues have high connectivity ( betweenness ) with respect to all possible network connections in the transition states of the protein structures [12] . The concept has been extended to the protein-folding process [13 , 14] , the protein–protein interface [15] , protein structure [16] and stability [17] , protein dynamics [18] , and key functional residues in enzymes [19] . By comparing the centrality of the residue interaction network , we found that ligand binding at the functional site has profound effects on the global network connections . Using the network connectivity index to distinguish between bonded and nonbonded connections , we found that breaking the chain connection in folding elements has a greater effect on active site loops than does breaking the chain in nonfolding regions . This leads us to conclude that the native sequence was selected to maximize the coupling between the protein fold and its functional dynamics . A folding–function interrelationship might particularly be the case for a fold like the DHFR , which currently has only been observed in the DHFR family , fulfilling a single function .
We have carried out a systematic network analysis and MD simulations of the native DHFR and 19 of its circularly permuted variants . Starting with the crystal structure of the native DHFR , we constructed circular permutations by linking the N-termini and C-termini with five glycine residues , and , following the experiment , introduced breaks in the chain connections in ten folding element regions and in nine nonfolding element regions . The 5-ns MD simulations provided average structures for the native DHFR and for the circularly permuted mutants that are able to fold according to experimental data [5] . For ten circularly permuted variants ( five from the folding element group and five from the nonfolding element group ) , additional 5-ns simulations provided dynamics effects reflecting the changes in the closeness in the functional loop regions . Analysis of the structural variations illustrated that there are no distinctive structural features indicating whether particular circularly permuted mutants with cuttings within the folding elements will be able to fold . Simulations of experimentally determined permuted mutants that are unable to fold sample only the local minimum on the protein-folding energy landscape and do not reveal the true folding properties of such circularly permuted variants . Nevertheless , our combined MD and network studies with breaks in the chain connection in the folding element or nonfolding element regions present different patterns of perturbation of the network properties near the active site . Our network analysis further leads us to propose that ligand binding induces network-bridging effects in the protein structure . We observed that substrate binding has profound effects on the DHFR network centrality , especially on closeness . The protein active site is near or coincides with residues through which the shortest paths in the residue interaction network tend to go . Our results suggest that cofactor binding leads to a modification of the interaction network , with most of the interactions now passing through the cofactor . The average shortest path is shorter with cofactor binding . Our findings are consistent with experimental observations that substrate binding increases DHFR folding stability [23 , 24] . In conclusion , our analysis demonstrates that active site dynamics of DHFR are communicated to the whole protein via both the peptide backbone and the nonbonded residue contacts . Such communication is indicated by the closeness change accompanying ligand binding , by breaking chain connections in the folding element region , and by breaking chain connections in the nonfolding element region . Even though the native and the circularly permuted proteins have similar overall folds , breaking the chain connections at different regions and ligand binding can change the properties of the network .
The structure of DHFR from E . coli ( closed state ) was taken from the Protein Data Bank ( http://www . rcsb . org/pdb ) . Both crystallographic waters and substrates were deleted . Circular permutation of the protein consists of connecting the native N- and C-termini covalently with a peptide linker and cleaving the peptide backbone at one specific site . Because a five-glycine peptide was shown to be the most favorable linker in the circularly permuted DHFR [5] , this peptide linker was used in all variants in our study . The MOE software ( Chemical Computing Group , http://www . chemcomp . com ) was used to obtain all the circularly permuted variants of the native DHFR , which now has 164 residues . A total of 19 variants were selected for simulation: ten in the proposed folding element region and nine in the nonfolding element region ( Table 1 ) . The CHARMM program [25] ( version 30b1 ) was used for all computations with the CHARMM force field version 22 using all atom representation [26] . The native DHFR and its circularly permuted variants were simulated in a 60 × 60 × 60 Å3 explicitly solvated periodic box . TIP3P water molecules were introduced . The simulations were carried out with a distance cutoff of 13 Å and a constant dielectric constant of 1 . Each simulation was initialized with adopted basis Newton-Raphson ( ABNR ) minimization followed by 3 ps system heating and 17 ps system equilibration . A production simulation run was carried out for each of the protein structures described above with a 1-fs time step . The coordinates were saved at 1-ps time intervals . Each simulation was run at 300 °K for 5 ns . The average structures from the last 2 . 5-ns simulations were used for network analysis . Patchdock [27] , a geometry-based molecular docking algorithm , was used to generate clusters for the DHFR–ligand interactions . We kept those clusters with at least four docked conformations with RMSDs within 3 . 5 Å . The amino acid network is defined by all residues within a contact distance . Residue interaction network analysis often uses uniformed distance as long as the contacting residues are within a cutoff distance . Such an approach does not distinguish between chain connection and nonbonded interaction . A recent network permutation analysis of DHFR based on such a definition led to two separate regions [22] . The nature of the chemical bond argues for a strong communication when two residues are sequentially linked by a peptide bond . In Figure 7 , we show that the most connected residue Ile5 has two bonded connections with Leu4 and Ala6 and the two closely packed residues Ile94 and Tyr111 . Since residues connected by a peptide bond have shorter Cα distances , a weighting of distance can distinguish between chain connection and nonbonded interaction . Thus , we define contact distances based on the Cα of all residues within 6 . 0 Å and use the integer of the distance as weight . Therefore , 6 is coded for a distance of 6 . 0 Å , 5 for a distance between 5 . 0 and 6 . 0 Å , 4 for a distance between 4 . 0 and 5 . 0 Å , and 3 for a distance between 3 . 0 and 4 . 0 Å . The chain connection can have a distance index of 3 or 4 , and a nonbonded contact can have a distance index of 5 or 6 . This definition can effectively reflect the effects of chain cleavage on the network properties . The algorithm by Pape [28] was used to calculate the shortest path lengths between nodes . Two network properties are computed to characterize network properties of a given protein structure . The betweenness [29] is one of the standard measures of node centrality . The betweenness bi of a node i is defined as where njk is the number of shortest paths connecting j and k , while njk ( i ) is the number of shortest paths connecting j and k and passing through i . The closeness centrality of node x is the inverse of the average distance between x and other nodes: The z-score of the closeness is calculated by z-score = ( C ( x ) − μ ) / σ , where μ is the average value of closeness and σ is the standard deviation .
The Protein Data Bank ( http://www . rcsb . org/pdb ) accession numbers for the structures discussed in this paper are open state ( 1ra9 ) ; closed state ( 1rx1 ) ; and occluded state ( 1rx5 ) .
|
The cooperative movements within a protein concerning the binding and dissociation of the reactants and products could be important for protein function . Communication among the various parts of an enzyme can be achieved by the networks connecting amino acids through peptide backbone connections and nonbonded amino acid contact . We used dihydrofolate reductase ( DHFR ) , a clinically important enzyme , as an example to explore the effects of amino acid communication on protein functions . We found that the peptide chain itself is an efficient “telephone wire” to transfer the communications . Breaking the telephone wire ( peptide chain ) at different points leads to differentiated behavior near the enzyme active site . The important points to keep the peptide chain communication are coupled with the place where protein folding occurs . On the other hand , ligand binding to the enzyme active site provides a “short cut” to the communication networks , with most of the interaction networks now passing through the added ligand and shortening the average communication path . We considered the short cuts to be “network-bridging effects” in the protein structure . The enzyme active site is the place where the short cut has the most dramatic effect in modifying protein communication networks .
|
[
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] |
[
"none",
"molecular",
"biology",
"computational",
"biology"
] |
2007
|
Ligand Binding and Circular Permutation Modify Residue Interaction Network in DHFR
|
Chromatin assembly mutants accumulate recombinogenic DNA damage and are sensitive to genotoxic agents . Here we have analyzed why impairment of the H3K56 acetylation-dependent CAF1 and Rtt106 chromatin assembly pathways , which have redundant roles in H3/H4 deposition during DNA replication , leads to genetic instability . We show that the absence of H3K56 acetylation or the simultaneous knock out of CAF1 and Rtt106 increases homologous recombination by affecting the integrity of advancing replication forks , while they have a minor effect on stalled replication fork stability in response to the replication inhibitor hydroxyurea . This defect in replication fork integrity is not due to defective checkpoints . In contrast , H3K56 acetylation protects against replicative DNA damaging agents by DNA repair/tolerance mechanisms that do not require CAF1/Rtt106 and are likely subsequent to the process of replication-coupled nucleosome deposition . We propose that the tight connection between DNA synthesis and histone deposition during DNA replication mediated by H3K56ac/CAF1/Rtt106 provides a mechanism for the stabilization of advancing replication forks and the maintenance of genome integrity , while H3K56 acetylation has an additional , CAF1/Rtt106-independent function in the response to replicative DNA damage .
Problems in DNA replication are a direct cause of genetic instability and are associated with early tumor development [1] . This instability is linked to a high susceptibility of the replication forks to become stalled , damaged or even broken , and for this reason understanding of the scenarios that threaten replication fork integrity is crucial , but also the mechanisms that promote replication fork repair and restart . Cells are endowed with a complex network of checkpoints mechanisms that coordinate DNA damage repair with cell cycle progression [2] . Thus , during S phase , arrested or damaged forks trigger a signal transduction cascade ending up in the phosphorylation of effector kinases ( e . g . , Rad53 in Saccharomyces cerevisiae ) that lead to specific responses such as the maintenance of replication fork stability , inhibition of late replication origins , DNA repair modulation and cell cycle arrest [3] . The presence of a sister chromatid provides a unique opportunity to repair and rescue the forks by homologous recombination ( HR ) , even though the molecular mechanisms by which HR repairs and/or tolerates replicative DNA damage remain unclear [4] . In eukaryotes DNA is packaged into a highly specialized and dynamic nucleoprotein structure called chromatin , which is actually the substrate for cell machineries that deal with DNA . The repetitive unit of chromatin , the nucleosome , is formed by ∼146 base pairs of DNA wrapped 1 . 65 times around an octamer of histones . Nucleosome assembly of the replicated DNA is conducted by histone chaperones and chromatin assembly factors that first deposit two heterodimers of histones H3 and H4 to form a core ( H3/H4 ) 2 tetramer to which an H2A/H2B dimer binds on each side [5] . This provides the substrate for a plethora of ATP-dependent remodeling and histone modifier complexes that will eventually set up the specific chromatin structures required for the regulation of each DNA metabolic process . Replication coupled ( RC ) -chromatin assembly occurs rapidly after the passage of the replication fork and involves physical interactions between components of the replisome with chromatin assembly and remodeling factors; e . g . , the replication processivity factor PCNA interacts with the chromatin assembly factor CAF1 [6] , [7] , the PCNA loader RFC with the histone chaperone Asf1 [8] and the MCM helicase complex with Asf1 and the chromatin remodeling complex FACT [9]–[11] . These interactions may facilitate nucleosome assembly but also help disrupt chromatin ahead of the fork . Besides , these interactions have been proposed to coordinate the flow of histones ensuring the exact supply at the fork [10] , a process that is also regulated at the level of DNA and histone synthesis during the cell cycle [12]–[14] . Newly synthesized histones H3 and H4 are acetylated before being deposited at the fork , and this modification is required for nucleosome assembly [15]–[19] . Histone H4 is acetylated at lysines 5 and 12 by the acetyltransferase Hat1 , this acetylation pattern being highly conserved from yeast to humans [15] , [20] , [21] . Histone H3 is also acetylated at its amino terminal tail , though the pattern is more variable among organisms . In the budding yeast H3 is acetylated at lysines 9 and 27 by the acetyltransferases Rtt109 and Gcn5 [22] . Additionally , histone H3 and H4 are acetylated in their globular domains at positions K56 and K91 by Rtt109 and Hat1 , respectively [19] , [23]–[26] . A detailed molecular analysis in yeast has recently deciphered part of the mechanisms of H3/H4 deposition during DNA replication . Thus , Asf1 binds to newly synthesized H3/H4 dimers [27] and presents them for acetylation of H3K56 by Rtt109 [23] , [24] . This histone modification enhances the binding affinity of H3 to the chromatin assembly factors CAF1 and Rtt106 and of CAF1 to PCNA , thus promoting histone deposition at the proximity of the fork [17] . This process is also facilitated by direct interactions between CAF1 with Asf1 and Rtt106 and Asf1 with Rtt109 [26] , [28]–[30] . Similarly , lysine acetylation at the amino terminal tail of H3 by Gcn5 enhances histone binding to CAF1 and Rtt106 and promotes RC chromatin assembly [16] , suggesting that lysine acetylation might be a general mechanism to regulate the interaction of histones with chromatin assembly factors . In addition to newly synthesized histones , cells recycle parental histones that result from the disassembly of the chromatin ahead of the replication fork , a process in which Asf1 is also involved [10] . A number of results have clearly shown over the last few years that defective chromatin assembly causes genetic instability . In plants and human cells , the absence of CAF1 causes inhibition of DNA synthesis , accumulation of DNA damage and activation of the S-phase checkpoint [31] , [32] . In yeast the disruption of a Gcn5-containing complex causes an accumulation of recombinogenic DNA damage [16] , while the absence of H3K56 acetylation in asf1Δ , rtt109Δ and H3K56R mutants increases the frequency of HR and gross chromosomal rearrangements ( GCRs ) [23] , [24] , [33] , [34] . Similarly , defective chromatin assembly by partial depletion of histones causes replication defects and hyper-recombination [35] , [36] . In addition to the accumulation of DNA damage , chromatin assembly mutants are usually sensitive to genotoxic agents that impair DNA replication; thus , acetylation of H3K56 and lysines at the amino terminal tails of H3 and H4 prevent DNA damage sensitivity by non-redundant mechanisms [17] , [23]–[25] , [27] , [37] , [38] . Similarly , a mutant lacking Cac1 – the largest subunit of CAF1 – and Rtt106 is defective in RC-chromatin assembly and replicative DNA damage repair/tolerance [17] . However , the mechanisms by which chromatin assembly prevents the accumulation of DNA damage and the sensitivity to replicative DNA damage remain unknown . This is in part due to the fact that many of the players functioning in RC-chromatin assembly do it as well in replication independent chromatin assembly processes like DNA repair and checkpoint recovery; e . g . , Asf1 and CAF1 are required for chromatin assembly and checkpoint turning off upon DNA double-strand break ( DSB ) repair [39]–[41] . In addition , it is difficult to discern whether the role of a histone mark in the DNA damage response ( DDR ) is prior or subsequent to histone deposition and whether it has a coding or a structural role . We have recently shown that defective chromatin assembly by partial depletion of H4 is rapidly followed by the collapse of replication forks , which are efficiently rescued via HR , suggesting that correct nucleosome deposition is required for replication fork stability [42] . This approach , however , needs to be validated for specific chromatin assembly mutants . Here we have dissected the H3K56ac-dependent CAF1 and Rtt106 chromatin assembly pathways in terms of HR , checkpoint activation , replication fork stability and response to different genotoxic agents . Our results indicate that defective nucleosome assembly by impairment of H3K56ac-dependent CAF1 and Rtt106 pathways increases HR by affecting the integrity of advancing , but not stalled , replication forks . In contrast , H3K56ac is required after replicative DNA damage for CAF1/Rtt106-independent DNA repair/tolerance mechanisms that are likely to occur after its incorporation into chromatin .
The histone chaperone Asf1 interacts with the histone acetyltransferase Rtt109 , and both proteins are required for acetylation at lysine 56 of newly synthesized histone H3 [23] , [24] , [26] , [43] . Consistent with a role for this histone modification in preventing DNA damage accumulation , the absence of H3K56 acetylation in asf1Δ , rtt109Δ and H3K56R mutants increases the frequency of genetic recombination and budded cells with foci of the recombination protein Rad52 fused to the yellow-fluorescence protein ( Rad52-YFP ) ( Figure 1; [23] , [24] , [34] ) . As previously shown for rtt109Δ [24] , we confirmed that the increase in recombination mediated by asf1Δ was due to its incapability acetylating H3 on lysine 56 , as the frequency of genetic recombination and Rad52-YFP foci in asf1Δ H3K56R was as in the single mutants ( Figure 1A and 1B ) . Histone H3K56 acetylation marks nucleosomes incorporated into chromatin via both RC and replication independent mechanisms [44] , [45] . Thus , we first assessed whether the observed increase in recombination was linked to defects in replication-independent chromatin assembly . In this regard , Asf1 interacts with the HIR complex ( formed by Hir1 , Hir2 and Hir3 in yeast ) [46] with which promotes replication-independent chromatin assembly [47] . We analyzed recombination in the absence of Hir1 since this subunit is required for the integrity and histone deposition activity of Asf1/HIR [47] . As shown in Figure 1C and 1D , disruption of the HIR complex in hir1Δ did not affect recombination . Acetylation of H3K56 is also involved in RC-nucleosome assembly . It promotes both the transfer of H3/H4 to the chromatin assembly factors CAF1 and Rtt106 and the binding of CAF1 to PCNA [17] . Consequently , hyper-recombination in asf1Δ , rtt109Δ and H3K56R could be associated with defective histone deposition but also with a loss of structural and/or coding information because of the absence of H3K56ac at chromatin . To distinguish between these possibilities we analyzed the role of CAF1 and Rtt106 in preventing the accumulation of recombinogenic DNA damage; CAF1 and Rtt106 have redundant chromatin assembly functions as shown by the fact that cac1Δ rtt106Δ , but not cac1Δ and rtt106Δ , is defective in histone deposition [17] . Besides , the levels of H3K56ac are not affected and its deposition at chromatin is delayed but not prevented in cac1Δ rtt106Δ [17] . While the single mutants cac1Δ and rtt106Δ were not affected in HR , the double mutant cac1Δ rtt106Δ increased the frequency both of genetic recombination and budded cells with Rad52-YFP foci as compared to the wild type ( Figure 1C and 1D ) , indicating that CAF1- and Rtt106-dependent chromatin assembly pathways prevent the accumulation of recombinogenic DNA damage . Besides , the triple mutant asf1Δ cac1Δ rtt106Δ displayed the same frequency of genetic recombination as asf1Δ and cac1Δ rtt106Δ , suggesting that H3K56ac avoids hyper-recombination through its function in CAF1/Rtt106-dependent chromatin assembly . Nevertheless , the triple mutants displayed a slight but significantly higher frequency of cells with Rad52 foci than asf1Δ and cac1Δ rtt106Δ , suggesting the existence of additional , non-overlapping functions of H3K56ac and CAF1/Rtt106 in preventing the accumulation of DNA damage . Another feature of asf1Δ , rtt109Δ and H3K56R is the activation of the DNA damage checkpoint in the absence of DNA damaging agents as determined by partial phosphorylation of Rad53 [23] , [48] , [49]; as shown in Figure 1E , only the simultaneous absence of CAF1 and Rtt106 led to the activation of Rad53 . Therefore , our results indicate that defective RC-nucleosome assembly causes accumulation of recombinogenic DNA damage and checkpoint activation . However , and strikingly , the absence of Rad52 did not increase the amount of phosphorylated Rad53 in asf1Δ as determined by western blot and in situ kinase assays ( Figure 1E and 1F ) , suggesting that accumulation of recombinogenic DNA damage and checkpoint activation are not genetically linked . Histone deposition and DNA synthesis are tightly connected during DNA replication . We therefore hypothesized that defective nucleosome assembly in asf1Δ , rtt109Δ , H3K56R and cac1Δ rtt106Δ mutants might affect replication fork integrity , which in turn would generate genetic instability . To address this possibility we followed the fate of replication intermediates ( RIs ) in wild type and mutants by 2D-gel electrophoresis . For this , cells were synchronized in G1 with α-factor and released into S phase , and DNA samples were analyzed at different times to follow the progression of replication forks from the early replication origin ARS305 ( Figure 2A ) . Replication initiation and early elongation can be followed with probe Or by the formation of a bubble arc that reverts to a single Y-arc of large Y-shaped molecules when forks cross the nearest restriction site ( Figure 2B , left panel ) , while replication fork progression along adjacent restriction fragments can be followed with specific probes by the accumulation of a complete arc of single Y-shaped molecules ( Figure 2B , central panel ) . Finally , converging forks and Holliday junction ( HJ ) -like structures can be detected by the accumulation of double Y- and X-shaped molecules , respectively ( Figure 2B , right panel ) . The amount of RIs at the origin during the kinetics ( i . e . , the sum of bubbles , Ys and Xs at region Or of all time points combined ) , taking the total amount of wild-type RIs as 100 , was reduced to ∼50% in asf1Δ and rtt109Δ ( Figure 2C ) . In agreement with this defect being mediated by the lack of acetylation at H3K56 in asf1Δ and rtt109Δ , the total amount of RIs in a H3K56R mutant was 33% ( Figure 2D ) . An increased drop in RIs was noticed in H3K56R as compared to asf1Δ and rtt109Δ ( Figure 2C and 2D ) , which might be due to either an additional effect by reduced levels of histones – strains in Figure 2D have one instead of two H3/H4 genes – or the specific change to arginine . Therefore , the absence of H3K56 acetylation causes a loss of RIs . It should be noted that this reduction was also observed at adjacent DNA fragments , even though the effect became less evident at fragment B because of the loss of synchrony in the peak of RIs as the forks move away from the origin . Next , we decided to address whether the loss of RIs in mutants defective in H3K56 acetylation was due to defective chromatin assembly as previously shown for recombination and checkpoint activation . For this , the amount of replication forks from cac1Δ , rtt106Δ and cac1Δ rtt106Δ mutants synchronized in G1 and released into S phase was analyzed . As shown in Figure 3A , whereas the single mutants cac1Δ and rtt106Δ accumulated wild-type levels of RIs , the double mutant cac1Δ rtt106Δ displayed a ∼50% reduction in the amount of RIs at the origin , indicating that CAF1- and Rtt106-mediated chromatin assembly pathways have redundant roles in preventing the loss of replication forks . Besides , the levels of RIs in cac1Δ rtt106Δ were the same as in asf1Δ and rtt109Δ ( ∼50% ) , suggesting that the major role of H3K56 acetylation in replication fork stability is through its function in chromatin assembly . Consistently , the reduction in RIs in the triple mutant asf1Δ cac1Δ rtt106Δ was neither synergistic nor additive as compared to asf1Δ ( 69±3%; Figure 3B ) , though this drop opens the possibility that H3K56ac and CAF1/Rtt106 have also additional , non-overlapping functions in preventing the loss of RIs . Finally , we observed that the total amount of RIs at the replication origin ARS315 was also significantly reduced in asf1Δ and cac1Δ rtt106Δ as compared to wild type ( ∼64 and ∼44%; Figure S1 ) , indicating that the loss of RIs was not restricted to ARS305 . In order to determine why defective chromatin assembly causes a loss of RIs , we first assessed the possibility that forks break during DNA extraction . Contrary to this , the loss of RIs in asf1Δ determined by collecting and digesting the DNA in agarose plugs to preserve its integrity was similar to that obtained with standard DNA extraction protocols ( Figure S2 ) . Alternatively , this loss of RIs might be due to differences in replication initiation , either in the efficiency or in the synchrony of the firing . As a first approach to assess this possibility we analyzed cell cycle progression in chromatin assembly mutants . FACS and budding analyses showed that most G1 cells reached G2/M in all mutants ( Figure 4A , 4B and 4C ) . Besides , neither asf1Δ nor rtt109Δ displayed a significant delay in completing S phase compared to the wild type ( Figure 4A and 4C ) , suggesting that the loss of RIs in these mutants is not due to defects in replication initiation; in contrast , H3K56R was clearly retarded as compared with its wild type . Also , while cac1Δ and rtt106Δ were not affected , cac1Δ rtt106Δ mutants displayed a slight but significant delay ( Figure 4A and 4C ) that might influence the amount of RIs . However , the reduction in RIs in the triple mutant asf1Δ cac1Δ rtt106Δ was neither synergistic nor additive as compared to asf1Δ ( Figure 3B ) , which is not affected in cell cycle progression . Therefore , the delay in the progression through S phase seems not to be the main cause for the loss of RIs in cac1Δ rtt106Δ , even though the 30% drop in the triple asf1Δ cac1Δ rtt106Δ versus the single asf1Δ mutant leaves open the possibility that a fraction of the drop in RIs reflects some defects in replication initiation . Since FACS and budding analyses estimate whole genome duplication , we cannot rule out the possibility that cells progress normally through S phase but having problems in the firing of some specific origins that could be compensated with altered programs of initiation and/or elongation . Likewise , a slow advance through S phase does not necessarily reflect a defect at a specific replication origin . Therefore , we first asked whether the loss of RIs was a consequence of inefficient ARS305 firing . In this regard , a defect in replication initiation would lead to a complete single Y-arc indicative of passive replication of the ARS305 fragment by forks coming from a neighbor origin . Even though the shape of the single Y-arc in the mutants was as in the wild type ( Figure 2 and Figure 3 ) , we cannot discard that the region were replicated later either from ARS305 or from a fork originated elsewhere . Therefore , we decided to determine the efficiency of replication initiation of the origin ARS305 . Previous works have shown that asf1Δ , rtt109Δ and H3K56R are proficient in the activation of this origin [8] , [50] . We studied replication initiation in our strains with a similar approach [42]; cells arrested in G1 with α-factor were released into S phase in the presence of hydroxyurea ( HU ) for 50 minutes , which causes the stalling of the forks in the proximity of the origin by depletion of available dNTPs . RT-PCR quantification of the total amount of DNA at the origin relative to an unreplicated fragment both in G1 and HU-arrested cells showed no significant defects in the firing of ARS305 in any of the mutants tested ( Figure 4D ) . Next , we asked whether the loss of RIs was due to differences in the synchrony of the firing of replication from ARS305 . Contrary to this possibility , chromatin assembly mutants displayed the same kinetics of RI accumulation as the wild type , with a peak for the ARS305 region at 20–30 minutes upon G1 release ( Figure 2 and Figure 3 ) . This was not the case for H3K56R , in which the slow accumulation of RIs might explain its difference with asf1Δ and rtt109Δ ( Figure 2D ) . Importantly , chromatin assembly mutants displayed a similar drop in RIs when released into S phase for 1 and 2 hours – what ensures that most cells have fired ARS305 ( Figure 4D ) – in the presence of HU ( see below ) , which stalls forks close to the origin and thereby minimizes putative differences in synchrony . Consequently , the loss of RIs in chromatin assembly mutants is not associated with defective replication initiation and therefore may reflect a loss of integrity of the replication forks as they move away from the origin . We have shown that chromatin assembly mutants display both a loss of RIs and an increase in recombination . Indeed , the stronger is the loss of RIs the higher is the percentage of cells with recombination foci . This correlation led us to hypothesize that the increase in recombination might result from the repair of collapsed replication forks . To address this possibility , we analyzed the role of Rad52 , essential for DNA repair by HR [51] , in the replication of cells lacking Asf1 . As shown in Figure 5A , the amount of RIs dropped from about 54% in asf1Δ and rad52Δ to 14% in asf1Δ rad52Δ , being this drop not associated with defects in the kinetics of RI accumulation or in the firing of ARS305 ( Figure 4D ) . This synergistic reduction of RIs in asf1Δ rad52Δ suggests that HR participates in the rescue of collapsed forks from ARS305 in asf1Δ . Consistently , asf1Δ rad52Δ cells displayed a delay in completing S phase ( Figure 4A and 4C ) . These results provide an explanation for the accumulation of recombinogenic DNA damage in chromatin assembly mutants and the slow growth of asf1Δ rad52Δ cells ( Figure 5B; [49] ) . Defective H3K56 acetylation in asf1Δ , rtt109Δ and H3K56R causes a reduction in the amount of ChIP-detected replisome components in the presence of HU that has been thought to be responsible for their high sensitivity to drugs that stall replication forks [8] , . Those experiments , however , do not provide information about the integrity of DNA at the fork and cause of the collapse , which could be a defect in chromatin assembly but also the absence of H3K56 acetylation at chromatin . Besides , our previous results suggest a role for this modification in keeping the stability of unperturbed replication forks , leaving its role unresolved on stalled replication forks . Therefore , we followed the fate of RIs in cells synchronized in G1 and released into the S phase in the presence of HU , which leads to the stalling of the wild-type forks at the proximity of the origin with a peak of RIs at 60 minutes upon α-factor release ( Figure 6A; [42] , [52] ) . A similar kinetics of replication fork stalling was observed in asf1Δ ( Figure 6A ) , indicating that synchrony was not affected; however , and consistent with previous ChIP analysis [8] , [26] , [50] , the total amount of stalled RIs over the whole region ( i . e . , the sum of bubbles , Ys and Xs of all fragments , either of all time points combined ( Figure 6A ) or at 1 hour ( Figure 6B ) ) , taking the total amount of wild-type RIs as 100 , dropped to ∼30% in asf1Δ and rtt109Δ ( Figure 6A and 6B ) and this reduction was not due to a distinctive distribution of the stalled forks along the DNA ( Figure S3 ) . Also , a similar drop in RIs was observed in cac1Δ rtt106Δ ( Figure 6B ) , indicating that proper chromatin assembly is required to prevent the loss of RIs in the presence of HU . Therefore , HU further decreases the amount of RIs in chromatin assembly mutants from approximately 50 to 30% . In principle , this enhanced loss of RIs in the presence of HU might be linked to a role for chromatin assembly in keeping the stability of both advancing and stalled replication forks , but also to a defect in resuming DNA replication upon HR-dependent fork rescue as a consequence of the HU-induced depletion of available dNTPs . In this case , however , the HU would not have any additional effect on replication fork stability in the absence of Rad52 . As previously shown [42] , the amount of RIs in rad52Δ was not affected by HU ( ∼50%; Figure 5A and Figure 6B ) , indicating that Rad52 is not required for the stability of stalled replication forks but likely for the rescue of damaged replication forks . Importantly , the amount of RIs in asf1Δ rad52Δ was not affected by the presence of HU ( ∼15%; Figure 5A and Figure 6B ) , suggesting that Asf1 , and by extension H3K56 acetylation , has a minor role in the stability of stalled replication forks . In addition , and consistent with the idea that HU partially prevents the restart of replication forks , asf1Δ cells released into S phase in the presence of HU displayed a 2-fold increase in X-shaped molecules ( Figure 6C ) . Unfortunately , the slight accumulation of X-shaped molecules in rad52Δ leaves an insufficient margin to determine the Rad52 dependency of the X-shaped molecules accumulated in asf1Δ . These results argue against a defect in the stability of stalled replication forks as a causative factor of the high sensitivity of asf1Δ , rtt109Δ and H3K56R to HU . Accordingly , the double mutant cac1Δ rtt106Δ was not sensitive to HU ( Figure 7A ) , despite this strain displaying a similar loss of RIs as asf1Δ and rtt109Δ . In agreement with the growth assay , cac1Δ rtt106Δ was not required for stalled forks restart as determined by treating G1 released cells with 200 mM HU for 1 hour and checking their ability to resume DNA replication by FACS analysis ( Figure 7B ) ( note that cac1Δ rtt106Δ displayed a similar delay during the S phase in the absence of HU ( Figure 4A ) ) . Strikingly , asf1Δ cells also resumed DNA replication after 1 hour in 200 mM HU and progressed to the following cell cycle without previous arrest ( Figure 7B ) ; consistently , asf1Δ cells did not display defects in checkpoint recovery and were viable ( data not shown; [53] , [54] ) . In summary , H3K56ac/CAF1/Rtt106-mediated chromatin assembly has no role in the stability and restart of forks stalled by HU , and therefore the loss of RIs observed in HU has to be of advancing replication forks . Our previous results indicate that the role of H3K56ac in preventing sensitivity to chronic treatment with HU is independent of CAF1/Rtt106 , suggesting that is a function separate from chromatin assembly and likely subsequent to its deposition at chromatin . A global epistatic analysis of pairs of gene deletions revealed a connection between Asf1 and Rtt109 with the Rtt101 ubiquitin ligase complex [53] , which appear to promote fork progression through damaged DNA by HR [55]–[57] . However , as previously shown and in contrast to asf1Δ and rtt109Δ , rtt101Δ was not sensitive to HU ( Figure 7A; [56] , [57] ) . H3K56ac , and by extension Asf1 and Rtt109 , are also required for growth in the presence of drugs that impair the advance of the replication forks by DNA damage , such as the topoisomerase I inhibitor camptothecin ( CPT ) or the DNA alkylating agent methyl methane sulfonate ( MMS ) ( Figure 7A; [23]–[25] , [27] , [39] ) . Again , these sensitivities could be associated with the role of H3K56ac in chromatin assembly . A comparative analysis showed that although the double mutant cac1Δ rtt106Δ was sensitive to both drugs , in particular to high concentrations , this sensitivity was much milder than that displayed by asf1Δ and rtt109Δ ( Figure 7A , see CPT at 7 . 5 µg/ml and MMS at 0 . 005% ) , suggesting that the main role of H3K56ac in response to CPT and MMS is also independent of CAF1/Rtt106 and subsequent to its deposition into chromatin . The ubiquitin ligase complex Rtt101 has been shown to be required for MMS- and CPT-induced HR [55] and for checkpoint recovery ( Figure 7C; [53] , [55] , [57] ) . Our comparative analysis showed that rtt101Δ was not as sensitive to MMS and CPT as asf1Δ and rtt109Δ ( Figure 7A ) ; thus , these results suggest that H3K56ac promotes fork progression through damaged DNA via Rtt101-mediated HR and , to a lesser extent , CAF1/Rtt106-mediated chromatin assembly . To further understand the role of the CAF1/Rtt106 chromatin assembly pathway on MMS and CPT resistance , we analyzed the ability of cac1Δ rtt106Δ to resume DNA replication upon the treatment of G1 released cells with a high concentration ( 0 . 033% ) of MMS . cac1Δ rtt106Δ cells resumed and completed DNA replication but remained partially arrested in mitosis ( Figure 7B ) as a consequence of a delay in checkpoint deactivation ( Figure 7C ) , being these phenotypes much stronger in asf1Δ and rtt101Δ in agreement with the sensitivity assay .
H3K56 acetylation is a histone modification required for chromatin assembly . Notably , mutants defective in H3K56 acetylation ( asf1Δ , rtt109Δ and H3K56R ) accumulate recombinogenic DNA damage as determined by genetic recombination , cells with Rad52 foci and molecular analysis of sister-chromatid exchange [23] , [24] , [34] . How H3K56 acetylation prevents DNA damage accumulation is not predictable , however , because its role in chromatin assembly is associated not only with replication but also with other processes that influence HR , such as transcription , silencing , DSB repair or DNA damage tolerance [58] . We first ruled out a role for replication-independent chromatin assembly as a disruption of the HIR/Asf1 complex in hir1Δ exhibited wild-type levels of recombination . Alternatively , and in agreement with a model in which spontaneous genetic instability stems from defective DNA damage repair/tolerance , hyper-recombination might result from defective repair/tolerance and channelling to HR of spontaneous DNA lesions . In this case , DNA damage induction with genotoxic agents to which these mutants are sensitive should further increase their levels of recombination . In contrast , Asf1 , Rtt109 and the Rtt101 complex are required for HR induced by MMS and CPT [55] . Given that Asf1 and Rtt109 are not required for DSB-induced HR , both ectopic and sister-chromatid recombination [34] , [49] , [55] , hyper-recombination in cells defective in H3K56 acetylation may be associated with the generation of DSBs . Accordingly , GCRs are mediated by the DSB-repair pathway of non-homologous end-joining and are prevented by HR in asf1Δ [33] . H3K56 acetylation enhances the binding affinity of H3 to CAF1 and Rtt106 , two factors with redundant histone deposition functions during replication [17] . We show that only the RC-chromatin assembly defective cac1Δ rtt106Δ , but not the RC-chromatin assembly proficient cac1Δ and rtt106Δ , leads to recombinogenic DNA damage and checkpoint activation , and that the main role of H3K56ac in preventing hyper-recombination is mediated by CAF1 and Rtt106 . Therefore , RC-chromatin assembly prevents the accumulation of recombinogenic DNA damage . We show that chromatin assembly mutants display a loss of RIs that is not due to defects in replication initiation , and that there is a correlation between the loss of RIs and the increase in HR . Besides , the absence of Rad52 , essential for HR , further increases the loss of RIs in asf1Δ . These results , together with the reported loss of replisome integrity in H3K56 acetylation mutants in the presence of HU [8] , [26] , [50] despite the fact that they are not affected in the stability and rescue of stalled replication forks ( Figure 6 and Figure 7 ) , strongly suggest that defective RC-chromatin assembly causes a loss of integrity of the advancing replication forks , and that HR participates in the rescue of these forks using the sister chromatid . Consistent with this , asf1Δ accumulates spontaneously sister-chromatid exchange products [34] . This loss of integrity may end up in the collapse of some of the forks , which can render unprotected DNA ends susceptible of being processed by HR [59]–[62] but that are difficult to be detected by 2D-gel analysis unless a homogeneous and stable population of intermediates accumulates . In particular , the detection of broken intermediates is not easy because the breakage of single Ys leads to linear molecules , while the breakage of bubbles leads to a mixture of asymmetric Ys that do not run at a defined arc . Additionally , defective chromatin assembly might generate DNA structures that are lost due to the running conditions required for the visualization of the RIs by 2D-gel analysis . Similarly , the reduction in the total amount of detectable RIs in chromatin assembly mutants in spite of the fact that they complete replication opens the possibility that the rescue of the collapsed forks and subsequent completion of DNA replication are not associated with the formation of a canonical replication fork [63] or reflects an asynchronous fork rescue along the DNA region . Finally , we cannot rule out that a fraction of the drop in the amount of RIs to be a consequence of problems in the initiation of replication of a subpopulation of cells as suggested by the analysis of cell cycle progression in cac1Δ rtt106Δ mutants . Strikingly , defective chromatin assembly hardly affected ( asf1Δ , rtt109Δ ) or delayed just 10–20 minutes ( H3K56R , cac1Δ rtt106 , asf1Δ cac1Δ rtt106Δ ) the time required for DNA duplication despite the loss of RIs . Replication fork rescue by HR cannot account for completion of DNA replication because asf1Δ rad52Δ cells are also capable of completing DNA duplication ( Figure 4 ) . Additional mechanisms may operate in the rescue of the collapsed replication forks; in this regard , it has recently been shown that asf1Δ accumulates ribosomal DNA repeats by a novel mechanism that is independent of HR but needs replication processivity functions known to be required for break-induced replication [64] . This work is consistent with our proposal that chromatin assembly mutants accumulate broken forks and that there may be mechanisms other than HR involved in the repair of these breaks . We have observed that the loss of RIs is not specific of forks coming from ARS305 ( Figure S1 ) ; however , we cannot rule out the possibility that not all chromatin regions display the same replication defects , that a proportion of the forks are functional but are lost during the 2D-gel analysis , and that chromatin assembly mutants counteract the instability of the replication forks by altering the program of replication initiation and/or increasing the rates of replication elongation . In this frame , it is possible that an “open” chromatin structure in these mutants favors alternative outputs of collapsed fork rescue and DNA replication as suggested above . Genome-wide analyses have to be conducted to address these possibilities . Why are replication forks unstable under conditions of defective RC-chromatin assembly ? These mutants are proficient in checkpoint activation ( Figure 1 and Figure 7; [23] , [34] , [48] , [49] , [53] , [65] ) , ruling out a defect in this mechanism of replication fork stability as responsible for the loss of RIs . In fact , the absence of checkpoint proteins in asf1Δ affects cell progression during the S phase [54] , suggesting that chromatin assembly and replication checkpoints have non-redundant functions in replication fork stability . In principle , the loss of RIs and the increase in HR could be associated with defects in chromatin structure as a consequence of the lack of H3K56 acetylation at chromatin . This modification breaks a water-mediated histone-DNA interaction at the point of entry and exit of the nucleosomal DNA that modulates chromatin compaction [25] , [66]–[68] . Also , this modification might recruit chromatin factors required for fork stability . We do not favor these possibilities in cac1Δ rtt106Δ because this mutant expresses acetylable H3K56 , although its deposition at chromatin appears to be delayed and might generate regions behind the fork with reduced H3K56ac [17] . Alternatively , replication fork instability might result from defective chromatin disassembly and/or transfer of parental histones ahead of the fork . In this regard , Asf1 , which is also a nucleosome disassembly factor [69] , interacts with MCM to coordinate fork progression and parental histone supply ahead of the fork [10] . However , asf1Δ and H3K56R mutants share similar defects in replication fork stability and HR and the effect of asf1Δ is due to defective H3K56 acetylation as determined by epistatic analysis . Since this modification marks preferentially newly synthesized histones [25] , our results point to defects in the pathway of newly synthesized histone deposition as the main cause of fork collapse and subsequent repair by HR . DNA synthesis and histone deposition are physically and genetically connected to ensure the exact supply of histones at the fork [6]–[11] . Histone excess is toxic and cells are endowed with different mechanisms to get rid of non-incorporated histones [12] . The opposite situation , a reduction in the pool of available histones , is also deleterious and phenocopies the defects in fork stability and HR reported here with RC-chromatin assembly mutants [42] . The current study provides additional support to the idea that , under conditions of defective H3/H4 deposition during replication , DNA synthesis and nucleosome assembly could become uncoupled exposing DNA fragments behind the fork . This uncoupling might favor the formation of unstable secondary DNA structures , as it has been proposed to explain the high levels of DNA breakage and contractions at CAG/CTG tracts displayed by asf1Δ and rtt109Δ but not rtt101Δ [70] . Although these structures could be targeted by nucleases , we failed to find single nuclease mutants that alter the frequency of RI loss in asf1Δ ( data not shown ) , a result that is not unexpected because of the redundancy of DNA nucleases in DNA damage repair [71] , [72] . Finally , the loss of RIs and the increase in HR could be due to defective stability of stalled forks , as suggested by the observation that the replisome is unstable in the presence of HU in H3K56 acetylation mutants [8] , [26] , [50] . Here , we present some evidence indicating that only advancing , but not stalled forks , are affected in RC-chromatin assembly mutants . First , the total amount of RIs in chromatin assembly mutants defective in fork rescue by HR ( asf1Δ rad52Δ ) is not affected by the presence of HU . Second , RC-chromatin assembly mutants ( asf1Δ , rtt109Δ and cac1Δ rtt106Δ ) are proficient in stalled fork stability and restart upon an acute treatment with HU as determined by FACS analysis , checkpoint recovery and cell viability . Therefore , our results point to defects in the stability of advancing forks as the cause of the genetic instability in RC-nucleosome assembly mutants , further supporting the idea that defective histone deposition uncouples DNA synthesis and nucleosome assembly . Notably , asf1Δ cells treated with HU also exhibited an accumulation of Polα at the fork and an uncoupling of the MCM helicase [8] . We speculate that these alterations in the replisome structure might also occur in the absence of HU . Indeed , Asf1 interacts with MCM [10] and with RFC – which loads PCNA and in this way replaces Polα with Polε and Polδ – [8] , and H3K56 acetylation regulates the function of the RFC [73]; it is thereby possible that the absence of Asf1 and/or H3K56ac could specifically alter the distribution of the polymerases and the MCM helicase at the fork . H3K56 acetylation – and by extent Asf1 and Rtt109 – is required for promoting resistance to replicative DNA damage [17] , [23]–[25] , [27] . Indeed , there is a correlation between the levels of H3K56 acetylation and the degree of DNA damage sensitivity to genotoxic agents [43]; consistently , H3K56Q , which mimics constitutive acetylation , suppresses asf1Δ sensitivity to HU and CPT [39] , [43] . In contrast to H3K56 acetylation mutants , cac1Δ rtt106Δ is only sensitive to high concentrations of MMS and CPT and is not sensitive to chronic treatment with HU , suggesting that the function of H3K56ac in the replicative DNA damage response can be separated from its role in CAF1/Rtt106-mediated chromatin assembly . This points to a role subsequent to its deposition into chromatin . In agreement with this idea , it has recently been shown that a change of lysine 56 to glutamic acid in H3 generates a histone proficient in binding to CAF1 and Rtt106 but sensitive to replicative DNA damage [74] . An epistatic analysis has included Asf1 , Rtt109 and the Rtt101 ubiquitin ligase complex into a functional group involved in DNA repair [53] . Rtt101 is recruited to chromatin in response to DNA damage in a process that requires Rtt109 [75] , and Asf1 , Rtt109 and Rtt101 promotes the repair of replicative DNA damage – but not DSBs – by SCE [34] , [49] , [55] , suggesting that H3K56 acetylation might facilitate the repair of fork-associated DNA lesions other than DSBs by recruiting Rtt101 , which in turn would promote HR . This model , however , would not be valid for HU sensitivity , which is Rtt101 independent , and may be related with sustained replication under conditions of low levels of dNTPs . Besides , our comparative analysis shows that H3K56 acetylation mutants are slightly more sensitive to DNA damage than rtt101Δ , suggesting an additional function for this histone modification in response to replicative DNA damage . This role could be to open the chromatin and facilitate the access of repair proteins to DNA . Other possibility is that H3K56 acetylation promotes checkpoint deactivation via CAF1/Rtt106-chromatin assembly upon the repair of the replicative DNA damage , as previously demonstrated for DSB repair [39] , [40] . This is supported by the fact that cac1Δ rtt106Δ becomes temporally arrested at mitosis by sustained phosphorylation of Rad53 upon DNA damage release , even though this defect might also be a consequence of an incomplete accumulation of H3K56ac behind the fork of the double mutant . Our results in yeast anticipate a similar role for chromatin assembly in the stability of advancing replication forks through the more demanding chromatin structure of mammalian genomes . It will thereby be well worth the effort to address replication fork integrity in human cells defective in RC-chromatin assembly , which are known to arrest in the S phase and accumulate DNA damage [13] , [32] , [36] . Finally , the results presented here reveal the process of RC-chromatin assembly as a potential target against cell proliferation in cancer therapy , as also suggested by a recent observation showing that human Asf1b is overexpressed in breast tumours [76] .
Yeast strains used in this study are listed in Table 1 . They all are isogenic to BY4741 , except for H3K56R mutants that are isogenic to MSY421 . pRS316-SU [77] and pWJ1344 ( kindly provided by R . Rothstein , Columbia University ) are centromeric plasmids containing the SU inverted-repeat recombination system and RAD52-YFP , respectively . Yeast cells were grown in supplemented minimal medium ( SMM ) , except for nocodazole ( NCD ) synchronization that were grown in YPD medium [78] . For G1 synchronization , cells were grown to mid-log-phase and α factor was added twice at 1 . 5 hours intervals at either 0 . 5 µg/ml ( asf1Δ rad52Δ , cac1Δ rtt106Δ and asf1Δ cac1Δ rtt106Δ ) or 0 . 25 µg/ml ( rest of strains ) . Then , cells were washed three times and released into the S phase at different times in fresh medium with or without 0 . 2 M HU and 50 µg/ml pronase . Cell cycle progression was followed by DNA content analysis ( data not shown ) . To prevent cells from re-entering a new cell cycle in Figure 4B and 4C ( bottom ) , G1-synchronized cells were shifted to YPD with α factor for 1 hour and released into the S phase in fresh YPD medium with 50 µg/ml pronase and 15 µg/ml NCD . The frequency of Leu+ recombinants generated by recombination between inverted repeat sequences was determined in cells transformed with plasmid pRS316-SU by fluctuation tests as the median value of six independent colonies [77] . DNA damage sensitivity was determined by plating ten-fold serial dilutions from the same number of mid-log phase cells onto medium without or with genotoxic agents at the indicated concentrations . The proportion of budded cells with Rad52-YFP foci was performed as described previously [34] . Mid-log-phase cells transformed with pWJ1344 were visualized with Leica CTR6000 fluorescence microscope . DNA content analysis was performed by fluorescence-activated cell sorting ( FACS ) as reported previously [35] . The percentage of budded cells was determined by counting 200 cells at each time point . Each replication kinetic was conducted in parallel with the mutants and the wild type . Cell cultures were arrested with sodium azide ( 0 . 1% final concentration ) and cooled down in ice . Total DNA was isolated either in agarose plugs or with the G2/CTAB protocol as previously reported [42] , digested with restriction enzymes , resolved by neutral/neutral two-dimensional-gel electrophoresis as described [79] , blotted to nylon membranes and analysed by sequential hybridization of the same membrane with different 32P-labelled probes ( for probes along the ARS305 region see [42]; probe for ARS315 was PCR amplified with oligos AACAGCTTCTCTTGCCGTAG and TGTACTGAACCTACCGCTCC ) . All signals were quantified using a Fuji FLA5100 and ImageGauge as analysis program . Quantification of the RIs was normalized to the total amount of DNA , including linear monomers ( n ) , to the size of the restriction fragment , and to the percentage of cells synchronized in G1; thus , the total amount of RIs at each specific region and time point was calculated as [ΣRIs/Σ ( RIs+n×g ) ]×f , where f is the ratio between the size of the DNA fragment containing the origin and the size of the specific DNA fragment , and g is the proportion of cells in G1 after α-factor synchronization . Total DNA from mid-log phase cells synchronized in G1 and released into S phase in the presence of 0 . 2 M HU for 50 minutes was extracted and the amount of DNA at the origin ARS305 and a non-replicated control region ( located at ∼7 kb from the late replicating origin ARS609 ) determined by qPCR ( ARS305: oligos CGCCCGACGCCGTAA and GAGCGGCCTGAAATACTGTCA; control region: oligos TACACCAGCCCGGATTTAAG and GACCAGTGGCTGAGTCACAA ) . The efficiency of replication initiation was calculated as the ratio between the amount of DNA in HU-arrested cells and the amount of DNA in G1-arrested cells at the origin normalized to the same ratio at the control DNA region . Yeast protein extracts were prepared from mid-log-phase cultures using the TCA protocol as described [35] and run on a 8% and 10% sodium dodecyl sulfate-polyacrilamyde gel for western blot and in situ kinase assay , respectively . Rad53 was detected either with rabbit polyclonal antibody JDI47 [80] ( Figure 1E ) or with goat polyclonal antibody ( yC19 ) ( Santa Cruz Biotechnology , INC ) ( Figure 7C ) . The autophosphorylation reaction was performed as described [81] .
|
Loss of replication fork integrity is a primary source of genetic instability . In eukaryotes DNA synthesis is rapidly followed by its assembly into chromatin , and these two processes are tightly connected . Defective chromatin assembly mutants accumulate DNA damage and are sensitive to genotoxic agents , even though the mechanisms responsible for this genetic instability remain unclear because chromatin assembly also plays essential roles in transcription , silencing , DNA repair , and checkpoint signaling . A good example is the acetylation of histone H3 at lysine 56 , which promotes histone deposition by the chromatin assembly factors CAF1 and Rtt106 . In this case , the absence of this modification also causes a loss of structural and/or coding information at chromatin . Here we show that defective replication-coupled chromatin assembly leads to an accumulation of recombinogenic DNA damage by affecting the integrity of advancing , but not stalled , replication forks . Therefore , we propose that H3K56ac/CAF1/Rtt106-dependent chromatin assembly provides a mechanism for the stabilization of replication forks . Besides , H3K56 acetylation promotes replicative DNA damage repair/tolerance through a function that is independent of CAF1/Rtt106 and likely subsequent to its deposition at chromatin , revealing this modification as a key regulator of genome integrity .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"biology",
"molecular",
"cell",
"biology",
"genetics",
"and",
"genomics"
] |
2011
|
Histone H3K56 Acetylation, CAF1, and Rtt106 Coordinate Nucleosome Assembly and Stability of Advancing Replication Forks
|
Joint genetic models for multiple traits have helped to enhance association analyses . Most existing multi-trait models have been designed to increase power for detecting associations , whereas the analysis of interactions has received considerably less attention . Here , we propose iSet , a method based on linear mixed models to test for interactions between sets of variants and environmental states or other contexts . Our model generalizes previous interaction tests and in particular provides a test for local differences in the genetic architecture between contexts . We first use simulations to validate iSet before applying the model to the analysis of genotype-environment interactions in an eQTL study . Our model retrieves a larger number of interactions than alternative methods and reveals that up to 20% of cases show context-specific configurations of causal variants . Finally , we apply iSet to test for sub-group specific genetic effects in human lipid levels in a large human cohort , where we identify a gene-sex interaction for C-reactive protein that is missed by alternative methods .
Understanding genetic interactions with external context ( GxC ) , including environment , is a major challenge in quantitative genetics . Linear mixed models ( LMMs ) have emerged as the framework of choice for many genetic analyses , mainly because the random effect component in this class of models provides robust control for population structure [1 , 2] and other confounding factors [3–5] . More recently , random-effect models have also been shown to be effective to test for polygenic effects from multiple causal variants that are in linkage [6–9] ( variant sets ) . Additionally , multivariate formulations of LMMs have been developed to test for genetic effects across multiple correlated traits , both in single-variant analyses [10 , 11] and more recently for joint tests using variant sets [12] . However , these existing multivariate LMMs have primarily been designed to increase the statistical power for detecting association signals , whereas methods to test for interactions are only beginning to emerge [10 , 13] . Classical single-variant models for GxC use fixed effects to test for differential effect sizes of individual variants between contexts , either using an ANOVA [14–16] or LMMs [10 , 17] . The main advantages of set-based tests compared to single-variant models are twofold . First , set tests reduce the effective number of tests and can account for effects due to multiple causal variants , thus increasing power for detecting polygenic effects [7 , 8 , 12 , 18] . Second , we here show that joint tests across multiple contexts and sets of variants allow for characterizing the local architecture of polygenic-GxC interactions . One way to generalize single-variant interaction tests to variant sets is using a model that assumes that context differences cause the same fold-differences in effect size across all genetic variants , such that all genetic effects in one context are proportional to the effects in a second context; a criterion that has also been considered to assess co-localization of multiple traits [19] ( Fig 1A , middle ) . We denote this class of interactions rescaling-GxC . More generally , however , there may also be differences in the configuration of causal variants between contexts ( Fig 1A , right ) , such that not all genetic variants show the same fold-difference between contexts , as some become more prominent in particular contexts and others less so . We denote these complex interactions heterogeneity-GxC . These two classes of interactions have different functional implications–the former suggest no difference in causal variants between contexts , and the latter suggest otherwise . Distinguishing between them is only possible using multi-variant models such as set tests , and is important for identifying different potential causal variants in different contexts . We here propose a multivariate LMM to test for interactions test between Sets of genetic variants and categorical contexts ( iSet ) and to distinguish between rescaling-GxC and heterogeneity-GxC . We find that iSet yields increased power for identifying interactions and uniquely is able to robustly differentiate between rescaling-GxC and heterogeneity-GxC . We first validate iSet using simulations before applying the model to test for gene-by-sex interactions in blood lipid levels [20] as well as gene-by-environment interactions in an expression quantitative trait loci ( eQTL ) study [21] . We identify up to 20% of the stimulus-specific eQTLs as cases of heterogeneity-GxC , suggesting that context-specific causal variants are common .
iSet generalizes previous multi-trait set tests [12] , while considering the same trait measured in two ( environmental ) contexts . For a fully observed design , where the trait is measured in N individuals and each context , the phenotype matrix Y is modeled as the sum of a genetic effect from a set component and residual noise: Y=FB⏟fixedeffects+Us⏟setcomponent+ψ⏟noise . Here , F and B denote the design and the effect size matrices of additional fixed effect covariates and Us and ψ are random effects that follow matrix-variate normal distributions: Us~MVN ( 0 , Cs , Rs ) , ψ~MVN ( 0 , Cn , IN ) , where Rs corresponds to a local realized relatedness matrix [22] of the set of interest s , and IN denotes a diagonal covariance , which corresponds to independent and Identically distributed residuals . The trait-context covariance matrices Cs and Cn model correlations between contexts due to the set component ( Cs ) , and residual noise ( Cn ) . A key insight derived here is that different assumptions on the structure of the trait-context covariance CS correspond to alternative genetic architectures that can be explained by a polygenic model ( Fig 1B , Methods ) . Persistent genetic effects across contexts ( no GxC ) can be modeled using an LMM with a constant block covariance ( Fig 1B ) ; rescaling GxC , where effect sizes in different contexts are proportional to each other , can be captured by a trait-covariance with rank one . Note that genetic effects that act only in one context are a special case of this model and corresponds to a zero-rescaling coefficient . Finally , the most general architectures with different relative effect sizes between contexts ( heterogeneity-GxC ) can be captured by an LMM with a full-rank trait-context covariance ( Methods ) . By comparing LMMs with these alternative covariance structures , it is possible to define set tests for general associations ( mtSet ) , which identifies both persistent and context-specific effects , a test for genetic interactions , both with or without changes in the configuration of causal variants ( iSet ) , and finally a test for heterogeneity-GxC effects ( iSet-het ) , which is specific to differences between contexts that cannot be explained by rescaling ( Fig 1B ) . These multivariate LMMs can be fit using principles that were previously derived for multivariate set tests [12] , and hence , provided the computations are suitable parallelized ( Methods ) , iSet can be applied to large cohorts with tens of thousands individuals ( S1 Fig ) . Permutation schemes are not well defined for interaction models [23] , so we use a parametric bootstrap procedure [23] to estimate P values . An important advantage compared to previous interaction tests [13 , 24–28] ( Methods ) , is that iSet can be applied both to study designs where all individuals have been phenotyped in each context and when stratifying populations into distinct subgroups using a context variable ( S2 Fig ) . iSet also provides control for population structure , either using principal components that are included as fixed covariates , or using an additional random effect term ( Methods ) . Finally , iSet can also be used to estimate the total phenotypic variance explained by variant sets and the relative proportions captured by persistent , rescaling-GxC and heterogeneity-GxC effects ( Methods ) . To illustrate the polygenic interactions that can be detected using iSet , we first considered a basic simulated example ( Fig 1C ) . We simulated genetic effects for one quantitative trait in two contexts , considering polygenic effects at three distinct loci ( Methods ) : a region with persistent genetic effects , a region with rescaling-GxC and a region with heterogeneity-GxC effects . We tested consecutive regions ( 30kb region , 15kb step ) using the three tests provided by our model ( mtSet , iSet , iSet-het ) , finding that by combining these results , it was indeed possible to resolve the architecture of each of the simulated regions ( Fig 1C and 1D ) . In particular , this example illustrates that , unlike single-variant tests , iSet-het can be used to discern heterogeneity-GxC effects specifically . Next , we used simulations based on genotypes from the 1000 Genomes project [29] to assess the statistical calibration and power of iSet . We generated a population of 1 , 000 individuals based on genotype data from European populations , initially simulating one quantitative trait measured in two distinct contexts in all individuals ( Methods ) . First , we considered data with simulated persistent polygenic effects , confirming that both iSet and iSet-het are calibrated when no interaction effects are simulated ( Fig 2A , S1 Table ) . Analogously , we also confirmed that iSet-het is calibrated when only rescaling-GxC effects are considered ( S3 Fig ) , and we assessed the robustness of iSet to different types of model misspecification ( S4 Fig , S1 Table ) . We compared iSet to single-variant interaction tests [10] ( mtLMM-int ) ( Methods ) , considering a wide range of different settings ( S2 Table , Methods ) . Because single-variant methods perform one test for each variant in the set ( S5 Fig ) , we adjusted for multiple testing using one of two approaches: i ) conservative Bonferroni adjustment ( Bonferroni ) or ii ) a recently proposed method that estimates the effective number of independent tests based on the local structure of linkage disequilibrium ( LD ) [30] ( eigenMT ) . Note that existing set-based interaction tests cannot be applied to complete designs with repeat measurements and hence were not considered ( Methods , S1 Text ) ; see below and Fig 5 for additional experiments where these methods were used . As expected , the power advantages of iSet compared to single-variant models were largest when multiple causal variants were simulated ( Fig 2B , for constant total genetic variance , Methods ) . However , iSet was better powered than mtLMM-int even for a single causal variant . Identical simulations based on synthetic independent genotypes ( S6 Fig ) revealed that this effect is predominantly due to local LD and advantages due the reduced number of total tests . We also considered the impact of different proportionality factors of genetic effects between contexts . All models were best powered to detect GxC for negative proportionality factors ( opposite effects ) , or when the proportionality factor was close to zero ( context-specific effects ) ( Fig 2C ) . Next , we simulated traits with context-specific causal variants ( heterogeneity-GxC ) . Heterogeneity-GxC is detectable when there is a change in the local causal configuration , which corresponds to the absolute correlation of local genetic effects between contexts ( r ) smaller than 1; the greater the heterogeneity GxC effects , the smaller the absolute correlation . Presence of GxC effects under tightly correlated genetic effects ( r ≈ ±1 ) cannot be distinguished from rescaling-GxC . To simulate these different settings , we randomly selected two causal variants in each context and varied the extent of correlations of the genetic effect between contexts ( Fig 2D ) . When using iSet-het for detecting heterogeneity-GxC effects , the model was best powered when there is a moderate to large change in causal configuration , corresponding to low correlated genetic effects ( >70% power for r2 < 0 . 16 , Fig 2D ) . We also considered additional settings with larger numbers of causal variants ( S8 Fig ) , and we assessed the accuracy of iSet-het to classify interaction effects into heterogeneity-GxC or rescaling-GxC effects ( S7 Fig , Methods ) . Taken together , these results confirm that iSet-het is a robust test for heterogeneity-GxC . We also investigated the proportion of local genetic variance that can be explained by models with persistent , rescaling-GxC and heterogeneity-GxC for the corresponding simulations ( Fig 2C and 2D , Methods ) . The persistent effect model explained large proportions of the simulated genetic variance , even in the presence of positively correlated GxC , but could not capture variance due to GxC effects with negative rescaling ( Fig 2C and 2D ) . An LMM that models rescaling-GxC did account for negative and positive rescaling , and captured some of the heterogeneity-GxC effects ( Fig 2D ) . Finally , variance contributions that were exclusively captured by a heterogeneity-GxC model were largest for uncorrelated context-specific genetic effects , the same regime where the corresponding test is best powered ( Fig 2D ) . We also confirmed that the most flexible heterogeneity-GxC model yields unbiased estimates of the total genetic variance in genomic regions , whereas other models were biased for some simulated architectures ( S9 Fig ) . Finally , we considered simulations where we varied both the size of the testing region and the simulated causal region , using a sliding window analysis ( S10 Fig , Methods ) . Overall , iSet was markedly robust to the window size , and was best powered when the sizes of the testing region approached the size of the simulated causal region , which is in line with previous findings for set-based association testing [12] . We also observed that iSet-het is best powered for small causal regions ( up to 100kb ) , and the power for detecting heterogeneity-GxC deteriorated when analyzing larger regions . We next applied iSet to test for stimulus-specific genetic effects in a monocyte stimulus eQTL study [21] . We considered gene expression profiles for 228 individuals in four stimulus contexts: naive state ( no stimulation ) , stimulation with interferon-γ for 24 hours ( IFN ) , and stimulation with lipopolysaccharide ( LPS ) for two and 24 hours . We applied iSet to test for pairwise interaction effects , considering the naive monocyte state and each stimulus condition in turn , performing a single test using proximal cis acting variants ( plus or minus 50kb from the transcription start site; Methods ) . After quality control , we considered 12 , 677 probes and tested for cis associations ( mtSet ) , GxC interactions ( iSet ) and for heterogeneity-GxC effects ( iSet-het ) . For comparison , we also considered a conventional multi-trait LMM [10] and tested for associations and interactions in the same genomic regions , using eigenMT [30] to adjust for multiple testing ( Methods ) . Although there was substantial overlap of the probes and stimulus conditions for which different methods identified significant interactions ( Fig 3B ) , iSet was better powered ( 32 . 7% power increase; 5 , 068 versus 3 , 818 probes and stimuli with an interaction; FDR<5% , Fig 3A , S11 and S12 Fig , S4 Table ) . Additionally , iSet-het identified 1 , 135 probes and stimulus contexts with significant heterogeneity-GxC effects ( Fig 3A and 3B ) . This shows that a substantial proportion of stimulus-specific eQTLs are associated differences in the configuration of causal variants , suggesting context-specific regulatory architectures . Although on average the proportion of variance explained by GxC tended to be smaller than for persistent effects ( median 3 . 7% for GxC versus median 9 . 5% for persistent effects , for probes with significant GxC , Fig 3C ) , GxC was the driving genetic source of variation for 11 . 8% of the significant cis eQTLs ( Fig 3D; defined as explaining 50% or more of the cis genetic variance ) . Consistent with previous reports [31 , 32] , we observed that genes with large relative GxC effects were associated with weak overall cis effects , whereas eQTLs with large effect sizes tended to be persistent across stimuli ( Fig 3D ) . To better understand the mechanisms that underlie genes with detected heterogeneity-GxC effects , we used an LMM with step-wise selection [33] , identifying 15 , 756 , 2 , 690 and 457 eQTLs ( across all probes and contexts ) with a single significant association , significant secondary and significant tertiary associations respectively ( FDR < 5% , Methods , S4 Table ) . Probes with significant heterogeneity-GxC were more likely to harbor multiple independent associations ( Fig 4A ) , confirming that heterogeneity-GxC eQTLs have complex genetic architectures . When overlaying heterogeneity GxC eQTLs detected using iSet-het with the results obtained from the single-variant step-wise LMM , we could attribute 46 . 2% of the heterogeneity-GxC effects ( 524 out of 1 , 135 ) to context-specific lead variants ( defined using r2<0 . 2 , FDR<5% , Fig 4B and 4C , Methods ) . For an additional 14 . 6% of the heterogeneity eQTLs ( 166/1 , 135 ) the lead variants from a single-variant analysis were in high LD ( r2>0 . 8 ) , with context-specific secondary effects ( Fig 4B and 4D ) . The remaining 445 heterogeneity eQTLs ( 39 . 2% ) could not be annotated using single-variant models . One reason why heterogeneity GxC effects cannot be annotated using a single-variant model are differences in power . Indeed , for 22 . 2% of the heterogeneity-GxC cases without a single-variant interpretation ( 99/445 ) , the single-variant LMM did not yield a significant effect in either of the two contexts ( S13A Fig ) . For an additional 58 . 2% of the unannotated heterogeneity GxC effects ( 259/445 ) , the single-variant LMM lead variants were in weak linkage ( 0 . 2<r2<0 . 8 example in Fig 4E ) , which neither confirms nor rules out distinct genetic effects . One explanation for these instances are distinct polygenic architectures in both contexts . Consistent with this possibility , we observed that genetic effects captured by a polygenic model in both contexts ( best linear unbiased predictor , Methods ) were markedly less correlated for probes with significant heterogeneity-GxC ( S13B and S13C Fig , Methods ) . Finally , we explored the relationship between probes with heterogeneity GxC and opposite effects as defined using conventional single-variant models . We classified associations as opposite effects when context-specific lead variant were in high LD ( r2>0 . 8 ) and the effect on gene expression was in different directions ( Methods ) . This approach identified 67 eQTLs with reversed effect directions between contexts . iSet-het detected significant heterogeneity-GxC for 8 of these eQTLs , a 2 . 2 fold enrichment ( P<5e-2 ) compared to eQTLs with consistent effect directions between contexts ( 238 gene/stimulus pairs with significant heterogeneity-GxC out of 4 , 119 eQTLs with consistent direction , Fig 4F ) . Similar enrichments were also observed when considering individual stimulus contexts , resulting in significant enrichments for two out of three stimulus contexts ( P<5e-2 , fold change>4 in naïve/IFN and naïve/LPS-24h , S5 Table ) . Among the genes with significant heterogeneity-GxC are OAS1 , LMNA and PTK2B , opposite-effect eQTLs that have been reported in the primary analysis of the same data [21] ( S14 Fig ) . Thus far , we have considered settings with repeat measurements , where the same phenotype is measured in all individuals and contexts . Next , we considered applications of iSet to studies where individuals are phenotyped in only one of the two contexts ( S2 Fig , Methods ) . This is a common strategy in investigation of genotype-context interactions , where a population is stratified using a context variable . We considered simulations analogous to those for complete designs ( Fig 2 ) to validate iSet for this design . We again confirmed statistical calibration of iSet ( S15A Fig ) and found similar power benefits as for complete designs ( Fig 5A and 5B , S15B and S15C Fig ) . In addition to single-variant LMMs , we also compared to a recently proposed set test for interactions ( GESAT; [13] ) , which is designed for stratified populations . Notably , iSet was consistently better powered than GESAT , most likely because GESAT does not model correlations of the local genetic effect between contexts ( Methods , S1 Text ) . Next , we applied iSet to test for genotype-sex interactions in four lipid-related traits ( fasting HDL and LDL cholesterol levels , triglycerides and C-reactive protein ) measured in 5 , 256 unrelated individuals from the Northern Finland Birth Cohort ( NFBC1966 [20] ) . We tested consecutive 100kb regions ( step size 50 kb; 52 , 819 genome-wide tests ) , and compared iSet to GESAT and the single-variant interaction test ( Methods ) . iSet retrieved one genome-wide significant interaction ( C Reactive protein , chr1:40 , 450 , 000; P = 1 . 47x10-6; FWER<10% ) , whereas alternative set tests and the single-variant models did not yield significant effects ( Fig 5C , S16 and S17 Figs , S6 Table ) , even when using dense genotypes derived using imputation strategies ( S18 Fig ) . This interaction was located within 400kb of an interaction identified in a large meta study ( 66 , 185 individuals [34] ) , which reports both an association for C-reactive protein at the same locus ( P<6x10-11 ) as well as a nominally significant interaction with sex ( P<5x10-3 ) . Finally , a local single-variant analysis , separately for female and male individuals , provided evidence that this interaction reflects a male-specific genetic effect ( S19 Fig ) . iSet revealed a second suggestive interaction with sex for LDL cholesterol levels ( chr3:121 , 850 , 000 , S16 Fig ) . Although this effect failed genome-wide significance ( FWER<20% ) , iSet again yielded stronger evidence than other methods ( PiSet = 3 . 7x10-6 , PGESAT = 4 . 8x10-6 , PmtLMM-int = 3 . 2x10-5 ) . Among the genes at this locus is ADCY5 , which has been linked to blood glucose levels in large meta-analyses [35 , 36] and hence is a plausible candidate to affect LDL via glucose regulation [37] . Finally , we note that context stratification of quantitative traits can increase power for detecting associations rather than interactions , which is similar to previous strategies applied for single-variant analyses of quantitative [38] and categorical traits [39 , 40] . Using this generalized association test , we identified three additional associations that were missed by conventional set tests and other methods ( S16 Fig , S6 Table ) . These include the same locus with a sex-specific effect on C-reactive protein ( chr1:40 , 450 , 000 , P = 1 . 42x10-7 using mtSet , P = 1 . 89 x10-3 using a standard set test ) , and two associations for HDL cholesterol levels and triglycerides , both of which were replicated in larger meta analyses [41] .
We have here proposed iSet , a method based on linear mixed models to test for gene-context interactions using variant sets . On simulated data as well as in applications to gene expression and human lipid-related traits , we have demonstrated that iSet yields increased power and improved interpretation for interaction effects compared to previous methods . Methods for the joint analysis of multiple traits , including tests for genetic interactions , are not new per se . Most previous studies have used set-based methods to test for associations [7 , 8 , 12 , 18] , whereas tests for genotype-context interactions are still primarily carried out using single-variant models [10 , 17] . iSet unifies several previous models ( Methods ) , and uniquely offers set-based interaction tests on phenotypes in different contexts under the same or different ( stratified ) genetic backgrounds . Additionally , we have shown that set-based interaction tests can be useful to disentangle the genetic architecture of such loci , discerning consistent changes of genetic effects between contexts ( rescaling-GxC ) and changes in the configuration of causal variants ( heterogeneity-GxC ) . The heterogeneity GxC test we propose is related to co-localization tests [19 , 42 , 43] , however with a different objective . In applications to a stimulus eQTL study , iSet has yielded increased power compared to single-variant tests ( Fig 3A and 3B ) , and we have shown that approximately 20% of the gene-stimulus interactions are associated with significant heterogeneity-GxC . This suggests that changes in the genetic architecture between stimulus contexts are relatively common . Additionally , we have observed that genes with opposite effects are enriched for heterogeneity-GxC . This finding points to a possible bias whereby opposite effects identified using single-variant models may in part be due to context-specific causal variants that are LD-tagged by a shared lead variant . Notably , although iSet is better powered overall , there may be true interactions that can be detected using single-variant models and are missed by set tests ( Fig 3B , S11 and S12 Figs , S4 Table ) . Hence , iSet should be considered as a complementary method and not as a replacement of conventional single-variant tests . The proposed iSet model is not free of limitations . First , scalable inference in our model is achieved by exploiting the low-rank structure of variant sets , meaning that the number of variants in the analyzed region is typically small compared to the number of individuals . Similar to previous set-based tests [12] , there are trade-offs between power and resolution , in particular when analyzing data from densely imputed or sequenced cohorts . General strategies for the design of optimal testing regions , for example using genome annotations and LD information , are an important area of future work . iSet is computationally efficient in cohorts with fully observed designs , or when stratifying a cohort using a context variable . Intermediate designs , for example in fully observed designs with missing phenotypes , may also be considered , however currently require the use of separate imputation schemes [11 , 44] . It is also worth noting that the test for heterogeneity-GxC ( iSet-het ) will be most accurate if all individuals are phenotyped in each context . Although in principle the model can also be used in stratified designs , there may be concerns that false positive heterogeneity GxC effects can arise due to technical factors , for example due to differences in genotyping accuracy or variant allele frequencies in the corresponding sub populations . A related issue is the need to choose the size of the region-set appropriately . While we find that the model is overall robust across a wide range of region sizes ( S10 Fig ) , the model will be best powered if the size of true causal regions approximately matches the testing region size , in particular for identifying heterogeneity-GxC effects . Finally , we have here focused on pairwise analyses of different contexts . In principle , the model could also be applied to analyze multiple related context and different traits , and the model could be extended to handle continuous environmental states , which currently require discretization . A related extension of the model is to test for genetic effects that are exclusive to one of the considered contexts . Developments in these directions are future work .
iSet is freely available as part of the LIMIX package ( https://github . com/limix/limix ) . Tutorials for using iSet either as command line tool or via a Python API can be found at https://github . com/limix/limix-tutorials/tree/master/iSet . To derive the model , we start assuming a fully observed design , where phenotypic measurements are available for all individuals and for each context . Briefly , the N × C phenotype matrix Y for N individuals and two or more contexts ( C ) is modeled as sum of fixed effects of K covariates , effects from S genetic variants in the region of interest ( set component ) and residual noise: Y=FB⏟fixedeffects+GW⏟setcomponent+ψ⏟noise . ( 1 ) Here F ( N × K ) and G ( N × S ) denote respectively the fixed-effect covariates and the standardized genotypes of the variant set and B ( K × C ) and W ( S × C ) denote the corresponding effect sizes . The noise component ψ is assumed to follow a matrix-variate normal distribution , ψ∼MVN ( 0 , Cn , IN ) , where Cn is a C × C covariance matrix that models residual covariances between traits . Note that in this formulation , population structure can be accounted for by including the leading principal component of the N × N ( global ) realized relatedness matrix [22] into the model as fixed effects [12] . In human populations , 10–20 principal components are typically sufficient to adjust for such structure [45] . Note that iSet can also account for population structure using an additional random effect term into the model ( see S1 Text ) . While computationally more expensive , this approach provides for additional robustness and calibration when analyzing cohorts with related individuals ( see [12] for a discussion ) . All experiments reported here have been carried out using adjustment based on principal components , considering 10 PCs . Simulations were carried out using a synthetic cohort of 1 , 000 individuals derived from genotypes of European populations in the 1000 Genomes project [29] ( phase 1 , 1 , 092 individuals , 379 Europeans ) . Following [12 , 47] , we composed synthetic genotypes as a mosaic of real genotypes from individuals of European ancestry , while preserving population structure ( S1 Text ) . We considered single-nucleotide polymorphism with a minor allele frequency of at least 2% ( S4 Fig ) . In all simulations , we simulated two contexts , modeled as the sum of a genetic contribution from a 30kb causal region , effects due to population structure , hidden covariates and identically distributed Gaussian noise . Effects due to population structure and hidden confounders were simulated with partial correlations across contexts , explaining variable proportions of the total phenotypic variance in each context ( S2 Table , S1 Text ) .
|
Genetic effects on phenotypes can depend on external contexts , including environment . Statistical tests for identifying such interactions are important to understand how individual genetic variants may act in different contexts . Interaction effects can either be studied using measurements of a given phenotype in different contexts , under the same genetic backgrounds , or by stratifying a population into subgroups . Here , we derive a method based on linear mixed models that can be applied to both of these designs . iSet enables testing for interactions between context and sets of variants , and accounts for polygenic effects . We validate our model using simulations , before applying it to the genetic analysis of gene expression studies and genome-wide association studies of human blood lipid levels . We find that modeling interactions with variant sets offers increased power , thereby uncovering interactions that cannot be detected by alternative methods .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"c-reactive",
"proteins",
"immune",
"cells",
"protein",
"interactions",
"built",
"structures",
"engineering",
"and",
"technology",
"immunology",
"random",
"variables",
"covariance",
"simulation",
"and",
"modeling",
"mathematics",
"research",
"and",
"analysis",
"methods",
"white",
"blood",
"cells",
"animal",
"cells",
"proteins",
"genetic",
"loci",
"probability",
"theory",
"biochemistry",
"cell",
"biology",
"monocytes",
"phenotypes",
"genetics",
"biology",
"and",
"life",
"sciences",
"structural",
"engineering",
"physical",
"sciences",
"cellular",
"types"
] |
2017
|
Joint genetic analysis using variant sets reveals polygenic gene-context interactions
|
Bacteria can survive antibiotic treatment without acquiring heritable antibiotic resistance . We investigated persistence to the fluoroquinolone ciprofloxacin in Escherichia coli . Our data show that a majority of persisters to ciprofloxacin were formed upon exposure to the antibiotic , in a manner dependent on the SOS gene network . These findings reveal an active and inducible mechanism of persister formation mediated by the SOS response , challenging the prevailing view that persisters are pre-existing and formed purely by stochastic means . SOS-induced persistence is a novel mechanism by which cells can counteract DNA damage and promote survival to fluoroquinolones . This unique survival mechanism may be an important factor influencing the outcome of antibiotic therapy in vivo .
Persistence is the ability of a subpopulation of susceptible bacteria to survive lethal doses of antibiotics . It is a transient and non-hereditary phenotype unlike resistance , which is due to genetic modification . The transient nature of persistence makes it inherently difficult to study therefore the underlying molecular mechanisms are still poorly understood . Persisters are thought to be slow growing , non-growing or dormant cells , which escape the lethal action of antibiotics because their drug targets are inactivated due to the physiological state . In an Escherichia coli high-persistence mutant , persisters to ampicillin were shown to be non-growing prior to the addition of the antibiotic [1] . In addition , a fraction of non-growing cells was isolated from untreated exponentially growing E . coli and was shown to be enriched in persisters to ofloxacin [2] . These studies demonstrated that persisters can form independently of antibiotics . The switch from growing to non-growing state or dormancy , is thought to be a purely stochastic process [1] , [3] , [4] . Both genetic and phenotypic variability can have important consequences on bacterial survival of antibiotic treatment . One of the most prescribed broad spectrum antibiotics today are the fluoroquinolones ( FQ ) , which target gyrase and topoisomerase . These essential enzymes regulate supercoiling of genomic DNA during replication and transcription [5] , [6] . FQs prevent ligation reactions of gyrase and topoisomerase resulting in double-strand breaks ( DSB ) [7] . DSBs are potentially lethal DNA lesions that occur under physiological conditions through collapse of stalled replication forks , overlapping repair tracts , spontaneous breakage of DNA , and other mechanisms . E . coli efficiently repairs DSBs through a series of reactions carried out by enzymes participating in homologous recombination and replication [8] . Processing of DSBs leads to the induction of the SOS response . SOS is a complex network composed of more than 40 genes [9] , [10] . Many of these genes are essential for efficient repair of various DNA lesions , including DSBs [11] , [12] . Even though fluoroquinolones are potent bactericidal antibiotics they cannot sterilize a bacterial culture . The bulk of the population rapidly dies in response to fluoroquinolones but a small fraction persists . According to one model , persisters might survive if gyrase and topoisomerase are inactivated due to cellular dormancy [3] . Dormant cells might be expected to form stochastically during growth of a culture , prior to the antibiotic exposure [2] , [13]–[15] . Alternatively , the persister state might be inducible in a cell subpopulation by exposure to the antibiotic , not stochastic and pre-existing . This could be because either dormancy is inducible , or persisters might be active and have more efficient drug efflux or more efficient repair of DSBs due to the stochastic overexpression of the genes involved in those pathways or due to the physiological events leading to the activation of the same pathways . In order to distinguish between these possibilities , we measured numbers of persisters to the fluoroquinolone ciprofloxacin in various genetic backgrounds with altered capacity for SOS induction and DSB repair . The majority of persisters were found to be formed upon exposure to the antibiotic and formation was dependent on the SOS DNA damage response . Contrary to the current view , a majority of surviving persisters to ciprofloxacin are not pre-existing , but induced by this antibiotic .
Fluoroquinolones ( FQ ) induce DSBs by interfering with the action of gyrase and topoisomerase [16] . The cellular response to DSBs primarily consists of induction of the SOS-regulon and ultimately in repair through recombination [17] , [18] . According to the prevailing model [3] , persisters are dormant and are formed stochastically prior to the addition of antibiotic . This suggests that persisters would not experience DSBs , would not induce an adaptive response to that type of lesion , and therefore would not need repair functions to survive . In order to test these predictions , we wanted to determine whether persisters experience DSBs and induce SOS . We measured the persister levels in different genetic backgrounds diagnostic of specific molecular events linked with DSBs and SOS induction . The surviving fraction of a wild-type culture treated with ciprofloxacin produces a typical biphasic pattern ( Figure 1A ) . This reflects the rapid killing of the bulk of the cells , and a surviving persister subpopulation . We examined some of the well-known DNA-repair pathways in order to probe their possible role in formation of persisters . RecA and RecBC are essential for repair of DSBs in E . coli [19] . In strains lacking RecA and RecBC , DSBs are lethal . As expected , the bulk of cells is more rapidly killed in both recA and recB backgrounds , compared to the wild type , presumably because DSBs could not be repaired ( Figure 1A ) . However , the persister fraction was also greatly reduced ( 40-fold in recA , 35- to 103-fold in recB ) . In recB , persisters were extremely rare or entirely absent after 6 hours of incubation . This shows that the persisters experience DSBs and hence depend on the repair functions . RecA and RecB functions are essential not only for DSB-repair but for SOS induction following processing of DSBs as well [18] , [20]–[22] , so in order to test whether persisters induce SOS we constructed strains unable to induce SOS but proficient for homologous recombination; one carrying a non-inducible SOS-repressor ( lexA3 ) [23] and the other a mutant RecA able to function as a recombinase but unable to induce cleavage of LexA ( recA430 ) [24] . In both backgrounds the bulk of cells dies more rapidly than in the wild type , confirming that SOS is efficiently induced following exposure to ciprofloxacin and contributes to the survival ( Figure 1B ) . Interestingly , the persister level is decreased 43-fold and by 6 hours it is as low as in recA background . This shows that the persistence to ciprofloxacin is largely dependent on a functional SOS response . XerCD site-specific recombinase resolves chromosome dimers at a dif site [25]–[27] . Chromosome dimers are formed by an odd number of recombination events . The absence of xerCD function does not affect the proficiency for SOS induction , but is lethal in cells in which chromosme dimers have formed . In xerC and xerD mutants the persister level is reduced ( 7- and 9-fold , respectively , taking into account the 3-fold reduction in viability of xerC and xerD mutants compared to the wild-type ) , suggesting that most persisters have undergone at least one successful recombination event , most likely repairing a ciprofloxacin-induced DSB ( Figure 1C ) . Taken together these results show that the formation of the majority of persisters in the presence of ciprofloxacin is dependent on the SOS-response . They also suggest that this antibiotic-tolerant state is induced , rather than pre-existing . The formal possibility that an SOS controlled function is essential for reaching or exiting a pre-existing multidrug-tolerant state can be ruled out because tolerance to ampicillin and streptomycin were not affected in recA , recB or lexA3 strains ( Figure 1D ) . We cannot rule out the possibility that spontaneous SOS induction was required for creating a pre-existing ciprofloxacin-tolerant state . A strain lacking the SOS-inducible RecN protein is also SOS proficient but partially deficient in DSB repair . recN mutant exhibited increased sensitivity of the bulk whereas persistence was largely unaffected ( Figure 1F ) . The entire population exposed to ciprofloxacin is expected to induce SOS yet only a fraction survives . SOS is a gradual response where the strength of induction reflects the extent and the persistence of the damage [28] , [29] . Upon addition of ciprofloxacin the number and the chromosomal location of DSBs will vary across the population depending on the activity and position of gyrase and topoisomerase molecules in any given cell . The resulting distribution of DSBs is expected to translate into a gradient of SOS induction . Therefore , it is possible that a specific level of SOS induction is required for persistence . If this is the case , persister levels are expected to change along with ciprofloxacin concentration and the overall level of SOS induction . We measured both the persister level and the induction of β-galactosidase under the control of an SOS-inducible recA promoter [30] as a population average of SOS induction for cultures exposed to increasing concentration of ciprofloxacin . Indeed , as shown in Figure 2 , increased concentration of ciprofloxacin led to an increased average SOS induction ( Figure 2A ) and decreased persister level ( Figure 2B ) . A strain constitutively expressing SOS functions ( lexA300 ( Def ) ) , also led to a 20-fold increase in persister level compared to the wild type ( Figure 1E ) . In order to examine a difference in SOS induction between persisters and the bulk at the single cell level , we followed a cI-cro gal reporter strain after addition of ciprofloxacin [31] . In this strain the cleavage of λ repressor CI leads to a heritable genetic switch rendering a cell gal+ . gal+ cells can be detected as red colonies on MacConkey galactose plates . Unlike LexA , which undergoes auto-cleavage early in SOS induction , CI cleavage occurs only if there is a high level of DNA damage and activated RecA [32] . Therefore , the cI-cro system reports conditions of only strong SOS induction . Following the addition of ciprofloxacin ( >0 . 5 µg/ml ) the proportion of cells giving rise to gal+ colonies increases , peaking at around 20 minutes and declines thereafter ( Figure 3 ) . This timing means that the massive amount of DNA damage occurs readily leading to a strong SOS induction . Cells undergoing strong SOS induction are able to withstand and repair the damage , if the ciprofloxacin is removed by plating . However , upon extended exposure gal+ cells become fewer ( Figure 3 ) , indicating that additional damage occurs and eventually becomes lethal . The persister subpopulation consisted almost entirely of gal− cells ( Figure 3 ) showing that persisters were not SOS-induced prior to ciprofloxacin treatment ( at least not highly induced ) and also that they did not experience high level of DNA damage nor strong SOS induction even in the presence of the antibiotic . Because the persister level is greatly reduced in strains unable to induce SOS ( lexA3 , recA430 , Figure 1B ) we conclude that persisters undergo weak SOS induction . This is in contrast to the bulk of cells , probably because fewer DSBs occur in eventual persisters . Increased sensitivity of the bulk and minimally affected persistence in a recN strain also supports this conclusion ( Figure 1F ) . SOS-inducible RecN protein promotes efficient repair of DSBs . While it is dispensable for the repair of a single break , it is essential for the repair of simultaneous multiple DSBs [33] . Next we exposed cells treated with a range of ciprofloxacin concentrations to a higher dose ( 1 µg/ml ) of the same antibiotic ( Figure 4 ) . Control cultures were exposed to 1 µg/ml of ciprofloxacin for the duration of the experiment . The persister fraction surviving exposure to 1 µg/ml was 10- to 40-fold higher in the cultures pretreated with a low concentration of ciprofloxacin ( 0 . 05–0 . 2 µg/ml ) , compared to the control ( Figure 4B ) . A dramatic , 1200-fold increase was found in cultures pretreated with sub-MIC ( minimal inhibitory concentration ) concentration of ciprofloxacin ( Figure 4B; compare full bar at 0 . 03 µg/ml and the second dashed bar of the control ) . This shows that many of the persisters are formed upon ciprofloxacin treatment rather than pre-existing . If they were pre-existing , the fraction surviving the exposure to the high concentration of ciprofloxacin ( 1 µg/ml ) would be the same regardless of the pretreatment . It was important to learn whether SOS induction caused by treatments other than FQ is able to induce persistence to ciprofloxacin . In order to test this we measured persistence to ciprofloxacin in cells exposed to mitomycin C . Mitomycin C interacts with DNA by intercalation and adduct formation , resulting in inter-strand crosslinks [34] . The cellular response is a potent SOS-induction dependent on RecFOR pathway [35] . We exposed exponentially growing cells to a sub-MIC concentration of mitomycin C and compared the persister levels at two different time points during the treatment . The results in Figure 5 show a 180-fold increase in persistence to ciprofloxacin in the culture treated with mitomycin C for 4 versus 2 hours , confirming the link between SOS induction and persistence to FQs , irrespective of the nature of the SOS inducing treatment . Persister levels are very low in early exponential phase and are maximal in stationary phase [14] . We treated aliquots of growing cultures with ciprofloxacin at different time points in order to determine the persister levels between these two extremes and establish the role of growth phase in SOS-induced persistence . Figure 6 shows an exponential increase in persister levels when cell densities reach around 5×107 CFU/ml in both the wild-type and the strain unable to induce SOS ( lexA3 ) . We conclude that the SOS-induced persisters make up the majority of persisters to ciprofloxacin regardless of the growth phase .
The processes leading to genetic variability in bacteria , mutagenesis and recombination , have been studied extensively [36]–[38] and their role in evolution of bacterial antibiotic resistance by generating and disseminating mutations is well established [39]–[46] . On the other hand , processes leading to phenotypic variability , which is also an important factor influencing bacterial ability to survive antibiotic treatments [47] , [48] have only recently become a subject of systematic investigation . In contrast to the well-understood mechanisms of bacterial resistance to antibiotics , molecular mechanism ( s ) of persistence have so far remained elusive . The current model of persistence assumes that persisters are non-growing or dormant cells , formed by stochastic process ( es ) independently of any physiological responses normally elicited by antibiotics [1] , [3] , [4] . Studies involving persistence to two different classes of antibiotics , a b-lactam ampicillin [1] and a fluoroquinolone ofloxacin [2] are consistent with this model , which was therefore presumed to hold universally . Here we show a mechanism of persister formation triggered by DNA damage inflicted by the fluoroquinolone ciprofloxacin . Formation of persisters in response to DNA damage reveals a deterministic component in this bistability phenomenon . Bistability is the stochastic production of two phenotypically distinct cell types within a clonal population of genetically identical kin cells . Bistability is observed in sporulation , competence , and motility [49]–[51] . In all cases studied , there is both a stochastic and deterministic component of bistability . Previous studies have shown that persisters can form stochastically , prior to the addition of antibiotics [1] , [14] . The present findings show that persister formation can also be induced by an antibiotic , through an active process . This sheds an entirely new light on the problem of antibiotic tolerance and its role in infectious processes . We also show that mutants defective in persistence to ciprofloxacin have normal persister levels to amipicillin and streptomycin ( Figure 1D ) , therefore it is still possible that persistence to β-lactams is purely stochastic and not inducible . These results suggest that there are different mechanisms of persistence to different antibiotics . Ciprofloxacin induces DSBs in cells with active gyrase and/or topoisomerase , which in turn leads to the activation of the general DNA damage stress response , the SOS gene network . Our results show that the majority of persisters to ciprofloxacin are dependent on a functional SOS response . DSBs and other SOS-inducing lesions occur under physiological conditions so at any given time there is a fraction of a bacterial population undergoing a certain degree of SOS induction [52] , [53] . However , we demonstrate that the SOS-dependent persister state is induced upon exposure to ciprofloxacin . Manipulating the extent of SOS induction by different antibiotic concentrations or by sequential exposure to a higher dose dramatically affects persister levels ( Figure 2 , Figure 4 ) . This would not be the case if the persisters were pre-existing in the population . If they were , the bulk would be killed by any bactericidal concentration of antibiotic , revealing the same pre-existing persister population . In addition , increasing the basal level of expression of the SOS regulon by genetic manipulation ( Figure 1E ) or by induction with different treatment ( Figure 5 ) also leads to an increase in persister level . Essentially all actively growing cells exposed to ciprofloxacin induce SOS , but not all become persisters , suggesting that a specific level of SOS induction is required for persister formation . SOS is a gradual response and depending on the nature of the inducer , its concentration and the time of the exposure , different sets of genes are induced [10] , [54] , [55] . Our data indicate that persister formation requires a functional SOS response but a high level of induction is not required ( Figure 1B , Figure 3 ) . Persister formation also depends on functional DSB repair ( Figure 1A ) but does not need RecN ( Figure 1F ) , a function important for the repair of multiple DSBs [33] . This implies that persisters are cells that experienced few DSBs upon ciprofloxacin addition and underwent weak SOS induction . Consistent with this , constitutive , full expression of the SOS regulon ( equivalent to high induction ) does not lead to the tolerance of the entire population , but to an increased level of surviving persisters ( Figure 1E ) . Even in a lexA ( Def ) mutant , expression levels of SOS genes appear to fall short of being truly uniform throughout the population [52] , [53] . Persisters could be the cells that express a certain SOS function at a specific high or low level . Additionally , other regulatory pathways could allow a persister formation function to be expressed only in certain cells after induction . Turning on the SOS response constitutively would increase the number of cells being able to express this function . Persister levels are known to change with the growth phase [14] . It is low in early exponential phase and attains its highest level in stationary phase . An exponential increase in persister levels begins when the cell density reaches around 5×107 CFU/ml ( Figure 6 ) . The persister shoot up at similar cell density has been observed in other studies under different antibiotic and growth conditions [1] , [14] . A cell density of 5×107 CFU/ml coincides with the point at which the balanced growth of the culture ceases and a slowdown of growth rate is observed , even though the population as a whole still increases exponentially [56] . The extent of the DNA damage caused by ciprofloxacin would be expected to reflect the activity levels of gyrase and topoisomerase . These enzymes are active during replication and transcription [6] , [57]; therefore their maximal activity would occur in rapidly growing and replicating cells and would be lowest in the non-growing state of stationary phase . Lending support to this , transcription of gyrA and gyrB coding for gyrase subunits is at the peak in the early exponential phase and the lowest in the stationary growth phase [58] , [59] . It follows that ciprofloxacin would inflict maximal damage , the irreparable chromosome fragmentation , in the exponentially growing cells and fewer DSBs in the cells that slow down when the medium cannot support steady-state growth [60] . Indeed , no cells survive treatment to ofloxacin , another FQ , when the culture is kept at low density in constant exponential growth by repeated subculturing [14] , in other words no persisters are formed in that growth phase . On the other hand , the surviving fraction increases dramatically between the end of true exponential growth and stationary phase ( Figure 6 , [14] ) . During that time the growth rate of the population decreases from its maximum to zero , but because not all cells stop growing at the same time the heterogeneity of growth rates across the population is expected within that time frame . Those cells lacking steady state equilibrium might be the ones which experience few DSBs , weak SOS-induction and enter the tolerant state . Consistent with this , the difference in persister level in SOS proficient and deficient strains is minimal in early exponential phase , whereas it increases after the cessation of steady state growth ( Figure 6 ) . Conditions for unrestricted growth are rarely met in natural environments , and most bacteria are in a state of slow or no-growth [61]–[63] . However , physical and chemical agents capable of causing DNA damage are ubiquitous , therefore the SOS-induced persister state is probably quite common . Furthermore , in conditions of slow growth and frequent or lasting presence of DNA damaging agents , damage prevention would likely be advantageous over continuous active repair . The induction of the persister state in response to DNA damage seems like such a strategy - the avoidance of the damage build up as opposed to the costly repair . SOS is induced in aging colony biofilms of E . coli [64] and in intracellular biofilms formed by uropathogenic E . coli during cystitis [65] . Biofilms are notoriously hard to eradicate even with bactericidal fluoroquinolones , and this enhanced ‘resistance’ could in fact reflect the SOS-induced tolerance . Virtually all natural isolates of E . coli and many other bacteria are lysogens and many prophages are DNA-damage inducible [66]–[69] . Induction of λ prophage in E . coli is a late SOS function . In that light , SOS-induced tolerance could have evolved as a life-saving strategy preventing prophage induction upon DNA damage frequently encountered . There are at least 43 genes in the E . coli genome negatively regulated by LexA [9] , [10] . Many encode proteins participating in repair by homologous recombination and/or translesion synthesis and about one third are of unknown function . Among those are several genes encoding toxin-antitoxin modules that are attractive candidates for persistence genes , as the overexpression of some toxins has been shown to induce a dormant-like state [13] , [70] . Indeed , in a parallel study we identified an SOS inducible toxin/antitoxin module , tisAB , as a function needed for persister formation ( Dörr T . , Vulić M . , Lewis K . , submitted ) . However we cannot exclude that other LexA-regulated genes also contribute to SOS-induced tolerance . SOS has been shown to induce formation of a senescence-like state in which cells are viable but unable to form colonies [53] . Here we show SOS-dependant formation of persister cells . Both states could be formed through the common mechanism , such as expression of SOS-regulated toxins . In that case the strength of SOS induction and hence the toxin expression levels would determine which of these two states a cell reaches . In conclusion , we have discovered an active , regulated mechanism of persister formation , which is part of the SOS response . SOS has been known to contribute to the survival of antibiotic treatments by increasing the frequency of resistant mutants through its mutagenic activities [44] , [71] . Here we show a novel function of this response , the induction of a tolerant state . SOS-induced persistence having an immediate impact on bacterial survival is likely an important factor influencing the outcome of antibiotic treatment .
Bacterial strains are listed in Table 1 . Wild-type E . coli K-12 MG1655 was used as the parental strain . Different alleles were moved into the parental background by P1 transduction [72] . The kanamycin resistance cassette from the alleles originated from KEIO collection [73] was cured when needed by expressing the FLP recombinase from the helper plasmid pCP20 according to the protocol in [74] . Experiments were conducted at 37°C in Mueller Hinton Broth ( MHB ) supplemented with 10 mg/L MgSO4 and 20 mg/L CaCl2 according to NCCLS ( National Committee for Clinical Laboratory Standards ) guidelines for susceptibility testing and 0 . 1 M HEPES/KOH pH 7 . 2 . Persistence was measured by determining survival upon exposure to 0 . 1 µg/ml ciprofloxacin ( unless indicated otherwise ) , 100 µg/ml ampicillin and 25 µg/ml streptomycin during time indicated on corresponding graph axes . All antibiotics were purchased from Sigma . Prior to the addition of antibiotic overnight cultures were diluted 100-fold in 3 ml of fresh medium in 17- by 100-mm polypropylene tubes and incubated for 1 . 5 hrs with shaking , typically reaching ∼2×108 CFU/ml . For determination of CFU counts , cells were washed in 1% NaCl solution , serially diluted and plated on LB ( Luria-Bertani medium ) agar plates supplemented with 20 mM MgSO4 . Persister fraction , reflected as a plateau in CFU counts , was calculated as an average of CFU counts at 3- and 6-hour time points . In lexA3 and recA430 strains the CFU counts stabilize later than in the wild-type and in that case the CFU counts at 6-hour time point were used as a representative of the persister fraction . For plate assays using the CI-cro-gal construct , overnight cultures grown in LB medium at 37°C were diluted 1∶200 in 15 ml of fresh medium and incubated in 125 ml flasks for 1 . 75 hrs at 37°C with shaking . Ciprofloxacin was added and aliquots of the culture were taken at different time points , washed in 1% NaCl solution , serially diluted and plated on LB agar plates supplemented with 20 mM MgSO4 for total CFU counts and on MacConkey agar plates supplemented with 1% galactose in order to determine the fraction of gal+ cells . For β-galactosidase activity measurement , overnight cultures grown in supplemented MHB medium ( see above ) at 37°C were diluted 1∶100 in 3 ml of fresh medium in 17- by 100-mm polypropylene tubes and incubated for 1 . 75 hrs at 37°C with shaking . Ciprofloxacin was added and after 15 minutes an aliquot of culture was taken and recA::lacZ expression was measured as described in [72] . Overnight cultures in supplemented MHB medium were diluted 1∶1000 in 15 ml of fresh medium and incubated in 125 ml flasks for 1 hr at 37°C with shaking after which 0 . 25 µg/ml of mitomycin C ( Sigma ) was added to the cultures . This concentration did not inhibit the growth of the culture . After 2 hrs the total CFU counts were determined by dilution and plating and an aliquot of the culture was taken out and exposed to 0 . 3 µg/ml of ciprofloxacin for 3 hrs . The number of survivors was determined by plating on LB agar plates supplemented with 20 mM MgSO4 after washing in 1% NaCl solution . The same procedure was repeated after 4 hours of exposure to mitomycin C . Overnight cultures in MHB medium were diluted 1000-fold in 100 ml of fresh medium in 500 ml flasks and incubated at 37°C with shaking . At defined time intervals the cultures were serially diluted and plated on LB agar for determination of total CFU counts . In the same time 1 ml aliquots were transferred into 2 ml eppendorf tubes and 0 . 1 µg/ml ciprofloxacin was added . After 3 hrs at 37°C cells were washed with 1% NaCl solution , serially diluted and plated on LB agar plates supplemented with 20 mM MgSO4 . The colonies were counted after 40 hours incubation at 37°C .
|
The frequent failure of antibiotic treatments is an acute public health problem . Bacteria can escape the lethal action of antibiotics by a mutation in the cell's DNA , leading to antibiotic resistance . Alternatively , they can enter a physiological state in which the antibiotics do not affect them . This phenomenon , referred to as persistence , is different from resistance because there is no genetic modification and because it is transient . Persisters are believed to form stochastically prior to antibiotic treatment . The presence of persister cells in bacterial biofilms contributes to the difficulty in treating biofilm-related infections . We investigated the persistence of Escherichia coli to one of the most widely used antibiotics , ciprofloxacin . We show that the majority of persister cells are formed in response to this antibiotic , contrary to the prevailing view of persister formation . Ciprofloxacin kills bacteria by damaging their DNA . DNA damage activates a SOS gene network , the result of which is the production of various repair proteins . We uncovered a novel part of this network that leads to the formation of tolerant persister cells . The induced tolerance as a side effect of antibiotic treatment is an effective bacterial survival strategy and is likely to contribute to recalcitrance of infections .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"and",
"genomics/epigenetics",
"microbiology/microbial",
"evolution",
"and",
"genomics",
"infectious",
"diseases/antimicrobials",
"and",
"drug",
"resistance",
"microbiology"
] |
2009
|
SOS Response Induces Persistence to Fluoroquinolones in Escherichia coli
|
Plants with facultative crassulacean acid metabolism ( CAM ) maximize performance through utilizing C3 or C4 photosynthesis under ideal conditions while temporally switching to CAM under water stress ( drought ) . While genome-scale analyses of constitutive CAM plants suggest that time of day networks are shifted , or phased to the evening compared to C3 , little is known for how the shift from C3 to CAM networks is modulated in drought induced CAM . Here we generate a draft genome for the drought-induced CAM-cycling species Sedum album . Through parallel sampling in well-watered ( C3 ) and drought ( CAM ) conditions , we uncover a massive rewiring of time of day expression and a CAM and stress-specific network . The core circadian genes are expanded in S . album and under CAM induction , core clock genes either change phase or amplitude . While the core clock cis-elements are conserved in S . album , we uncover a set of novel CAM and stress specific cis-elements consistent with our finding of rewired co-expression networks . We identified shared elements between constitutive CAM and CAM-cycling species and expression patterns unique to CAM-cycling S . album . Together these results demonstrate that drought induced CAM-cycling photosynthesis evolved through the mobilization of a stress-specific , time of day network , and not solely the phasing of existing C3 networks . These results will inform efforts to engineer water use efficiency into crop plants for growth on marginal land .
Drought is the most pervasive abiotic stress and plants have evolved diverse strategies to mitigate the effects of water deficit [1] . Most water loss in plants occurs through transpiration as a byproduct of daytime stomata mediated CO2 uptake . Crassulacean acid metabolism ( CAM ) plants have evolved an alternative carbon assimilation pathway to store CO2 nocturnally when evapotranspiration rates are lower [2] . CAM plants store and concentrate atmospheric or respiratory CO2 at night through the carboxylation of phosphoenolpyruvate ( PEP ) by the enzyme phosphoenolpyruvate carboxylase ( PPC ) . The resulting four carbon acid , oxaloacetate , is subsequently reduced to malate by malate dehydrogenase ( MDH ) , and then transported to the vacuole as malic acid , producing the characteristic nighttime acidification observed in CAM plants [3] . During the day , malic acid is decarboxylated to release the CO2 for fixation by Rubisco . Because of this temporal separation , CAM plants have remarkably high water use efficiency , and use roughly 35% less water than C4 plants and up to 80% less water than comparable C3 species [4 , 5] . These traits make CAM an attractive model for engineering improved water use efficiency and drought tolerance into crop plants that may be grown on more marginal land prone to seasonal droughts[6] . The CAM pathway is highly plastic and occurs along a continuum ranging from plants that are predominately C3 with weak or inducible CAM activity to constitutive species with highly optimized and efficient CAM . This diversity reflects the multiple origins of CAM and its utility in different environments [7 , 8] . The biochemistry and physiology of CAM was worked out in species from the Crassulaceae , Cactaceae , and Bromeliaceae that exhibited strong constitutive CAM regardless of environmental conditions [9 , 10] . Constitutive CAM plants often inhabit arid and semi-arid environments or live in epiphytic habitats subject to seasonal drought . In regions with seasonal or periodic drought , facultative and CAM cycling CAM species often display a flexible system of switching from C3 to CAM under drought conditions . Facultative CAM species have high nocturnal stomatal conductance under drought and CAM-cycling plants display typical C3 diel stomatal conductance , but re-fix respiratory CO2 at night . The tremendous variation of CAM is partially explained by convergent evolution , since CAM has evolved independently at least 40 times across 35 diverse plant families [7] . Model CAM species have emerged for several families and a recent wealth of genetic and genomic resources have advanced our understanding of CAM pathway evolution . The genomes of three constitutive CAM species have been sequenced including pineapple ( Ananas comosus ) [11] , the orchid Phalaenopsis equestris [12] , and the eudicot Kalanchoe fedtschenkoi [13] . Draft genomes of the facultative CAM orchids Dendrobium catenatum [14] and D . officinale [15] are also available , as are transcript and expression studies across the Agavoideae [16] . Facultative CAM and CAM-cycling species provide an excellent system for dissecting the molecular basis of CAM as we can directly compare expression dynamics between C3 and CAM gene networks . Comparisons between C3 and CAM states can be used to identify metabolite , physiology , and gene expression changes associated with CAM evolution . Transcriptome and metabolite surveys of the facultative CAM species Talinum triangulare identified a set of core CAM pathway genes and candidate transcription factors mediating photosynthetic plasticity [17] . Though this study captured changes throughout drought progression , limited temporal resolution did not capture the circadian components of facultative CAM . Time series microarray data from ice plant ( Mesembryanthemum crystallinum ) also identified components of CAM induction , and suggested that there is a circadian regulated 4–8 hour phase shift underlying the C3 to CAM shift [18] . Taking an evolutionary approach comparing cycling genes in C3 , C4 and CAM ( Agave ) plants , it was found that CAM photosynthesis resulted from both accelerated evolution of specific protein domains as well as reprogramming of diel networks [19] . However , it is still unclear how these time of day networks interact with the circadian clock and how the C3 to CAM switch is mediated within a single system . The circadian clock has evolved to optimize the daily timing , or phase , of cellular biology with the local external light and temperature cycles[20] . In the model C3 species Arabidopsis thaliana , between 30–50% of genes cycle under diel conditions with control by at least three evolutionarily conserved transcriptional modules defined by the morning ( ME: CCACAC; Gbox: CACGTG ) , evening ( EE: AATATCT; GATA: GGATA ) , and midnight ( TBX: AAACCCT; SBX: AAGCCC; PBX: ATGGGCC ) [21] . Moreover , this time of day transcriptional network ensures development and environment-specific growth through opposing phytohormone influences[22] , and is conserved across distantly related C3 and C4 species [23 , 24] . The circadian clock directly controls components of the CAM pathway and is thought to coordinate the temporal oscillations of CAM . PPC is activated nocturnally via phosphorylation by the circadian clock activated phosphoenolpyruvate carboxylase kinase ( PPCK ) [25] . Silencing of PPCK in Kalanchoë fedtschenkoi not only reduces CAM activity by ~66% , but also perturbs central components of the circadian clock , suggesting PPCK may be essential for clock robustness [26] . Well-characterized CAM pathway genes in pineapple have acquired novel clock associated cis-elements compared to their orthologs from C3 and C4 species , suggesting a broad control of CAM by the core circadian clock [27 , 28] . Expression patterns of clock genes are conserved under C3 and CAM mode in ice plant [29] but the role of the clock in facultative and weak CAM induction remains unclear . Here we generated a draft genome for the CAM-cycling species Sedum album using long read Single Molecule Real-Time ( SMRT ) sequencing and surveyed high-resolution temporal gene expression , metabolite , and physiology changes across a diel time course in well-watered ( C3 ) and drought ( CAM-cycling ) conditions . This comparative approach revealed a massive time of day specific rewiring of the transcriptional networks in S . album where only 20% of cycling genes overlap between the two conditions . This work demonstrates for the first time that a phase shift in the core circadian clock underpins a massive time of day rewiring that enables S . album to switch to CAM photosynthesis and overcome water stress .
We assembled a draft genome of S . album to serve as a foundation resource to understand the genome-wide changes associated with the transition from C3 to drought induced CAM photosynthesis . We estimated the genome size to be 611 Mb based on flow cytometry , which was significantly larger than a previously reported value for S . album ( 142 Mb ) [30] . Karyotype analysis suggests S . album is tetraploid ( 2n = 4x = 68 ) with a deduced diploidized genome size of ~305 Mb ( S1A Fig ) . This is consistent with the kmer based genome size estimate , which revealed a heterozygous peak ( first ) with the full genome at 799 Mb and monoploid ( second peak ) at 256 Mb ( S1B Fig ) . Because of this complexity , we utilized a PacBio based SMRT sequencing approach to build the draft genome . We generated 4 . 4 million PacBio reads collectively spanning 33 . 7 Gb or 55x genome coverage . Raw PacBio reads were error corrected and assembled using the two leading PacBio assemblers: Falcon [31] and Canu [32] . Canu was able to accurately phase the highly similar homeologous regions and produced an assembly of 627 Mb across 15 , 256 contigs with an N50 of 47kb , designated as the S . album V2 genome . The average nucleotide similarity of homeologous regions is 99 . 6% based on the Canu assembly , suggesting S . album is either autotetraploid or allotetraploid with two highly similar genomes . Falcon collapsed the homeologous regions into a single haplotype with a total assembly size of 302 Mb across 6 , 038 contigs and an N50 of 93kb . The Canu and Falcon based contigs were polished to remove residual errors using high-coverage Illumina data with Pilon [33] . The S . album assemblies were annotated using the MAKER-P pipeline [34] . We utilized transcripts assembled from the timecourse RNAseq data ( described below ) , and protein datasets from other angiosperms as evidence for ab initio gene prediction . After filtering transposon derived sequences , the tetraploid and diploidized assemblies contained 93 , 910 and 44 , 487 gene models respectively . The two resulting polished assemblies were highly complete with a BUSCO score of 97% . The tetraploid assembly is highly fragmented with lower contiguity , and partial gene models may explain the greater than 2x number of genes compared to the diploid assembly . We identified homeologous genes between the diploid and tetraploid assembly and compared their relative expression . Homeologs within the tetraploid assembly have low sequence divergence and most RNAseq reads map ambiguously to both homeologs , leading to similar transcript quantification for each homeolog in the pair . Comparison of expression between homologous genes in the tetraploid and diploid assemblies revealed a strong correlation across all timepoints collected from well-watered ( C3 ) or drought ( CAM ) timecourses ( C3 r = 0 . 927 , CAM r = 0 . 901; S2 Fig ) . Given the similarity in expression and for simplicity of downstream analyses , we proceeded with the diploidized assembly produced from Falcon , and this version is referred to as the S . album V3 genome . We surveyed synteny and gene level collinearity between S . album and the closely related constitutive CAM plant Kalenchoe fedtschenkoi . Synentic regions between the two genomes were largely collinear , with conserved gene content and order ( Fig 1 ) . Roughly 71% of the S . album V3 genome had detectable synteny with K . fedtschenkoi , but this is likely an underestimation given the fragmented nature of both assemblies . Syntenic depth for each S . album region ranged from 1–5 to each region of K . fedtschenkoi , reflecting the two shared whole genome duplications events in the Crassulaceae [13] and polyploidy in S . album ( S3 Fig ) . Collinear regions were larger in K . fedtschenkoi compared to S . album , highlighting the relatively compact nature of the S . album genome ( Fig 1 ) . To estimate the divergence time of the S . album subgenomes , we calculated Ks ( synonymous substitutions per synonymous site ) between homoeologous gene pairs based on synteny with K . fedtschenkoi . We identified a single peak with a median Ks of 0 . 026 , corresponding to an estimated divergence time of ~853 , 000 years ( S4 Fig ) . This supports a relatively recent polyploid origin of S . album . We surveyed patterns of gene family dynamics across the genomes of seven CAM , five C3 and two C4 plants to identify expanded and contracted gene families in S . album . After normalizing for ploidy , we identified 33 gene families that were uniquely contracted in S . album and 41 that are uniquely expanded compared to the other species ( S1 Table ) . Among the contracted gene families in S . album are ABA responsive proteins , sucrose transporter 2 , Aluminum activated malate transporter 1 , and several flavonoid biosynthesis pathways . Among the expanded gene families in S . album are various transporters , photosynthesis related proteins , and early light induced proteins , among others ( S1 Table ) . When S . album is subjected to drought conditions , it switches from robust uptake of CO2 during the light phase to low rates of nocturnal carbon assimilation [35] , which is consistent with weak CAM-cycling induction ( Fig 2 ) . We collected parallel diel timecourse data under well-watered ( C3 ) and drought ( CAM-cycling ) conditions to identify genes , cis-elements and networks associated with the switch from C3 to CAM-cycling . The CAM-cycling timecourse was collected after 14 days of sustained drought when the CAM pathway was induced to conserve water usage ( Fig 2A ) . The well-watered ( C3 ) S . album had mild diel titratable acid and stomatal aperture changes , whereas nocturnal acid accumulation and reduced stomatal aperture were observed under drought ( Fig 2B and 2C ) . Sedum has relatively low rates of carbon assimilation under drought conditions [36] , so it is difficult to speculate on stomatal conductance based on aperture alone . However , the stomatal aperture was reduced under drought compared to well-watered samples and with no significant changes in nocturnal stomatal aperture . Together , this suggests that S . album is a drought-inducible CAM-cycling plant , and much of the nocturnal carbon assimilation is likely from recycled respiratory CO2 , as previously reported [37] . Parallel sampling of C3 ( well-watered ) and CAM-cycling ( 14-day drought ) plants were collected every two hours over a 24-hour diel timecourse in triplicate for RNA sequencing and metabolite measurements . Differentially expressed genes between pairwise C3 and CAM-cycling timepoints varied from 702 to 3 , 245 genes , reflecting massive transcriptional reprogramming ( S1 Table ) . GO enrichment analysis of these differential expressed genes indicated that most of the up-regulated genes under drought and CAM induction were related to abscisic acid responses , oxidative stress , and water deprivation at both day and night . Genes responding to ethylene were mainly enriched at night ( S2 Table ) . Down-regulated differential expressed genes in the CAM samples were enriched with genes involved in cell wall metabolism , organization and biogenesis reflecting the reduced growth under drought conditions . There was also an over-representation of down-regulated genes related to photosynthesis at night ( 10pm-4am ) in drought stressed S . album . After filtering genes with low expression ( total TPM < 5 ) , 29 , 372 genes were used to construct a differential co-expression network utilizing a weighted correlation network analysis ( WGCNA ) approach [38] . This approach grouped genes into 32 and 25 co-expression modules for the C3 and CAM-cycling networks respectively . Preservation of modules between the C3 and CAM networks was low ( Fig 3A ) , consistent with massive rewiring of the expression network under drought stress and CAM induction . The core CAM pathway genes ( described in more detail below ) were found in different modules in the CAM-cycling network , suggesting regulation of CAM and C3 pathways through distinct transcriptional programs and network rewiring . Photosynthesis and stress related pathways were co-regulated in the drought induced CAM-cycling co-expression network . The C3 network module 3 and CAM-cycling network module 7 were enriched in photosynthesis-related GO terms such as photosynthetic electron transport chain ( GO:0022900 ) , response to light stimulus ( GO:009416 ) , and photosynthesis light harvesting in photosystem I ( GO:0009765; S3 and S4 Tables ) . CAM-cycling network module 7 was also enriched with GO terms related to stress response ( GO:0006950 ) and oxidative stress ( GO:0006979 ) compared to the orthologous module in the C3 network , which had no enrichment . GO terms related to photosystem II assembly and function were over-represented in the C3 network module 3 but not in the corresponding CAM-cycling module 7 . This indicates photosystem I and electron transport chain pathways are co-regulated with stress related genes under drought but not photosystem II . This is likely due to drought induced damage on photosystem II and the tightly coordinated regulation of stress responsive and photosynthesis pathways to mitigate photooxidative and drought associated damage . This hypothesis is further supported by the enriched GO terms in CAM-cycling network module 15 and 19 , which contain ATP metabolic process , ATP hydrolysis coupled protein transport , and superoxide metabolic processes ( S3 and S4 Tables ) . These GO terms were not enriched in any well-watered , C3 network modules . Reactive oxygen species scavengers accumulate during water deficit to prevent oxidative damage [39] , and scavengers such as superoxide dismutase are up-regulated in S . album under drought stress ( S5 Fig ) . The comparison of gene co-expression networks between well-watered ( C3 ) and drought conditions ( CAM-cycling ) support the tight regulation between photosynthesis and stress responses . In pineapple , core CAM pathway genes have diel expression patterns along with enriched circadian associated cis-elements [27 , 28] . Most drought related genes are circadian regulated [40] and proper diurnal expression is associated with growth and resilience under water deficit [41–43] . Therefore , we also estimated the periodicity , phase ( peak time of expression ) and amplitude differences of each S . album gene under either C3 or CAM-cycling using the JTK_CYCLE program [44] . There were 10 , 278 ( 35% ) and 11 , 946 ( 41% ) genes cycling under C3 and CAM-cycling conditions respectively ( Table 1 ) . These numbers are similar to those reported for Arabidopsis [21] and other species [24] under diel conditions of 12 light/12 hrs dark ( LD ) and thermocycles ( HC: hot/cold ) . Surprisingly , only 22% ( 6 , 480 ) genes cycled under both conditions , and of these shared cycling genes , 39% ( 2 , 504 ) had a change in amplitude and 75% ( 4 , 579 ) had a change in phase . In addition , almost 50% of genes displayed a unique pattern of cycling in only one condition . These results corroborate the differential expression and network analysis , which is consistent with a massive time of day rewiring of the transcriptional machinery in S . album under drought and during the switch to CAM-cycling photosynthesis . To understand the shifts in the S . album circadian clock during C3-CAM conversion , we looked at cycling expression patterns ( S5 Table , S6 Fig ) . First , we found an expansion of core clock genes in S . album compared to Arabidopsis , with large changes in circadian clock associated 1/late elongated hypocotyl ( CCA1/LHY , 7 vs . 2 ) , PRRs ( 10 vs . 4 ) , early flowering 3 ( ELF3 , 5 vs . 2 ) and lux arrhythmo ( LUX , 7 vs . 5 ) and additional copies of gigantea ( GI ) and cryptochrome 1 ( CRY1 ) ( S6 and S7 Tables ) . Many of these expansions were shared with K . fedtschenkoi [13] , pineapple[27] , and orchid[12] , suggesting gene expansion may be linked to the circadian regulation of CAM pathways ( S7 Table ) . While none of the CAM species have a copy of the AtPRR9 ortholog , there were two PRRs only found in S . album and K . fedtschenkoi ( S7 Fig ) . In general , the phase of the cycling of S . album circadian clock gene orthologs matched those of Arabidopsis [21 , 24] but they showed a phase shift under drought stress ( S6 Table , S6 Fig ) . While GI , CCA1/LHY and ELF3 maintained the phase of expression between C3 and CAM in S . album , the amplitude of GI and ELF3 decreased under CAM ( 25–50% ) and CCA1/LHY amplitude increased ( 25% ) under CAM . In contrast , most of the PRR1s , PRR5s ( but not PRR7/3s ) displayed a 1–4 hr phase shift ( later ) under CAM-cycling conditions , which is consistent with the global 1–4 hr phase shift ( S5 Table ) . By comparing the phase and amplitude changes observed in drought treated S . album with the drought treated C3 plant Brassica rapa and two CAM systems ( A . comosus and K . fedtschenkoi ) ( S6 Table ) , we were able to untangle the circadian rhythmicity change unique to drought treated S . album . Shared amplitude changes between S . album and B . rapa may represent universal drought responses and those that are unique to S . album represent lineage specific drought responses and changes related to CAM induction . We identified several core clock genes with divergent changes in phase and amplitude . For instance , CRY1 genes were cycling in both conditions of S . album but only under well-watered conditions of B . rapa ( S6 Table ) , suggesting the circadian rhythm of CRY1 was conserved in S . album under drought stress . In addition , flavin-binding , kelch repeat , f box ( FKF ) displayed cycling pattern in all three CAM plants surveyed but not in the two conditions of B . rapa . Interestingly , we observed a phase shift where genes with rhythmic expression in well-watered conditions were over-represented at phases 11–17 while genes with rhythmic expression under drought were enriched in later phases ( phases 18–23 ) ( Fig 4A ) . This observed phase changing is not solely caused by drought or CAM photosynthesis but an interaction of both , as drought treated B . rapa and CAM performing A . comosus have different responses ( Fig 4A ) . Therefore , the phase shift of S . album genes under drought is possibly caused by the involvement of both circadian and drought-related cis-element , and expression shift of the circadian clock genes mentioned above . We identified a complete conservation of the time of day overrepresentation of the diel and circadian-related cis-elements , consistent with what we have found in other plant systems[21 , 23 , 24] . In addition , the time of day overrepresentation of both the evening specific cis-elements , evening elements and TBX , displayed a 1–4 hr phase shift between C3 and CAM-cycling , while the morning specific elements ME and Gbox did not ( S8A Fig ) . Furthermore , there were many more significant evening-specific cis-elements under CAM that were distinct from conserved elements ( S8B Fig ) , including 8 novel drought/CAM-specific cis-elements in S . album that are enriched in specific phases of drought but not well-watered ( Fig 4B , S8 Table ) . These elements were mainly enriched at the later phase , which is consistent with the phase shift of drought-specific cycling genes we observed . While all eight elements were over-represented in drought-specific cycling genes ( S9 Table ) , elements CAM-1 and CAM-7 were also enriched in cycling genes that had higher amplitude in drought condition ( Fig 4C ) . These eight novel cis-elements support our findings that the time of day networks are completely rewired under drought induced CAM photosynthesis , and that is not just a phasing of the C3 networks but instead a new drought and CAM specific network . In addition to the novel cis-elements , genes with CAM and drought specific cycling or higher amplitude in CAM were enriched with the stress-mediated ABRE ( YACGTGGC ) and ABRE-like ( BACGTGKM ) cis-element in their 2kb upstream promoter sequences ( Table 1 ) . CAM specific rhythmic genes were also enriched with the MYB transcription factor binding ( MACCWAMC ) cis-elements . In contrast , cycling genes with higher amplitude under C3 were enriched with the circadian related evening element [45] and MYC2 transcription factor binding site . To understand the biological functions of genes in each category , we performed a GO enrichment test . Genes with cycling expression in C3 only were mainly enriched with GO terms related DNA replication , translation and protein related metabolic process . Enriched GO terms in genes with CAM specific cycling were mainly related to transcription and gene expression regulation by RNA . Genes that are cycling in both conditions , but with higher amplitude in CAM were over-represented by GO terms involved in stress response including water deprivation ( S10 Table ) . Genes in the core CAM pathway belong to large gene families shared across plant genomes , and most have functions unrelated to CAM activity . Independent CAM lineages show evidence of convergent amino acid substitutions in core enzymes [13] , but independent recruitment of the same orthologs remains untested for most CAM genes . Through comparing diel expression patterns under C3 and CAM induction , we identified S . album genes associated with the core CAM pathway and putative vacuolar transporters involved in metabolite transport . We also compared the annotated S . album CAM genes with their pineapple and K . fedtschenkoi orthologs to identify similarities among the three sequenced CAM plants with diel expression data [13 , 27 , 28] . Putative CAM pathway genes in S . album had higher expression ( PPC , PPCK , NADP+-ME , MDH , PPDK ) or become oscillating ( PPCK , NAD-ME ) under drought induced CAM-cycling compared to C3 ( Fig 5 , Table 2 ) . Core CAM pathway genes were identified in the constitutive CAM plants K . fedtschenkoi and pineapple using a similar approach [13 , 27] . Beta carbonic anhydrase ( β-CA ) converts CO2 to HCO3- in the first step of nocturnal carboxylation in CAM photosynthesis . None of the annotated β-CA genes in S . album had high nocturnal expression under drought induced CAM-cycling compared to C3 ( S6 Fig ) . Low β-CA expression was also observed in the drought inducible CAM plant Talinum triangulare [17] and constitutive CAM species Yucca aloifolia [46] . In contrast , β-CA from pineapple and K . fedtschenkoi had strong nocturnal expression , which may support a higher conversion rate of CO2 to HCO3- ( S9 Fig ) . Nocturnal carbon assimilation in the weak CAM plant S . album likely occurs at a much lower rate than constitutive CAM species . Thus , β-CA may not be essential to provide additional bicarbonate for phosphoenolpyruvate carboxylation in the dark period during CAM-cycling in S . album . Phosphoenolpyruvate carboxylase ( PPC ) and Phosphoenolpyruvate carboxylase kinase ( PPCK ) are the core CAM genes that mediate the fixation of CO2 at night . Under CAM induction , these genes had up to 73x fold higher expression than C3 . PPC ( Sal_001109 ) has high expression during the day while PPCK genes ( Sal_021277 , Sal_045582 , Sal_050851 ) were expressed mainly nocturnally ( Fig 5 ) . PPC transcript expression was slightly shifted compared to pineapple ( Aco010025 . 1 ) and K . fedtschenkoi ( Kaladp0095s0055 . 1 ) , while PPCK expression pattern was more conserved across the three CAM species ( Fig 6A ) . Although there are multiple copies of PPC in each plant genome , S . album and K . fedtschenkoi recruited the same PPC ortholog for CAM ( Fig 6B ) . Given the unique monocot and eudicot specific duplications of PPC , it is not possible to assess if the same PPC was recruited in monocot and eudicot CAM plants . Based on the conserved diurnal expression of PPC , we surveyed cis-element enrichment across CAM , C4 , and C3 orthologs to identify cis-elements related to circadian regulation . The morning element [21] circadian clock motif was present in the PPC promoters from all three CAM plants , as well as the two C4 plants ( Fig 6B ) . In addition , TCP and MYB transcription factor binding sites were also present in the three CAM PPC promoters . Based on expression dynamics between conditions , diurnal decarboxylation in S . album is likely driven by chloroplastic NADP+-malic enzyme ( NADP+-ME , Sal_015733 ) and pyruvate orthophosphate dikinase ( PPDK , Sal_044029; Fig 5 ) . K . fedtschenkoi also utilizes the ME and PPDK pathway but pineapple primarily utilizes the phosphoenolpyruvate carboxykinase pathway ( Aco010232 . 1 ) for decarboxylation [13 , 27] . Nocturnally fixed CO2 is converted to malate and stored in the vacuole to avoid cytosolic pH fluctuations and prevent feedback inhibition of PPC [10] . While the proton pumps required to establish a gradient in the vacuole are well-characterized [47] , the malate transporters remain uncharacterized in CAM plants . The most likely transporters responsible for malate influx and efflux are homologs of the Arabidopsis clade II aluminum activated malate transporter ( ALMT 3/4/5/6/9 ) and tonoplast dicarboxylate transporter respectively [28 , 48 , 49] . However , under CAM-cycling conditions , no S . album ALMT genes shared a conserved transcript pattern with pineapple and K . fedtschenkoi . The tonoplast dicarboxylate transporter ortholog in S . album had peak expression in the early morning in both C3 and CAM-cycling mode , but the expression drops to almost zero at dusk ( ZT12 ) in CAM-cycling . The sudden expression decrease coincided with the titratable acid accumulation ( Fig 1B ) . This expression pattern was similar in pineapple and K . fedtschenkoi where their expression decreases sharply ( >10 fold reduction ) before sunset and gradually increases before sunrise ( S10 Fig ) . This supports a potential conserved role of tonoplast dicarboxylate transporter as a malate exporter across CAM plants . To elucidate the potential regulators of CAM-cycling photosynthesis , gene regulatory networks were constructed and compared between the C3 and CAM timecourses and putative activators and suppressors controlling core CAM genes were identified ( Table 3; Fig 3B and 3C ) . Core CAM pathway genes shared few interactors between C3 and CAM-cycling supporting independent networks . PPC expression in the CAM-cycling network was associated with the appearance of 12 putative activators and the disappearance of 2 suppressors . The three copies of phosphoenolpyruvate carboxylase kinase ( PPCK ) showed similar regulatory patterns . Interestingly , one of the PPCK copies ( Sal_021277-RA ) interacted with the ABA biosynthesis gene 9-cis-epoxycarotenoid dioxygenase ( Sal_006713-RA ) , an ABA exporter ( Sal_009291-RA ) and an ABA receptor regulator ( Sal_007192-RA ) under C3 conditions and this interaction disappeared when S . album switched to CAM-cycling under drought . PPCK phosphorylates PPC at night to reduce the allosteric inhibition of PPC by malate . This phosphorylation process is under circadian control [26] and the changes of interactions of core clock genes with PPCK are evident in our gene interaction networks . Two CCA1/LHY gene copies ( Sal_047094-RA and Sal_051928-RA ) interacted with PPCK ( Sal_021277-RA ) under both conditions but two additional CCA1/LHY copies ( Sal_050684-RA and Sal_051929 ) only interacted with PPCK under well-watered condition while CCA1/LHY_C1 ( Sal_047093-RA ) and PPCK only interacted in drought treated S . album . The only other CAM core gene that had putative gene interactions with core clock gene CCA1/LHY was malic enzyme ( ME ) , where some of these interactions were only observed in the drought network ( S11 Table ) . Malate dehydrogenase ( MDH ) was potentially controlled by 312 activators and 2 suppressors , including phosphoglycerate mutase ( Sal_015714-RA ) , which is involved in glycolysis and gluconeogenesis . Four additional glycolysis and gluconeogenesis pathway genes ( Sal_003892-RA , Sal_043275-RA , Sal_048516-RA , and Sal_050408-RA ) were also predicted to be involved in activating ME expression in CAM photosynthesis . In addition , starch synthesis pathway genes including ADP glucose pyrophosphorylase large subunit ( Sal_005475-RA and Sal_005955-RA ) and starch branching enzyme ( Sal_041721-RA ) were potential activators for ME and pyruvate orthophosphate dikinase ( PPDK ) respectively . Our gene regulatory networks demonstrate a direct link between CAM photosynthesis and carbohydrate metabolism ( Fig 3B , c ) and the drought induction of CAM photosynthesis requires the rewiring of large number of existing genes , similar to patterns observed in pineapple [27] . Nocturnal synthesis of PEP is essential for maintaining high levels of CAM activity . PEP is supplied through the degradation of transitory starch or soluble sugars [50] , and mutants deficient in starch degradation have impaired CAM activity [51] . Transitory starch accumulated in S . album throughout the day and was depleted nocturnally in both C3 and CAM , but the total amount of starch synthesized was higher when S . album was operating in CAM-cycling mode ( Fig 7 ) . A large proportion of genes in the starch biosynthesis pathway ( phosphoglucoseisomerase , starch branching enzymes , granular bound starch synthase and glucan water dikinase ) had higher expression in CAM-cycling than C3 ( S11A Fig ) . Starch is mainly synthesized in the stroma from glucose 6-phosphate ( G6P ) supplied from the Calvin Benson cycle through the conversion of F6P to G6P via glucose-6-phosphate isomerase ( PGI ) . G6P can also be imported from the cytosol through a glucose 6-phosphate/phosphate translocator ( GPT ) . The chloroplastic PGI is tightly regulated by multiple factors in Arabidopsis and this can limit the flux of starch synthesis [52 , 53] . When the required flux of starch synthesis exceeds the limit set by PGI ( such as the switch to CAM ) , plants may bypass the rate limiting plastidic PGI pathway by exporting carbon as triose phosphates . Carbon from cytosolic PGI can then be remobilized to the plastid as hexose phosphate using the GPT . Genes involved in this PGI bypass pathway were upregulated in CAM compared to C3 including the triose phosphate/phosphate translocator ( Sal_024485-RA ) ( S11B Fig ) . Most of the transcripts for genes involved in the branch of the pentose phosphate pathway from glucose 6-phosphate to ribulose 5-phosphate ( G6P shunt ) in both the plastid and cytosol were down-regulated in the CAM mode ( S11C Fig ) . Together , these data suggest that metabolism in CAM-cycling S . album is shifted to increase starch synthesis while avoiding carbon loss through the G6P shunt , which is consistent with the hypothesis previously proposed by Sharkey and Weise [54] . There are plastidic GPTs in plants , however they are not normally expressed in autotrophic tissue . It is hypothesized that this transporter is not expressed to prevent the pool of G6P from reaching the Km threshold of glucose-6-phosphate dehydrogenase in the chloroplast [54 , 55] . A large pool of G6P in the chloroplast could result in a futile cycle and loss of carbon as CO2 through the G6P shunt [54] . A S . album ortholog to the Arabidopsis GPT2 gene ( Sal_018693-RA ) had peak daytime expression in both C3 and CAM-cycling but showed drastically increased expression ( ranged from 11 to 205-fold higher ) in CAM-cycling condition ( Fig 8 , S12 Table ) . A similar diurnal expression pattern was observed in the constitutive CAM plants pineapple and K . fedtschenkoi , but not the C3 plant Arabidopsis ( Fig 8B ) . In contrast , the S . album GPT1 ortholog had constitutive expression in both C3 and CAM-cycling , and is likely not associated with CAM activity ( S13 Table ) . Expression dynamics between CAM and C3 indicates that the starch synthesis from G6P in S . album might be regulated at multiple pathways including starch synthesis , plastidic PGI bypass , G6P shunt , as well as triose phosphate and G6P transporters as summarized in Fig 8C . At night , transitory starch is broken down into either maltose and glucose via the hydrolytic degradation pathway , or to G6P via the phosphorolytic degradation pathway [56] . It is thought that starch degradation shifts from hydrolytic to phosphorolytic pathway during the transition from C3 to CAM in ice plant [56 , 57] . Chloroplasts isolated from the ice plant under C3 photosynthesis mainly export maltose during starch degradation and switch to G6P export under CAM . We measured the total G6P and maltose contents of S . album across the C3 and CAM diurnal timecourses . C3 performing S . album had a higher abundance of maltose than CAM-cycling plants at night but the G6P content was similar in both C3 and CAM-cycling plants ( Fig 7 ) . We also examined the transcript level of genes proposed to be involved in the two starch degradation pathways ( hydrolytic and phosphorolytic ) [56] . Homologs of Arabidopsis glucan phosphorylase ( LSF2 ) , β-amylase 6 and β-amylase 9 were expressed higher in well-watered conditions but glucan phosphorylase ( SEX4 ) , β-amylase 2 , β-amylase 3 , β-amylase 7 and isoamylase 3 were expressed strongly under drought ( S11D Fig ) . Hydrolytic pathway genes were not universally upregulated in either photosynthetic condition . This indicates that either S . album does not switch its starch degradation pathways during the C3-CAM conversion , or the regulation of this shift was not controlled at the transcriptional level .
Facultative CAM and CAM-cycling plants are optimized for rapid growth under favorable C3 conditions and sustained resilience under drought through CAM related water conservation . This reversible mechanism makes facultative CAM and CAM-cycling plants an ideal model to dissect the genetic basis of CAM and for engineering improved water use efficiency and drought tolerance in crop plants . Engineering inducible CAM photosynthesis into C3 crops could improve water use efficiency and resilience under prolonged drought stress while maintaining high yields when field conditions are ideal . Establishing high quality genomic resources are an essential foundation for downstream functional genomics of CAM . The complex polyploidy of S . album presented a computational challenge , but through a combination of long read sequencing and optimized assembly we produced a high quality , diploidized reference . High-resolution diurnal expression data for both C3 ( well-watered ) and CAM-cycling ( drought ) provided a comparative system for identifying genes and pathways involved in CAM . Utilizing a comparative differential co-expression approach , we found a strong link between stress responses and photosynthesis , supporting the tight regulation of CAM photosynthesis by drought networks . The CAM-cycling and C3 co-expression networks have a low degree of overlap , supporting massive reprogramming of genes related to drought and CAM photosynthesis induction . This is further supported by the gene regulatory networks , which showed major shifts in gene interaction under CAM induction . In addition , the large-scale rewiring of gene rhythmicity and phase underlies the drought protective mechanism in S . album . The time of day reprogramming is comprised of phase and amplitude change of core clock genes , and expression regulation under known and novel drought/CAM-specific cis-elements , which in turn lead to a phase difference between well-watered and drought cycling genes . The genome level phase shift of transcript expression is uniquely observed in S . album and this pattern was not observed in drought treated B . rapa and the constitutive CAM plant A . comosus . This highlights the complexity of CAM-cycling induction and the tight coordination of multiple metabolic pathways including stress responses , photosynthesis , and circadian rhythm . We observed conserved diurnal transcript expression of core CAM genes ( PPCK , ME , PPDK ) and potential transporters among independent CAM lineages , but patterns are not conserved for every gene ( i . e . PPC , MDH ) . Though PPC is highly expressed in all CAM plants , the diel oscillation varies between species . PPC orthologs from the same clade ( PPC-1 and not PPC-2 ) are always recruited for CAM related carboxylation [58] and the copies from the CAM species surveyed have common cis-regulatory elements related to circadian regulation . Based on our transcript data , β-CA likely has minimal function in CAM-cycling S . album compared to the constitutive CAM pathway where strong nocturnal expression of β-CA might be essential to rapidly assimilate and store carbon . The non-cycling expression of β-CA was also reported in the inducible CAM plant Talinum triangulare [17] . We hypothesize that the nocturnal β-CA expression is coupled with the high nocturnal stomatal conductance in constitutive CAM plants , which enhances the nocturnal CO2 fixation rate . β-CA involved in normal metabolic processes in the cytosol is probably sufficient for CAM activity in weak , low assimilating species , but enzymatic activity remains to be surveyed . β-CA is not essential for efficient operation of the C4 pathway in maize [59] , but the dispensability of β-CA in CAM is untested . Transitory starch content is high in CAM plants to provide enough substrate to regenerate PEP . CAM induced expression changes suggest S . album increases the transitory starch content through multiple pathways . Starch biosynthesis pathway genes have increased expression under CAM-cycling compared to C3 . During CAM-cycling , G6P transporters and triose phosphate transporters are upregulated and genes involved in the G6P shunt are downregulated . The diurnal induction of GPT transcripts and transport activity under CAM has been reported in ice plant previously [60 , 61] . Our high-resolution transcriptome data showed that expression of GPT2 increased after CAM-cycling induction while GPT1 had similar expression in both conditions . Therefore , GPT2 is the likely controller of G6P translocation in CAM-cycling photosynthesis . In C3 plants , the G6P pool in the stroma is 3 to 20 times lower than the cytosol [62 , 63] . We thus speculate that CAM plants activate GPT2 in the daytime to import G6P into the stroma for transitory starch synthesis , this would bypass the plastidic PGI which can be kinetically limiting to starch synthesis and result in higher stromal G6P concentration . Interestingly , the S . album GPT2 had a low but consistent diurnal oscillation under C3 conditions while the Arabidopsis GPT2 has low constitutive expression . Inducible CAM plants might have C3-CAM intermediate transcript expression for certain genes in order to rapidly switch between photosynthetic modes in response to abiotic stress . Concentrations of total cellular G6P and maltose and transcript expression of the starch degradation pathway genes revealed the complexity of PEP regeneration during the C3-CAM transition . The minimal change of G6P content between well-watered and drought conditions could be due to high cytosolic G6P content masking the G6P produced through phosphorolytic starch degradation . It could also indicate that there is no increase in phosphorolytic starch degradation during C3-CAM conversion . The increase in transcript expression of many hydrolytic pathway genes might suggest that there is no switch of starch degradation pathways from hydrolytic to phosphorolytic in S . album , contrasting pattern observed in ice plant [60 , 61] . However , the starch degradation patterns observed in ice plant were assayed with isolated chloroplasts , and in vivo degradation patterns may differ when concentrations of G6P and maltose in both the chloroplast and cytosol become kinetically relevant . Alternatively , the shift of starch degradation pathways could be species specific . In S . album , C3 to CAM-cycling switching results from a complete rewiring of the time of day networks . We found that there was a specific drought/CAM network that leverages a novel set of cis-elements . While the circadian cis-elements are completely conserved under C3 conditions , an alternative transcriptional module is leveraged to engage CAM photosynthesis . Underlying this alternative pathway may be a new repertoire of core circadian genes since we observed a significant increase in these genes as well as distinct phasing and amplitude under CAM-cycling conditions . Core circadian genes have been shown to be retained after whole genome duplication ( WGD ) and these new gene copies may provide a means to augmented stress responses . Drought responsive and CAM pathways are intrinsically linked in facultative and CAM-cycling plants . Disentangling these two processes would ideally require a C3 Sedum outgroup with no detectable CAM-cycling activity . CAM occurs along a continuum within Sedum , ranging from CAM-cycling and facultative to constitutive[36] . Sedum is highly polyphyletic[64] and the C3 Eurasian species are in phylogenetically distinct clades , making detailed comparisons with CAM lineages challenging . Furthermore , CAM anatomy and gene expression predate CAM origins within Yucca and more broadly across Agavoideae [46 , 65] , and CAM-like signatures of gene expression under drought may be observed in C3 Sedum species . The phase shift during CAM-cycling induction in S . album was not observed in similar drought timecourses from B . rapa [43] , and few genes with cycling expression are overlapping . This suggests the massive shifts of expression in S . album are driven by CAM related activities and not simply conserved drought responses . More detailed comparisons are needed to parse pathways related to CAM and drought in S . album . Genomic and high-resolution transcriptomic data from S . album provide a valuable foundation for downstream functional genomics work on the molecular basis of drought induced CAM-cycling photosynthesis . These pathways may be useful for engineering improved water use efficiency and drought tolerance into C3 or C4 crop plants .
Sedum album plants were vegetatively propagated via cuttings in growth chambers under 12-hour photoperiod with day/night temperatures of 24°C and 20°C respectively . The light period started at 8am ( ZT0 ) and ended at 8pm ( ZT12 ) . To induce CAM activity , plants were subjected to moderate drought by withholding water for 14 days . Well-watered ( C3 ) control plants were watered every 2 days for the duration of the experiment . For the 24-hour diurnal timecourse experiments , leaves from 5 pots of randomly sampled plants were pooled for each replicate and three biological replicates were collected for each timepoint . The well-watered ( C3 ) 6am ( ZT22 ) and drought ( CAM ) 8am ( ZT0 ) samples had two and one replicates respectively due to loss of samples . Samples were collected every 2 hours over the 24-hour experiment . All tissues were frozen into liquid nitrogen immediately and stored at -80C . Carbohydrate and titratable acidity data was collected for each of the samples used in the C3 and CAM timecourse experiment . Carbohydrate and titratable acidity measurements were taken from the same tissue samples used for RNAseq experiment . Frozen leaf tissue ( 0 . 1–0 . 5g ) was ground , weighed and transferred into 1 . 5mL tubes . 500uL of 3 . 5% perchloric acid was added to the leaf tissue and mixed by vortexing . The resulting supernatant ( perchloric acid extract ) was neutralized to pH 7 . 0 using neutralizing buffer ( 2M KOH , 150mM Hepes , and 10mM KCl ) . The samples were then frozen to precipitate salts , centrifuged , and the supernatant was transferred to a new 1 . 5mL tube for carbohydrate assays . Sucrose , glucose , fructose and fructose-6-phosphate , β-maltose , and glucose-6-phosphate were measured . The pellet was resuspended in 1mL of 80% ethanol and mixed by vortexing followed by two rounds of washing . The pellet was air-dried for 30 min , resuspensed in 200mM KOH and incubated at 95°C for 30 min . 1M acetic acid was used to adjust the pH to 5 . 0 and 5 ul of an enzyme cocktail containing 5 units α-amylase ( E-ANAAM Megazyme , Bray , Wicklow , Ireland ) and 6 . 6 units amylogucosidase ( E-AMGDF Megazyme ) was added to each tube and incubated at room temperature for 24 hours . The resulting supernatant containing glucose was transferred to a new tube for starch content measurement using the glucose assay . Glucose content was measured using a 96 well plate in a Filter Max F5 plate reader ( MDS Analytical Technologies , Sunnyvale CA , USA ) at 340 nm with an NADP ( H ) -linked assay . Wells were filled with 200 μl of 150 mM Hepes buffer pH 7 . 2 containing 15 mM MgCl2 , 3 mM EDTA , 500 nmol NADP , 500 nmol ATP and 0 . 4 units glucose-6-phosphate dehydrogenase ( G8529 Sigma St . Louis MO , USA ) . Five μl of sample was added to each well and the reaction was started by adding 0 . 5 units of hexokinase ( H4502 Sigma ) . Absolute glucose content was determined using an extinction coefficient of 6220 L mol-1 cm-1 for NADPH at 340 nm[66] . Sucrose content was assayed as reported above with 20U of invertase ( Sigma-aldrich , I4505 ) and 0 . 5U of hexokinase . Other carbohydrates ( β-maltose , glucose-6-phosphate , glucose , fructose and fructose-6-phosphate ) were measured as previously reported by Weise et al . [63 , 67] . All enzymes used were purchased from Sigma-Aldrich and 4 unit of maltose phosphorylase was used . To measure titratable acidity , ~0 . 3g of fresh ground tissue was mixed with 3mL of 80% ethanol and boiled at 80°C for 60 min . The supernatant was cooled to room temperature and titrated with 0 . 1N sodium hydroxide until an endpoint pH of 8 . 3 . Titratable acid ( in μ Eq per gram of fresh weight ) was calculated as volume of 0 . 1N sodium hydroxide X 0 . 1 X 1000 / Fresh weight ( gram ) . Epidermal peels from well-watered ( C3 ) and drought stressed ( CAM ) S . album leaves were prepared according to Wu et al . [68] . Microscope slides were visualized using a Nikon Eclipse Ni-Upright microscope with 40X differential interference contrast objective lens and images were captured with Nikon DS-Fi3 camera . Stomatal width and length was measured using program NIS-Elements and recorded in an Excel file . A minimum of 10 individual stoma from 5 leaves were examined for each time point ( ZT06 and ZT22 ) and condition ( C3 and CAM-cycling ) . A Student’s t-test was performed to test for significant changes in stomatal aperture . High molecular weight ( HMW ) genomic DNA was isolated from well-watered S . album leaf tissue for PacBio and Illumina sequencing . HMW gDNA was isolated using a modified nuclei preparation [69] followed by phenol chloroform purification to remove residual contaminants . PacBio libraries size selected for 25kb fragments on the BluePippen system ( Sage Science ) and purified using AMPure XP beads ( Beckman Coulter ) . The 25kb PacBio libraries were sequenced on a PacBio RSII system with P6C4 chemistry . In total , 4 million PacBio reads were sequenced , collectively spanning 33 . 7 Gb or 55x genome coverage . Illumina DNAseq libraries were constructed from the same batch of HMW gDNA using the KAPA HyperPrep Kit ( Kapa Biosystems ) followed by sequencing on an Illumina HiSeq4000 under paired end mode ( 150 bp ) . In total , 38 Gb of Illumina data was generated for error correction , representing ~62x coverage . RNA was extracted using the Omega Biotek E . Z . N . A . Plant RNA kit according to the manufacturer’s protocol . RNA quality was examined on a 1% agarose gel and RNA concentration was quantified using the Qubit RNA HS assay kit ( Invitrogen , USA ) . 2μg of total RNA was used to construct stranded RNAseq library using the Illumina TruSeq stranded total RNA LT sample prep kit ( RS-122-2401 and RS-122-2402 ) . Multiplexed libraries were pooled and sequenced on HiSeq4000 using paired end 150nt mode . Given the complex polyploidy of S . album , several long read assembly algorithms were tested for their ability to resolve and collapse homeologous regions . Raw PacBio reads were error corrected using Falcon ( V0 . 2 . 2 ) [31] and Canu ( V1 . 4 ) [70] . Parameters for Falcon were left as default . The following parameters for Canu were modified: minReadLength = 2000 , GenomeSize = 612Mb , minOverlapLength = 1000 . The Canu assembly accurately separated the homeologous regions and produced an assembly of 627 Mb across 15 , 256 contigs with an N50 of 47kb . The Canu assembly graph was visualized in Bandage [71] and most of the nodes ( contigs ) were highly interconnected . The assembly graph complexity is likely caused by a combination of polyploidy and heterozygosity . This Canu based assembly is referred as tetraploid genome version of S . album and named as Sedum V2 genome . Homeologous regions in the Falcon based assembly were highly collapsed and the total assembly size was 302 Mb with 6 , 038 contigs and an N50 of 93kb . This assembly is about half of the estimated genome size , and represents a diploidized version of the tetraploid S . album genome . We named this Falcon version as diploid genome version and designated as Sedum album V3 genome . We assessed the completeness of the tetraploid and diploidized assemblies using the benchmarking universal single-copy orthologs ( BUSCO; v . 2 ) [72] with the plant specific dataset ( embryophyta_odb9 ) . ~97% of the 1 , 440 plant specific genes were identified in both S . album assemblies , supporting they contained an accurate representation of the gene space . Contigs from the Falcon and Canu assemblies were polished to remove residual errors with Pilon ( V1 . 22 ) [33] using 62x coverage of Illumina pared-end 150 bp data . Illumina reads were quality-trimmed using Trimmomatic [73] followed by aligning to the assembly using bowtie2 ( V2 . 3 . 0 ) [74] with default parameters . The total alignment rate of Illumina data was 92 . 3% for the Falcon assembly and 93 . 1% for Canu , suggesting both were largely complete . Pilon parameters were modified as followed and all others were left as default:—flank 7 , —K 49 , and—mindepth 15 . Pilon was run four times to remove any residual errors . The S . album diploid and tetraploid assemblies were annotated using the MAKER-P pipeline [34] . A custom library of long terminal repeat [75] retrotransposons was constructed for repeat masking . LTR harvest ( genome tools V1 . 5 . 8 ) [76] and LTR Finder ( v1 . 07 ) [77] were used to predict putative LTRs and this candidate list was refined using LTR retriever ( v1 . 8 . 0 ) [78] . Parameters for LTR harvest and LTR finder were assigned based on suggestions from the LTR retriever package . This high quality library was used as input for RepeatMasker ( http://www . repeatmasker . org/ ) [79] with implementation in the MAKER pipeline . RNAseq data from the C3 and CAM-cycling timecourses was used as transcript evidence . Representative transcripts from the RNAseq data were assembled using Trinity [80] with default parameters . Protein sequences from Arabidopsis [81] and the UniprotKB plant databases [82] were used as protein evidence . Ab initio gene prediction was done using SNAP [83] and Augustus ( 3 . 0 . 2 ) [84] with two rounds of training . Transposable element derived gene models were filtered using a library of representative transposases . To identify homologs between the diploidized and tetraploid genome assemblies , a reciprocal blast approach was used . First , blastp was run using the tetraploid gene model sequences as query against the diploid gene models with an e-value cutoff of 0 . 001 , maximum target sequence of 1 , HSP ≥ 100 , protein similarity ≥60% , and match length > 50 amino acids . A reciprocal blastp search was performed with the reversed query and database using the same parameters except without maximum target sequence option . Only gene pairs retained in both blastp results are identified as homologs between the two genome versions . Expression profiles between the tetraploid and diploid assemblies were compared to assess if the diploid assembly contained a representative gene set for downstream analyses . Expression levels ( in TPM ) of the RNA-seq samples were quantified using Kallisto v0 . 43 . 0[85] with the tetraploid and diploid gene model sets . Expression values from all timepoints collected under C3 or CAM were averaged and plotted for homologous gene pairs in the diploid and tetraploid genome assemblies . Pearson correlation of all homologous gene pairs were calculated to assess the transcript expression similarity between two genomes . Based on the high expression correlation , the diploid version of the assembly and annotation was used for downstream analyses . Paired end raw reads were trimmed using Trimmomatic v0 . 33 [73] to remove adapters and low quality bases . Quality trimmed reads were pseudo-aligned to the S . album gene models to quantify expression using Kallisto v0 . 43 . 0[85] . Default parameters were used for Kallisto and 100 bootstraps were run per sample . Transcript expression was quantified in transcripts per million ( TPM ) and an averaged TPM from the three replicates was used for gene co-expression and single gene analyses . Differential expressed genes of each timepoint between C3 and CAM samples are identified using R program sleuth . Likelihood ratio test and Wald test were used and only genes with q-value < 0 . 05 and b-value > |1| in both tests were categorized as differentially expressed ( DE ) gene . The timecourse RNAseq data was clustered into gene co-expression networks using the R package WGCNA [38] . Prior to network construction , genes were filtered based on TPM and any gene with total TPM < 5 across all samples or 25% of datapoints have zero expression was removed . In total , 29 , 372 genes were used to construct separate networks for the well-watered ( C3 ) and drought ( CAM-cycling ) timecourses . Parameters for the C3 network were as follows: power = 7 , networkType =“signed" , corType = "bicor" , maxPOutliers = 0 . 05 , TOMType = "signed" , deepSplit = 3 , mergeCutHeight = 0 . 01 . For the CAM-cycling network , the following parameters were used: power = 10 , networkType = "signed" , corType = "bicor" , maxPOutliers = 0 . 05 , TOMType = "signed" , deepSplit = 3 , mergeCutHeight = 0 . 15 . The JTK_CYCLE program [44] was used to detect rhythmic expression patterns of genes across the two timecourses . JTK_CYCLE was implemented in the R package MetaCycle ( v . 1 . 0 . 0 ) [86] . The 29 , 372 genes used to construct co-expression networks were tested for rhythmicity under the C3 and CAM-cycling conditions . Genes with Bonferroni adjusted p-value < 0 . 05 were classified as cycling and changes in amplitude were assessed by subtracting the amplitude under C3 from that for CAM-cycling . Gene with amplitude difference less than 3 was categorized as having no changes between the two conditions . The same 29 , 372 genes used in WGCNA and JTK_CYCLE analysis were used for gene interaction network construction using default setting of program CMIP [87 , 88] . The threshold for the C3 and CAM networks was 0 . 47 and 0 . 52 , respectively . The gene interaction pairs between C3 and CAM networks were compared . Gene interacting pairs that were found in the C3 network but not in the CAM network were classified as “repressor” and gene interactions that were predicted in the CAM network but not in C3 were categorized as “activator” . For circadian rhythmicity , JTK_CYCLE analysis across the four species ( S . album , B . rapa , A . comosus and K . fedtschenkoi ) was performed as stated above . The B . rapa well-watered and drought datasets , were downloaded from NCBI GEO GSE90841 and only the second 24 hours of data from the well-watered and drought conditions were used for analysis . For A . comosus , the leaf green tip and white base datasets published in [28] were used and K . fedtschenkoi leaf timecourse dataset were downloaded from the NCBI SRA database ( BioSample SAMN07453940-SAMN07453987 ) . An expression matrix table was provided by Dr . Xiaohan Yang ( Oak Ridge National Laboratory , USA ) . These expression datasets were filtered using the same criteria as S . album as stated . For phase enrichment analysis , Bonferroni adjusted p-value of each phase ( 0–23 ) was calculated using R program and phases with an adjusted p-value ≤ 0 . 05 was classified as enriched . Gene Ontology ( GO ) terms of S . album protein sequences were annotated using InterProScan 5 [89] . GO terms for genes without InterProScan annotations were inferred using corresponding GO term from the top BLAST hit to Arabidopsis orthologs . The GO terms from both methods were merged and 33 , 362 genes had at least one annotated GO term . GO enrichment analysis was performed using the R package TopGO [90] with Fisher’s exact test and Bonferroni adjusted p-values . OrthoFinder ( v1 . 1 . 9 ) [91] was used to identify gene families between sequenced CAM and representative C3 and C4 species . The CAM species Sedum album , Phalaenopsis equestris , Dendrobium catenatum , Dendrobium officinale , Apostasia shenzhenica , Ananas comosus , and Kalanchoe fedtschenkoi , C3 species Brassica rapa , Oryza sativa , Solanum lycopersicum , Populus trichocarpa and Arabidopsis thaliana and C4 species Sorghum bicolor , Setaria italica were used for orthogroup identification . Protein sequences from each species were downloaded from Phytozome V12 . Orthofinder was run using default parameters . Only orthologs from the CAM species pineapple and K . fedtschenkoi were included in downstream analyses as timecourse data for orchid ( Phalaenopsis equestris ) is currently unavailable . To assess the phylogenetic relationships between orthologs , protein sequences of genes within the same orthogroup identified by OrthoFinder were extracted and aligned with MUSCLE ( v . 3 . 8 . 31 ) [92] and maximum likelihood phylogenetic trees were built using RAxML ( v8 . 2 . 10 ) [93] with bootstraps of 100 and the–m PROTGAMMAJTT flag . Orthogroups were further analyzed to identify expansions and contractions unique to S . album compared to other CAM as well as C3 and C4 species . The number of genes for each species was normalized by the total number of genes in the genome to account for differences in gene number and ploidy between species , as previously described [94] . Orthogroups were classified as contracted with a ratio of less than 0 . 2 in S . album compared to other species , and orthogroups with a ratio greater than 3 in S . album compared to other species were classified as expanded . cis-element analysis was carried out as previously described [21] . The S . album promoters ( 500 , 1000 , 2000 bp ) were parsed using the gene models . Gene lists were generated for each phase as called by JTK_CYCLE with a cycling significance cut-off of 0 . 05 and for specific conditions: CAM only cycling , C3 only cycling , C3 and CAM cycling with equal expression , C3 and CAM cycling with greater C3 expression , C3 and CAM cycling with greater CAM expression , Overrepresentation of cis-element was calculated for all of 3–8 mers using the ELEMENT program [95] with a p-value cut off of 0 . 05 .
|
Crassulacean acid metabolism ( CAM ) photosynthesis represents an important adaptation to arid environments as CAM plants take up CO2 at night when evapotranspiration rates are lower . Genomes and large-scale datasets are available for several plants with constitutive CAM activity , but they provided little insight on how this trait evolved from C3 . Here we sequenced the CAM-cycling plant Sedum album , which switches from C3 to CAM photosynthesis under drought conditions . We performed a global gene expression analysis sampling every two hours over 24 hours in C3 ( well-watered ) and CAM ( drought ) conditions . This comparative approach allowed us to identify components of the CAM pathway that were previously unidentified in constitutive CAM plants such as pineapple and orchid . Our results reveal a massive time of day specific rewiring of the transcriptional networks in Sedum where only 20% of cycling genes overlap between the C3 and CAM conditions . This time of day reprogramming results in broad network changes linking stress pathways and photosynthesis under CAM .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"ecology",
"and",
"environmental",
"sciences",
"c3",
"photosynthesis",
"genetic",
"networks",
"water",
"resources",
"chemical",
"compounds",
"carbohydrates",
"departures",
"from",
"diploidy",
"organic",
"compounds",
"circadian",
"oscillators",
"plant",
"science",
"tetraploidy",
"network",
"analysis",
"chronobiology",
"photosynthesis",
"starches",
"computer",
"and",
"information",
"sciences",
"natural",
"resources",
"gene",
"expression",
"chemistry",
"polyploidy",
"biochemistry",
"plant",
"biochemistry",
"circadian",
"rhythms",
"organic",
"chemistry",
"gene",
"identification",
"and",
"analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences"
] |
2019
|
Time of day and network reprogramming during drought induced CAM photosynthesis in Sedum album
|
Robustly predicting outcome for cancer patients from gene expression is an important challenge on the road to better personalized treatment . Network-based outcome predictors ( NOPs ) , which considers the cellular wiring diagram in the classification , hold much promise to improve performance , stability and interpretability of identified marker genes . Problematically , reports on the efficacy of NOPs are conflicting and for instance suggest that utilizing random networks performs on par to networks that describe biologically relevant interactions . In this paper we turn the prediction problem around: instead of using a given biological network in the NOP , we aim to identify the network of genes that truly improves outcome prediction . To this end , we propose SyNet , a gene network constructed ab initio from synergistic gene pairs derived from survival-labelled gene expression data . To obtain SyNet , we evaluate synergy for all 69 million pairwise combinations of genes resulting in a network that is specific to the dataset and phenotype under study and can be used to in a NOP model . We evaluated SyNet and 11 other networks on a compendium dataset of >4000 survival-labelled breast cancer samples . For this purpose , we used cross-study validation which more closely emulates real world application of these outcome predictors . We find that SyNet is the only network that truly improves performance , stability and interpretability in several existing NOPs . We show that SyNet overlaps significantly with existing gene networks , and can be confidently predicted ( ~85% AUC ) from graph-topological descriptions of these networks , in particular the breast tissue-specific network . Due to its data-driven nature , SyNet is not biased to well-studied genes and thus facilitates post-hoc interpretation . We find that SyNet is highly enriched for known breast cancer genes and genes related to e . g . histological grade and tamoxifen resistance , suggestive of a role in determining breast cancer outcome .
Metastases at distant sites ( e . g . in bone , lung , liver and brain ) is the major cause of death in breast cancer patients [1] . However , it is currently difficult to assess tumor progression in these patients using common clinical variables ( e . g . tumor size , lymph-node status , etc . ) [2] . Therefore , for 80% of these patients , chemotherapy is prescribed [3] . Meanwhile , randomized clinical trials showed that at least 40% of these patients survive without chemotherapy and thus unnecessarily suffer from the toxic side effect of this treatment [3 , 4] . For this reason , substantial efforts have been made to derive molecular classifiers that can predict clinical outcome based on gene expression profiles obtained from the primary tumor at the time of diagnosis [5 , 6] . An important shortcoming in molecular classification is that ‘cross-study’ generalization is often poor [7] . This means that prediction performance decreases dramatically when a classifier trained on one patient cohort is applied to another one [8] . Moreover , the gene signatures found by these classifiers vary greatly , often sharing only few or no genes at all [9–11] . This lack of consistency casts doubt on whether the identified signatures capture true ‘driver’ mechanisms of the disease or rather subsidiary ‘passenger’ effects [12] . Several reasons for this lack of consistency have been proposed , including small sample size [11 , 13 , 14] , inherent measurement noise [15] and batch effects [16 , 17] . Apart from these technical explanations , it is recognized that traditional models ignore the fact that genes are organized in pathways [18] . One important cancer hallmark is that perturbation of these pathways may be caused by deregulation of disparate sets of genes which in turn complicates marker gene discovery [19 , 20] . To alleviate these limitations , the classical models are superseded by Network-based Outcome Predictors ( NOP ) which incorporate gene interactions in the prediction model [21] . NOPs have two fundamental components: aggregation and prediction . In the aggregation step , genes that interact , belong to the same pathway or otherwise share functional relation are aggregated ( typically by averaging expressions ) into so called “meta-genes” [22] . This step is guided by a supporting data source describing gene-gene interactions such as cellular pathway maps or protein-protein interaction networks . In the consequent prediction step , meta-genes are selected and combined into a trained classifier , similar to a traditional classification approach . Several NOPs have been reported to exhibit improved discriminative power , enhanced stability of the classification performance and signature and better representation of underlying driving mechanisms of the disease [18 , 23–25] . In recent years , a range of improvements to the original NOP formulation has been proposed . In the prediction step , various linear and nonlinear classifiers have been evaluated[26 , 27] . Problematically , the reported accuracies are often an overestimation as many studies neglected to use cross-study evaluation scheme which more closely resembles the real-world application of these models [7] . Also for the aggregation step , which is responsible for forming meta-genes from gene sets , several distinct approaches are proposed such as clustering [23] and greedy expansion of seed genes into subnetworks [18] . Moreover , in addition to simple averaging , alternative means by which genes can be aggregated , such as linear or nonlinear embeddings , have been proposed [17 , 28] . Most recent work combines these steps into a unified model [8 , 29] . Recent efforts that extend these concepts to sequencing data by exploiting the concept of cancer hallmark networks have also been proposed [30] . Despite these efforts and initial positive findings , there is still much debate over the utility of NOPs compared to classical methods , with several studies showing no performance improvement [21 , 31 , 32] . Perhaps even more striking is the finding that utilizing a permuted network [32] or aggregating random genes [10] performs on par with networks describing true biological relationships . Several meta-analyses attempting to establish the utility of NOPs have appeared with contradicting conclusions . Notably , Staiger et al . compared performance of nearest mean classifier [33] in this setting and concluded that network derived meta-genes are not more predictive than individual genes [21 , 32] . This is in contradiction to Roy et al . who achieved improvements in outcome prediction when genes were ranked according to their t-test statistics compared to their page rank property [34] in PPI network [28 , 35] . It is thus still an open question whether NOPs truly improve outcome prediction in terms of predictive performance , cross-study robustness or interpretability of the gene signatures . A critical—yet often neglected—aspect in the successful application of NOPs is the contribution of the biological network . In this regard , it should be recognized that many network links are unreliable [36 , 37] , missing [38] or redundant [39] and considerable efforts are being made to refine these networks [38 , 40–42] . In addition , many links in these networks are experimentally obtained from model organisms and therefore may not be functional in human cells [43–45] . Finally , most biological networks capture only a part of a cell’s multifaceted system [46] . This incomplete perspective may not be sufficient to link the wide range of aberrations that may occur in a complex and heterogeneous disease such as breast cancer [47 , 48] . Taken together , these issues raise concerns regarding the extent to which the outcome predictors may benefit from inclusion of common biological networks in their models . In this work , we propose to construct a network ab initio that is specifically designed to improve outcome prediction in terms of cross-study generalization and performance stability . To achieve this , we will effectively turn the problem around: instead of using a given biological network , we aim to use the labelled gene expression datasets to identify the network of genes that truly improves outcome prediction ( see Fig 1 for a schematic overview ) . Our approach relies on the identification of synergistic gene pairs , i . e . genes whose joint prediction power is beyond what is attainable by both genes individually [49] . To identify these pairs , we employed grid computing to evaluate all 69 million pairwise combinations of genes . The resulting network , called SyNet , is specific to the dataset and phenotype under study and can be used to infer a NOP model with improved performance . To obtain SyNet , and allow for rigorous cross-study validation , a dataset of substantial size is required . For this reason , we combined 14 publicly available datasets to form a compendium encompassing 4129 survival labeled samples . To the best of our knowledge , the data combined in this study represents the largest breast cancer gene expression compendium to date . Further , to ensure unbiased evaluation , sample assignments in the inner as well as the outer cross-validations folds are kept equal across all assessments throughout the paper . In the remainder of this paper , we will demonstrate that integrating genes based on SyNet provides superior performance and stability of predictions when these models are tested on independent cohorts . In contrast to previous reports , where shuffled versions of networks also performed well , we show that the performance drops substantially when SyNet links are shuffled ( while containing the same set of genes ) , suggesting that SyNet connections are truly informative . We further evaluate the content and structure of SyNet by overlaying it with known gene sets and existing networks , revealing marked enrichment for known breast cancer prognostic markers . While overlap with existing networks is highly significant , the majority of direct links in SyNet is absent from these networks explaining the observed lack of performance when NOPs are guided by the phenotype-unaware networks . Interestingly , SyNet links can be reliably predicted from existing networks when more complex topological descriptors are employed . Taken together , our findings suggest that compared to generic gene networks , phenotype-specific networks , which are derived directly from labeled data , can provide superior performance while at the same time revealing valuable insight into etiology of breast cancer .
We first evaluated NOP performance for three existing methods ( Park , Chuang and Taylor ) and the Group Lasso ( GL ) when supplied with a range of networks , including generic networks , tissue-specific networks and SyNet . As a baseline model , we used a Lasso classifier trained using all genes in our expression dataset ( n = 11748 ) without network guidance . The Lasso exhibits superior performance among many linear and non-linear classifiers evaluated on our expression dataset ( see S3 for details ) . The AUC of the four NOPs , presented in Fig 2 , clearly demonstrates that SyNet improves the performance of all NOPs , except for the Park method in which it performs on par to the Correlation ( Corr ) network . Notably , SyNet is inferred using training samples only , which prevents “selection bias” in our assessments [50] . Furthermore , comparison of baseline model performance ( i . e . Fig 2 , rightmost bar ) and other NOPs supports previous findings that many existing NOPs do not outperform regular classifiers that do not use networks [8 , 21 , 32] . The GL clearly outperforms all other methods , in particular when it exploits the information contained in SyNet . This corroborates our previous finding [8] that existing methods which construct meta-genes by averaging are suboptimal ( see S1 for a more extensive analysis ) . The GL using the Corr network also outperforms the baseline model , albeit non-significantly ( p~0 . 6 ) , which is in line with previous reports [23] . It should be noted that across all these experiments an identical set of samples is used to train the models so that any performance deviation must be due to differences in ( i ) the set of utilized genes or ( ii ) the integration of the genes into meta-genes . In the next two sections , we will investigate these factors in more details . Networks only include genes that are linked to at least one other gene . As a result , networks can provide a way of ranking genes based on the number and weight of their connections . One explanation for why NOPs can outperform regular classifiers is that networks provide an a priori gene ( feature ) selection [32] . To test this hypothesis and determine the feature selection capabilities of SyNet , we compare classification performances obtained using the baseline classifier ( i . e . Lasso ) that is trained using enclosed genes in each network . While this classifier performs well compared to other standard classifiers that we investigated ( see S3 for details ) , it cannot exploit information contained in the links of given network . So , any performance difference must be due to the genes in the network . The number of genes in each network under study is optimized independently by varying the threshold on the weighted edges in the network and removing unconnected genes ( see section “Regular classifiers and Network based prediction models” for network size optimization details ) . The edge weight threshold and the Lasso regularization parameter were determined simultaneously using a grid search cross-validation scheme ( see S5 for details ) . Fig 3 provides the optimal performances for 12 distinct networks along with number of genes used in the final model ( i . e . genes with non-zero Lasso coefficients ) . We also included the baseline model where all genes ( n = 11748 ) are utilized to train Lasso classifier ( rightmost bar ) . The results presented in Fig 3a demonstrate that SyNet is the only network that performs markedly better than the baseline model which is trained on all genes . Interestingly , we observe that SyNet is the top performing network while utilizing a comparable number of genes to other networks . The second-best network is the Corr network . We argue that superior performance of SyNet over the Corr network stems from the disease specificity of genes in SyNet which helps the predictor to focus on the relevant genes only . It should be noted that the data on which SyNet and the Corr network are constructed are completely independent from the validation data on which the performance is based due to our multi-layer cross-validation scheme ( see Methods and S5 ) which avoids selection bias [50] . We conclude that dataset-specific networks , in particular SyNet which also exploits label information , provides a meaningful feature selection that is beneficial for classification performance . Our result show that none of the tissue-specific networks outperform the baseline . Despite the modest performance , it is interesting to observe that performance for these networks increases as more relevant tissues ( e . g . breast and lymph node networks ) are utilized in the classification . Additionally , we observe that tissue-specific networks do not outperform the generic networks . This may be the result of the fact that generic networks predominantly contain broadly expressed genes with fundamental roles in cell function which may still be relevant to survival prediction . A similar observation was made for GWAS where SNPs in these widely-expressed genes can explain a substantial amount of missed heritability [51] . In addition to classifier performance , an important motivation for employing NOPs is to identify stable gene signatures , that is , the same genes are selected irrespective of the study used to train the models . Gene signature stability is necessary to confirm that the identified genes are independent of dataset specific variations and therefore are true biological drivers of the disease under study . To measure the signature consistency , we assessed the overlap of selected genes across all repeats and folds using the Jaccard Index . Fig 3b shows that a Lasso trained using genes preselected by SyNet , identifies more similar genes across folds and studies compared to other networks . Surprisingly , despite the fact that the expression data from which SyNet is inferred changes in each classification fold , the signature stability for SyNet is markedly better than for generic or tissue-specific networks that use a fixed set of genes across folds . Therefore , our results demonstrate that synergistic genes in SyNet truly aid the classifier to robustly select signatures across independent studies . The ultimate goal of employing NOPs compared to classical models that do not use network information is to improve prognosis prediction by harnessing the information contained in the links of the given network . Therefore , we next aimed to assess to what extent also connections between the genes , as captured in SyNet and other networks , can help NOPs to improve their performance beyond what is achievable using individual genes . As before , we utilized identical datasets ( in terms of genes , training and test samples ) in inner and outer cross-validation loops to train all four NOPs as well as the baseline model which uses Lasso trained using all genes ( n = 11748 ) . Our results presented in Fig 4a , clearly demonstrate that compared to other NOPs under study , GL guided by SyNet achieves superior prognostic prediction for unseen patients selected from an independent cohort . To confirm that NOP performance using SyNet is the result of the network structure , we also applied the GL to a shuffled version of SyNet ( Fig 4a ) . We observe a substantial deterioration of the AUC , supporting the conclusion that not only the genes , but also links contained in SyNet are important to achieve good prediction . Moreover , this observation rules out that the GL by itself is able to provide enhanced performance compared to standard Lasso . The result of a similar assessment for the Corr network is given in S12 . Additionally , we found that SyNet remains predictive even when the dataset is down sampled to 25% of samples ( see S13 for details ) . We also evaluated a recently developed set of subtype-specific networks for breast cancer [52] and found that SyNet markedly outperforms these networks in predictive performance ( see S18 for details ) . We next assessed the performance gain of the network-guided model compared to a Lasso model that cannot exploit network information . To this end , the GL was trained based on each network whereas the Lasso is was trained based on the genes present in the network . Fig 4b demonstrates the results of this analysis . We find that the largest gain in GL performance is achieved when using SyNet ( Fig 4b , x-axis ) , indicating that the links between genes in SyNet truly aid classification performance beyond what is obtained as a result of the feature selection capabilities of Lasso . Fig 4c provides the Kaplan-Meier plot when each patient is assigned to a good or poor prognostic class according to frequency of predicted prognosis across 10 repeats ( ties are broken by random assignment to one of the classes ) for Lasso as well as Group Lasso . Result of this analysis suggests that superior performance of the GL compared to the Lasso is mostly stemming from GLs ability to better discern the patients with poor prognosis . An important property of an outcome predictor is to exhibit constant performance irrespective of the dataset used for training the model ( i . e . performance stability ) . This is a highly desirable quality , as concerns have been raised regarding the highly variable performances of breast-cancer classifiers applied to different cohorts [7 , 53] . To measure performance stability , we calculated the standard deviation of the AUC for Lasso and GL . The y-axis in Fig 4b represents the average difference of standard deviation for Lasso and GL across all evaluated folds and repeats ( 14 folds and 10 repeats ) . Based on this figure , we conclude that a NOP model guided by SyNet not only provides superior overall performance , it also offers improved stability of the classification performance . Finally , we investigated the importance of hub genes in SyNet ( genes with >4 neighbors ) and observe that a comparable performance can be obtained with a network consisting of hub genes exclusively at the cost of reduced performance stability ( see S14 for details ) . Moreover , we did not observe performance gain for a model that is governed by combined links from multiple networks ( either by intersection or unification , see S15 for details ) . We further confirmed that the performance gain of the network-guided GL is preserved when networks are restricted to have equal number of links ( see S7 for details ) , or when links with lower confidence are included in the network ( see S16 for details ) . We also considered the more complex Sparse Group Lasso ( SGL ) , which offers an additional level of regularization ( see S1 Text for details ) . No substantial difference between GL and SGL performance was found ( see S8 for details ) . Likewise , we did not observe substantial performance differences when the number of genes , group size and regularization parameters were simultaneously optimized in a grid search ( see S9 for details ) . Together , these findings can be considered as the first unbiased evidence of true classification performance improvement in terms of average AUC and classification stability by a NOP . Many curated biological networks suffer from an intrinsic bias since genes with well-known roles are the subject of more experiments and thus get more extensively and accurately annotated [54] . Post-hoc interpretation of the features used by NOPs , often by means of an enrichment analysis , will therefore be affected by the same bias . SyNet does not suffer from such bias , as its inference is purely data driven . Moreover , since SyNet is built based on gene pairs that contribute to the prediction of clinical outcome , we expect that the genes included in SyNet not only relate to breast cancer; they should play a role in determining how aggressively the tumor behaves , how advanced the disease is or how well it responds to treatment . To investigate the relevance of genes contained in SyNet in the development of breast cancer and , more importantly , clinical outcome , we ranked all pairs according to their median Fitness ( Fij ) across 14 studies and selected the top 300 genes ( encompassing 3544 links ) . This cutoff was frequently chosen by the GL as the optimal number of genes in SyNet ( see section “SyNet improves NOP performance” ) . Fig 5 visualizes this network revealing three main subnetworks and a few isolated gene pairs . We performed functional enrichment for all genes as well as for the subcomponents of the three large subnetworks in SyNet using Ingenuity Pathway Analysis ( IPA ) [55] . IPA reveals that out of 300 genes in SyNet , 287 genes have a known relation to cancer ( 2e-06<p<1e-34 ) of which 222 are related to reproductive system disease ( 2e-06<p<1e-34 ) . Furthermore , according to IPA analysis , the top five upstream regulators of genes in SyNet ( orange box , Fig 5 ) are CDKN1A , E2F4 , RABL6 , TP53 and ERBB2 , all of which are well known players in the development of breast cancer[56–60] . The mean degree of the 300 genes in SyNet is 24 , but there are 12 genes which have a degree of 100 or above: ASPM [61] , BUB1[62] , CCNB2 [63] , CDKN3 [64] , CENPA [65] , DLGAP5 [66] , KIF23 [67] , MCM10 [68] , MELK [69] , RACGAP1 [70] , TTK [71] and UBE2C [72] . All these genes play a vital role in progression through the cell cycle and mitosis , by ensuring proper DNA replication , correct formation of the mitotic spindle and proper attachment to the centromere . In addition to a clear involvement of genes linked to breast cancer generically , IPA also finds clear indications that the genes in SyNet are relevant to clinical outcome and prognosis of the disease . For instance , the most highly enriched cluster ( Fig 5; green cluster ) is found by IPA to be associated to histological grade of the tumor ( p = 6e-201 ) . The histological grade , which is based on the morphological characteristics of the tumor , has been shown to be informative for the clinical behavior of the tumor and is one of the best-established prognostic markers [73] . Notably , the blue cluster is enriched for genes involved in tamoxifen resistance ( p<2e-3 ) , one of the important treatments of ER-positive breast cancer . Two other sub-clusters ( yellow and purple in Fig 5 ) , contain genes from distinctly different biological processes than the main cluster . In these clusters we also observe clear hub genes: SLC7A7 and CD74 in the yellow and ACKR1 and MFAP4 in the purple cluster . ACKR1 is a chemokine receptor involved in the regulation of the bio-availability of chemokine levels and MFAP4 is involved in regulating cell-cell adhesion . The recruitment of cells , as regulated by chemokines , and reducing cell-cell adhesion both play an important role in the process of metastasis . CD74 has also been linked to metastasis in triple negative breast cancer [74] . Metastasis , and not the primary tumor , is the main cause of death in breast cancer [3] . IPA highly significantly identifies the SyNet genes as upstream regulators of canonical pathways implicated in breast cancer ( Fig 5 ) , such as Cell Cycle Control of Chromosomal Replication ( 8e-18 ) , Mitotic Roles of Polo-Like Kinase ( 4e-15 ) , Role of CHK Proteins in Cell Cycle Checkpoint Control ( 6e-12 ) , Estrogen-mediated S-phase Entry ( 2e-11 ) , and Cell Cycle: G2/M DNA Damage Checkpoint Regulation ( 5e-10 ) . Although all cancer cells deregulate cell cycle control , the degree of dysregulation may contribute to a more aggressive phenotype . For instance , it is recognized that the downregulation of certain checkpoint regulators is related to a worse prognosis in breast cancer[75 , 76] . In summary , SyNet predominantly appears to contain genes relevant to two main processes in the progression of breast cancer: increased cell proliferation and the process of metastasis . Although many genes have not previously been specifically linked to breast cancer prognosis , their role in regulating different stages of replication and mitosis points to a genuine biological role in the progression and prognosis of breast cancer . We next sought to investigate the similarity between SyNet and existing biological networks that directly or indirectly capture biological interactions . To enable a comparison with networks of different sizes , we compare the observed overlap ( both in terms of genes as well as links ) to the distribution of expected overlap obtained by shuffling each network 1000 times ( while keeping the degree distribution intact ) . Overlap is determined for varying network sizes by thresholding the link weights such that a certain percentage of genes or links remains . Results are reported in terms of a z-score in Fig 6 . Fig 6a shows that for the majority of networks a significantly higher than expected number of SyNet genes is contained in the top of each network . The overlap is especially pronounced for the tissue-specific networks , in particular the Breast-specific and Lymph node-specific networks , supporting our observation that SyNet contains links that are relevant for breast cancer . The enrichment becomes even more significant when considering the overlap between the links ( Fig 6b ) . In this respect , SyNet is also clearly most similar to the Breast-specific and Lymph node-specific networks . We confirmed that these enrichments are not only driven by the correlation component of SyNet by repeating this analysis with a variant of the SyNet network without the correlation component ( i . e . only average and synergy of gene pairs are used for pair-ranking; see S10 for details ) . It should moreover be noted that , although a highly significant overlap is observed , the vast majority of SyNet genes and links are not present in the existing networks , explaining the improved performance obtained with NOPs using SyNet . Specifically , out of the 300 genes in SyNet , only 142 are contained within the top 25% of genes ( n = 1005 ) in the Breast-specific network , and 151 in the top 25% of genes ( n = 1290 ) in the Lymph node-specific network . Similarly , out of the 3544 links in SyNet , only 1182 are contained within the top 25% of links ( n = 12500 ) in the Breast-specific network , and 617 in the top 25% ( n = 12500 ) of the Lymph node-specific network ( see S11 for details ) . We further confirmed that the overall trend in observed overlaps between SyNet and other networks does not change when the size of these networks ( in terms of the number of links ) are increased or reduced ( see S17 for details ) . In addition to direct overlap , we also aimed to investigate if genes directly connected in SyNet may be indirectly connected in existing networks . To assess this for each pair of genes in SyNet , we computed several topological measures characterizing their ( indirect ) connection in the biological networks . We included degree ( Fig 7a ) , shortest path ( Fig 7b ) and Jaccard ( Fig 7c ) ( see S1 Text for details ) . To produce an edge measure from degree and page rank ( which are node based ) , we computed the average degree and page rank of genes in a pair respectively . Furthermore , we produced an expected distribution for each pair by computing the same topological measures for one of the genes and another randomly selected gene . The results from this analysis supports our previous observation that the information contained in the links of SyNet is markedly—yet only partially—overlapping with the information in the existing networks . Notably , the similarity increases for networks of increased relevance to the tissue in which the gene expression data is measured ( i . e . breast tissue ) . Encouraged by the overlap with existing biological networks , we next asked whether links in SyNet can be predicted from the complete collection of topological measures calculated based on existing networks . To this end , we characterized each gene-pair by a set of 12 graph-topological measures that describe local and global network structure around each gene-pair . In addition to the degree , shortest path and Jaccard , we included several additional graph-topological measures including direct link , page rank ( with four betas ) , closeness centrality , clustering coefficient and eigenvector centrality ( see S1 Text for details ) . While converting node-based measures to edge based measures , in addition to using the average , we also used the difference between the score for each gene in the pair , similar to our previous work [77] . We applied these measures to all 10 networks in our collection yielding a total of 210 features . The gene-pairs are labeled according to their presence or absence in SyNet . Inspection of this dataset using the t-SNE [78] reveals that the links in SyNet occupy a distinct part of the 2D embedding obtained ( Fig 8a ) . We trained a Lasso and assessed classification performance in a 50-fold cross validation scheme where in each fold 1/50 of pairs in SyNet is kept hidden and the rest of pairs is utilized to train the classifier . To avoid information leakage in this assessment , we removed gene pairs from the training set in case one of the genes is present in the test set . Based on this analysis we find that a simple linear classifier can reach ~85% accuracy in predicting the synergistic gene relationships from SyNet ( Fig 8b , rightmost bar ) . The contribution from generic networks is notably smaller than for the tissue-specific networks . In particular the networks relevant to breast cancer are highly informative , to the extent that combining multiple networks no longer improves prediction performance . Further investigation of feature importance revealed that the page rank topological measure was commonly used as a predictive marker across folds . Apparently , while direct overlap between SyNet and existing networks is modest , the topology of the relevant networks ( i . e . breast-specific and lymph node-specific networks ) are highly informative for the links contained in SyNet . This corroborates findings from Winter et al . in which the page rank topological measure was proposed to identify relevant genes in outcome prediction [34 , 35 , 79] .
Although the principle of using existing knowledge of the cellular wiring diagram to improve performance , robustness and interpretability of gene expression classifiers appears attractive , contrasting reports on the efficacy of such approach have appeared in literature [21 , 28 , 35] . Consensus in this field has particularly been frustrated by an evaluation of a limited set of sub-optimal classifiers [21 , 23 , 28 , 35] , small sample size [18 , 24 , 26] , or the use of standard K-fold cross-validation instead of cross-study evaluation schemes , which results in inflated performance estimates [24 , 26] . For this reason , it remained unclear if network-based classification , and in particular network-based outcome prediction , is beneficial . Here , we present a rigorously cross-validated procedure to train and evaluate Group Lasso-based NOPs using a variety of networks , including tissue-specific networks in particular , which have not been evaluated in the context of NOPs before . Based on our analyses , we conclude that none of the existing networks achieve improved performance compared to using properly regularized classifiers trained on all genes . In this work we therefore present a novel gene network , called SyNet , which is computationally derived directly from the survival-labeled samples . The links in SyNet connect synergistic gene pairs . We followed a cross-validation procedure in which the inference of SyNet and validation of its utility in a NOP is strictly separated . We find that SyNet-based NOPs yields superior performance with higher stability across the folds compared to both the baseline model trained on all genes as well as models that use other existing gene networks . We therefore conclude that at least in outcome prediction problem , network guidance can improve model performance , but only if this network is phenotype-specific . Supporting this conclusion , we also show that a correlation network , which is dataset-specific but not phenotype specific , also improved performance but much less compared to SyNet . A major benefit of SyNet over manually curated gene networks is that its inference is purely data driven , and therefore not biased to well-studied genes . Post-hoc interpretation of the genes selected by a NOP that utilized SyNet is therefore expected to provide a more unbiased interpretation of the important molecular players underlying breast cancer and patient survival . Analysis of the genes contained in SyNet shows strong enrichment for genes with known relevance to breast cancer . More importantly , the largest subcomponent of SyNet is strongly linked to patient prognosis as it includes many genes with a known relation to the histological grade of the tumor . To investigate if SyNet captures known biological gene interactions , we extensively compared SyNet with existing networks . We find highly significant overlaps between links , indicating that SyNet connects genes that also have a known biological interaction . Despite this significant overlap , the majority of the SyNet links are not recapitulated by direct links in the existing networks . However , we find that accurate prediction of links in SyNet are possible if more complex graph topological descriptions of the indirect connections in the existing networks are employed . Interestingly , accurate predictions are only obtained when using the breast specific networks . Apparently , although the information contained in SyNet is similar to other gene interaction networks , the wiring of SyNet much better supports GL-based classification . This might explain why using existing biological networks in NOPs directly is unsuccessful and why graph topological measures have been successful in identifying relevant genes in outcome prediction [34 , 35 , 79] . Taken together , our results underline that network-based outcome prediction is a promising approach to improving patient prognosis prediction and therefore can provide an important contribution towards more personalized healthcare . At the same time , the SyNet approach provides an unbiased interactome which makes the NOP more amenable for model interpretation , thus providing important insights into the etiology of the disease under study .
We hypothesized that , in order to improve outcome prediction by network-based classification , interconnections in the network should correspond to gene pairs for which integration yields a performance beyond what is attainable by either of the individual genes ( i . e . synergy ) . Accordingly , we formulated the synergy Sij between gene i and gene j as Sij=AijMax ( Ai , Aj ) where Ai , Aj and Aij respectively represent the Area Under Curve ( AUC ) of gene i , the AUC of gene j and the AUC of meta-gene ij formed by aggregation of gene i and gene j . Meta-gene formation is carried out by a linear regression model which demonstrated superior performance in our experiments ( see S1 for details ) . Cross-validation performance of the linear regression ( see section “Cross validation design” for details ) is obtained and the median of 65 AUCs ( 13 folds and 5 repeats ) is used as the final score Aij for each pair . The AUC of the individual genes ( i . e . Ai and Aj ) is obtained in a similar fashion . Defining the synergy as a function of AUC yields a phenotype-specific ( i . e . label-specific ) measure which effectively ignores extraneous relationships between gene pairs that are not relevant in outcome prediction . The synergy measure Sij depends on the performance of individual genes where poorly performing genes tend to achieve higher degree of synergy compared to two predictive genes ( see S2 for corresponding analysis ) . In order to account for this effect , the average AUC of individual genes is included as a second criterion . Furthermore , our preliminary tests confirmed previous findings [8 , 23 , 80] , that integrating highly correlated genes ( which reduces meta-gene noise ) may improve survival prediction . For this reason , we added correlation of pairs as a third criterion . To combine these three measures , each measure is normalized independently between [0 , 1] and then combined into an overall fitness score Fij for gene pair ij: Fij=- ( 1-Sij¯ ) 2+ ( 1-Mij¯ ) 2+ ( 1-Cij¯ ) 2 Here , Mij and Cij represent mean AUC and absolute spearman correlation of gene i and j respectively . Bars above letters indicate that the corresponding values are normalized to the [0 , 1] interval . Employing the Dutch grid infrastructure , we quantified the fitness for all 69 million possible pairs of genes ( n = 11748 ) . Fig 1c visualizes the fitness of all pairs in a three-dimensional space . Finally , the top 50 , 000 pairs with highest fitness are considered as SyNet . Accurately estimating survival risk and identifying markers relevant for progression of a complex disease such as breast cancer requires a large number of samples [11] . To this end , samples from METABRIC [81] ( n = 1981 ) are combined with 12 studies collected in ACES [21] ( n = 1606 ) as well as samples from the TCGA breast invasive carcinoma dataset [82] ( n = 532 ) ( see S1 Text for details ) . Collectively , these datasets , spanning 14 distinct studies , form a compendium encompassing 4129 samples . To the best of our knowledge , the data combined in this paper represents the largest breast cancer gene expression compendium to date . As a result , our compendium should capture a large portion of the biological heterogeneity among breast cancer patients , as well as technical biases originating from the variability in platforms and study-specific sample preparations [83] . This variability will assist the trained models to achieve better generalization which is crucial in real world application of the final classification model [9 , 13 , 84] . To correct for technical variations that may arise during the library preparation , initially the expression data within each study is quantile normalized and then batch-effect corrected using Combat [85] where the outcome of patients was modeled as an additional covariate to maintain the variance associated with the prognostics . This procedure was shown to perform well among many batch effect removal methods [86 , 87] . Successful removal of batch effects was confirmed using t-SNE visualization [78] ( See S4 for details ) . The label for each patient corresponds to overall survival time ( or recurrence free survival if available ) with respect to a 5-year threshold ( good vs . poor outcome ) . Ascertaining the relevance of networks in outcome prediction should be performed using a robust predictor capable of providing adequate performance in prognostic prediction . Previous assessments in this regard have been limited to only few classifiers [21 , 23 , 28 , 35] . To identify the optimal predictor , we have compared performance of wide range of linear and nonlinear classifiers ( see S3 for details ) . Supporting our previous findings [8] , this evaluation demonstrates that simple linear classifiers outperform the more complex ones , with the regularized linear classifier ( Lasso ) reaching the highest AUC . This classifier supports both classical and well as network-based prediction by its derivative called Group Lasso ( GL ) [88] . The GL is structurally analogous to standard Lasso with the exception of the way in which the regularization is performed; Lasso applies regularization to genes while GL enforces selection of groups of genes ( See S1 Text for details ) . In order to incorporate network information in the GL , similar to our previous work [8] , each gene in the corresponding network is considered as seed gene and together with its K neighbors the group structure provided to the GL . Priority of neighbor selection is determined by edge weights between each neighbor and corresponding seed gene . The hyperparameters for each classifier ( e . g . K in the GL ) are determined by means of a grid search in the inner cross validation loop ( see S5 for schematic overview ) . For comparison , we include three well-known NOPs in our analysis . Park et al . utilized hierarchical clustering to group highly correlated genes [23] . Each group is summarized into a meta-gene by averaging the expression profile of the genes in that group . These meta-genes are then employed as regular features to train a Lasso classifier . The optimal cluster size for hierarchical clustering is identified by iterative application of Lasso in an inner cross-validation . Chuang et al . employs a greedy search to define subnetworks [18] . This is done by iteratively expanding a sub-network initiated from a seed gene guided by a supervised performance criterion which halts when performance no longer increases ( in the training set ) . After groups are formed , the meta-genes are constructed by averaging expression of each gene within each group similar to Park et al . Finally , Taylor et al . focus on hubs ( i . e . highly connected genes , degree>5 ) in a network [24] . To identify dysregulated subnetworks , the change in correlation between each hub and its direct neighbors across two classes of outcome ( poor vs . good ) is assessed . Meta-genes are formed from candidate subnetworks similar to the procedure employed by Park et al . In addition to SyNet , we considered a range of publicly available networks , including generic networks ( HumanInt , BioPlex , BioGRID , IntAct and STRING ) as well as a correlation network ( Corr ) which was previously shown to be an effective network in outcome prediction [8 , 23] . Additionally , we assessed five tissue-specific networks ( including brain , kidney , ovary , breast , lymph node ) that are recently introduced by Greene et al . [44] . These tissue-specific networks are inferred by integrating protein-protein interactions collected from Human Protein Reference Database [89] and tissue-specific information from BRENDA tissue ontology [90] and then filtered using expert-selected Gene Ontology ( GO ) terms . The tissue-specificity of each network is then validated by a comprehensive collection of expression and interaction datasets encompassing about 38000 conditions collected from approximately 14000 publications . To the best of our knowledge , our study is the first to evaluate tissue-specific networks in the context of NOPs . To maintain a reasonable network size , we utilized only the top 50 , 000 links ( based on the link weight ) in each network ( similar to number of links in SyNet ) . For the only unweighted network , HumanInt [38] , all interactions ( n = ~14k ) were included and links were weighted according to the average degree of the two interacting genes . Moreover , a randomized version of each network is constructed by shuffling nodes in the network which destroys the biological information of the links while preserving the overall network structure ( see S1 Text for full details on preparation of networks ) . In order to ascertain if network information truly aids outcome prediction , the evaluation should be based on a rigorous cross-validation that closely resembles the real-world application of these models . To this end , we perform cross-study validation in order to mimic a realistic situation in which a classifier is applied to data from a different hospital than it was trained on [7] . Briefly , one study is taken out for validation of the final performance ( outer loop test set ) . SyNet inference and NOP training are carried out on the 13 remaining studies ( outer loop training set ) . Within each fold of the outer loop training set , again one study is left out to obtain the inner loop test set and the rest of studies for inner loop training set . The inner loop training set is sub sampled ( with replacement ) to 70% and regression is performed for every gene as well as gene pairs ( identical set of samples are used across all genes and pairs ) . The AUC scores ( Ai , Aj and Aij ) are calculated on the inner loop test set . This is repeated 5 times . To train a NOP for this fold , a new inner loop training set is formed by redrawing 70% of the samples from the outer loop training set . This set is also used to infer correlation network . To assess the final performance of the NOP the outer loop test set is used ( see S5 for a detailed schematic ) . Our initial experiments showed a large variation of performance across studies ( see S6 for details ) . To prevent this variation from influencing our comparisons , assignment of samples to folds in both inner and outer cross-validation loops are kept identical across all comparisons throughout the paper . We used Area Under the ROC Curve ( AUC ) as the main measure of performance in this paper .
|
Cancer is caused by disrupted activity of several pathways . Therefore , to predict cancer patient prognosis from gene expression profiles , it may be beneficial to consider the cellular interactome ( e . g . the protein interaction network ) . These so-called Network based Outcome Predictors ( NOPs ) hold the potential to facilitate identification of dysregulated pathways and delivering improved prognosis . Nonetheless , recent studies revealed that compared to classical models , neither performance nor consistency ( in terms of identified markers across independent studies ) can be improved using NOPs . In this work , we argue that NOPs can only perform well when supplied with suitable networks . The commonly used networks may miss associations specially for under-studied genes . Additionally , these networks are often generic with low coverage of perturbations that arise in cancer . To address this issue , we exploit ~4100 samples and infer a disease-specific network called SyNet linking synergistic gene pairs that collectively show predictivity beyond the individual performance of genes . Using a thorough cross-validation , we show that a NOP yields superior performance and that this performance gain is the result of the wiring of genes in SyNet . Due to simplicity of our approach , this framework can be used for any phenotype of interest . Our findings confirm the value of network-based models and the crucial role of the interactome in improving outcome prediction .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"genetic",
"networks",
"breast",
"tumors",
"statistics",
"protein",
"interaction",
"networks",
"cancers",
"and",
"neoplasms",
"oncology",
"mathematics",
"forecasting",
"network",
"analysis",
"genome",
"analysis",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"genomics",
"mathematical",
"and",
"statistical",
"techniques",
"gene",
"expression",
"breast",
"cancer",
"proteomics",
"prognosis",
"biochemistry",
"diagnostic",
"medicine",
"gene",
"regulatory",
"networks",
"gene",
"identification",
"and",
"analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"gene",
"prediction",
"statistical",
"methods"
] |
2019
|
A data-driven interactome of synergistic genes improves network-based cancer outcome prediction
|
Myxococcus xanthus cells self-organize into aligned groups , clusters , at various stages of their lifecycle . Formation of these clusters is crucial for the complex dynamic multi-cellular behavior of these bacteria . However , the mechanism underlying the cell alignment and clustering is not fully understood . Motivated by studies of clustering in self-propelled rods , we hypothesized that M . xanthus cells can align and form clusters through pure mechanical interactions among cells and between cells and substrate . We test this hypothesis using an agent-based simulation framework in which each agent is based on the biophysical model of an individual M . xanthus cell . We show that model agents , under realistic cell flexibility values , can align and form cell clusters but only when periodic reversals of cell directions are suppressed . However , by extending our model to introduce the observed ability of cells to deposit and follow slime trails , we show that effective trail-following leads to clusters in reversing cells . Furthermore , we conclude that mechanical cell alignment combined with slime-trail-following is sufficient to explain the distinct clustering behaviors observed for wild-type and non-reversing M . xanthus mutants in recent experiments . Our results are robust to variation in model parameters , match the experimentally observed trends and can be applied to understand surface motility patterns of other bacterial species .
Myxococcus xanthus is a model organism for studying self-organization behavior in bacteria [1] . These rod-shaped bacteria are known for their ability to collectively move on solid surfaces . Depending on environmental conditions , this collective movement allows cells to self-organize into a variety of dynamic multi-cellular patterns [2 , 3] . For instance , when nutrients are abundant , cells collectively swarm into surrounding spaces [1] . When cells come into direct contact with other bacteria that can serve as their prey , M . xanthus cells self-organize into ripples , i . e . , bands of traveling high-cell-density waves [4–6] . Alternately , if nutrients are limited , cells initiate a multi-cellular development program resulting in their aggregation into 3-dimensional mounds called fruiting bodies [7 , 8] . Self-organization in M . xanthus requires coordination among cells and collective cell motility [1 , 5 , 6 , 9 , 10] . Despite decades of research , the mechanisms that allow for motility coordination in M . xanthus are not fully understood . In particular , the ability of cells to collectively move in the same direction is crucial to the observed multi-cellular behavior at various stages of their lifecycle [11–13] . Given that individual rod-shaped M . xanthus cells move along their long axis , coordination of cell direction in a group can be achieved by forming aligned cell clusters . Such clusters are observed in a variety of environmental conditions: low-density swarming [13] , aligned high-cell-density bands in ripples [12] and long streams of aligned cells during the initial stages of aggregation [14 , 15] . However , the mechanisms responsible for this collective cell alignment are not completely clear . Another important aspect of M . xanthus cell motility is the periodic reversal of its travel direction by switching the cell’s polarity i . e . , flipping the head and tail poles . Recent experiments indicate that the clustering behavior of M . xanthus cells is dramatically affected by variation in cell reversal frequency [16 , 17] . Starruẞ et al . [16] observed that , above a certain cell density , non-reversing M . xanthus mutants ( A+S−Frz− ) form large moving clusters , whereas reversing wild-type cells organize into an interconnected mesh-like structure . In a recent study , Thutupalli et al . [17] observed that starving wild-type M . xanthus cells increased their reversal frequency with time , which resulted in a change in their clustering behavior from aggregates ( large clusters ) to streams ( elongated clusters ) . In addition , this study indicated that reversing and non-reversing cells differ in their dynamic behavior inside clusters . Reversing ( wild-type ) cells form stream-like clusters that appear stationary , and the cells move within the clusters . In contrast , non-reversing ( ΔfrzE ) mutants form flock-like isolated clusters that move around , and the cells inside clusters appear to be moving with the same velocity as the clusters . Therefore , our ability to explain cell alignment into clusters and variation of cell clustering behavior with changes in reversal frequency is essential for successful models of all self-organization phenomena . Several prior studies [16 , 18 , 19] attempted to understand the cell clustering process in M . xanthus using mathematical and computational approaches . Starruẞ et al . [16] developed a kinetic model , inspired from coagulation theory for colloidal particles , in which cell clusters’ dynamics resulted from their fusion , splitting , and growth-decay processes . Using this model , they were able to explain the observed cluster size distribution for non-reversing cells . However , this model could not explain the cell clustering behavior for wild-type ( reversing ) cells . In another study , Harvey et al . [18] showed symmetry breaking between free cells ( uniform gas phase ) and nematically ordered cell clusters ( dense phase ) using a multi-phase continuum model . However , this model did not explicitly study the effects of changing reversal frequency on clustering , and the equations developed are limited to 1D and quasi-1D settings . Furthermore , both the models follow phenomenological approaches and do not provide a clear relationship between the model assumptions and individual cell behavior . In this study , we overcome the limitations of previous approaches by connecting the individual cell behavior with collective cell motility through a biophysical agent-based model . Our overarching hypothesis is that cell clustering can be explained solely via mechanical interactions among cells and between cells and substrate . In other words , the observed patterns do not rely on biochemical signals such as chemotaxis . To test this hypothesis , we simulate interactions among a large number of cells through an agent-based simulation ( ABS ) framework . Using this framework , we first study the formation of aligned cell clusters in non-reversing M . xanthus cells and later extend our investigation to reversing cells . Furthermore , we investigate the effect of cell-substrate interactions , such as slime-trail-following , on the clustering patterns . The results of our simulation are compared with experimental data from the literature and can be applicable to other bacteria that display surface motility .
First , we investigated whether mechanical interactions among M . xanthus cells would be sufficient to induce aligned cell cluster formation . This approach was motivated by our previous study [20] , which demonstrated alignment in cell pairs as a result of head-to-side collision , and soft-condensed matter models for clustering in self-propelled rigid rod particles [21–24] . We hypothesized that successive collisions of cells with previously aligned cell clusters will result in the formation of even larger clusters . Thus , we simulated mechanical interactions among non-reversing cells , similar to self-propelled rod models , but with realistic cell flexibility values . For this step , we have used the bending stiffness value ( kb ) for M . xanthus cells from our previous study [20] , which reproduces realistic pair-wise cell collision behavior in model agents . Under these estimates of kb , we studied clustering behavior of the model M . xanthus cells in our ABS framework at different cell densities ( η , defined as the fractional area occupied by all cells in the simulation region ) . To simulate mechanical interactions of cells moving on a 2D surface , we used our previously developed framework—briefly described below ( see Methods for further details ) . In this framework , each agent consists of multiple segments , enabling a realistic mechanical model of a single M . xanthus cell . To this end , we use a connected string of nodes with linear and angular springs between nodes to simulate elastic behavior . Agents move forward through propulsive forces acting on the nodes tangential to the cell length ( towards the next node ) . This is similar to the force generation through multiple motor protein complexes distributed along the cell length as observed by recent models of M . xanthus gliding motility [25–28] . Agents experience drag forces opposing their motion due to the surrounding fluid . Adhesive attachments to the underlying substrate at nodes resist lateral displacement of agents during collisions ( the focal adhesion model of gliding motility [26] ) . At low densities , M . xanthus are known to move as a monolayer of cells . Therefore , collisions among agents are resolved by applying appropriate forces on nodes that keep agents from overlapping . Agents move over a 2D simulation space with periodic boundary conditions according to the net forces acting on their nodes . We introduce random noise in agent travel direction by altering the direction of propulsive force on the front node . We observe the agent behavior by solving Newton’s equations of motion on nodes to obtain their position and velocity at each time step of the simulation . We use the Box2D [29] physics library to solve these equations of motion and efficiently handle the excluded-volume forces . We start the simulation by initializing the cells one by one in the simulation region at random positions and with random orientations until the desired cell density is reached . While initializing , we accept only the cell configurations that do not result in cell overlap . As soon as the simulation begins , cells start moving and colliding with their neighboring cells and , as a result , align along their major axis [20] . This alignment is nematic [30]: aligned cells can move in the same or opposite directions depending on the initial orientation of cells . When aligned cells move in the opposite directions , they separate; however , when they move in the same direction , a small cluster of aligned cells is formed . These clusters grow in size as more cells join through collisions or due to merging with other cell clusters . Clusters shrink in size as peripheral cells leave the cluster due to random change in their travel direction ( S1 and S2 Movies ) . We quantify the evolution of clusters through cluster size distribution ( CSD , see S1 Text ) . After approximately 180 min of simulation time , the CSD is stabilized ( S1 Fig ) , and we observe that cells in the simulation regions are distributed among clusters of different sizes , while few cells remain isolated . Depending on the cell density ( η ) , we observe a variation in the cluster size distribution and in the number of isolated cells . Cells form stable clusters ( containing > 102 cells ) only for sufficiently high cell densities ( η ≥ 0 . 16 ) ( Fig 1A–1D ) , while cells largely remain isolated for lower densities ( η = 0 . 08 ) . We have quantified the effect of increasing cell density ( η ) on clustering behavior by measuring the mean cluster size 〈m〉 ( refer to S1 Text for details on the quantification procedures ) at each cell density value . We observe that an increase in cell density results in an increase in mean cluster size ( Fig 1E ) . We have quantified the alignment within the cell clusters using a mean cell orientation correlation , C ( r ) = 〈cos ( 2Δθr ) 〉 , as a function of the neighbor cell distance r ( Fig 1F ) . Here , Δθr is the angle deviation between the orientations ( θ ) of a pair of agents whose center nodes are separated by a distance r ( see Methods ) . We use 2Δθ to ensure that correlation values in parallel and anti-parallel alignment configurations remain the same [31] . The orientation correlation results confirm that , in comparison with the initial distribution , clustering results in longer-distance orientation correlation for high cell densities . We observe that , immediately after the start of the simulation ( 1 min ) , cells exhibit very low correlation with their immediate neighbors ( r = 2 − 3 μm ) . However , after a long simulation time ( 180 min ) , we observe a large increase in cell orientation correlation with neighbor distances ( except for η = 0 . 08 , Fig 1F ) , indicating the formation of larger aligned clusters ( refer to S2 Fig for the evolution of orientation correlation with time ) . To test the robustness of our results , we have varied the cell flexibility ( kb ) values over a wide range ( 0 . 1x – 10x ) and studied the cell clustering behavior in our simulations . We observed that our model agents formed clusters except for the case of very high cell flexibility values ( 0 . 1x , kb = 10−18 N . m ) ( S3A–S3F Fig ) . Furthermore , mean cluster sizes increased with increases in cell densities for all cell flexibility values ( S3G Fig ) . Interestingly , increases in cell flexibility decreased the mean cluster sizes . Thus , we observe that flexible agents can form aligned clusters through mechanical collisions for sufficiently high cell densities ( η ≥ 0 . 16 ) , similar to self-propelled hard rods [19] . Furthermore , these cell clusters from our simulations are very similar to the isolated cell clusters experimentally observed for non-reversing M . xanthus ( frz− ) cells [16 , 19] . Next , we investigated the effect of cell reversals on clustering behavior . We introduced periodic reversals of cell travel direction ( reversal period = 8 min [32] ) in our model agents . Similar to M . xanthus cells , each reversal results in a switch of the agent polarity i . e . , flipping of the head and tail nodes . Surprisingly , with the addition of periodic cell reversals , cells failed to form large clusters even after a long simulation time ( 180 min ) ( Fig 2A , S3 Movie ) . Furthermore , we observed that increases in cell density did not improve the mean cluster sizes significantly ( Fig 2B , black line ) . Even when we started with cells that initially formed clusters by simulating non-reversing cells first for 90 min and then turned on cell reversals , we observed the destruction of existing cell clusters within approximately 30 min ( S4 Fig ) . Thus , our simulation results indicate that steric alignment is not sufficient for formation of large aligned clusters in a population of periodically reversing agents . However , given that wild-type M . xanthus cells reverse their polarity but still form clusters , additional interactions must be included in our model to explain M . xanthus clustering behavior . In our first attempt to correct this , we tested whether cohesive interactions among M . xanthus cells [33] can restore clustering . Studies on colloidal particles indicate that adhesion between particles can lead to their clustering [34] . M . xanthus cells secrete exopolysaccharide ( EPS ) proteins and fibrils on their surface , and these are observed to form a network with the surface fibrils of other cells that are in close contact , resulting in cell-cell cohesion [35 , 36] . These cohesive interactions can keep cells together and thus may lead to clustering in reversing M . xanthus cells . We investigated this mechanism by introducing lateral adhesion forces between neighboring agent nodes in our simulations ( Refer to Methods ) . However , we observed that adhesive interactions between neighbor cells did not lead to significant cell clustering for reversing cells , even with high adhesion forces ( S5 Fig ) . Thus , lateral adhesions are not sufficient to stabilize the clusters of reversing cells . To understand the rationale behind why cell reversals prevent the formation of large clusters , we examined the cell clustering dynamics in our simulations with and without cell reversals . For non-reversing cells , we observe that clusters grow in size due to collisions with new cells and that cells inside the clusters are unable to leave their cluster . At steady-state , cluster size is determined by a balance between the flux of peripheral cells leaving the cluster and new cells joining the cluster , similar to the kinetic theory developed in Ref . [19] . In contrast , for reversing cells , we observed that , even though mechanical collisions often lead to the transient formation of small clusters , these clusters fail to grow and stabilize . This occurs because , upon reversal , cells from the cluster interior move past the other cells in the opposite direction and leave the cluster . Furthermore , random changes in their travel direction prevent them from returning to their original clusters after another reversal . This also explains why adhesive cell interactions failed to result in the clustering of cells in our simulation . Lateral adhesive interactions do not stop cells from leaving the clusters after reversal and cannot influence the direction of cell movement once it leaves the existing cluster . Based on the results thus far , we conclude that an additional mechanism that could reduce random orientation changes in the cells could help overcome the destabilizing effects of reversals on clustering . A possible mechanism for this is suggested by the observation of slime-trail-following by M . xanthus cells . M . xanthus cells secrete slime , a polymeric gel , from their surface , and it is deposited on the underlying substrate as long trails during cell movement [37] . Furthermore , cells tend to follow their own trails after reversal , and , when in contact with slime trails deposited by others , cells can reorient and follow these [38] . Accordingly , we hypothesize that slime trails act as an orientation memory that reduces cells’ ability to randomly change travel direction and assists in clustering for reversing cells . We investigated the above mechanism of cell clustering based on slime-trail-following using our ABS framework . As the mechanistic basis of slime-trail-following by M . xanthus cells is not fully clear , we opt for a phenomenological model of slime-trail-following by reorienting part of the propulsive force on a cell’s leading pole ( head node ) parallel to the slime trail it is crossing ( Refer to Methods for more details ) . The results of these simulations indicate that the slime-trail-following mechanism restored clustering for reversing cells ( Fig 2C , S4 Movie ) . This is reflected by a significant increase in mean cluster sizes ( green line in Fig 2B ) for slime-trail-following cells compared to cells that do not follow slime trails ( dashed line ) . Additionally , slime-trail-following also increased large-distance orientation correlations of cells , indicating the formation of aligned cell clusters ( Fig 2D ) . Notably , the cell clusters in our simulations for reversing cells with the slime-trail-following-mechanism resemble an interconnected mesh-like structure ( Fig 2C ) . These clusters are distinct from the freely moving isolated cell clusters of non-reversing cells ( Fig 1C ) . However , these interconnected cell clusters in our simulations are very similar to the interconnected mesh-like structure observed for wild-type ( reversing ) M . xanthus cells in experiments [16] . To investigate the robustness of clustering to the values of unknown parameters and to demonstrate key features of the model that are essential for clustering , we investigated effects of variation in the slime-trail-following ability of cells . For this , we perturbed the parameters that affect the slime-trail-following mechanism in our model: the slime effectiveness factor ( εs ) , which controls the ability of a cell to follow a slime trail , and the slime trail length ( Ls ) , which controls the memory effect of a cell path ( refer to Methods for details ) . High εs values decrease a cell’s chance of escaping from the slime trail , whereas high Ls values increase the chance of a cell to encounter slime trails from other cells . We have varied both parameters over a wide range in our simulations: εs ( 0 . 1 to 1 . 0 ) and Ls ( 0 . 16 to 11 μm ) . For short slime trail length ( Ls = 0 . 16 μm ) and a low slime effectiveness value ( εs = 0 . 1 ) , reversing cells show a dispersed cell pattern with minimal cell clustering ( Fig 3A ) . This dispersed cell pattern is very similar to the situation for cells without slime-trail-following ( Fig 2A ) . The underlying pattern of slime distribution in the inset shows minimal slime paths in the simulation , which do not effectively result in cells following others . Increasing the slime trail length to a higher value ( Ls = 11 μm ) but keeping the slime effectiveness value low ( εs = 0 . 1 ) did not improve cell clustering significantly ( Fig 3B ) . Although cells are able to leave longer slime trails , creating an interconnected slime network ( inset ) , the low slime effectiveness ( εs ) value allows cells to easily escape from the slime paths , and the slime-trail-following cannot effectively stabilize the formed clusters . In the same fashion , an increased slime effectiveness value ( εs = 1 . 0 ) but a low slime trail length ( Ls = 0 . 16 μm ) also did not result in significant cell clustering ( Fig 3C ) . Here , even though cells are able to follow slime trails effectively , slime trails are not long enough for other cells to follow , and thus cells are more or less separated except for small cell clusters . However , with high slime effectiveness ( εs = 1 . 0 ) and long slime trails ( Ls = 11 μm ) , cells are able to produce the normal cell clustering pattern for reversing cells ( Fig 3D ) . Here , long slime trails allow for cells to follow other cells’ slime trails , thus producing an interconnected slime network , and the high slime effectiveness factor prevents cells from escaping from slime paths and thereby results in a mesh-like clustering of cells . Thus , we observe that high slime-trail-following efficiency and sufficiently long slime trails allow for reversing cells to form cell clusters . To further investigate the robustness of the slime-trail-following mechanism on agent clustering behavior , we have measured the mean cluster sizes via simulation for variations in slime effectiveness and slime trail length over a wide range of values ( εs = 0 . 1−1 . 0; Ls = 0 . 2−11 μm- 64x change in slime production rate; see Methods for details ) . Our results indicate that except for very short slime trails ( Ls ≤ 1 μm ) , increases in the slime effectiveness value increased the mean cluster sizes ( Fig 3E ) . Similarly , increases in the slime trail length resulted in significant increases of mean cluster sizes except for very low slime effectiveness values ( Fig 3F ) . Thus , reversing agents along with the slime-trail-following-mechanism can form clusters over a wide range of model parameters . To further assess our clustering model , we decided to quantitatively compare our model predictions with the available experimental data on clustering behavior for both reversing and non-reversing strains of M . xanthus . To this end , we quantified the cell clustering behavior in our simulations by measuring the cluster size distribution , cell path maps , and cell visit frequency distribution from our simulations and compared our results with experiments reporting similar metrics [16 , 17] . First , we compared the cell cluster size distribution from our simulations with experiments of Starruẞ et al . [16] . For this , we performed simulations with the same cell density as in the experimental conditions for both reversing and non-reversing cells . We measured the cluster size distribution ( CSD ) from our simulations and plotted the probability , p ( m ) , of finding a cell in a cluster of size m as a function of cluster size ( solid lines in Fig 4A and 4B ) and compared with the experimentally observed distribution ( symbols ) . We observe that our simulation results qualitatively follow a similar trend to that of the experimental data . We chose model parameters ( slime effectiveness , εs; slime trail length , Ls ) to produce an approximate match . Global parameter optimization could further improve the agreement but was not performed . At small cell densities ( η = 0 . 08 ) , both reversing and non-reversing cells show a monotonically decreasing CSD with a large number of cells either being isolated or belonging to small clusters ( m ∼ 10 − 102 ) . However , no clusters larger than 102 cells are observed . Nevertheless , with increases in cell density ( η ) , non-reversing cells show a power-law distribution for CSD ( mβ , β = −0 . 90 – closely matches with the result β = −0 . 88 from Starruẞ et al . [16] ) , and a significant number of cells now belong to large clusters ( m ∼ 102 − 103 ) . In contrast , reversing cells show a decreasing CSD with increases in cluster size , and the largest clusters formed are limited to < 400 in size even at high cell densities . Next , inspired by recent experimental studies indicating that wild-type ( reversing ) and ΔFrzE ( non-reversing ) M . xanthus mutants form distinct cell clusters that differ in their shape and dynamic behavior [17] , we investigated these phenomena in our simulations . For this , we traced the cell paths over time and plotted the cell visit frequency of sites in the simulation region as a heat map for 2 consecutive hours after an initial transition period of 60 min ( Fig 4C and 4D ) . We observed localized high-frequency visit areas and changing shapes of cell trace paths over time for non-reversing cells ( Fig 4C ) , indicating the formation of large clusters that move all over the simulation region ( S5 Movie ) . In contrast , reversing cells organized into interconnected clusters that resemble a mesh-like structure , and the shape of the structure itself remained approximately the same over time ( Fig 4D , S6 Movie ) . Furthermore , the gap regions in the mesh structure ( white areas ) mostly remain free of cells or show very low visit frequency , indicating that reversing cells are confined within the cluster network ( clearly seen for high-slime-trail-following-efficiency parameters , e . g . , Ls = 11 μm , εs = 1 . 0; see S4 Movie ) . Additionally , we have quantified the probability of cell visits , p ( N ) , as a function of visit frequency , N , in our simulations for both reversing and non-reversing agents ( Fig 4E ) . We observe that simulations with reversing cells show a large fraction of sites with high visit frequencies ( N = 20 – 50 visits for a 60-min interval ) compared to non-reversing cells . Thus , reversing cells in the simulation region frequently visit specific sites , indicating stationary cluster structures . These results are qualitatively consistent with the observations of Thutupalli et al . [17] on the dynamic behavior of clusters .
Aligned cell clusters are crucial for formation of the multicellular structures observed during the M . xanthus lifecycle [12–15] . However , the mechanisms responsible for the cell alignment and clustering were not completely understood . Inspired by the studies of clustering in self-propelled hard-rods through mechanical collisions [21–24] , we have developed an agent-based simulation framework to investigate mechanical collision-based cell clustering in M . xanthus . In this framework , each agent is based on a biophysical model of an individual M . xanthus cell that realistically mimics flexible cell motility behavior . The results from our simulations show that non-reversing flexible model agents can form clusters through mechanical collisions alone under realistic cell bending stiffness values of M . xanthus cells . However , the addition of periodic cell reversals eliminated the cell clusters in our simulations . Thus , we observe that mechanical collisions alone are insufficient for cell clustering of reversing cells . We hypothesized an additional mechanism of cell clustering based on slime-trail-following by M . xanthus cells . As expected , slime-trail-following by cells restored clustering for reversing cells . By varying the parameters in our model , we observe that effective slime-trail-following and long slime trails are required for cell clustering using the slime-trail-following mechanism . We quantified cell clustering behavior from our simulations and compared our results with experiments for both non-reversing and reversing cells . We observe that our simulation results qualitatively agree with experimental cell clustering behavior . Thus , our analysis shows that M . xanthus cells can form aligned clusters through mechanical collisions and slime-trail-following . We believe that the following mechanism enables the reversing M . xanthus cells to form clusters through slime-trail-following ( Fig 5A ) : a single M . xanthus cell leaves a slime trail while moving on a substrate and traces back its own trail while reversing and thus reinforces its own slime trail . When other cells cross this trail , they reorient and align with this slime trail and start following it . This results in a positive feedback mechanism where newly joined cells in the slime trail further reinforce the trail with their own slime , causing more cells to join the trail . Thus , more cells aligned with the original slime trail are recruited into the trail , resulting in a cluster of aligned cells . Within a cluster , cells maintain alignment with neighbor cells through mechanical interactions . In the current study we limited cell densities ( η ) to 0 . 32 due to the limited availability of experimental data [16] . However , to extrapolate our conclusions , we have simulated the clustering behavior of cells for higher densities ( up to η = 0 . 60 ) . Results from these simulations indicate that cell alignment and clustering trough mechanical interactions also occur at these high densities ( S6 Fig ) . Interestingly we observe clustering of reversing cells at high cell densities even without slime-trail-following by cells ( S6B Fig ) . These results suggest diminished role of slime trails in collective cell alignment at these conditions as the whole area covered by cells is likely to contain slime . However , we have opted not to investigate these conditions at greater depth due to limitations of our current 2D simulation framework and cluster quantification metrics for such conditions . At high densities cells in our simulations form large continuous clusters such that separating and characterizing individual clusters is practically impossible . Moreover at high cell densities real M . xanthus cells are capable of moving on top of one another resulting in a multi-layered biofilm whose dynamics are different from that of low cell density scenario . These effects would be explored in depth elsewhere . Our simulations show that distinct clustering behaviors observed in M . xanthus mutant strains can be explained through mechanical interactions alone . Quantitative results from our simulations ( CSD , cell visit frequency ) follow the general trend as observed in experimental data [16 , 17] . Although our results do not exactly match with the experiments , this is understandable , as we were aiming to explain the observed cell clustering phenomena with a minimal interaction model . In our current model , we ignored many other interactions that exist among M . xanthus cells ( e . g . , the twitching of M . xanthus that uses type-IV pili to pull cells together ) . The addition of these processes along with further optimization of immeasurable parameters and choosing other model parameters from direct experimental observations ( e . g . , distribution of cell orientation changes , reversal time distribution ) could further improve our current model but are beyond the scope of this study . During development , M . xanthus cells exhibit circular aggregates , some of which later serve as initial fruiting body seed centers [14] . A recent study by Janulevicius et al . [39] , using an agent-based-model similar to our current model , concluded that cells form circular aggregates when the end parts of leading and lagging cell pairs interact through short-range active forces that keep the distance between cell pairs constant . They reasoned that such active forces can come through type-IV pili at the leading end of a cell interacting with the other cell surface or through adhesive interactions between cell poles . However , in our current simulations , we occasionally observed such circular aggregates ( Fig 5B ) without using any active interactive forces between end-to-end cell pairs . Furthermore , in contrast to the predictions of [39] , we observe that these aggregates do not rotate as rigid bodies as the agents inside the aggregate slide past one another ( S7 Movie ) . In our simulations , agents move with approximately the same speed , and , as a result , the angular velocity is higher for cells near the aggregate center . Thus , we argue that the circular aggregates observed in M . xanthus cells can be explained by slime-trail-following without active attractive forces between cells and propose that tracking cells in such aggregates can discriminate between the alternative models of their formation . Cell clustering and the alignment of cells inside the clusters play a major role in M . xanthus physiology . M . xanthus are predatory bacteria that feed on other bacteria by secreting proteolytic enzymes into their surroundings . To maximize their predation , these cells form groups that move together . The alignment of cells inside these groups allows for a dense packing of cells per a given area , thereby increasing their predation efficiency . Furthermore , the variations in cell-clustering behavior observed by Thutupalli et al . [17] with concomitant changes in cell reversal frequency may enable starving cells to optimize their search for nutrients . During the initial phase of starvation , M . xanthus cells exhibit a low reversal frequency that allows them to form flock-like clusters that explore their surroundings for nutrients . Once nutrients are found , cells switch to a high reversal frequency , thus enabling cells to form stationary cluster structures that allow them to conduct optimal nutrient gathering . Notably , cell clustering via slime following is observed in other bacterial systems . A recent study by Zhao et al . [40] showed that P . aeruginosa also uses a slime-trail-following mechanism to form initial cell clusters . Using cell-tracking algorithms and fluorescent staining of the secreted Psl exopolysaccharides ( slime ) , they concluded that P . aeruginosa cells form cell clusters by depositing slime trails that influence the motility of their kin cells that encounter these trails , to follow and further strengthen the trails . These processes results in a positive feedback loop reinforcing the trails . Our study shows that M . xanthus cells use a similar mechanism to form aligned cell clusters . Furthermore , our results show that differences in surface motility mechanisms ( e . g . , reversals or the ability to follow trails ) lead to distinct cell-clustering behaviors . These distinctions can be used to identify the nature of cell motility from snapshot images of bacteria for which direct observations on individual cells are difficult . Therefore , the mechanistic model of cell clustering and alignment developed here can be applicable to a wide class of bacteria displaying surface motility .
|
Many bacterial species are capable of collectively moving and reorganizing themselves into a variety of multi-cellular structures . However , the mechanisms behind this self-organization behavior are not completely understood . The majority of previous studies focused on biochemical signaling among cells . However , mechanical interactions among cells can also play an important role in the self-organization process . In this work , we investigate the role of mechanical interactions in the formation of aligned cell groups ( clusters ) in Myxococcus xanthus , a model organism of bacterial self-organization . For this purpose , we developed a computational model that simulates mechanical interactions among a large number of model agents . The results from our model show that M . xanthus cells can form aligned cell clusters through mechanical interactions among cells and between cells and substrate . Furthermore , our model can reproduce the distinct clustering behavior of different M . xanthus motility mutants and is applicable for studying self-organization in other surface-motile bacteria .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion"
] |
[] |
2015
|
Mechanism for Collective Cell Alignment in Myxococcus xanthus Bacteria
|
All organisms have developed mechanisms to respond to organ or tissue damage that may appear during development or during the adult life . This process of regeneration is a major long-standing problem in Developmental Biology . We are using the Drosophila melanogaster wing imaginal disc to study the response to major damage inflicted during development . Using the Gal4/UAS/Gal80TS conditional system , we have induced massive cell killing by forcing activity of the pro-apoptotic gene hid in two major regions of the disc as defined by Gal4 inserts in the genes rotund ( rn ) and spalt ( sal ) . The procedure ensures that at the end of a 40–48 hrs of ablation period the great majority of the cells of the original Rn or Sal domains have been eliminated . The results indicate that the damage provokes an immediate response aimed to keep the integrity of the epithelium and to repair the region under ablation . This includes an increase in cell proliferation to compensate for the cell loss and the replacement of the dead cells by others from outside of the damaged area . The response is almost contemporaneous with the damage , so that at the end of the ablation period the targeted region is already reconstructed . We find that the proliferative response is largely systemic , as the number of cells in division increases all over the disc . Furthermore , our results indicate that the Dpp and Wg pathways are not specifically involved in the regenerative response , but that activity of the JNK pathway is necessary both inside and outside the ablated domain for its reconstruction .
During the development or during the adult life organisms may suffer different kinds of insults , ranging from minor injuries to massive tissue damage or physical amputations . In consequence they have developed response mechanisms to reconstruct the damaged tissue . The phenomenon of regeneration is a major long -standing topic in Developmental Biology [1] . A principal factor in regeneration is the induction of the additional cell proliferation necessary to compensate for the lost tissue . Moreover , the new cells have to integrate with the pre-extant ones within the frame of the regenerating organ . Ultimately , regeneration is a matter of tissue homeostasis; it requires cooperative interactions between the old and the new parts of the tissue so that at the end the damaged organ is fully reconstructed in size and shape . This already points to the existence of overall control mechanisms that orchestrate the interactions of the different cell populations . In Drosophila melanogaster , studies on regeneration have been carried out principally in the imaginal discs , the precursors of adult cuticular structures . These are named after the adult structure for which they are determined , wing disc , leg disc , eye-antennal disc and so forth . The imaginal discs are well-defined developmental systems formed by small groups of progenitor cells - ranging 10-50 , depending on the disc [2] . They grow during the larval period , isolated from the larval tissues , and eventually differentiate sets of specific adult derivatives in stereotyped patterns . The imaginal discs are classical objects in developmental biology as they establish a clear distinction between determination and differentiation [3] . One important advantage of the imaginal discs is that after many years of study their development is very well known . Not only the basic growth parameters , but also the relevant developmental genes and signalling pathways have been described in detail [2] , [4] Our work has focused on regeneration in the wing disc , which is arguably the best characterised . The initial and final number of cells , the length of the cell proliferation period , the evolution of the cell division rates during development , etc , are known very precisely ( see [5] and references therein ) . Moreover , during development it undergoes a process of gradual determination by which different cell populations progressively adopt defined developmental fates . This process is visualised by the appearance of distinct cell linage blocks ( compartments ) associated with the localised expression of developmental genes such engrailed ( en ) , hedgehog ( hh ) , decapentaplegic ( dpp ) , apterous ( ap ) , wingless ( wg ) and several others of more restricted expression . At the end of the growth period the wing disc cells have attained a high degree of determination . This was demonstrated by experiments of transplanting disc fragments from fully developed discs into larvae close to metamorphosis , thus precluding further growth of the transplant . The transplanted fragments undergo metamorphosis along with the host larva and differentiate the adult patterns for which they were determined . These experiments showed that the various regions of the discs were strictly committed to differentiate specific adult patterns , such that it was possible to build maps of the presumptive adult structures , called “fate maps” , in the mature disc prior to differentiation [6] . This strict regionalization correlates with the localised expression in the mature disc of developmental genes such pannier [7] , vestigial , Distal-less , optomoter-blind , spalt [8] , cut , achaete/scute [2] that determine the differentiation and pattern of the adult structures . Classically , the regenerative potential of the imaginal discs was assayed by transplanting fragments from mature discs into adult hosts , which permitted almost indefinite growth ( Hadorn and Buck , 1962 ) . These experiments showed that fully determined discs possess a high regenerative potential: a disc fragment can , under appropriate circumstances , regenerate the rest of the disc [9] . Work , especially from the Schubiger laboratory [10] , [11] has established that ectopic activation of the wg gene is a major player in the process . Interestingly , the regeneration of Hydra head after amputation involves ectopic activation of the wg homologue Wnt in the region close to the cut [12] . Moreover , the regeneration of the zebrafish tail fin is also associated with Wnt/ß-catenin signalling , mediated by the Wnt10a and Wnt5a ligands [13] . The introduction of new methods of tissue ablation has provided a convenient approach to study the response of Drosophila discs to massive damage [14] . These methods rely on employing the Gal4/UAS system to force expression of pro-apoptotic genes that cause cell killing in specific disc domains . The usage of the temperature-sensitive form of the Gal4 -dominant suppressor Gal80 allows manipulation of the activity of the pro-apoptotic genes . It permits the control of the ablation time and the recovery process . The work of Smith-Bolton et al . , 2009 has established that after ablation of a large part of a growing disc , it is able to regenerate the damaged tissue to form normal adult structures . They find increased proliferation in the region close to the injury , associated with ectopic expression of the wg and dMyc genes in the damaged tissue . The activities of the Jun-N terminal Kinase ( JNK ) and the Hippo pathway are also required for regenerative growth [15] , [16] , [17] . Considered together , all the preceding experiments suggest a mode of regeneration in which the damage elicits a local response characterised by the activation of Wg/Wnt signalling pathways and associated with increased levels of cell proliferation . However , there are examples of full regeneration in which there is no evidence of activation of signalling pathways . After massive cell death induced by irradiation , the Drosophila wing disc is able to compensate for the loss of cells and differentiates adult structures of normal size and shape . This phenomenon is known as compensatory proliferation [9] and , unlike the experiments cited above , it does not depend on wg activation [18] . This suggests that there may be alternative mechanisms to respond to damage . Using a procedure as in Smith-Bolton et al . , 2009 , we describe the response of the wing imaginal disc to the ablation of two regions , as defined by the expression domains of the genes rotund ( rn ) [19] and spalt ( sal ) [20] . We find that the damage provokes a very rapid response of cells from outside the ablated tissue . These cells immigrate into the damaged territory and maintain the continuity of the epithelium and the genetic identity of the affected domain during the ablation time . These experiments reveal a very powerful homeostatic mechanism that is able to repair damaged domains almost contemporaneously with the damaging process . Unlike previous reports [14] , [15] we find that the proliferative response after localised damage is largely systemic , as all regions contribute to the regeneration by increasing the number of cells in division . This suggests the existence of an overall control mechanism that measures the cell population and triggers the regenerative response . Our results also indicate that the JNK pathway is involved in the repair process , but that the Dpp and Wg pathways do not play a specific role .
We targeted for massive cell killing two regions of the wing disc , defined by the rotund-Gal4 and salEPv-Gal4 lines . The Rotund ( Rn ) domain covers most of the wing pouch , approximately 40% of the entire disc . The SalEPv domain is smaller , occupies the central region of the wing pouch and represents about 16% of the disc ( Figure S1 ) . These lines were also used in previous studies [14] , [15] . We paid special attention to the choice of the pro-apoptotic vectors . It was important to develop a procedure of cell killing that is very effective and also that allows a precise definition of the chosen domain , so that the damage is restricted to the target area . We tested six different UAS vectors , directing the expression of head involution defective ( hid ) , reaper ( rpr ) , Drosophila inhibitor of apoptosis1-RNAi ( diap1-IR ) , eiger ( egr ) , Drosophila homolog of p53 ( dp53 ) and the activated form of hemipterous ( hepCA ) . We find that the damage inflicted by these pro-apoptotic vectors can be quite different , depending on the vector ( details of the comparison are provided in Figures S2 and S7 ) . A particularly important aspect is the capacity of the disc to regenerate fully after the treatment with the pro-apoptotic vector . In the standard conditions used in our experiments , 40–48 hrs of ablation , the UAS-hid construct allows full regeneration of the disc , whereas after treatment with the other vectors the regeneration is incomplete ( Figure S2 ) . Consequently , we have used the UAS-hid vector in our experiments . The fact that different vectors may cause distinct kinds of damage raises the possibility of different regenerative responses ( see below ) . A summary of the comparison of the various vectors is presented in Table S1 To induce massive but conditional cell death in the Rn domain we constructed larvae of genotype rn-Gal4 tub-Gal80TS UAS-hid UAS-Flp act>stop>lacZ ( see Exp . Proc . for details ) . In this genotype there is no significant apoptosis in the wing disc at 17°C , ( permissive temperature for Gal80TS ) but massive apoptotic levels at 29°C ( restrictive temperature ) in the Rn domain . In addition , the UAS-Flp transgene will be activated to produce large amounts of Flipase that will recombine the act<stop<lacZ cassette in the cells of the domain . As a result all or nearly all the cells of the target domain are indelibly labelled by lacZ activity ( Figure S3 ) . This system is very similar to that utilised by Smith-Bolton et al . , 2009 . To induce cell death in the SalEPv domain we used a similar genotype , just replacing the rn-Gal4 line for salEPv-Gal4 . We first tested the efficacy of the method by allowing continuous activity of Hid in the Rn and Sal domains , achieved by rearing the rn-Gal4 tub-Gal80TS UAS-hid UAS-Flp act>stop>lacZ and salEPv-Gal4 tub-Gal80TS UAS-hid UAS-Flp act>stop>lacZ animals all the time at 29°C . This results in adult flies in which the regions corresponding to either the Rn or the SalEPv domain are lacking ( Figure S1 ) . In the conditional ablation experiments temperature shifts ( 17 to 29°C and vice versa ) were used to control the stage and the timing of the cell killing process . The standard protocol was to allow normal larval growth ( 17°C ) until day 7 after egg-laying , which corresponds to interface between the second and the third larval period . Then the larvae were shifted to 29°C for various lengths of time before been shifted back to 17°C . The first experiments were aimed to find out the maximal ablation time that was compatible with full reconstruction of the targeted domain . The results are illustrated in Figure S1 . Treatments of 48 hrs at 29°C still allow nearly full recovery of the size and pattern of the wing . After 72 hrs of Hid activity there is some recovery but the wings are smaller than normal and show pattern defects . Consequently , in the majority of the experiments the ablation time was of 40–48 hrs . To follow the reconstruction process , the discs were fixed at different times after the initiation and the end of the treatment . As indicated above , the experimental system allows the identification of the cells of the ablated domain . We checked the efficacy of the marking system in control experiments in which the UAS-hid construct was substituted by UAS-GFP . In this case the temperature shift does not produce cell killing , but the activity of the UAS-Flp construct will induce flip out of the act<stop<lacZ cassette thus marking the cells of the Rn domain . The result of this experiment is that all the cells of the domain become marked with lacZ activity ( Figure S3 ) . This is a significant result , for it indicates that in the experimental series we can mark all the original cells of the ablated domain and therefore we can trace the lineage of the reconstructed tissue . One important parameter to estimate was the percentage of cells of the domain killed in the experiments . In our experiments we find a co-extensive expression of caspase-positive and lacZ-expressing cells ( Figure S3 ) , suggesting that we are eliminating the great majority of the cells of the domain . Nevertheless , this estimate is problematic because , as we discuss below , there is a continuous immigration of cells into the domain during the ablating procedure . A careful counting of cells marked for caspase and Topro after 24 hrs of Hid activity indicates a 55% of cell killing in the Rn domain . It suggests that after 48 hrs of ablating time most of the original cells of the domain have been killed . We have followed the evolution of the Rn domain during the ablation period and also after the end of it . Unexpectedly , it turned out that the most significant events take place contemporaneously with the massive cell killing . In Figure 1F-1J we illustrate the changes that occur during a 48 hrs Hid treatment in comparison with control discs in which hid is not activated ( Figure 1A–1E ) . 16 hrs after the initiation there already is noticeable apoptosis and also many of the cells in the domain express lacZ . The cells in apoptosis are extruding towards the basal surface of the disc . At 24 hrs lacZ-expressing cells occupy most or all the domain but only a fraction of them are in apoptosis . Since both Flipase and Hid activities depend on Gal4 , this suggests that flip-induced recombination takes shorter than Hid-induced apoptosis . Both in control ( Figure 1B ) and in experimental discs ( Figure 1G ) , after 24 hrs of Flp activity , the great majority of the cells in the Rn domain are labelled with lacZ . After 32 hrs of Hid treatment many of the lacZ-expressing cells are disappearing and their remains accumulate in the basal surface . However , the apical section of the disc appears normal and contains numerous cells that do not express lacZ . This indicates that cells from outside the targeted domain ( we refer to them as immigrant cells ) are being incorporated . The process continues by 40 hrs , and finally by 48 hrs the Rn domain appears to be constituted by a majority of cells not labelled with lacZ . Its size is approximately 75% of the normal Rn domain . We have assayed the genetic identity of the repaired domain at the end of the ablation period by examining the expression of rn and nubbin ( nub ) , genes that can be considered markers of the wing pouch identity [21] , [22] . The result ( Figure 1K–1N′ ) is that these genes are expressed as in a normal disc , indicating that the immigrant cells , which have repopulated the domain , have acquired the appropriate identity . The cellular debris , still containing caspase activity , accumulates basally . There are three remarkable aspects of this process , i ) during the ablation process there always is continuity in the epithelium , ii ) at the end of the ablation process the Rn domain is almost completely reconstituted ( Figure 1J ) . The expression of rn and nub also indicates a full genetic reconstruction ( Figure 1L–1N′ ) , iii ) The Rn domain has been largely reconstructed by immigrant cells . The results obtained in the experiment ablating the SalEPv domain reinforce the conclusions of the rn>hid experiment . At the end of 40 hrs of Hid activity the SalEPv domain appears largely or completely reconstructed; the apical surface is populated by a majority of immigrant cells , whereas the apoptotic debris accumulate in the basal surface ( Figure S4 ) We have estimated the contribution of immigrant cells to the repaired domain in discs fixed 48 hrs after the end of Hid treatment . At that time the Rn domain is fully normal . The lacZ-marked cells cover 25–30% of the repaired domain . Thus , the average contribution of immigrant cells is about 70% ( Figure 2B , 2C ) . However , there are cases in which the non-lacZ territory is about 90% ( Figure 2E ) , indicating that the domain can be almost entirely rebuilt by immigrant cells . The 70% value is probably an underestimate because some of the LacZ expressing cells may be late immigrants into the domain that performed the flip out , but in which the apoptotic program was not executed . To explore further the possibility of the immigration of cells into the Rn domain we have used the QF/QUAS system [23] as an independent method to mark cells . The Q line GH146 is expressed predominantly in part of the hinge region of the wing disc ( Figure 2 ) , although there is some overlap with the rotund domain . The comparison of the contribution of the GH146 cells to the normal Rn domain or after massive cell killing ( Figure 2F , 2G , 2H ) clearly indicates that many cells originally in the hinge region contribute to the repaired tissue . Moreover , it has to be considered that only a fraction of hinge cells are marked with the GH146 line , thus the contribution of unmarked outside cells cannot be scored . We also observe that in the reconstructed domain many cells of the GH146 lineage appear to be scattered ( Figure 2G ) . The amount of scattering in greater than in control discs ( Figure 2I ) , suggesting the existence of cell movements during the reconstruction process . One of the principal features of the regeneration process is the induction of a proliferative response necessary to generate new cells [24] , [25] . We have used several methods to measure this response . The comparison of the mitotic index , calculated by the number of cells containing the phosphorylated form of Histone 3 ( PH3 ) in rn>hid and in rn>GFP ( control ) discs at the end of a 24 hrs period at 29°C , indicates an increase of cells in mitosis in the rn>hid discs ( Figure 3A–3B′ ) . In addition , the examination of the distribution of mitotic cells reveals some significant features of the regeneration process . In the first place the density of cells in mitosis , though higher than in control , is not localised near the damaged tissue , but it is uniform over the disc , including the Rn domain ( Figure 3B ) . We have calculated the mitotic index in three different regions of the control and the regenerating disc ( Figure 3D–3F ) : region 1 covers the Rn domain , region 2 the area surrounding the domain and region 3 the rest of the disc . We find that already after 24 hrs of ablation the mitotic index is significantly higher ( 50% ) in the disc under ablation than in the control , and that the increase is observed in the three regions ( Figure 3D–3F ) . Altogether these data indicate that the localised cell killing elicits a systemic proliferative response that affects all regions of the disc . This response begins very soon after the damage has started . Interestingly , after 48 hrs the increase in the mitotic index appears restricted to region 1 , corresponding to the damaged territory . This suggests that the initial systemic response is followed by a more localised one in the affected zone . A second interesting feature is that the mitotic cells in the Rn domain under reconstruction localise in the apical section of the disc ( Figure 3B ) and are absent in the basal section ( Figure 3B′ ) , where the apoptotic debris accumulates . The presence of actively dividing cells in the domain at a time in which the pro-apoptotic Hid protein is still active provides additional support to the idea that the reconstruction process is concomitant with cell killing . It is very likely that many of the cells in division in the Rn domain are immigrant cells from outside . The finding that there is an overall response to the localised damage is also supported by the results obtained examining BrdU incorporation , another indicator of actively dividing cells . We have compared BrdU levels in ablated and control discs ( Figure 3G–3J′ ) after 40 hrs at 29°C . In both genotypes there is a homogenous incorporation in all regions of the discs , although the intensity of BrdU incorporation is higher in the ablated discs . The levels of BrdU incorporation after 24 and 48 hrs of Hid activity are similar inside and outside the reconstructed domain . We note the discrepancy between our results indicating a systemic response and those of Smith-Bolton et al 2009 and Bergantiños et al 2010 , which suggest a local response . As discussed below , it suggests the possibility of different regenerative responses , perhaps depending on the kind of damage . We have performed a clonal analysis of discs in which the Rn domain is under ablation . The genotype of the larvae is described in the Methods section . The clones were induced at the beginning of a 48 hrs ablation period and the discs fixed at the end . That is , the clones had been growing during the ablation time . Their size and shape was compared with control clones generated in larvae in which the temperature shift activated UAS-GFP instead of UAS-hid . We also considered the position of the clones in relation to the ablated domain . The results of this experiment are shown in Figure 4 . The average size of the clones in the regenerating discs is about twice that of the controls . This is expected considering that the ablated Rn domain is about 40% of the disc total and the surviving cells have to perform additional divisions to approximately double the amount of tissue . In addition we find that the size of the clones is approximately the same in all the regions of the disc , including that of the clones inside the ablated domain . This can be visualised by plotting clone size with respect to the distance to the border of the Rn domain ( Figure 4C ) . The average size of clones located closer that 50 µm to the Rn border is comparable to that of those located further than 50 µm . Thus the growth rate of the clones is similar in all regions of the disc . These observations strongly support the notion that there is a systemic response to the localised damage . We have investigated the activity of the Dpp , Wg and JNK pathways in response to the ablation of the Rn and SalEPv domains . These pathways are known to be involved in growth control and response to damage in imaginal discs . The Dpp pathway has a principal role as growth inducer in the wing disc; loss of Dpp activity inhibits cell division and reduces wing size , whereas high activity levels results in increased proliferation and in discs and wings of very large size [26] , [27] , [28] . Therefore we expected that the Dpp pathway would be involved in the growth response after injury . We have first examined dpp expression during and after the ablation period . It was visualised by in situ hybridization with a specific probe and also using a dpp-lacZ insert . As shown in Figure 5 and Figure S5 , after 20 or 40 hrs of ablation dpp expression remains normal; no significant alteration was found ( Figure 5B″ , 5C″ ) , even though after 40 hrs of ablation there is a great deal of apoptosis in the Rn domain ( Figure 5C′ ) . We then examined the expression levels of the Dpp mediator pMad [4] and of two well known Dpp targets , spalt ( sal ) and optomoter-blind ( omb ) . The expression of sal and omb is essentially normal ( Figure 5F′ , 5H′ ) but , surprisingly , the activity of pMad appears uniformly elevated in the Rn domain ( Figure 5D′ ) . Since sal and omb activities are mediated by pMad [4] , it is hard to reconcile the normal expression of sal and omb seen in the damaged domain with the extended and elevated pMad expression . However , we find that the higher levels of pMad localise to the basal section of the disc , where the apoptotic cells accumulate ( Figure 5D′ ) . Also , double staining for TUNEL and pMad ( Figure 5D ) indicates that the nuclear fragments labelled with TUNEL contain the pMad protein . This suggests that the high levels of pMad observed in the experiments reflect a special feature of cells in apoptosis , but devoid of functional significance . We have explored the association of apoptosis and pMad accumulation in heavily irradiated discs , in which we observe high pMad levels associated with cells in apoptosis , but the expression of dpp remains normal ( Figure S6 ) . Thus , considering all the evidence , our results indicate that the ablation of the Rn domain does not alter the activity of the Dpp pathway . It is worth pointing out that the normal Dpp activity is indeed required for the domain reconstruction; in experiments in which the UAS-hid transgene is replaced by UAS-dpp-RNAi , the elimination of dpp activity in the Rn domain for 48 hrs produces a phenotype similar to that of larvae that suffered continuous ablation ( not shown ) . Regarding the Wg pathway , it has been reported [14] that massive damage caused by over-expressing the TNF ligand Eiger [29] , [30] causes strong wg up-regulation in the damaged tissue . However , in our experiments inducing apoptosis with Hid we do not find a significant alteration of wg expression . As shown in Figure 6 , after 40 hrs of Hid activity wg expression appears normal , both inside and outside the ablated domain . Two target genes of the Wg pathway , Distalless ( Dll ) and vestigial ( vg ) [8] are also expressed normally ( Figure 6D–6G ) . Experiments of ablation of the SalEPv domain yield similar results ( Figure S4 ) . Since there was the possibility of a function of the Wg pathway not easily detectable with the current labelling methods , we assayed the ability to regenerate of the Rn domain in conditions in which wg activity is inhibited , using a UAS-wg-RNAi line . We first showed that the UAS-wg-RNAi line effectively suppresses wg function ( Figure 6H ) . The result , shown in Figure 6H , is that the Rn domain regenerates in the absence of wg function . Thus , our results indicate that there is no significant response of the Wg pathway to the damage caused by Hid-dependent apoptosis . This result is in contrast with that reported by Smith-Bolton et al . , 2009 . In their experiments they used similar Gal4 drivers but different pro-apoptotic vectors , the TNF ligand Eiger ( UAS-egr ) and the pro-apoptotic gene rpr . We repeated their experiment in order to ascertain the experimental difference . We find that overexpression of Eiger and of Rpr indeed induce wg activity , although not too extensive in the case of Rpr ( Figure S7 ) . After 40 hrs of Eiger or Rpr activity the discs exhibit a distorted morphology ( Figures S2 and S7 ) , and do not recover a normal morphology even 48 hrs after the end of Eiger activity . This contrasts with our experiments using Hid , in which we observe full recovery after massive apoptosis . We believe that , especially Eiger overexpression , may cause other effects in addition to apoptosis and these effects interfere with the normal repair process . Finally , we analysed the response of the JNK pathway to Hid-induced apoptosis . To monitor JNK activity we have used a LacZ insert in the puckered ( puc ) gene , a target of the pathway [31] . Normally the JNK pathway is not active in the Rn domain ( Figure 7 ) , but in the domain under reconstruction we observe many cells expressing the puc-LacZ transgene . These cells can be detected already 16 hrs after the commencement of the ablation and accumulate preferentially in the apical section of the disc ( Figure 7 ) , where the healthy cells are . We reasoned that these cells could be involved in the repair process . Besides , Bergantiños et al . , 2010 have provided evidence that the JNK pathway is required for the regeneration of the patch domain . We tested the functional role of this adventitious JNK activity by inducing Hid activity in the Rn domain and at the same time inhibiting JNK by overexpressing the negative regulator puc [31] . By itself the overexpression of puc does not have any effect on the development of the disc ( not shown ) . The degree of rescue was studied in adult wings , comparing the wings in which JNK can be up regulated with those in which the elevated levels of puc suppress JNK activity . The result is that the rescue of the Rn domain is significantly diminished ( Figure 7F ) , supporting a role of the JNK pathway in the repair process . We observed that some of the puc-expressing cells were not labelled with GFP , suggesting they come from outside the Rn domain ( Figure 7C–7E′ ) . Therefore we checked the effect of JNK in the immigration of cells towards the domain under repair . As mentioned above , the QF line GH146 is expressed preferentially in cells outside the Rn domain ( Figure 2F ) and the contribution of cells derived from the GH146 domain is increased significantly after massive damage in the Rn domain ( Figure 2G , 2H ) . We have inhibited JNK activity in the GH146 cells by overexpressing puc . The result is that the contribution of these cells to the repopulation of the damaged Rn domain is now not increased after damage ( Figure 7G ) . This experiment clearly suggests a requirement for JNK pathway activity in the immigration process .
This reveals the existence of a powerful mechanism of tissue homeostasis that can cope with major lesions during development . Very shortly after the beginning of massive apoptosis in the Rn domain it is possible to observe immigrant cells ( Figure 2 ) –identified by the marking method and in the lineage experiment using the QF system- that are occupying the place of the dying ones , which have been extruded to the basal section of the disc . The incorporation of these cells ensures the integrity of the epithelium and also that the domain develops normally , even during the ablation process . This is clearly manifest by examining the morphology of the disc and the pattern of expression of key developmental genes such dpp , wg , vg , Dll , nub , rn , omb , sal in the domain after 48 hrs of Hid activity and without allowing any time for recovery ( Figure 1 , Figure 5 , Figure 6 ) . All these genes are expressed normally , indicating that the immigrant cells have acquired the appropriate identity corresponding to the normal Rn domain . This is a striking result , as it demonstrates that the reconstruction process has occurred concomitantly with the massive cell killing . Previous studies [14] , [15] have failed to observe this phenomenon , likely because those authors did not study the events during the ablation time; their analysis of the regeneration process was focused on the post-ablation period . This result was also unexpected in view of published reports [14] , [15] indicating that the ablation produces a local proliferative response around the damaged region . Our results , on the contrary , suggest that the response to the ablation of the Rn domain is largely systemic , as the increase in cell proliferation is generalised in the disc . The experimental support is as follows: 1 ) proliferation markers such as BrdU incorporation and PH3 staining are homogenously distributed over the regenerating disc ( Figure 3 ) . Also , accurate accounting of the number of mitotic ( PH3 labelled ) cells in three distinct regions reveals a significant increase of mitotic cells during ablation that affects the three regions ( Figure 3 ) , 2 ) Clonal analysis of the disc under ablation shows that while the average clone size is greater than in controls , the size increase is not localised , but general in the disc ( Figure 4 ) . However , the evolution of the mitotic index during and after the ablation ( Figure 3 ) suggests that the initial systemic response in cell proliferation is followed by a local increase in the damaged zone . We believe that these results can be interpreted in the following manner . There is a mechanism of overall control that continuously monitors compartment size during development and that blocks growth once the compartment has reached the final stereotyped size ( see [34] ) . A massive damage like the ablation of the entire Rn domain is interpreted as a substantial diminution in size . This triggers a growth response in both the anterior and posterior compartments ( the Rn domain extends to both ) , which is aimed to achieve normal size . Since the mechanism in question measures the overall size , the response is also general , increasing cell proliferation in all regions of the disc , including the domain under restoration . We are aware that data from previous reports 14 , 15 suggest that the proliferative response is localised close to the damaged region . One possible reason for the difference is that the strong systemic response occurs during the first 24 hrs of ablation time ( Figure 3 ) , and is later followed by a more localised effect . Smith-Bolton et al 2009 restricted their analysis to post-ablation time , whereas Bergantiños et al 2010 , only analysed the proliferative response in part of the disc . We do not see major alterations in the Dpp and Wg signalling pathways . The expression of dpp and wg remain essentially unaffected during and after ablation ( Figure 5 and Figure 6 ) . Moreover , their target genes sal , omb , vg and Dll are also expressed normally after 40 hrs of ablation . Intriguingly , we find elevated levels of pMad in the ablated tissue , but we believe that these high pMad levels are of little functional significance . One reason is that up regulation of pMad would be expected to induce a raise in the expression sal and omb , which is not observed ( Figure 5 ) . Besides , this accumulation of pMad appears associated only with cells in apoptosis , suggesting that it may be a peculiar property of apoptotic cells devoid of functional consequence . We wish to emphasise however that the reconstruction of the ablated domain does require the regular activity of the Dpp pathway , because in its absence there is no recovery of the ablated domain . This simply reflects the normal requirement of Dpp activity for the development of the wing structures . The results obtained on wg deserve special attention . This gene has been implicated in the regeneration of the imaginal discs after transplantation [10] and there is evidence that Wnt genes are involved in regeneration in Hydra and in other organisms [12] , [35] . Moreover , in experiments similar to ours , Smith-Bolton et al 2009 have shown that wg is up regulated , an observation that we have confirmed ( Figure S7 ) . Yet in the rn>hid regeneration experiments its expression is not altered and its function is not required ( Figure 6 ) . These observations raise the possibility of different regenerative responses depending on the kind and perhaps the developmental status of the damaged tissue . Contrasting with the lack of response of the Dpp and Wg pathways , the activity of the JNK pathway is activated after damage and its function is necessary for the repair process ( Figure 7 ) , confirming prior work by Bergantiños et al ( 2010 ) . In addition , our experiments inhibiting JNK in the cells close to the Rn domain ( Figure 7G ) strongly suggest that its activity is necessary to repopulate the domain under ablation , possibly facilitating the migration of cells . It is known that JNK activity plays a role in developmental processes like the embryonic dorsal closure or the sealing of the left and right sides of the discs [36] , [37] , that require cell movements , thus it may perform a similar function during the reconstruction of the damaged Rn or sal domains .
The Drosophila stocks used were rn-Gal4 [19]; salEPv-Gal4 ( gift of JF de Celis , CBMSO , Madrid , Spain ) ; tub-Gal80ts [38]; puc-lacZ line ( pucE69 ) and UAS-puc14C line [31]; UAS-GFP ( BDSC ) , UAS-shmi Dpp2 [39]; UAS-wgRNAi ( VDRC 13352 ) ; hs-Flp112 [40];QH146-QF line , QUAS-Flp line and QUAS-GFP line [23]; UAS-Flp ( BDSC ) ; act>stop>lacZ ( gift of G Struhl ) ; ubi>stop>GFP [41]; UAS-Diap1RNAi [42]; UAS-eiger [30]; UAS-dp53 . EX , UAS-hepCA , UAS-hid , UAS-rpr and P{dpp-lacZ . B} lines are described in Flybase . In the experiments using the Gal4/UAS system the general rule was to use only two UAS transgenes . This was to avoid the possibility of titrating the amount of Gal4 protein available for the UAS vectors . To induce neutral clones in the domain under ablation we heat shocked ( 12 min 37°C ) larvae of genotype hs-Flp; UAS-hid tub-Gal80TS/+; rn-Gal4/ubiP63E>stop>GFP at the beginning of the 48 hrs ablation period . The discs were fixed immediately after the end of the Hid treatment . As controls we induced clones in hs-Flp; UAS-hid tub-Gal80TS/+; ubiP63E>stop>GFP/+ larvae in which the temperature shift did not activate Hid . The Q system is a binary expression system that is independent of the Gal4/UAS system [23] . In our experiments , we used the Gal4/UAS system to ablate the Rn domain , whereas with the GH146-QF line we traced the lineage of a group of cells preferentially located in the hinge area . The genotypes used were UAS-hid tub-Gal80TS/QUAS-Flp ubi>stop>GFP ; +/GH146-QF ( control ) and UAS-hid tub-Gal80TS/QUAS-Flp ubi>stop>GFP ; rn-Gal4/GH146-QF ( hid induction in Rn domain In the experiment to overexpress puckered ( puc ) the genotypes were UAS-hid tub-Gal80TS/QUAS-puc QUAS-Flp ubi>stop>GFP ; +/GH146-QF ( control ) and UAS-hid tub-Gal80TS/QUAS-puc QUAS-Flp ubi>stop>GFP ; rn-Gal4/GH146-QF ( experimental ) The puckered cDNA was obtained from the UAS-puc2A line [31] by PCR amplification and cloned into the pQUAST vector ( Addgene plasmid 24349 ) as described in Potter et al . , 2010 . To enhance the construct expression and avoid positional effects of the integration , the BamH1 fragment from the resulting vector was cloned into pCa4B2G [43] at the BamH1 site . This vector flanks the insert with gypsy insulators and allows PhiC31-mediated integration [44] . The construction was integrated at the 51D locus . Immunostaining and double antibody staining/in situ hybridization was performed as described previously [45] . Images were captured with a Leica ( Solms , Germany ) DB5500 B confocal microscope or a LSM510 Vertical ( Zeiss , Thornwood , NY , USA ) confocal microscope . The following primary antibodies were used: rabbit anti-Caspase 3 ( Roche ) 1∶50; mouse anti-ß-Galactosidase ( DSHB 40-1a ) 1∶50; mouse anti-BrdU ( Roche ) 1∶10; rabbit anti-Vestigial [46] 1∶500; mouse anti-Distalless [47] 1∶400; rabbit anti-Omb ( gift of G . O . Pflugfelder ) 1∶100; rat anti-Spalt ( gift of JF de Celis , CBMSO , Madrid , Spain ) 1∶50; rabbit anti-pMAD [48] 1∶100; mouse anti-Wingless ( DSHB 4D4 ) 1∶50; anti-Digoxigenin-AP ( Roche , Basel , Switzerland ) 1∶4000; anti-Phospho-Histone H3 ( Millipore ) ; rabbit anti-Rotund 1∶250 and rabbit anti-Nubbin 1∶250 ( gifts of F . Diaz-Benjumea , CBMSO , Madrid , Spain ) Fluorescently labelled secondary antibodies ( Molecular Probes Alexa ) were used in a 1∶200 dilution . TO-PRO3 ( Invitrogen ) was used in a 1∶600 dilution to label nuclei; Phalloidin-TRITC ( Sigma ) and Phalloidin-Cy5 were used in a 1∶200 dilution to label the F-actin network . The RNA probe for the dpp transcript was obtained using the BDGP cDNA clone RE20611 as a template . Activity of the lacZ gene was visualised by staining for ß-gal activity , as described in [49] . For Bromo-2′-deoxyuridine ( BrdU ) incorporation , wing imaginal discs were cultured in PBS 1x medium supplemented with 1 mg/mL BrdU ( Roche ) for 30 minutes at room temperature . Discs were subsequently washed in PBS 1x , fixed in 4% Paraformaldehyde , 0 . 1% Triton , 0 . 1% DOC for 1 h at room temperature , treated with RQ1 DNase ( Promega ) for 2 hours and fixed again in 4% Paraformaldehyde , 0 . 1% Triton , 0 . 1% DOC for 30 minutes . Discs were then immunostained as described above . Apoptotic cells were detected based on labeling of DNA strand breaks using TUNEL technology ( In situ Cell Death Detection kit , TMR red , Roche ) . Adult wings were mounted in Euparal Mounting medium after having dissected the flies in a mixture of alcohol/glycerine . Images were captured with a Leica microscope
|
The study of how organs or tissues regenerate after damage is a classic topic in Developmental Biology . We are studying this process in the developing wing imaginal disc of Drosophila melanogaster , using genetic methods to inflict massive damage in the region destined to form the wing blade . We find that the lesion provokes a very strong and rapid reaction in the remaining disc aimed to reconstruct the lost tissue , both in size and in shape . The response includes an increase of cell proliferation to compensate for the loss of cells and the immigration of cells from neighbouring areas to replace the dead ones . The immigrant cells change their original identity and acquire that of the cells they are replacing . We propose that these experiments reveal the existence of a powerful homeostatic mechanism that is able to cure massive injuries that may appear during development .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"developmental",
"biology",
"growth",
"control",
"organism",
"development",
"regeneration",
"biology",
"morphogenesis",
"cell",
"fate",
"determination",
"limb",
"development"
] |
2013
|
Tissue Homeostasis in the Wing Disc of Drosophila melanogaster: Immediate Response to Massive Damage during Development
|
Nonsense-mediated mRNA decay ( NMD ) prevents the accumulation of transcripts bearing premature termination codons . Here we show that Saccharomyces cerevisiae NMD mutants accumulate 5′–extended RNAs ( CD-CUTs ) of many subtelomeric genes . Using the subtelomeric ZRT1 and FIT3 genes activated in response to zinc and iron deficiency , respectively , we show that transcription of these CD-CUTs mediates repression at the bona fide promoters , by preventing premature binding of RNA polymerase II in conditions of metal repletion . Expression of the main ZRT1 CD-CUT is controlled by the histone deacetylase Rpd3p , showing that histone deacetylases can regulate expression of genes through modulation of the level of CD-CUTs . Analysis of binding of the transcriptional activator Zap1p and insertion of transcriptional terminators upstream from the Zap1p binding sites show that CD-CUT transcription or accumulation also interferes with binding of the transcriptional activator Zap1p . Consistent with this model , overexpressing Zap1p or using a constitutively active version of the Aft1p transcriptional activator rescues the induction defect of ZRT1 and FIT3 in NMD mutants . These results show that cryptic upstream sense transcription resulting in unstable transcripts degraded by NMD controls repression of a large number of genes located in subtelomeric regions , and in particular of many metal homeostasis genes .
A large fraction of eukaryotic genomes is transcribed , even in the non-coding regions ( reviewed in [1] ) . One of the major questions that arise from these observations is to understand whether the RNAs expressed from these regions serve any functional purpose or whether they correspond to genomic noise that is ultimately routed for degradation . Transcription of non-coding ( nc ) RNAs nearby protein-coding genes has emerged as a means of transcriptional control ( reviewed in [1] , [2] ) . In the yeast S . cerevisiae , the first and best-documented example is the SRG1 ncRNA , which regulates transcription of the SER3 gene involved in serine metabolism through transcriptional interference [3] . While the SRG1 ncRNA is transcribed in the sense direction upstream of its target gene , most ncRNAs found to regulate S . cerevisiae transcription are antisense transcripts , such as the ones described for the PHO84 , Ty-1 and GAL10 loci [4]–[7] , which control chromatin modification marks at these genes . However , upstream sense transcription resulting in the ICR1 ncRNA has been found to regulate the FLO11 gene [8] . Cryptic transcripts can also be generated through the transcription of elements that control the expression of bona fide protein coding genes . There is ample evidence that promoter regions are associated with bidirectional transcription in S . cerevisiae [2] , [9] . In addition , it has been shown that enhancers are transcribed by RNA polymerase II [10] . In S . cerevisiae , many transcripts associated with intergenic or promoter transcription are unstable under normal conditions . They are hardly detectable in wild-type strains because of their rapid degradation by nuclear RNA turnover . Indeed , many transcripts associated with cryptic transcription are detectable only when the activity of the nuclear exosome , or that of the TRAMP-complex , which stimulates exosome activity , is inhibited [2] , [11] , [12] . Therefore , these transcripts have been labeled “CUTs” for Cryptic Unstable Transcripts . Most of the degradative activities targeting CUTs seem to be concentrated in the nucleus . However , the cytoplasmic exonuclease Xrn1p can degrade the antisense RNAs that regulate the Ty-1 gene [5] . In addition , several CUTs can be degraded by the cytoplasmic degradation machinery [12]–[13] , showing that the degradation of RNAs arising from transcription in non-coding regions can result from both nuclear and cytoplasmic RNA degradation pathways . Nonsense mediated decay is an RNA surveillance mechanism that recognizes transcripts containing premature translation termination codons ( PTCs; [14] , [15] ) . This degradation system is used to prevent accumulation of aberrant mRNAs that would encode potentially toxic proteins [16] , [17] , but is also used for gene expression control . For example , NMD degrades alternatively or inefficiently spliced mRNAs that contain PTCs [18]–[21] , and transcripts containing long 3′-UTRs [22] . In S . cerevisiae , NMD also controls Mg++ uptake by degrading the transcript encoding the major Mg++ transporter [23] . Previous studies using microarray analysis of S . cerevisiae NMD mutants have shown that NMD influences directly or indirectly the expression of many genes , including some that do not exhibit PTCs [24]–[26] . The identification of RNAs that associate with the NMD factor Upf1p by RNA pull-down has proved to be an efficient way to discriminate direct and indirect NMD targets [27] . However , many NMD targets do not accumulate to high levels in the presence of a functional NMD system , which renders this approach difficult for low expression transcripts . In this study , using tiling microarrays , we show that NMD degrades 5′-extended transcripts generated by cryptic transcription upstream from bona fide promoters , and that upstream transcription is responsible for the repression of metals homeostasis genes in conditions of metal repletion . These results show that NMD controls the expression of a large number of genes by modulating the expression of 5′-extended RNAs that control transcription .
We previously used tiling microarrays to show that NMD controls the degradation of inefficiently spliced S . cerevisiae pre-mRNAs [21] . Further analysis of these microarrays in intergenic areas revealed that many subtelomeric regions accumulated RNA signal outside and upstream from the open-reading frames ( ORF ) in the upf1Δ and xrn1Δ mutants . Examples shown in Figure 1 include the subtelomeric ZRT1 gene encoding the primary zinc transporter [28] , the FIT3 gene involved in siderophore-iron transport facilitation [29] , and the FLO5 gene encoding a cell wall protein ( Figure 1 ) . Many genes located in subtelomeric regions are involved in the response to adverse growth conditions [30] . In the case of the FIT3 and FLO5 genes , the array profiles suggested that these species originate 5′-to the normal transcriptional start site and extend in the open-reading frame ( which was confirmed by northern analysis , see below ) . In the case of ADH4 and ZRT1 , increase of signal in the upstream regions correlated with a decrease of signal in the downstream regions ( Figure 1 ) . Northern blot analysis using antisense riboprobes confirmed that these species correspond to 5′-extended transcripts originating upstream from the bona fide promoters and partially or completely overlapping the open-reading frame ( Figure 2 and see below ) . Based on the array profiles , some of these species were very long , containing 5′-extensions up to several kb for ZRT1 ( Figure 1 , Figure S2 ) . In the case of FIT3 , a previous study had mapped in detail the 5′-extended species that extends into the ORF [31] . Thus , these species are different from the short antisense CUTs transcribed divergently from the promoters [2] , [9] . Accumulation of 5′-extended transcripts in NMD mutants has been reported for a few transcripts , but was interpreted as the result of transcriptional noise [27] . We hypothesized that some of these 5′-extended unstable transcripts might be used for regulatory purposes , as shown in other systems [3] , [32] , [33] . Because the accumulation of 5′-extended transcripts relied on the inactivation of cytoplasmic degradation pathways , we named these 5′-extended species CD-CUTs ( Cytoplasmically Degraded Cryptic Unstable Transcripts ) . In the case of subtelomeric genes involved in zinc metabolism ( ZRT1 , ADH4 , VEL1 , YOR387C , Figure 1; ZPS1 , Figure S1 ) , the accumulation of CD-CUTs was correlated with a decreased signal in the ORF regions , suggesting that they might play a negative role in the expression of the bona fide mRNAs . Interestingly , zinc regulon genes not located in subtelomeric areas , such as the one encoding the low affinity zinc transporter ZRT2 [34] did not exhibit CD-CUTs ( data not shown ) . We found that the growth conditions used for the microarrays ( synthetic complete minimal medium ) corresponded to mild zinc deficiency conditions , in which some of the zinc regulon genes were partially induced ( data not shown ) . Therefore the decrease of signal observed in the ORF regions for ZRT1 and ADH4 ( Figure 1 ) might be due to a defect in the partial induction of these genes in minimal medium . We further analyzed the expression of zinc responsive mRNAs and their corresponding CD-CUTs in wild-type and upf1Δ strains grown in the presence or absence of zinc , using northern blots and antisense riboprobes covering the open-reading frames ( ORFs ) or the upstream ( UP ) regions ( Figure 2A ) . This analysis confirmed the accumulation of CD-CUTs of ZRT1 , ADH4 and ZPS1 in the upf1Δ mutant strain . For ZPS1 , CD-CUTs were readily detectable in the wild-type as well as in the upf1Δ strain , explaining why only a modest increase in upstream signal was observed for this gene on the arrays , which compare the transcripts levels of the upf1Δ mutant to the wild-type ( Figure S1 ) . Some CD-CUTs were detected only with the upstream riboprobes , demonstrating that they are independent from the ORFs ( Figure 2A ) . However , most CD-CUTs were also detected using the ORF riboprobes ( Figure 2A and see below ) , indicating that they extend through the ORFs . A detailed characterization of the ZRT1 and ADH4 CD-CUTs using various upstream probes and ORF probes is shown in Figure S2 and described below . ORF riboprobes detected the induction of the normal ZPS1 , ADH4 , and ZRT1 mRNAs in wild-type cells grown in a medium lacking zinc ( Figure 2A ) . However , these mRNAs were less abundant in the upf1Δ mutant ( Figure 2A ) . This observation suggested that the accumulation of CD-CUTs due to NMD inactivation was deleterious to the expression of the bona fide mRNAs . We further characterized the expression of these genes in the upf1Δ mutant by monitoring their kinetics of induction ( Figure 2B ) . This experiment showed that the upf1Δ mutant exhibited a delay in the induction of the subtelomeric ZPS1 , ZRT1 , ADH4 ( Figure 2B ) and VEL1/YOR387C genes ( Figure S3A ) . The induction defect was also observed in the upf2Δ and upf3Δ strains ( Figure S3A ) , indicating that it is a general feature of NMD mutants . In contrast , the induction of other zinc regulon genes such as ZRT2 , GPG1 ( YGL121C ) and PHM7 ( YOL084W ) , which are not localized in subtelomeric regions and do not exhibit CD-CUTs ( data not shown ) was unaffected by Upf1p absence ( Figure 2B ) . A previous study found that NMD can influence gene expression by controlling the level of transcriptional activators [35] . We found that the mRNA levels of the Zap1p activator ( which activates zinc regulon genes ) were similar in wild-type and upf1Δ mutants ( Figure S3B ) , suggesting that the delay of induction in the absence of Upf1p was not due to reduced Zap1p levels . Overall these results show that the defective induction of subtelomeric zinc-responsive genes in the upf1Δ mutant is not due to a global deficiency in zinc sensing in this mutant , but is specific to genes that exhibit CD-CUTs . Defective induction was also observed for the FIT3 iron depletion-responsive gene in a strain lacking Xrn1p or Upf1p activity ( Figure S4 ) . Overall these observations suggest that the presence or transcription of CD-CUTs represses the expression of many subtelomeric genes , some of which are involved in zinc or iron homeostasis . To further investigate the mechanisms of turnover of CD-CUTs by the different RNA degradation machineries , we analyzed the accumulation of CD-CUTs of ZRT1 and ADH4 in strains lacking Upf1p , the nuclear exosome component Rrp6p , or both . These two genes were chosen because the 5′-extensions found in the CD-CUTs of ZRT1 and ADH4 are very different in size ( extension of 2 kb for ZRT1 vs . a few hundred nucleotides for ADH4 ) . This analysis showed that the major ZRT1 CD-CUT ( CD-CUT1 , Figure 2C ) accumulated dramatically in the upf1Δ strain , and to a much lesser extent in the rrp6Δ mutant . Accumulation of the ZRT1 main CD-CUT-1 was not increased in the upf1Δrrp6Δ double mutant compared to the upf1Δ single mutant ( Figure 2C ) . These observations suggest that the degradation of the longer CD-CUT of ZRT1 relies mostly on NMD , consistent with the long upstream region lacking any extended ORF . In contrast , the short ADH4 CD-CUT accumulated to similar levels in the upf1Δ and rrp6Δ strains ( Figure 2D ) . Strikingly , the accumulation of this CD-CUT was dramatically increased in the upf1Δrrp6Δ double mutant ( Figure 2D , time zero ) , suggesting that this CD-CUT is degraded by the cooperative action of the nuclear exosome and of NMD . Interestingly the induction of the bona fide ADH4 mRNA was completely defective in this double mutant strain , while the single mutants were only partially delayed . The correlation between the strong accumulation of the ADH4 CD-CUT in the upf1Δrrp6Δ double mutant and the severe induction defect of the bona fide ADH4 mRNA provides further evidence for the repression of this subtelomeric gene by its CD-CUT . The CD-CUT of the subtelomeric iron responsive gene FIT3 was previously shown to accumulate in the xrn1Δrat1-1 double mutant [31] , raising the question of which exonuclease was primarily responsible for its degradation . We analyzed the expression of this CD-CUT in strains lacking Xrn1p , Upf1p , Rrp6p , in the xrn1Δrat1-1 mutant strain and in other double mutant strains . This analysis revealed that the long CD-CUT of FIT3 accumulated to the highest level in the strain lacking Xrn1p and to a lesser extent Upf1p ( Figure 2E ) , and its accumulation was not dramatically increased by Rrp6p inactivation , in contrast to what was found for ADH4 . Thus , based on this steady-state analysis , the long CD-CUT of FIT3 is also primarily targeted by cytoplasmic turnover pathways that include Xrn1p and Upf1p . Because CD-CUTs exhibit a lack of extended ORFs , we interpreted their accumulation in the upf1Δ and xrn1Δ mutant strains as the result of a lack of degradation by NMD . Alternatively , we could not rule out that transcription of these upstream regions might be indirectly up-regulated in these mutants . To test this hypothesis , GFP-HIS3 cassettes [36] were inserted upstream from the normal ZRT1 or ADH4 promoters in wild-type and upf1Δ strains , such that the expression of the GFP mRNA was under the control of the upstream regions ( Figure S5A ) . Northern blot analysis of GFP inserted upstream from ZRT1 or ADH4 showed that this reporter transcript was expressed at similar levels in the wild-type and upf1Δ strains ( Figure S5B ) . These results suggested that the accumulation of CD-CUTs in NMD mutants is not due to an increased transcription of these upstream regions , but to the absence of ORF in the 5′- extension of the CD-CUTs . To gain further evidence that the turnover of CD-CUTs is directly dependent on NMD , we replaced the region upstream from the site of transcription initiation of the ZRT1 and FIT3 CD-CUTs with a galactose inducible promoter . This allowed us to measure the rate of decay of these CD-CUTs in the presence or absence of functional NMD . The kinetics of turnover of the ZRT1 CD-CUT ( Figure 2F ) or of the FIT3 CD-CUT ( Figure 2G ) showed that these species are much more unstable in the presence of functional NMD ( t1/2 = 2–3 min . ) than in the absence of Upf1p ( t1/2 = 20–30 min . ) . Because the turnover rate of these species is strongly decreased when NMD is inactivated , we conclude that these CD-CUTs are directly targeted by NMD for their degradation . Previous microarray analysis of a strain inactivated for the histone deacetylase Rpd3p showed that ZRT1 is derepressed in the rpd3Δ strain and that Sir2p played a role antagonistic to Rpd3p in ZRT1 expression [37] . To investigate whether Rpd3p or Sir2p control ZRT1 by modulating the expression of its CD-CUT , we inactivated Upf1p in rpd3Δ or sir2Δ backgrounds and studied the induction of ZRT1 in these strains . ZRT1 was strongly derepressed in the rpd3Δ strain ( Figure 3A , time zero ) , in agreement with previous data [37] . Strikingly , inactivation of Rpd3p in the upf1Δ strain completely rescued the induction defect of this NMD mutant , and resulted in a strong derepression of ZRT1 in normal zinc conditions ( Figure 3A ) . Inactivation of Rpd3p also resulted in the almost complete disappearance of the ZRT1 CD-CUT observed in the upf1Δ strain . These results show that Rpd3p positively controls the expression of CD-CUT of ZRT1 , and suggest that the derepression of ZRT1 in the rpd3Δ mutant [37] is due to the absence of the CD-CUT . In contrast , Sir2p inactivation reduced ZRT1 levels ( Figure 3A ) , in agreement with the previous results [37] . Combining the sir2Δ deletion to the upf1Δ deletion exacerbated the ZRT1 induction delay when compared to the upf1Δ mutant ( Figure 4A ) , but the sir2Δupf1Δ mutant did not exhibit higher levels of CD-CUT than the upf1Δ single mutant ( Figure 4A ) . Therefore , the negative effects of Sir2p inactivation on ZRT1 expression are unlikely to be directly linked to its effect on the ZRT1 CD-CUT . To investigate the specificity of the effect observed with Rpd3p on the ZRT1 CD-CUT levels , we performed the same genetic analysis with the Hos1p , Hda1p , Hda2p and Hda3p deacetylases . The hos1Δ strain showed a slight delay in the induction of ZRT1 , correlated with an increase of CD-CUT levels , but the hos1Δupf1Δ double mutant showed no additive effect when combined with the upf1Δ deletion ( Figure 3B ) . Neither Hda1p ( Figure 3C ) , nor Hda2p or Hda3p ( Figure S6 ) were found to affect ZRT1 induction or repression . These results show that the major effect observed with Rpd3p on the ZRT1 CD-CUT is specific to this deacetylase . We tried to corroborate these results by monitoring the presence of Rpd3p in the region 5′ to ZRT1 but could not obtain reproducible evidence for enrichment by ChIP ( data not shown ) . However the genetic data shown above strongly suggest that Rpd3p mediates the repression of ZRT1 through the modulation of the transcription of the CD-CUT . If CD-CUTs are involved in the repression of the ZRT1 gene , we predicted that the replacement of its upstream region by the GFP-HIS3 coding cassette ( Figure S5A ) might alleviate its repression . Northern blot analysis of the strain carrying one of the insertions upstream from ZRT1 ( zrt1-up1; inserted 1978 to 578 nucleotides upstream from the ZRT1 ATG; Figure S5A ) showed a four-fold derepression of ZRT1 in zinc repletion conditions , both in the wild-type and upf1Δ backgrounds ( Figure 4A ) . Insertion of this cassette eliminated the detection of the main CD-CUT of ZRT1 , with the exception of the short CD-CUT3 ( Figure 4A ) . This insertion also partially suppressed the induction defect of the upf1Δ strain during zinc deficiency ( Figure S5C ) . The kinetics of disappearance of ZRT1 upon shifting back to zinc-containing medium was also monitored in these strains after 4 hours of induction , but we found no difference in the rate of ZRT1 shutoff in the presence or absence of its main CD-CUT ( Figure S5C ) . A similar derepression was observed for FIT3 in a strain carrying a 3 Kb deletion of the region upstream of the FIT3 gene ( from −4 kb to −1 kb upstream FIT3; fit3-upΔ , Figure 4B ) . Analysis of the fit3-upΔxrn1Δ double mutant strain showed that the CD-CUTs of FIT3 were eliminated in this double mutant , which confirmed that the derepression was due to the absence of the CD-CUT . Thus , deleting the regions encoding the CD-CUTs is sufficient to trigger derepression of the bona fide mRNAs , even in a wild-type context . To provide more direct evidence that transcription of CD-CUTs is responsible for repression of the downstream promoters , we inserted the ADH1 transcription terminator ( ADH1t ) at 3 positions upstream from ZRT1 ( tA: −1839 , tB: −771 and tC: −184 bp; Figure 4C , 4D ) . If CD-CUT transcription or accumulation prevents the binding of RNA polymerase or of the transcriptional activator Zap1p , we hypothesized that terminating transcription of the CD-CUTs prior to the ZRT1 transcriptional control elements could derepress ZRT1 and/or rescue of the induction defect of NMD mutants . We first assessed ZRT1 mRNA and CD-CUTs levels in these strains in normal zinc conditions ( Figure 4C ) . A sample from a strain grown in low zinc conditions was included as a control for the ZRT1 mRNA . Insertion of ADH1t at position A did not result in major ZRT1 derepression , probably because terminating transcription at this site results in activation of an alternative CD-CUT downstream from that site ( labeled CD-CUT1′ . Figure 4C ) . However insertion of this terminator resulted in a much shorter transcript that was now insensitive to a upf1 deletion , further showing that the sensitivity of CD-CUTs to NMD is dependent on their long size , and potentially on the lack of extended ORF in the 5′-extension . Strikingly , insertion of the ADH1t at positions −771 ( B ) or −184 ( C ) resulted in a derepression of ZRT1 ( Figure 4C ) . The strongest effect was observed for ADH1t-C , possibly because this terminator stops all CD-CUT transcription immediately before the ZRT1 TATA box . In the ADH1t-B strain , an increased accumulation of the CD-CUT3 is observed , while this species disappears in the ADH1t-C strain . These results show that terminating transcription of the CD-CUTs upstream from the ZRT1 promoter is sufficient to derepress ZRT1 in conditions of non-induction . Additionally , insertion of these terminators allowed us to map in further detail the CD-CUTs upstream from ZRT1 . Based on the hybridization pattern with the different probes ( Figure S2A ) , the effect of the various terminators on their mobility in northern blots ( Figure 4C ) , the approximate architecture of these CD-CUTs is shown in Figure 4D . Transcription of the zinc regulon genes is activated by binding of the transcriptional activator Zap1p to their promoters during zinc deficiency [28] , [38] . The terminator sequence inserted at position 771 is located upstream from the three major Zap1 binding sites ( ZRE; [38] ) , while the terminator sequence inserted at position 184 is inserted downstream from them ( Figure 4D ) . Based on this , we hypothesized that if transcription of the CD-CUTs prevents binding of Zap1p , the two strains containing terminators at the two positions might behave differently during a shift into low zinc conditions . Indeed , insertion of the ADH1t at position B derepressed ZRT1 , and also fully rescued the induction defect of the upf1Δ strain ( Figure 5A ) . However strains carrying an insertion of the ADH1t at position C failed to induce ZRT1 in conditions of induction , even in a context of active NMD . This result suggests that the region located between positions B and C , which contains most of the Zap1p binding sites must be accessible for ZRT1 induction . It is unclear why the strain containing the ADH1t at position C failed to induce ZRT1 , even when NMD is active . It is possible that the higher levels of expression of the ZRT1 transporter in non-induction conditions resulted in higher cellular zinc levels prior to induction , thus delaying the response . Additionally we cannot rule out that inserting the ADH1t at site C might have changed the chromatin structure , and thus perturbed the induction of ZRT1 . To corroborate these results , we studied RNA Polymerase II occupancy in two regions , upstream from ZRT1 ( −1223 to −1123 ) , and within the ZRT1 ORF ( +905 to +1021 ) , using chromatin immunoprecipitation ( ChIP ) of the Rpb3p subunit . We found slightly above background levels of occupancy of the polymerase in conditions of zinc repletion in both regions in wild-type and upf1Δ strains ( Figure 5C ) . Interestingly , Rpb3p occupancy was similar for both strains in the upstream region , indicative of a similar level of transcription of the CD-CUTs . This result further indicates that accumulation of the CD-CUTs in the upf1Δ strain is due to a lack of degradation rather than increased transcription . Upon a shift to low zinc medium , Rpb3p occupancy increased for the wild-type strain in the ZRT1 ORF , but not in the upstream region , reflecting the induction of the ZRT1 gene . However this increase was not observed in the upf1Δ strain , corroborating the results observed by northern analysis . To show that the insertion of the terminator upstream from ZRT1 rescues the induction defect of the upf1Δ strain by allowing polymerase binding , we performed the same analysis by comparing Rpb3p occupancy in the upf1Δ and upf1Δ-ADH1t-B strains . Strikingly , Rpb3p levels were increased upon insertion of ADH1t at position B in the upf1Δ strain , both in normal zinc medium and after 3 hrs of induction , showing that the derepression of ZRT1 and the rescue of the induction defects are due to increased RNA Polymerase occupancy . We also analyzed binding of the transcriptional activator Zap1p by ChIP in wild-type and upf1Δ strains using a myc-tagged version of Zap1p inserted at the chromosomal locus . We found background levels of Zap1p occupancy to its binding sites ( ZREs ) in conditions of repression in both strains ( Figure 5D ) . However increased occupancy was observed in the wild-type strain upon a shift to low zinc ( Figure 5D ) . Binding was reduced in the upf1Δ mutant , further showing that the accumulation of CD-CUTs perturbs Zap1p binding during ZRT1 induction . Zap1p enrichment was highly specific , as it was not observed in the ZRT1 coding region ( Figure 5D ) . Overall the differences in RNA polymerase II and Zap1p occupancies in the ZRT1 gene are consistent with the results described above by Northern blot , showing that the effects observed in NMD mutants upon accumulation of CD-CUTs are indicative of transcriptional defects of the ZRT1 gene . The previous result showed that binding of the Zap1p activator is deficient in the NMD mutant upf1Δ during the low zinc response . If so , we predicted that overexpressing Zap1p might suppress the induction delay in this strain . Indeed , overexpressing Zap1p in the upf1Δ mutant was sufficient to rescue ZRT1 induction to levels comparable to those observed in the wild-type strain ( Figure 6A ) . This result shows that defective binding of Zap1p to the ZREs in the upf1Δ mutant can be overcome by overexpressing this activator . To extend these results to another gene induced in different conditions and controlled by a different activator , we monitored the expression of FIT3 in wild-type and xrn1Δ strains expressing aft1-up , a constitutively active version of Aft1p ( kindly provided by J . Kaplan; [39] , [40] . Aft1p is one of the two major transcriptional activators involved in the low iron response [39] . As expected , expression of the aft1-up allele resulted in derepression of the FIT3 mRNA in normal iron conditions in the wild-type strain ( Figure 6B ) . However , similar levels of the mature FIT3 transcript were observed in an xrn1Δ background when the aft1-up allele was expressed , showing that the presence of a constitutively activated form of Aft1p can overcome CD-CUT-mediated repression . The accumulation of the FIT3 CD-CUT was not affected by expression of the aft1-up construct ( Figure 6B ) , showing that this effect was not due to a decrease of expression of CD-CUT . We also monitored FIT3 induction in these strains upon a shift to low iron conditions ( Figure 6C ) . Interestingly , FIT3 levels did not increase in the aft1-up strain upon a shift to low iron conditions ( Figure 6C ) . However FIT3 accumulation was higher in the xrn1Δ aft1-up double mutant , possibly because of a reduced degradation of the FIT3 mRNA in the absence of Xrn1p . To investigate the specificity of the effects described above , we searched for conditions in which the induction of ZRT1 or FIT3 was uncoupled from activation by their transcriptional activators . Mutation of the Med2p tail component of the Mediator complex into a non-phosphorylated isoform ( med2-S208A ) was shown to result in a constitutive expression of FIT3 [41] . This observation led us to investigate the effect of the accumulation of the FIT3 CD-CUT on the derepression of FIT3 induced by this Mediator component mutation . We inactivated Xrn1p in a strain carrying the med2-S208A mutation ( kind gift of F . Holstege ) and analyzed the expression of FIT3 . FIT3 derepression was observed in the med2-S208A mutant grown in normal medium ( Figure 6D; time zero ) , in agreement with previous findings [41] , but this mutant did not exhibit any further induction in low iron until 3 hours after the shift . Strikingly , inactivating Xrn1p in the med2-S208A strain abolished the derepression of FIT3 observed in the med2-S208A strain ( Figure 6D ) . However the xrn1Δmed2-S208A mutant strain was not as defective for induction as the xrn1Δ strain , since the double mutant showed kinetics of FIT3 induction comparable to the wild-type strain . The result obtained in normal iron conditions ( time zero , Figure 6D ) shows that the accumulation of the FIT3 CD-CUT can inhibit the activation of FIT3 that results from a mediator component mutation . These results contrast with the result observed previously with the aft1-up mutation , which can activate expression of FIT3 even when CD-CUTs accumulate due to the inactivation of Xrn1p . Taken together , these results suggest that the function of the FIT3 and ZRT1 CD-CUTs is to prevent the premature binding of the RNA polymerase or of transcriptional activators such as Zap1 and Aft1p when these genes are transcriptionally repressed .
The precise mechanism by which CD-CUTs mediate transcriptional repression is not fully understood . Accumulation of CD-CUTs in NMD mutants negatively interferes with production of the normal transcripts and with RNA polymerase II and transcriptional activator binding ( Figure 5 , Figure 6 , Figure 7 ) . We do not know whether acting in cis is strictly required , which would be consistent with an SRG1-like transcriptional interference model [3] . We tried to express the ZRT1 CD-CUTs from a plasmid to test for the possibility of a trans effect , but could not detect any reproducible effect on ZRT1 induction ( data not shown ) . A recent study showed that transcription of the sense upstream ncRNA ICR1 mediates transcriptional control of the subtelomeric FLO11 gene [8] . We found that the region upstream from the FLO11 gene encoding ICR1 shows elevated RNA levels in the upf1Δ strain ( Figure S6 ) . Like ICR1 , expression of the ZRT1 CD-CUT is under the control of the histone deacetylase Rpd3p ( Figure 4 and [8] ) . Based on these similarities , it is possible that the mechanisms of transcriptional control of the FLO11 gene mediated by ICR1 described in [8] may be applicable to the action of the CD-CUTs that control other subtelomeric regions . The experiments in which we inserted transcription terminators upstream from ZRT1 do not allow us to differentiate between the cis and trans-acting models for CD-CUTs . Insertion of the terminator at position B relieves repression and allows the upf1Δ strain to induce ZRT1 in low zinc conditions ( Figure 5 ) . However in this strain , the CD-CUTs terminate before the Zap1 binding sites , so the results could be interpreted either way ( transcriptional interference or trans-acting ) . Insertion of the terminator at position C relieves repression , but also inhibits ZRT1 induction even when NMD is active ( Figure 5 ) . Thus , we cannot conclude whether the CD-CUTs act in trans or are only the product of transcription that generates transcriptional interference . Because NMD mutants show higher CD-CUTs levels without a higher level of RNA Polymerase in the CD-CUT transcribed region ( Figure 5C ) and also result in stronger repression , we favor the hypothesis that these RNAs act in trans . However , further work is required to fully prove this point . Another unanswered question is to understand how the transcriptional machinery overcomes CD-CUT mediated repression in conditions of induction . It is possible that CD-CUT transcription is decreased in these conditions , but neither northern analysis nor the ChIP data seem to indicate that this is the case . Another alternative is that activation of the transcriptional activators is so potent during induction that it can overcome CD-CUTs mediated repression , even if the level of transcription of CD-CUTs does not change . The results obtained with the Zap1p overexpression or the constitutive Aft1p allele strains ( Figure 6 ) are consistent with this model . One of the paradoxes raised by our observations is that 5′-extended species of subtelomeric genes are degraded by NMD , which is a cytoplasmic degradation pathway ( Figure 7 ) , yet , these CD-CUTs mediate transcriptional repression , and must therefore be localized to the nucleus if they mediate repression . Interestingly , many subtelomeric genes exhibit a perinuclear localization in S . cerevisiae [48] . In addition , connections have been made between nuclear activation of genes and localization at the periphery of the nuclear envelope near the nuclear pores ( reviewed in [49] ) . Finally , Upf1p has been shown to interact with two nucleoporins localized on the outer side of the nuclear envelope , Nup100p and Nup116p ( [50]; Figure 7 ) , suggesting that at least part of the NMD process might occur at the vicinity of the nuclear envelope ( Figure 7 ) . Therefore if both subtelomeric genes and NMD components are localized close to the nuclear envelope but on opposite sides of the nuclear pores , the physical distance between the sites of transcription and action of CD-CUTs , and their site of degradation might be closer than thought from just considering the nuclear/cytoplasmic distribution ( Figure 7 ) . This would possibly allow retrograde transport of these CD-CUTs from their site of degradation to their site of action , and would allow a regulation of transcription by ncRNAs primarily degraded in the cytoplasm ( Figure 7 ) . It is also possible that CD-CUTs might be degraded when they emerge out of the nuclear pore complex , and that failure to degrade induces an increase of their nuclear localization , explaining a higher level of repression in NMD mutants . A better analysis of the mechanisms of nuclear/cytoplasmic trafficking of these CD-CUTs will ultimately allow a full understanding of their mechanisms of action . Despite these unanswered questions , our results have uncovered a novel function for NMD in controlling the accumulation of transcripts that negatively interfere with transcription of genes involved in zinc and iron homeostasis . Previous work has shown that upstream ncRNAs are involved in controlling gene expression related to metals homeostasis . It was shown that Zap1p activates the transcription of ncRNAs which mediate the repression of zinc-dependent alcohol dehydrogenases by transcriptional interference during zinc deficiency [51] . However it is unclear whether or not these ncRNAs are targeted by NMD , in a manner similar to CD-CUTs that regulate ZRT1 , ADH4 and FIT3 . A potential transcriptional interference mechanism involving long unstable upstream sense ncRNA has also been described in Chlamydomonas during copper deficiency [52] , suggesting that this mechanism has been conserved during evolution to contribute generally to metal homeostasis genes regulation . Thus , there seems to be prevalent use of ncRNA transcription to control gene expression during metals homeostasis in different organisms . In addition to the potential crosstalk with transcription described here , NMD was also recently shown to regulate Mg++ cellular levels [23] by degrading the transcript encoding the main Mg++ transporter . The fact that this RNA surveillance system is so intimately implicated in the regulation of metals homeostasis in general might be linked to the prevalence of these metals in the ribosome and in their function of translation and in its fidelity [23] , possibly revealing another layer of co-evolution between NMD and translation .
Most strains were derived from BY4741 or 4742 ( Open Biosystems ) . Strains in which the GFP-HIS cassette in the upstream region of ZRT1 or ADH4 gene were obtained by homologous recombination [36] . The strain carrying the deletion of the region upstream FIT3 and the strains containing the terminator insertions were obtained by delitto perfetto [53] . Double mutants in which the UPF1 or XRN1 genes were knocked out were obtained by direct disruption of these genes in other mutant strains , as described [21] . Insertion of the myc-tag for Zap1p was performed as described [36] . Strains were grown in conditions of non-induction in either YPD or Synthetic Complete medium ( SC ) supplemented with 2 mM ZnCl2 . Growth in condition of low zinc gene induction was performed in either a Chelex-treated synthetic complete medium ( CSC ) or a SC medium containing 1 mM EDTA , pH adjusted to 4 . 4 with 20 mM citrate . CSC was prepared as described [28] except that all amino acid required were added and pH adjusted to 4 . 4 . Growth in conditions of low iron gene induction was performed by adding BPS chelator as described [31] . Strains were grown in YPD ( pre-low iron shifts ) or SC+2 mM Zn ( pre-low zinc shifts ) until OD600 = 0 . 5 , washed twice in sterile water , and shifted into YPD medium with BPS ( low iron shift ) or SC+EDTA medium ( low zinc shift ) for the indicated times . For the experiments including overexpression of Zap1p , WT or upf1Δ strains containing the vector pUG35 or pIT31 ( see below ) were grown in SC medium without Uracil ( SC-URA ) supplemented with 2 mM Zn . At OD600 = 0 . 45 , cells were washed twice in sterile water and shifted in SC-URA without methionine ( SC-URA-MET ) supplemented with 2 mM Zn for 2 hours to overexpress Zap1p prior to zinc starvation . After 2 h , cells were washed twice in sterile water and maintained in log phase in SC-URA-MET medium containing 1 mM EDTA . Kinetics of induction were performed as described above . A PCR product corresponding to the ZAP1 gene was generated from genomic DNA with primers containing the restriction sites ClaI and SalI and inserted in the vector pUG35 digested by the same restriction enzymes . After transformation and amplification in E . coli , the plasmid ( pIT31 ) was confirmed by sequencing . The aft1-up expression plasmid was obtained from J . Kaplan ( U . of Utah ) . Tiling Arrays used in this study were described previously [21] , [31] and are accessible in the GEO database ( accession number GSE11621 ) . Northern blot hybridization analysis was performed as previously described [21] , [31] . All riboprobes were synthesized with the T3 MAXIscript kit ( Ambion ) . Riboprobes were hybridized at 67°C except for the ADH4 ORF probe ( 65°C ) . ChIPs using anti-Rpb3 RNA polymerase II subunit and a myc-tagged version of Zap1p inserted at the chromosomal locus were performed as described [54] , [55] .
|
Precise expression of genes relies on their accurate and timely transcription and on turning off these genes when production of the proteins is not required . Our study describes that in the baker's yeast , S . cerevisiae , repression of many genes relies on transcription of long extended RNAs upstream from where transcription normally initiates . These long extended RNAs are degraded by a machinery that recognizes that these RNAs do not encode any functional proteins . Using genes involved in controlling the uptake of the essential elements zinc and iron , we find that transcription of these long extended non-coding RNAs represses transcription by preventing the binding of the transcriptional machinery to the normal transcriptional control elements . Our findings show that repression of transcription of many genes relies on the transcription of unstable long RNAs and that this mechanism of control is particularly prevalent for genes involved in controlling the level of metals inside cells .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genetics",
"biology",
"genomics",
"genetics",
"and",
"genomics"
] |
2011
|
Cryptic Transcription Mediates Repression of Subtelomeric Metal Homeostasis Genes
|
The p75 neurotrophin receptor , a member of the tumor necrosis factor receptor superfamily , is required as a co-receptor for the Nogo receptor ( NgR ) to mediate the activity of myelin-associated inhibitors such as Nogo , MAG , and OMgp . p45/NRH2/PLAIDD is a p75 homologue and contains a death domain ( DD ) . Here we report that p45 markedly interferes with the function of p75 as a co-receptor for NgR . P45 forms heterodimers with p75 and thereby blocks RhoA activation and inhibition of neurite outgrowth induced by myelin-associated inhibitors . p45 binds p75 through both its transmembrane ( TM ) domain and DD . To understand the underlying mechanisms , we have determined the three-dimensional NMR solution structure of the intracellular domain of p45 and characterized its interaction with p75 . We have identified the residues involved in such interaction by NMR and co-immunoprecipitation . The DD of p45 binds the DD of p75 by electrostatic interactions . In addition , previous reports suggested that Cys257 in the p75 TM domain is required for signaling . We found that the interaction of the cysteine 58 of p45 with the cysteine 257 of p75 within the TM domain is necessary for p45–p75 heterodimerization . These results suggest a mechanism involving both the TM domain and the DD of p45 to regulate p75-mediated signaling .
The neurotrophin receptor p75 is a member of the tumor necrosis factor receptor ( TNFR ) superfamily and has four extracellular cysteine rich domains , a single transmembrane ( TM ) domain , and an intracellular domain ( ICD ) comprising a juxtamembrane and a death domain ( DD ) [1]–[5] . Depending on co-receptor partners and cellular contexts , p75 may play seemingly opposing effects in multiple systems . For example , p75 interacts with Trk receptors to promote neurotrophin-dependent nerve growth . In contrast , p75 has been shown to play a role in apoptosis when binding to pro-neurotrophins and with the co-receptor sortilin [4] . In addition , p75 inhibits nerve growth mediated by myelin-associated inhibitors via functioning in part as a co-receptor for the GPI-linked neuronal Nogo-66 receptor ( NgR ) [6] or another non-NgR molecule that is yet to be identified [7] , [8] . Elucidation of the mechanisms that modulate p75-mediated signaling may increase our understanding of neural development and nerve injury . Upon nerve injury in adult mammals , factors at the injury site such as myelin-associated inhibitors inhibit regeneration of injured axons , resulting in permanent disability . Axon regeneration is blocked by the presence of multiple types of nerve growth inhibitors , such as myelin-associated inhibitors from damaged myelins , chondroitin sulphate proteoglycans , and repulsive axon-guidance molecules expressed by reactive glial cells [9]–[12] . The structurally dissimilar myelin-associated inhibitors Nogo66 , MAG , and OMgp inhibit axon growth by binding to the NgR , a GPI-linked protein , which then transduces the inhibitory signal into the cell by binding to co-receptors with intracellular signaling domains , such as p75 [13] , [14] or TROY [15] , [16] . LINGO-1 also plays a role in NgR signaling [17] . Downstream from their receptor binding , these myelin inhibitors trigger inhibition of axonal growth through the activation of the small GTPase Rho [18]–[21] in a protein kinase C ( PKC ) -dependent manner [22] . Targeting this complex has been described to lead to the promotion of neurite outgrowth , oligodendrocyte proliferation and differentiation , and inhibition of cell death . p45 is highly homologous in sequence to p75 . It is also called neurotrophin receptor homologue 2 ( NRH2 ) [25] , neurotrophin receptor alike DD protein ( NRADD ) [24] , or p75-like apoptosis inducing DD protein ( PLAIDD ) [23] . P45 displays strong sequence similarity to p75 in the TM , juxtamembrane , and DD regions [26] . P45 contains a truncated and short extracellular domain ( ECD ) with no neurotrophin binding domain . It has been shown previously that p45 associates with p75 and with TrkA receptors [23] , [25] , [27] . In addition , p45 participates in the trafficking of sortilin to the plasma membrane [27] . However , its role in other p75-regulated signaling pathways has not been studied . In this study , we have explored the modulation of p75/NgR signaling . The results indicate that p45 heterodimerizes with p75 and , thereby , impedes the formation of p75 homodimer that is required for the p75/NgR complex formation and its downstream activation of RhoA GTPase . In addition , we found that p45 binds p75 through both the TM and the ICDs . Furthermore , we showed that a cysteine–cysteine interaction within the TM domain of p45 and p75 is required for stabilization of their heterodimer formation . The results reveal a new mechanism of modulating p75-mediated inhibitory signaling via heterodimer formation with a member of the TNFR superfamily such as p45 .
p45 contains a DD ( Figure 1A ) . Because the function of DDs is to bind the DD of other members of the DD superfamily in order to transduce signals [28] , we investigated whether p45 can interact with other members of the DD superfamily , such as p75 , FADD , TNFR1 , and Fas . As shown in Figure 1B , p45 can be co-immunoprecipitated with p75 or FADD when co-expressed in 293 cells , but not with caspase-8 , Fas , or TNFR1 . Recently we characterized the interaction of p45 with FADD and its role upon spinal cord injury [29] . In the present study , we focus on the interaction of p45 with p75 . p45 and p75 share a high degree of amino acid similarity in their TM domain ( 94% ) , including conserved cysteine residues [30] , and the ICD ( 50% ) ( Figure 1A ) . However , the ECD of p45 is short and has no binding sites for neurotrophins ( Figure 1A ) . p45 and p75 are expressed in both the peripheral nervous system ( PNS ) and central nervous system ( CNS ) during development [26] . We found that p75 and p45 form an immunocomplex in cerebellum extracts ( Figure 1C ) . The expression level of p45 is high in embryonic but is significantly reduced in adult tissues ( Figure 1D ) . However , p45 is up-regulated after sciatic nerve injury in the spinal cord and sciatic nerve in a similar fashion as p75 ( Figure 1E , F ) . Because p45 shares a similar protein sequence with p75 and has similar up-regulated expression patterns with p75 after injury , we decided to investigate whether p45 regulates signaling mechanisms involving p75 . p75 mediates nerve growth inhibition by myelin-associated inhibitors via functioning in part as a co-receptor for GPI-linked neuronal NgR [6] . Because p45 and p75 form a stable complex ( Figure 1B ) , we investigated whether p45 interferes with or enhances the formation of the complex between p75 and NgR and subsequent signaling through the p75/NgR complex . When HEK293 cells were co-transfected with p75 and Flag-tagged human NgR ( Flag-hNgR ) expression constructs followed by immunoprecipitation with anti-p75ICD antibodies , we found that p45 markedly reduced the levels of complex formation between p75 and NgR ( Figure 2A ) and it does in a concentration-dependent manner ( Figure 2B ) . As a control , p45 is not able to bind to NgR directly ( Figure S1 ) . To further characterize the domains of p45 involved in such inhibition , we made several deletion constructs of p45 ( Figure S2 and Table S1 ) . As shown in Figure S2 , both p45 intracellular and TM domains are necessary for p45 inhibition of p75/NgR complex formation . We examined the possibility that p45 would antagonize signaling through the p75/NgR complex . Previous results demonstrated that p75 is required for MAG-induced RhoA activation through NgR [20] , [31] . RhoA activation is necessary for neurite outgrowth inhibition mediated by myelin-associated inhibitors . Thus , we examined whether overexpression of p45 blunts signaling through the p75/NgR complex . Postnatal day 7 ( P7 ) cerebellar granule neurons ( CGNs ) that express low levels of endogenous p45 were transfected with full-length , capped p45 RNAs containing poly ( A ) tails generated by an in vitro transcription system . As shown in Figure S3A , the level of p45 protein was markedly elevated 24 h following RNA transfection . The cultures were then serum-starved and treated with Fc or MAG-Fc proteins . The level of activated RhoA was measured . As illustrated in Figure S3B , overexpression of p45 blocks MAG-Fc–induced RhoA activation . We also quantitatively measured RhoA activation using the G-LISA kit ( Cytoskeleton Inc . ) on CGN cultures from wild-type mice and Thy1-p45 transgenic mice that consistently overexpress p45 under a Thy1 promoter [29] . As shown in Figure 2C , MAG-Fc treatment of the WT CGNs induced 50% increase in the RhoA activity , whereas the MAG-Fc–induced RhoA activation is completely abolished in Thy1-p45 CGNs ( Table S1 ) . These results suggest that p45 is capable of effectively blocking RhoA activation through the p75/NgR complex . We then examined whether overexpression of p45 prevents neurite outgrowth inhibition induced by Nogo66 , MAG , or CNS myelin . P7 CGNs were transfected with p45 RNAs , plated onto dishes previously coated with different substrates , and allowed to grow overnight . The cultures were double immunostained with antibodies against p45 ( green ) and neurotubulin ( TuJ1 , red ) ( Figure 2D ) . The neurite length of control CGNs and transfected CGNs that display increased p45-immunoreactivity over control CGNs was measured . As shown in Figure 2E , neurite outgrowth inhibition elicited by Nogo66 is alleviated by p45 overexpression ( Table S1 ) . Similarly , p45 overexpression significantly promotes neurite outgrowth that was otherwise inhibited when cultured on dishes coated with CNS myelin or MAG-expressing cells ( Figure 2E ) . These results support the idea that p45 promotes neurite outgrowth . It is worth noting that despite NgR has been implicated in mediating nerve growth inhibition induced by myelin inhibitors in culture , neurite outgrowth of CGNs from NgR null mutants is still inhibited by myelin inhibitors [7] , [8] . Recent results suggest that NgR is required only for the acute growth cone-collapsing but not chronic growth-inhibitory actions of myelin inhibitors [32] . Furthermore , no measurable corticospinal tract regeneration was observed in mice lacking all Nogo isoforms [33]–[35] ( but see Cafferty et al . [36] ) . In contrast , inhibition of nerve growth by myelin inhibitors is significantly reduced in p75-deficient CGNs [7] , [8] . These results raise the possibility that a yet to be identified receptor mediates myelin inhibitor activity through p75 . To understand the mechanism by which the p75–p45 interaction regulates p75-dependent signaling , we first characterized biophysically the ICD domain of p75 . Because the DD of p75 is a protein–protein interaction motif and often is involved in functionally essential homo- and hetero-associations [37]–[40] , we studied the oligomerization behavior of p75ICD comprising residues 290–418 . We used gel filtration chromatography of purified p75ICD to analyze the oligomerization state of p75ICD in phosphate buffer at pH 8 . 0 . Purification of p75ICD from bacteria yielded two peaks in the elution profile of gel filtration ( Figure 3A , black lines ) with the presence of some high molecular weight aggregates ( Figure 3A , asterisk ) . After running SDS-PAGE of the peak fractions in reducing and nonreducing conditions , we found that the elution fraction I corresponds to a covalently disulfide bond dimer and fraction II to a monomer molecular weight ( Figure 3B ) . When we purified p75ICD with the presence of DTT , a single peak was observed that eluted between the monomer and dimer peaks ( Figure 3A ) . The same behavior is observed in the presence of iodoacetamide , which blocks free cysteines , in the lysis buffer ( Figure S4 ) . The estimated molecular weight of this fraction from the gel filtration yields a mass of 35 kDa ( p75ICD MW is 16 kDa ) , suggesting the presence of a dimer . We carried out an analytical ultracentrifugation analysis of purified p75ICD in the presence of DTT , using equilibrium sedimentation and velocity experiments of recombinant p75ICD in the same buffer ( Figure 3C ) . Our ultracentrifugation confirms that p75ICD behaves in solution as a single species with a molecular weight of 30 . 7 kDa , close to the theoretical dimer of p75ICD ( see Figure 3C legend ) . Altogether , we conclude that p75ICD is a noncovalent dimer that during purification or in oxidative conditions dimerizes through a disulfide bond ( Figure S5 ) . Recently the crystal structure of a covalent disulfide p75ICD dimer , purified from bacteria in oxidative conditions ( in the presence of DTNB ) , has been described [41] . The dimer is mediated by a covalent disulfide bond through Cys379 . We think that the dimer found in this work is the same dimer , because when we made p75-C379A mutant , gel filtration gives a monomer peak ( Figure S6 ) . Whether this covalent disulfide dimer is formed in the reducing conditions encountered inside cells and what its biological function is need to be further investigated . NMR titration experiments were performed at different p75ICD concentrations in the presence of DTT to avoid formation of the disulfide dimer . Concentration-dependent chemical shift changes in [42] TROSY spectra [43] of 15N-labeled p75ICD were then used to map the homodimer interface ( Figure 4A ) . Figure 4B shows several examples of concentration-induced chemical shift changes . The appearance of only one peak suggests a slow monomer–dimer equilibrium for NMR time scale measurements . This pattern of change is indicative of a weak binding that we can estimate from plotting changes in the chemical shift of interface residues . We obtained a Kd of ∼100 µM ( Figure 4C ) . Such low binding affinities are typically observed for DD-type interactions [44] . Residues with the most pronounced chemical shift changes—that is , L360 , E363 , Q367 , H370 , D372 , F374 , T375 , C379 , H376 , E377 , A383 , L384 , L385 , and W388 , ( Figure 4A ) —were then mapped onto the reported NMR structure of p75ICD ( Figure 4D ) [45] . These residues are all located on one side of the DD , in particular within helices α3 and α4 . Together with the presence of only one TROSY cross-peak per 15N-1H moiety , these results suggest the formation of a symmetrical p75-DD homodimer . Because helices α3 and α4 contain charged amino acid residues , the homodimer formation may be in part due to electrostatic interactions . Recently the crystal structure of an asymmetrical dimer of p75-DD has been described [41] . In that structure , R404 from monomer A interacts with S373 , H376 , and E377 of monomer B . In our NMR titration study , chemical shift changes of R404 are not observed ( Figure 4A ) , although changes in the chemical shift of S373 ( small ) and E377 and H376 ( big ) are clearly visible ( Figure 4A ) . These results suggest that in solution the symmetrical dimer is favored , although we cannot exclude the possibility that p75ICD with different conformations is present in solution . Sometimes crystallization favors conformations that are better packed but are not necessarily the prevalent conformations in solution . To further characterize and confirm the homodimer interface in the full-length p75 , we made p75 mutant constructs containing amino acid replacements at different residues , which showed high concentration-dependent chemical shift changes in the NMR studies of p75ICD ( Figure 4A ) . The emphasis of the amino acid replacements was charge changes due to the potential electrostatic nature of the interaction ( D372R , H376E , and E377R; Figure 4E ) . Wild-type or mutant p75 constructs were co-transfected with a Flag-tagged construct that contained only the TM domain and the ICD of p75 ( Flag-ΔECD-p75 ) in HEK293 cells . The presence of the p75 dimers was measured by co-immunoprecipitation with an anti-Flag antibody and detection of full-length p75 with an anti-p75 antibody . As shown in Figure 3G , wild-type p75 forms a dimer with Flag-ΔECD-p75 . In comparison to wild-type p75 , the p75 mutant E377R shows a significant decrease in dimer formation , suggesting the importance of this residue and the negative charge in the homodimer interface . In contrast , the mutant H376E has a stronger binding than wild-type p75 ( Figure 4E ) . The role of H376E mutation in dimer formation suggests that the dimer formation could be dependent on the ionization of H376 and then on the pH of the solution . The fact the mutation H376E favors the dimerization suggests that an electrostatic interaction plays a role in homodimerization . Very little is known about how p75 and NgR interact from a structural point of view . To shed light on this and to understand how p45 modulates p75/NgR signaling , we first investigated how p75 and NgR interact with each other . We used the following p75 constructs: ( 1 ) p75 dimerization mutants E377R , D372R , and H376E ( Figure 4 ) , which exhibit a significant reduction or increase of p75 homodimer formation , and ( 2 ) p75-C257A , a mutant in the TM domain of p75 , which although able to form dimers , is not functional upon NGF binding [30] . When co-transfected , mutants E377R and D372R showed less interaction with NgR ( Figure 5A ) . However , mutant H376E , which promotes p75 homodimer formation , displayed a significant increase in its capability to bind NgR ( Figure 5A and Table S1 ) . We then determined whether C257 plays a role in the interaction with NgR . When co-transfected with NgR , the p75-C257A/NgR interaction was markedly impaired , although some interaction was observed ( Figure 5B and Table S1 ) . When we performed the co-immunoprecipitation of NgR with p75-wt and ran a nonreducing SDS-PAGE , we found the majority of p75-wts that were co-immunoprecipitated with NgR were in the form of dimers , whereas NgR co-immunoprecipitated a small and similar amount of p75-wt and p75-C257A in the form of monomers ( Figure 5C ) . This suggests that NgR and p75 form a complex that is better stabilized with p75-wt than with p75-C257A . In addition , because p75-C257A is still able to form dimers as inferred by crosslinking [30] , [46] , these results suggest that p75-wt dimers mediated by C257 have a preferred conformation for binding to NgR . From these data we conclude that it is not the mere dimerization of p75 but the conformation stabilized by both disulfide bond and DD electrostatic interactions that is preferred for NgR interaction . Our data ( Figure S2 ) and previous published data from other authors have mapped the interaction between p45 and p75 to the TM and the ICD [27] . To map the binding interface of p45 and p75 ICDs , first we solved the three-dimensional solution structure of mouse p45ICD by NMR spectroscopy ( Figure 6 ) . The NMR studies showed that p45ICD contains a flexible domain at the N terminus ( residues 75–140 ) that could not be assigned because they display limited chemical shift dispersions , and a folded domain at the C terminus ( 141–218 ) ( Figure S7 ) . Figure 6A presents the three-dimensional NMR structure of the folded DD domain obtained from the experimental restraints ( Table S2 ) . The regions with a secondary structure are the best defined and typical for a DD , and p45DD is composed of six α-helix disposed in a specific orientation ( Figure 6A ) comprising residues 141–147 ( α1 ) , 154–167 ( α2 ) , 169–180 ( α3 ) , 182–190 ( α4 ) , 200–207 ( α5 ) , and 212–218 ( α6 ) . A DALI [47] search revealed p75DD as the closest structural relative with an rmsd of 2 . 7 Å , followed by other members of the DD family ( Table S3 ) . The DD of p75 and p45 share many structural features and the same arrangement of all the six α-helices , which is not surprising as p75DD and p45DD are homologues . Only the length of the loop between α1 and α2 is longer in p45DD because of an insertion of four amino acid residues in this segment . When compared with p75DD , the longer loop reorients α1 in respect to α2 and α3 and brings residue E153 of the loop in close neighborhood to other negative charged residues ( Figure 6B ) . Together with some additional amino acid residue differences , this small structural reorientation changes significantly the charge distribution around helix α3 of p45DD ( Figure 5B ) . The negative charged region of p45DD is formed by E153 , E160 , E170 , E173 , and D178 . In the equivalent region of p75DD ( Figure 6B ) , the negative charged E363 , E369 , D372 , and E377 are located in a more balanced environment surrounded by positive charged residues R358 , H370 , and H376 , which are positively charged depending of their specific pka . Another interesting feature of the sequence of p45DD is the presence of the RxDΦ motif ( x , any residues; Φ , a hydrophobic residue ) at the beginning of helix α6 ( Figure 5C ) , which is typically observed in death effector domains ( DEDs ) , such as the DED from PEA-15 , FADD , Caspase-8 , and others [48] , [49] . In DEDs , this motif has been suggested to stabilize the DED fold , because it participates in a salt-bridged network between the arginine side chain and the aspartic acid side chain of the RxDΦ motif , and a glutamic acid side chain located in the helix α2 ( for instance , R72 , D74 , and E19 in FADD-DED ) ( Figure 6C ) [48] . Such a charged network is also present in the three-dimensional structure of p45DD between residues R211 , D213 , and E160 , as indicated by the large downfield shift of the Hε for R211 of p45 , indicative of a charged interaction ( Figure 6A , the asterisk indicates the downshift of R211 ) . A similar shift has been observed for R72 of FADD-DED [48] . The presence of this salt bridge is an unexpected feature of p45DD , because this motif is not found in any other DD . p75DD has a similar RxDΦ sequence , RADI ( highlighted in black in Figure 6C ) , and from this argument , Park et al . [50] has suggested that p75DD is a DED , not a DD . However , in our hands , we did not see a large shift of the arginine of p75 by NMR , like in p45 and in PEA , suggesting that maybe this arginine is not forming a salt bridge . In fact , in the amino acid sequence of p75 , the analogue residue for E160 involved in the salt bridge in p45 is a His residue ( H217 ) ( Figure 6C ) . Nevertheless , the possibility that p45 and p75 DDs are actually DEDs or a chameleon between DD and DED should be considered in future research . We then started the characterization of the p75ICD–p45ICD interaction . Using NMR chemical shift perturbation experiments with p45ICD , the binding site of p75ICD on p45ICD was mapped to α2/α3 in the DD of p45 ( Figure S8 ) . To further characterize the interaction between p45DD and p75DD in the corresponding full-length proteins , p45 point mutations at some residues with significant chemical shift changes were constructed ( i . e . , H164E , E173R , C177H , D178A , and D178R ) . The presence of a full-length p45–p75 complex in correspondingly transfected HEK293 cells was measured by co-immunoprecipitation of p75 with wild-type or mutant p45 ( Figure 7A ) . Wild-type p45 formed a complex with p75 . Of the entire series of p45 mutant constructs studied , E173R and D178R showed a significant decrease in complex formation , whereas C177H and D178A showed a significant increase in complex formation ( Figure 7A ) . NMR experiments with purified p45 mutants showed that they were well folded ( Figure S9 ) . These data are consistent with the structural studies and suggest that p45 forms a complex with p75 through its helix α3 of the DD . Next , the amino acid residues in p75ICD that interact with p45ICD ( Figure S4 ) were characterized . In the chemical shift perturbation experiment , p45ICD is titrated against 10 µM of 15N-labeled p75DD . Chemical shift changes on p75DD are observed in amino acid residues located close to and on helix α3 of p75DD ( i . e . , L360 , E363 , Q367 , H370 , D372 , F374 , T375 , H376 , E377 , A378 , A383 , L384 , L385 , W388 ) , indicating helix α3 is the binding site for p45DD on p75DD ( Figure 7B ) . The effect on the interaction between p45 and p75 by some of the amino acid replacements located in this side of p75DD ( D372R , H376E , and E377R ) supports the NMR-derived interface ( Figure 7B ) . A comparison of the p45DD binding site on p75DD ( Figure 7A ) with the p75DD homodimer binding site ( Figure 4D ) shows that the two sites overlap . The presence of an overlapping binding site on p75 suggests that p45 is binding to a monomeric p75 by forming a heterodimer . We tried to purify a stable complex between p75ICD and p45ICD by gel filtration , but we were unsuccessful . It could be possible that p45ICD interaction inhibits the formation of p75 dimers or multimers , but the interaction is too weak to yield sufficient p75/p45 heterodimers . The presence of a partially overlapping binding site on p75DD for homodimerization with p75DD and heterodimerization with p45DD suggests that p45DD may break a p75DD dimer by competition . To get insights into this potential mechanism , the p45ICD-induced chemical shift perturbations of p75ICD were measured at high ( dimeric ) and low ( monomeric ) concentrations of p75ICD . p45ICD binds p75ICD even at high p75ICD concentrations , where p75ICD is mainly homodimeric ( Figure S6 ) . As demonstrated above , A378 is participating in the homodimer interface of p75DD because chemical shift differences between low and high concentrated p75ICD were observed ( Figure S10 ) . However , the addition of p45ICD at low concentrations of p75ICD did not result in any chemical shift changes of A378 , indicating that A378 is not part of the p45–p75 binding site . In contrast , the addition of p45ICD at a high ( dimeric ) concentration of p75ICD generated chemical shift perturbations of the 15N-1H moiety of A378 to values identical to p75ICD at a low ( monomeric ) concentration in absence of p45ICD ( Figure S10 ) . Our explanation of these findings is as follows . Although A378 is participating in p75DD homodimer formation , it is not involved in p45DD binding . However , the presence of p45 breaks the p75 dimer and forms a p45–p75 heterodimer and thus shifts A378 from a dimeric environment to a monomeric environment ( Figure S10 ) . Complementary information about the breaking of a p75 homodimer into a p45–p75 heterodimer can be extracted by the p75ICD- and p45ICD-dependent chemical shift perturbations of T375 . T375 has different chemical shifts at high and low concentrations of p75ICD , indicative of its participation in the p75DD homodimer formation . However , upon the addition of p45DD , the chemical shifts of T375 move to a new position that is independent of the p75ICD concentration ( at least at the concentration window studied here; Figure S10 ) . This result suggests that T375 is involved in both p45DD and p75DD binding . Furthermore , the data indicate that p45 is able to compete with the p75DD homodimer by the formation of a p45DD–p75DD heterodimer . Other residues of p75 showed similar behavior , suggesting that p45DD is able to break the relatively weak p75DD homodimer by forming a heterodimer p45DD–p75DD . The results from above suggest that the p45 modulation of p75 signaling is increased with the presence of the TM domain of p45 . The TM domain of p75 self-associates [30] . Because the TM domain of p45 is highly homologous to p75-TM ( Figure 7C ) , we asked if p45-TM was able to bind p75 through its TM . Co-immunoprecipitation experiments were conducted in HEK293 cells transfected with p75 wt and p45 , and analysis in nonreducing SDS-PAGE showed a band recognized by both p75 and p45 antibodies and with a molecular weight corresponding to a heterodimer p75–p45 ( Figure 7D ) . This band was lost when p75-C257A was co-transfected with p45 . These results suggest that p75 and p45 can form a heterodimer through C257 from p75 and another cysteine residue from p45 . Due to the homology in the TM domain , we mutated the cysteine equivalent in p45 , Cys58 ( Figure 7C ) . As shown in Figure 7E , p45-C58A did not form a complex with wt p75 . Only when the blot was exposed longer did we see an interaction between p45 and p75 , presumably by interaction through their intracellular DDs .
The presented biophysical and biological characterization of p45 and p75 interactions shows that a p45–p75 heterodimer is formed , using both the TM and the ICDs . p45 binding to p75 inhibits RhoA activation and increases neurite outgrowth . In light of these data , the following mechanistic model of p45 action is proposed ( Figure 8 ) . NgR binding to p75 recruits intracellular proteins that activate RhoA signaling . In the presence of p45 , however , p45–p75 heterodimers are formed , stabilized by the Cys257–Cys58 interaction within the TM domain and enhanced by the interaction of cytoplasmic domains . In this heterodimer conformation , the interaction of NgR with p75 is impaired and this translates into diminished RhoA signaling . p45 interaction with p75 has been described previously [23] , [27] . However , a prominent role of the TM domain in that interaction was not suggested . The interaction between p45 and p75 is decreased by a deletion of the p75ICD , but not entirely blocked , suggesting that the TM domain also has a prominent role in the p75/p45 interaction ( see figure 3C in [27] ) . It is also interesting to note that p45 has been suggested previously to interact with TrkA and modulate its activity [25] . In that publication , the authors made several deletion constructs of p45 and suggested that the TM domain of p45 was needed for the p45–TrkA interaction [25] , [51] . That observation , together with our data , suggests that the TM domains of neurotrophin receptors are starting to be recognized as very important for their function . One important question that remains to be answered is where and how in the cell the p45/p75 heterodimer could be formed . It is difficult to imagine that p45 could break a covalently formed p75 homodimer , however free monomeric p75 is present in the cell membrane [30] . One possibility is that the heterodimer is formed in the ER during the translation and maturation of both proteins . Kim et al . have shown that p45 is involved in the trafficking of sortilin to the plasma membrane , indicating that the p45/sortilin interaction takes place in the ER membrane [27] . The p75/p45 interaction could also occur in the ER membrane during receptor maturation . This implies that it requires the synthesis of new proteins for p45 activity , and only when new p75 and p45 molecules are synthesized will they have the opportunity to form the heterodimers . Such a situation is plausible because both p75 and p45 are up-regulated upon nerve injury and they are expressed in the same cells ( Figure S11 ) . In that situation , at the plasma membrane , it could be possible to have different oligomers of p75—namely , p75 monomers , p75 homodimers , and p75/p45 heterodimers . The function of those species may be different , contributing to the complexity of p75 signaling . We showed that p75ICD exists in a monomer–dimer equilibrium mediated by electrostatic interactions , and we postulate that p75 dimerization is pH-dependent or promoted by the presence of counterions , like phosphate in our buffer . This could reconcile the contradictory results reported by previous studies of p75ICD structure . NMR structural studies detected only the monomeric p75ICD species at a pH in the range of 6–7 and in plain water with no counterions [45] , solution conditions in which p75ICD will not be favored to self-associate . In contrast , X-ray structures of p75ICD revealed a dimer [41] , but the buffers used for crystallization were at or above pH 7 . 0 and contained 1 . 1–1 . 4 M sodium malonate or sodium citrate , as stabilizing counterions . Thus , we conclude that p75 dimerization is dependent on the pH and the presence of counterions . It is interesting to compare the homo- and hetero-dimeric interactions found here with other protein oligomerizations that involve DDs and DEDs . Although DDs and DEDs were first identified in proteins that mediate programmed cell death , they are now recognized to act as protein interaction domains in a variety of cellular signaling pathways [28] . In the DD subfamily , low sequence homology produces diverse interaction surfaces , enabling binding specificity within a subfamily [52] . Protein–protein interactions of DDs have been thereby thought to be predominantly homotypic among different adaptors , as shown here for p75DD homodimer formation , although some examples of heterotypic interactions have been demonstrated ( reviewed in [28] ) , including the p45–p75 presented here . These results suggest that DDs employ diverse mechanisms for interactions [53] . They can be classified into three types of interactions ( reviewed in [52] ) . Type I interaction is exemplified by the procaspase-9 CARD:Apaf-1 CARD complex [54] , whereas the type II interaction is represented by the Pelle DD:Tube DD complex [53] , and the type III interaction is proposed to exist in the Fas DD:FADD DD complex [55] . p75 and p45 DD interaction appears symmetrical based on chemical shift data , but one cannot exclude an asymmetrical interaction that uses the same regions of the interface; as p75 is a homodimer and p45 and p75 binding sites overlap only in parts , the p45–p75 interaction could not be totally symmetric . Because p75DD interacts with itself through residues located between helices α-3 and the loop connecting α-3 and α-4 , they belong to a type III interaction [48] . The asymmetric type I , II , and III interactions between DDs are conserved in all current structures of oligomeric DD signaling complexes [55]–[57] . These interactions likely represent the predominant mechanism of DD polymerization . Here , p45DD acts as an inhibitor of those interactions , shutting off or modulating the p75 signal strength . Interestingly , both p45DD and p75DD are promiscuous and can interact also with other proteins through their DD . Although p75DD is able to interact with downstream targets , such as the CARD of RIP-2 [58] , p45DD appears to interact with FADD , thereby reducing FADD-mediated cell death [29] . Recently the regions of p75DD involved in the three different p75 signaling paradigms has been mapped by mutagenesis—namely , apoptosis , NF-κB activation , and Rho signaling [59] . For p75/p45 heterodimers , p45DD will occupy the region very close to where the CARD domain of RIP-2 is binding to p75DD , according to previous data [59] . Strikingly , RIP-2 binding to p75 is necessary for Rho-GDI release and RhoA inhibition . The data suggest that p45DD binding to p75 will release RhoGDI and inhibit RhoA activation . This is in agreement with our data that p45 binding to p75 inhibits RhoA activity . Recently it has been described that p75 could adopt two different conformations , a symmetrical dimer , stabilized by a cysteine disulfide bond , and an asymmetrical dimer [41] . The authors proposed that p75 could be in equilibrium between both conformations unless oxidant conditions inside the cell promote the formation of the disulfide bond [41] . Our NMR data suggest that the symmetrical conformation is the predominant form at least in solution , because interaction between residues from helix 3 and helixes 5–6 are not seen in our conditions . Further investigation will be needed to understand which conformation belongs to the active receptor . The fact that p45 is able to bind and to block the symmetrical interface suggest a well-designed and potent p75 inhibitor . Apart from the p75 signaling , p45 might play additional roles of an inhibitory nature . In particular , because the DD of p45 appears to be important for p45/p75 interaction and several members of the TNFR family contain DDs such as TNF-R1 or CD95 , which have been shown to play important roles in SCI [60] , [61] , it is intriguing to speculate whether p45 may antagonize the activity of some of these receptors upon SCI by binding to their DD domains as well . Thus , the promiscuous structural nature of p45 may facilitate functional recovery after SCI by inhibiting multiple signaling pathways that are detrimental for neuronal survival and nerve regeneration . In summary , p45 presents an example of a new antagonizing mechanism by which an interaction mediated by the TM and cytoplasmic domains is able to inhibit p75 disulfide dimer formation and function . Such knowledge provides a new glimpse into our understanding of the multiple distinct activities and signaling capabilities of p75 .
In this study , animal protocols were approved by the Institutional Animal Care and Use Committee ( IACUC ) of the Salk Institute and the Council on Accreditation of the Association for Assessment and Accreditation of Laboratory Animal Care ( AAALAC International ) . Euthanasia will be performed by methods specified and approved by the IACUC Panel on Euthanasia . The following antibodies were used in the present studies . Anti-p75 antibodies ( anti-p75ECD epitope: 9651 , a gift from Dr . Moses Chao , New York University; anti-p75ICD epitope: Buster , a gift from Dr . Phil Barker , Montreal Neurological Institute ) , anti-p45 antibodies ( anti-p45ECD epitope , 6750 and antip45ICD epitope , 6655 were generated in house ) , and anti-V5 antibody were purchased from Life Technologies ( R960-25 ) ; anti-Falg M2 antibody was purchased from Sigma ( F1804 ) ; and anti-HA antibodies were purchased from Roche ( clone 3F10 , 11867423001 ) and Santa Cruz Biotechnology ( sc-805 ) . For the recombinant expression of p45ICD , Escherichia coli BL21 ( DE3 ) freshly transformed with the pGST-p45 expression vector , which encodes a N-terminal GST tag and a thrombin cleavage site , was used . Two liters of M9 minimal medium containing either ( NH4 ) 2SO4 or ( 15NH4 ) 2SO4 and either glucose or 13C-glucose as the sole nitrogen and carbon sources were inoculated in the presence of ampicillin ( 100 µg/ml ) with 30 ml of preculture of pGST-p45–containing E . coli BL21 ( DE3 ) cells , which had been grown at 37 °C overnight . At OD600 = 0 . 6 , expression was induced with 1 mM isopropyl L-D-galactopyranoside . The cells were harvested after 4 h , and the pellet was resuspended in 30 ml of PBS . After sonication , the cell lysate was centrifuged at 20 , 000 g for 30 min , and the supernatant was incubated with Glutathione Sepharose ( Amersham ) in PBS with DTT 1 mM at 4 °C for 1 h . After several washes with PBS , the GST fusion protein was removed by thrombin digestion overnight at room temperature while still bound to the sepharose resin . The digest was incubated with benzamidine resin to remove the thrombin ( Sigma ) . The supernatant contained free p45ICD with 95% purity according to SDS-PAGE . For the recombinant expression of human p75ICD , E . coli BL21 ( DE3 ) freshly transformed with the pHisp75 expression vector , which encodes a N-terminal 22-aa affinity tag containing a six-histidine sequence and a thrombin cleavage site , was used . Two liters of unlabeled or stable-isotope–labeled minimal medium ( see above ) containing kanamycin ( 50 µg/ml ) were inoculated with 30 ml of preculture of pHisp75-containing E . coli BL21 ( DE3 ) cells that had been grown at 37 °C overnight . At OD600 = 0 . 6 , expression was induced with 1 mM isopropyl L-D-galactopyranoside . The cells were harvested after 4 h , and the pellet was resuspended in 30 ml buffer A ( 25 mM Tris-HCl , pH 8 . 0/500 mM NaCl/10 mM β-mercaptoethanol ) . After sonication , the cell lysate was centrifuged at 20 , 000 g for 30 min , and the supernatant was applied to a Ni2+-charged NTA column ( Qiagen , Chatsworth , CA ) . The fusion protein was eluted with a stepwise gradient of 0–500 mM imidazole in buffer A . After dialysis against PBS ( pH 8 . 0 ) , the N-terminal fusion tail was removed by thrombin cleavage performed as described above . The supernatant contained free p75ICD with 95% purity according to SDS-PAGE . The mutant p75 constructs were expressed and purified accordingly . The protein constructs were concentrated using a 10 kDa centriprep amicon concentrator . The concentrations of all proteins used in this study were determined from their absorbance at 280 nm by using molar extinction coefficients calculated from the Expasy protein server software ( http://www . expasy . ch ) . If not stated otherwise , all biophysical experiments were measured in PBS pH 8 , 100 mM NaCl . Ni-NTA purified p75ICD and GST-sepharose purified p45ICD were loaded on an S200-Superdex gel filtration column at 4 °C and eluted isocratically in PBS pH 8 . 0 buffer at 0 . 7 ml/min . Sedimentation equilibrium measurements of samples of p75ICD and p45ICD were conducted at concentrations of 10 µM , 300 µM , and 700 µM . Data were collected at four different speeds ( 10 , 000 , 14 , 000 , 20 , 000 , and 28 , 000 rpm ) . The NMR experiments were carried out on a Bruker DRX700 spectrometer at 25 °C by using protein solutions that contained PBS pH 8 . 0 , 100 mM NaCl , 1 mM sodium azide , and 95%/5% H2O/D2O . Sequential assignment and structure determination was performed with the standard protocol for 13C , 15N-labeled proteins [62] . Hence , sequential assignments of backbone resonances of 15N , 13C-labeled p45ICD and 15N , 13C-labeled p75ICD were obtained from HNCAcodedCB , HNCAcodedCO [63] . HNCA [42] and 15N-resolved [64] NOESY [65] spectra . The side chain signals of p45ICD were assigned from HCCH-COSY [66] and 13C-resolved [64] NOESY experiments . Aromatic side chain assignments were obtained with 2D [64] NOESY in D2O [64] . Distance constraints for the calculation of the 3D structure were derived from 3D 13C- , 15N-resolved [64] NOESY and 2D [64] NOESY spectra recorded with a mixing time of 80 ms . [42]-TROSY [43] spectra with parameters as described below were measured for p75ICD mutants . The data were analyzed using the CARA software program ( www . nmr . ch ) . For the chemical shift perturbation experiments , [42]-TROSY spectra [43] of stable isotope-labeled p45ICD or p75ICD were measured with t1 , max = 88 ms , t2 , max = 98 ms , and a data size of 200×1 , 024 complex points . [42]-TROSY experiments of 13C , 15N-labeled p45ICD were performed at protein concentrations of 0 . 1 mM and 0 . 5 mM in order to study the oligomerization state of p45ICD . [42]-TROSY experiments of 15N-labeled p75ICD were performed at protein concentrations of 10 mM , 0 . 1 mM , 0 . 2 mM , 0 . 5 mM , and 2 mM to study the oligomerization state of p75ICD and to elucidate the homodimer interface . [42]-TROSY experiments of 15N-labeled p75ICD were performed at a protein concentration of 10 mM free and in the presence of 0 . 1 mM unlabeled p45ICD to study the p45ICD–p75ICD heterodimer interface . [42]-TROSY experiments of 15N-labeled p75ICD were performed at a protein concentration of 0 . 5 mM at a 1:0 mixture of 15N-labeled p75ICD and unlabeled p45ICD followed by stepwise addition of unlabeled p45ICD up to a p45ICD concentration of 2 mM ( protein ratios were 1:0 , 1:0 . 5 , 1:1 , 1:2 , and 1:4 ) . The same NMR setups were used in titration experiments performed to investigate the binding site of p75ICD on p45ICD . The titration experiments were started with a 1:0 mixture of 13C , 15N-labeled p45ICD , and unlabeled p75ICD was added stepwise from 0 mM to 2 mM ( protein ratios were 1:0 . 5 , 1:1 , 1:2 , and 1:4 ) . We observed 2 , 130 NOEs in the NOESY spectra , leading to 1 , 065 meaningful distance restraints and 372 angle restraints ( Table S2 ) . For the structure calculation , the program CYANA was used [67] , [68] , followed by restrained energy minimization using the program INSIGHT . CYANA initially generated 100 conformers , and the 20 conformers with the lowest energy were used to represent the three-dimensional NMR structure . The 20 refined conformers showed small residual constraint violations that are compatible with the observed NOEs and the short interatomic distances ( Table S2 ) . Similar energy values were obtained for all 20 conformers . The quality of the structures is reflected by the RMSD values of 0 . 65 Å relative to the mean coordinates of p45 residues 141–218 ( see Table S2 and Figure 5A ) . The bundle of 20 conformers representing the NMR structure is deposited in the PDB database under accession no . 2IB1 . Constructs containing NgR , full-length p45 or p75 , as well as deletion and amino-acid point mutants were transfected into HEK293 cells by TransFectin transfection reagents ( BioRad ) . Transfected cells were collected and lysed in RIPA buffer ( 150 mM NaCl , 1% NP-40 , 0 . 5% DOC , 0 . 1% SDS , 50 mM Tris pH 8 . 0 ) . Lysates were immunoprecipitated with antibodies described in the text . Samples were analyzed using SDS-PAGE and Western blots . For quantitative analysis , gel images were analyzed by Image J program ( NIH ) . Statistical analyses and data graph were done using Prism software . Full-length p45 RNA with capping at the 5′ end and poly ( A ) sequences was transcribed in vitro using the mMessage-mMACHINE kit ( Ambion ) . RNA was transfected into CGNs with the TransMessenger transfection reagent ( Qiagen ) . We found that this protocol achieves 50%–70% transfection efficiency . Neurite outgrowth assay using CGNs was carried out as previously described [14] . CGN culture was performed as previously described [13] , [14] . Briefly , p5–p7 cerebella were isolated from WT and Thy1-p45 mice . Neurons were plated on Poly-D-Lysine–coated six-well tissue culture dishes at the density of 5 million cells per well . Cells are allowed to grow for 48 h and then starved in basal medium eagle ( BME , Life Technologies ) for 7 h before being treated with preclustered MAG-Fc at the concentration of 2 µg/ml for 15 min . Preclustered human IgG was used as the control . The preclustering is achieved by incubating the MAG-Fc with an anti-human-Fc antibody at a 2:1 molar ratio in BME for 30 min in 37 °C . Cells are then lysed on ice according to the manufacturer's suggestion , and the RhoA assay was performed following the instructions of the G-LISA ( absorbance based ) kit . The RhoA activity is measured by reading the 490 nm absorbance using a 96-well plate reader . We have also used the Millipore ( Upstate ) RhoA assay kit for a pull-down of activated RhoA by Western blotting analysis .
|
Injuries to the brain and spinal cord often result in paralysis due to the fact that the injured nerves cannot regrow to reach their normal targets and carry out their functions . At the injury sites , there are proteins released from the damaged myelin that bind the Nogo receptor ( NgR ) on the nerve and inhibit its regeneration . The NgR needs to form a complex with the p75 neurotrophin receptor in order to mediate this inhibitory signal . Here we found that p45 , a homologue of p75 , can also bind to p75 and block its inhibitory activity when overexpressed . To perform its function , p75 needs to dimerize through both its transmembrane and intracellular domains , facilitating the recruitment of several proteins . Our structural and functional studies show that p45 binds specifically to conserved regions in the p75 transmembrane and the intracellular domain and that this blocks p75 dimerization along with its downstream signaling . Thus , this study demonstrates that altering the oligomerization of p75 is a good strategy to override p75's inhibitory effects on nerve regeneration , and it opens the door for the design of specific p75 inhibitors for therapeutic applications .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biology",
"and",
"life",
"sciences"
] |
2014
|
Heterodimerization of p45–p75 Modulates p75 Signaling: Structural Basis and Mechanism of Action
|
In January 2011 , human cases with hemorrhagic manifestations in the hospital staff were reported from a tertiary care hospital in Ahmadabad , India . This paper reports a detailed epidemiological investigation of nosocomial outbreak from the affected area of Ahmadabad , Gujarat , India . Samples from 3 suspected cases , 83 contacts , Hyalomma ticks and livestock were screened for Crimean-Congo hemorrhagic fever ( CCHF ) virus by qRT-PCR of which samples of two medical professionals ( case C and E ) and the husband of the index case ( case D ) were positive for CCHFV . The sensitivity and specificity of indigenous developed IgM ELISA to screen CCHFV specific antibodies in human serum was 75 . 0% and 97 . 5% respectively as compared to commercial kit . About 17 . 0% domestic animals from Kolat , Ahmadabad were positive for IgG antibodies while only two cattle and a goat showed positivity by qRT-PCR . Surprisingly , 43 . 0% domestic animals ( Buffalo , cattle , sheep and goat ) showed IgG antibodies in the adjoining village Jivanpara but only one of the buffalo was positive for CCHFV . The Hyalomma anatolicum anatolicum ticks were positive in PCR and virus isolation . CCHFV was isolated from the blood sample of case C , E in Vero E-6 cells and Swiss albino mice . In partial nucleocapsid gene phylogeny from CCHFV positive human samples of the years 2010 and 2011 , livestock and ticks showed this virus was similar to Tajikistan ( strain TAJ/H08966 ) , which belongs in the Asian/middle east genetic lineage IV . The likely source of CCHFV was identified as virus infected Hyalomma ticks and livestock at the rural village residence of the primary case ( case A ) . In addition , retrospective sample analysis revealed the existence of CCHFV in Gujarat and Rajasthan states before this outbreak . An indigenous developed IgM ELISA kit will be of great use for screening this virus in India .
Crimean-Congo hemorrhagic fever ( CCHF ) is a severe acute febrile illness caused by the CCHF virus ( CCHFV , family Bunyaviridae , genus Nairovirus ) , with overall case fatality of 9–50% [1] . CCHF was first recognized in the Crimean peninsula in the mid-1940s , [2] , but the virus was first isolated from a patient in Kisangani , Democratic Republic of Congo , in 1956 [3] . Person-to-person transmission of CCHFV occurs through direct exposure to blood or other secretions , and instances of nosocomial transmission are well-documented [1] . The virus is maintained in nature predominantly in the Ixodid tick vectors , particularly ticks of the genus Hyalomma [4] , [5] . CCHFV can persist in the tick throughout its life stages by transtadial transmission , and can be passed onto the offspring by transovarial transmission [5] . Among domestic animals , cattle , sheep , and goat play an important role in the natural cycle of the virus [6] . In these animals , CCHFV replicates to high titres in the lung , liver , spleen , and reticuloendothelial system in other organs [7] , but generally causes only subclinical disease . In contrast , human infections often result in severe hemorrhagic fever ( HF ) , with high levels of viral replication occurring in all major organs , including the liver [8] . In recent years , a number of zoonotic viral diseases have emerged in Southeast Asia as Nipah virus , and CCHFV [9] , [10] , [11] , [12] . CCHF was recently confirmed for the first time in India , although the virus had been identified nearby in Pakistan and western China [11] , [12] . In January 2011 , human cases with hemorrhagic manifestations in the hospital staff were reported from a tertiary care hospital in Ahmadabad , Gujarat . The clinical samples of three hospitalized patients were referred to National Institute of Virology ( NIV ) , Pune and laboratory investigations confirmed as CCHFV [13] . Here , we report detection and isolation of CCHFV associated with that nosocomial outbreak in Gujarat , India , and the presence of the virus in livestock and ticks in this region . The disease is newly recognized in India . It is required to create awareness about this disease in public health workers and physicians . CCHFV symptoms are difficult to distinguish not only from other HF , but that the real challenge is to distinguish the signs and symptoms from other , more common , febrile diseases .
NIV , Pune is responsible for investigation of viral disease outbreaks of human including Zoonosis in India . As per the mandate of our institute , collection of the clinical samples from different species of animals for viral isolation and detection is required . This institute has policy to take approval on the projects which involves animals from the national committee called “Committee for the Purpose of Control & Supervision of Experiments on Animals ( CPCSEA ) under the Ministry of Environment and Forests , Government of India . Our study project No . HCL01/NIV15/2010 is approved by the Institutional Animal Ethical Committee ( IAEC ) permitting the use of infant and adult mice as laboratory animals for isolation of virus and development of antibodies respectively . The present nosocomial CCHF outbreak was informed to Institutional Human Ethical Committee , NIV Pune . As per policy of institute the work conducted during epidemic situations is exempt from prior approval from the IEC . Due to this policy , this study was not pre-approved , but the Committee was notified after the outbreak . All study participants provided informed consent . Informed consent was in written format , both in English and local language ( Guajarati ) . All the record analyzed was anonymized . Every sample was registered in the central registry of the institute and allotted a NIV number , which was used throughout the study . After the request from the Gujarat Government Public Health authorities , a team from the NIV , Pune visited the private tertiary care hospital in Ahmadabad , along with local public health authorities . Medical records of hospitalised patients presented with HF manifestations were examined , and their family members or caretakers were interviewed . Blood and urine samples were collected from cases with HF manifestations and their contacts . All clinical samples were transported in cold chain to NIV , Pune . To ascertain whether additional CCHFV infections had occurred in contacts ( hospital staff or family members ) lacking CCHF infection symptoms , an additional 86 blood samples including CCHF suspected cases were tested by Vector-Best and indigenously developed ELISA for detection of IgM antibodies against CCHFV . Earlier , there were two serum samples referred to NIV , Pune from Rajkot , Gujarat , from patients suspected for Hantaan and Nipah virus in the month of February 2010 . These samples were from the patient and consulting physician . Both had succumbed to the unknown infection . These samples were also included in the present study . Blood samples and ticks were collected from the buffalo , cattle , sheep and goat from the vicinity of the index case ( case A ) , as well as from Kolat and surrounding villages , Jivanpura , Navapura and Changodar . A total of 138 ticks were collected from domestic animals from Kolat , were classified and pooled as per species . Eight tick pools were made and homogenized in MEM media . Tick homogenates were suspended in lysis buffer for viral RNA extraction . These were tested for CCHFV by qRT-PCR and nested RT-PCR . Hyalomma tick homogenates were also used for virus isolation in Vero E-6 cells and Swiss albino mice . Rodents ( n = 90 ) were also trapped from Kolat villege , morphologically identified and only blood samples from these animals were collected and transported to NIV , Pune . RNA was extracted from human ( serum and urine ) , and animal serum samples using Qiagen ( Valencia , CA , USA ) RNA extraction kit . Tick pools were homogenized in Minutesimum Essential Medium ( MEM ) . This homogenate was used for RNA extraction and for virus isolation . In the initial screening CCHFV-specific TaqMan based qRT-PCR was carried out on the RNA as previously described [14] . RT-PCR was performed with the SuperScript One-Step RT-PCR kit with Platinum Taq ( Invitrogen Corp . , Carlsbad , CA , USA ) . Two sets of primers were used for initial RT-PCR . The primer set CCHF-F2 ( TGG ACA CCT TCA CAA ACT C ) and CCHF-R3 ( GAC AAA TTC CCT GCA CCA ) amplified a 530 nt region of the nucleocapsid ( N ) gene of CCHFV , while nested PCR using the primers CCHF-F3 ( GAA TGT GCA TGG GTT AGC TC ) and CCHF-R2 ( GAC ATC ACA ATT TCA CCA GG ) amplified a 226 nt region [15] . PCR products were analyzed on 2% agarose gel electrophoresis and Ethidium bromide straining . Cyclic sequencing was carried out at PCR condition 96°C - 1 minute , 96°C - 10 sec , 45°C - 5 sec and 60°C - 4 minutes for 25 cycles using ABI Big-Dye 3 . 1 dye chemistry ( Applied Biosystems , Foster City , CA ) . These products were purified using Dyex 2 . 0 kit ( Qiagen ) according manufacturer's instructions and sequencing was performed using the ABI 3100 automated DNA sequencer . The sequences obtained were curated using KODON software for both the reads from both the ends . The curated sequences were aligned using program Clustal W and phylogenetic tree was constructed using neighbour joining algorithm with 500-bootstrap replicates as implemented in Mega v 4 . 0 software [16] . The tick pools were first tested by qRT-PCR . Homogenates of CCHFV-positive tick pools were inoculated into Swiss albino mice via intracerebral and intraperitoneal routes and into Vero E6 cells for virus isolation . Virus isolation was attempted from the CCHF positive human blood , serum , and urine samples . Two CCHF IgM ELISA kits were used; a ) commercial kit , b ) indigenously developed test for detection of IgM antibodies in the human serum samples . The inactivated animal serum samples were tested for evidence of anti-CCHFV IgG using an ELISA kit provided by CDC , Atlanta . The protocol followed was , ELISA plates were coated with anti-CCHFV hyper immune mouse ascetic fluid ( HMAF ) ( dilution 1∶1000 ) in phosphate buffer saline pH 7 . 4 overnight at 4°C . BPL inactivated CCHFV infected Vero E6 cell lysate antigen ( 1∶20 diluted , 100 µl/well ) was added as a positive antigen , normal Vero E6 cell lysate was used as negative antigen and incubated for one hr at 37°C . One hundred µl of 1∶100 diluted serum samples were added and incubated for one hr at 37°C . These wells were washed and anti-sheep IgG HRP conjugate ( 1∶4000 diluted , 100 µl ) was added and incubated for one hr at 37°C . ABTS substrate was added and incubated at 37°C for 30 minutes . The reaction was stopped by adding 1% SDS and plates were read at 414 nm . The plates were washed five times using 10 mM PBS pH 7 . 4 with 0 . 1% Tween-20 ( Sigma , USA ) at the end of each step . Appropriate controls were included in the test . During investigation , ELISA was also performed to detect CCHFV specific IgG antibodies in the animals from the residential area and surrounding villages of index case ( case A ) .
In December 2010 and January 2011 , a cluster of viral hemorrhagic fever ( VHF ) cases was identified in Ahmadabad , Gujarat , India , which was declared as a nosocomial outbreak of Crimean-Congo hemorrhagic fever [13] . The initial case identified was a 25 year old nurse ( case C ) who worked in a hospital in Ahmadabad , and presented on January 13th , 2011 with a three days history of an acute febrile illness characterized by fever , chills , vomiting and headache followed by hemorrhagic symptoms [13] . Her condition rapidly deteriorated with onset of delirium , multiple hemorrhagic symptoms ( palatal petechia , coughing up of blood , bleeding from lips , vaginal bleeding , hematuria , hematemesis , and pulmonary hemorrhage ) . She was placed in isolation on January 16th and given oral ribavirin based on suspicion of VHF . Despite treatment , she died on January 18th from multi-organ failure and disseminated intravascular coagulation ( Table 1 ) . Based on the suspicion of VHF , CCHFV was added to the diagnostic testing and the patient sera tested positive by qRT-PCR and RT-PCR in urine and serum [13] . Epidemiologic investigation revealed that case C had earlier provided care to a patient ( case A ) with similar VHF symptoms . Probable case A was a 32 year old housewife from Kolat village , approximately 20 km outside Ahmadabad . She had been admitted to the hospital on December 27th , 2010 and had died on December 31st with VHF-like symptoms ( Table 1 ) . While no specimens remained from case A to allow confirmatory testing for the presence of evidence of CCHF infection , case A was strongly suspected to be the source of nosocomial infection of the attending nurse ( case C ) given the similarity in clinical and laboratory findings ( Table 2 ) . Further investigation revealed a similar VHF-like illness in a 42 year old physician ( probable case B ) , who had treated probable case A and had subsequently presented with symptoms on January 6th 2011 ( Table 1 ) and died on January 13th . Virus transmission to two further contacts was documented . The husband ( case D ) of probable probable case A was admitted to the hospital on January 16th , displayed similar VHF-like symptoms , was treated with oral ribavirin , and recovered and was released on January 26th . Laboratory testing confirmed evidence of CCHF virus infection . The last identified case was a 25 year old doctor ( case E ) who had contact in the hospital with probable case B and case D . Given his onset of illness on January 26th , case E was likely exposed to case D , who had been admitted on January 16th and released on the 26th ( and case B had died on January 13th ) . Despite oral ribavirin treatment , case E progressed to multi-organ failure and died on January 31st [13] . In total , 3/5 of these hospitalized individuals were confirmed by qRT-PCR , nested RT-PCR & IgM ELISA as having been infected with CCHFV . The patients initially reported fever , headache , myalgia and vomiting . Death occurred within 5–9 days POD in 4 of the 5 patients . Analysis showed that all the probable cases presented with similar laboratory data as laboratory confirmed CCHF cases , including leukopenia , thrombocytopenia , increased SGOT and SGPT and serum LDH , and increased ferritin and prothrombin time ( Table 2 ) . None of blood samples were found positive by qRT-PCR from the family and hospital contacts . With the finding of CCHFV associated with acute hemorrhagic diseases in Gujarat , retrospective analysis was carried out on some samples collected in February , 2010 from patients that showed similar case descriptions and laboratory findings as reported here for these Ahmadabad cases . Acute blood samples from two patients from Rajkot , a town approximately 200 km west of Ahmadabad , were found positive for CCHFV by qRT-PCR and RT-PCR , retrospectively confirming them to be CCHF cases . Out of 86 samples one hospital and one family contact ( both asymptomatic ) was positive for presence of IgM antibodies against CCHFV . Case C and E had a very high titter of IgM in the serum samples while case D was negative in the early sample which was positive for qRT-PCR . The 2nd sample of case D was positive for CCHFV specific IgM and also qRT-PCR . The indigenously developed ELISA was compared with commercial Vector-Best assay . The sensitivity and specificity of indigenous ELISA was 75 . 0% and 97 . 5% respectively using Vector-Best assay as gold standard ( Table 3 ) . The serum samples of case C , D , and E were found positive for IgM antibodies while screening with both the kits . As positive samples were only four , further standardization of indigenously developed assay using more human serum samples is required . The blood and urine samples collected from the laboratory confirmed CCHF cases ( case C , D and E ) were inoculated into Vero E6 cells and Swiss albino mice [intra cerebral . and intra peritonial . routes] . Sickness and death was observed in mice on 9th post infection day ( PID ) . Cytopathic effect was observed in Vero E 6 cells at 3rd PID and virus was harvested on 7th PID . CCHFV was isolated from the blood samples in both mice and Vero E6 cells and sequence analysis confirmed that the PCR products were derived from CCHFV RNA . No virus isolate was obtained from the urine samples; the Ct values ( 34 ) of the qRT-PCR also suggested low viral titres in the urine . The initial virus source for this person to person chain of transmission appeared to be probable case A . She and her husband ( case D ) were reside in the rural village of Kolat , made a living by farming and have daily exposure to livestock and potentially ticks on these animals ( Figure 1 ) . In November and December 2010 , just prior to this outbreak , a serosurvey was performed to examine livestock for evidence of CCHFV using blood samples collected from slaughterhouses in the northern adjoining state of Rajasthan and some more distant areas from Maharashtra and West Bengal states . The serum samples from buffalo , goat and sheep from Sirohi district , in southern Rajasthan were positive for IgG antibodies against CCHFV . Serum samples from northern West Bengal or Pune area of Maharashtra were negative ( Table 4 ) . The area from which positivity was reported was approximately 200 kilometres north of Ahmadabad ( Figure 1 ) . Evidence of CCHFV infection ( IgG positive ) was also found in a follow-up study of livestock [including buffalo , cattle , goat , and sheep] from Kolat ( residence of case “A” and case “D” ) and the surrounding villages of Changodar , Jivanpara and Navapura ( Table 5 ) . Overall IgG antibodies positivity in the small sample sizes varied between villages , ( 10–43% ) ( Table 5 ) . qRT-PCR analysis showed evidence of CCHFV infection in the blood samples from three animals [two cattle and a goat] indicating circulation of the virus in these animal populations ( data not shown ) which was further confirmed by CCHFV specific nested RT-PCR . Rodent samples from Rattus rattus rufescens ( 72 ) , Suncus murinus ( 6 ) , Mus booduga ( 3 ) , Mus musculus ( 8 ) and Rattus norvegicus ( 1 ) were tested by qRT-PCR for detection of CCHFV . All the 90 blood samples were negative for the presence CCHFV . Out of eight tick pools collected from the livestock around the Kolat village two pools , both Hyalomma anatolicum were positive by qRT-PCR and RT-PCR specific to CCHFV ( Table 6 ) . These two Hyalomma tick pools also showed CPE in Vero E-6 cells and propagated in Swiss albino mice also . CCHFV specific nested RT-PCR products were generated from the blood samples from CCHF cases C , D and E of this outbreak and two retrospective serum samples of year 2010 . Along with this , sequences of CCHFV positive two tick pools , and four livestock samples collected in Kolat village were also included for the analysis . Sequencing analysis of the PCR products generated a 226 nt partial fragment of the virus S gene . Comparison of this virus sequence fragment with that of other CCHFVs showed that the strain detected in Ahmadabad had maximum nucleotide identity ( 98 . 0% ) with Tajikistan strain ( TAJ/H08966 ) of CCHFV which belongs in the Asian/middle east genetic lineage IV ( Figure 2 ) .
India did not report any CCHF cases until January 2011 [13] . However , the presence of CCHFV in India had been suspected , since it was detected in the neighbouring countries of Pakistan and western China , especially once CCHFV was first isolated from the tick species Hyalomma anatolicum and from a mixture of Hyalomma and Boophilus tick species collected in Pakistan [17] , [18] . In the past a serosurvey was conducted in Jammu and Kashmir , the western border districts of India ( 1976 ) , showed CCHFV antibodies in many of the animal sera , however , studies by Rodrigues et . al . , ( 1986 ) precipitated the antibody in 34 of the 655 domestic animal sera from different states/territories of southern India , and from Maharashtra state . Most of the positive sera were collected from goat in southern India [19] . CCHFV IgG antibodies were detected in animals in November and December 2010 , in the neighbouring Rajasthan state in buffalo , cattle , goat , and sheep . The border between India and Pakistan is porous for the entry of livestock animals , which might play an important role in transmission of CCHFV between these countries . Our investigation identified CCHFV in cattle , goat , and buffalo from surrounding villages also . These villages have a large buffalo population . Since this virus is isolated from the ticks collected from buffalo in the affected area , these animals may have an important role in infecting large number of ticks in these areas . The pool of male Hyalomma anatolicum ticks was positive for CCHFV , suggesting that this disease is not due to recent introduction of CCHFV to this area . The virus strain studied here showed highest homology ( 98–100 . 0% ) with the Tajikistan strain ( TAJ/H08966 ) , suggesting Asian/middle east origin ( IV clade ) . Among the CCHF confirmed cases , serum ferritin levels were high; in the fatal cases , such high ferritin levels have also been reported as indicator of severity of the disease [20] . Leukopenia , thrombocytopenia , increased SGOT , SGPT , LDH , and increased prothrombin time was in accordance to the World Health Organization ( WHO ) case definition of CCHF . The increased serum creatine phosphokinase is perhaps indicative of severe damage to the liver and other organs . Swiss albino mice inoculated with patient serum also showed intra-abdominal bleeding , multifocal lung necrosis , liver enlargement , and necrosis . Interestingly , blood smear examination reports from all patients showed erythophagocytosis , and investigations of mice inoculated with patient blood samples revealed similar findings . Case-D , who recovered 10 post onset days ( POD ) and was monitored for the presence of virus in the urine , and found positive for CCHF up to 13 days POD . This raises the question whether a recovered patient discharged before 13–14 days POD can still be a potential source of infection . The short incubation period and many of the nonspecific symptoms of CCHF , which overlap with the symptoms of other hemorrhagic fevers , raise the risk for misdiagnosis and person-to-person transmission of CCHFV . This puts close contacts , and healthcare providers at risk for secondary infection . The symptoms , signs , and laboratory abnormalities of CCHF are nonspecific and can overlap with those of other tropical infections ( dengue , Kyasanur Forest disease and leptospirosis ) . The efficacy of ribavirin is still debated , and it is difficult to say whether the only surviving CCHF case ( husband of index case ) in the present study was due to immediate ribavirin oral therapy . These features highlight the importance of quick diagnosis of the etiologic agent associated with HF cases . Moreover , once CCHF is suspected , patients and specimens should be handled with adequate biosafety measures . Authorities and contacts should be immediately notified , and patients' samples should be sent for specific laboratory diagnosis , keeping in view the precautions associated with BSL 4 risk group agent hazards . Nosocomial transmission in this outbreak as reported here and earlier by Mishra et . al . , ( 2011 ) is evident by the development of disease in close contacts of a CCHF patient [13] . This episode of nosocomial infection resulted in four deaths in a small population as of February 15 , 2011 . There are very few commercially available kits for detection of CCHFV antibodies from humans . Therefore , an attempt was made to develop indigenous assays for detection of CCHFV antibodies . The performance of indigenous assay was satisfactory as compared to commercially available Vector-Best assay . Since , the positive samples were only four , further standardization of indigenously developed assay using more human serum samples is required . Detection of IgM antibodies in one each hospital and family contacts suggest a low level of asymptomatic cases . The CCHF positivity of the two retrospective human samples from Rajkot , Gujarat also indicate that disease was prevalent in this state for a long time and probably might have contributed significant morbidity and mortality in the past in this state .
|
A nosocomial outbreak of CCHFV occurred in January 2011 , in a tertiary care hospital in Ahmadabad , Gujarat State in western India . Out of a total five cases reported , contact transmission occurred to three treating medical professionals , all of whom succumbed to the disease . The only survivor was the husband of the index case . These results highlight the importance of considering CCHFV as a potential aetiology for Hemorrhagic fever ( HF ) cases in India . This also underlines the need for strict barrier nursing and patient isolation while managing these patients . During the investigation presence of CCHFV RNA in Hyalomma anatolicum ticks and livestock were detected in the village from where the primary case ( case A ) was reported . Further retrospective investigation confirmed two CCHF human cases in Rajkot village 20 kilometres to the west of Ahmadabad in 2010 , and CCHFV presence in the livestock 200 kilometres to the north in the neighbouring State Rajasthan . This report shows the presence of CCHFV in human , ticks and animals in Gujarat , India . The fact of concern is the spread of this disease from one state to another due to trading of livestock .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"biology"
] |
2012
|
Detection, Isolation and Confirmation of Crimean-Congo Hemorrhagic Fever Virus in Human, Ticks and Animals in Ahmadabad, India, 2010–2011
|
An important step in understanding gene regulation is to identify the DNA binding sites recognized by each transcription factor ( TF ) . Conventional approaches to prediction of TF binding sites involve the definition of consensus sequences or position-specific weight matrices and rely on statistical analysis of DNA sequences of known binding sites . Here , we present a method called SiteSleuth in which DNA structure prediction , computational chemistry , and machine learning are applied to develop models for TF binding sites . In this approach , binary classifiers are trained to discriminate between true and false binding sites based on the sequence-specific chemical and structural features of DNA . These features are determined via molecular dynamics calculations in which we consider each base in different local neighborhoods . For each of 54 TFs in Escherichia coli , for which at least five DNA binding sites are documented in RegulonDB , the TF binding sites and portions of the non-coding genome sequence are mapped to feature vectors and used in training . According to cross-validation analysis and a comparison of computational predictions against ChIP-chip data available for the TF Fis , SiteSleuth outperforms three conventional approaches: Match , MATRIX SEARCH , and the method of Berg and von Hippel . SiteSleuth also outperforms QPMEME , a method similar to SiteSleuth in that it involves a learning algorithm . The main advantage of SiteSleuth is a lower false positive rate .
An important step in characterizing the genetic regulatory network of a cell is to identify the DNA binding sites recognized by each transcription factor ( TF ) protein encoded in the genome . A TF typically activates and/or represses genes by associating with specific DNA sequences . Although other factors , such as metabolite binding partners and protein-protein interactions ( for example , between a TF and RNA polymerase or a second TF ) , can affect gene expression [1] , it is important to identify the sequences directly recognized by TFs to the best of our ability to understand which genes are controlled by which TFs . A better understanding of gene regulation , which plays a central role in cellular responses to environmental changes , is a key to manipulating cellular behavior for a variety of useful purposes , as in metabolic engineering applications [2] . A number of computational methods have been developed for predicting TF binding sites given a set of known binding sites [3]–[10] . Commonly used methods involve the definition of a consensus sequence or the construction of a position-specific weight matrix ( PWM ) , where DNA binding sites are represented as letter sequences from the alphabet {A , T , C , G} . More sophisticated approaches further constrain the set of potential binding sites for a given TF by considering , in addition to PWMs , the contribution of each nucleotide to the free energy of protein binding [3] and additional biologically relevant information , such as nucleotide correlation between different positions of a sequence [8] or sequence-specific binding energies [6] . Perhaps not as widely used as sequence analysis , the idea of employing structural data for predicting TF binding sites has been considered [11]–[15] . Most of these methods use protein-DNA structures rather than DNA by itself . Acquiring training sets large enough to be useful is problematic for even well-studied TFs , for which only small sets of known binding sites ( on the order of 10 sites ) are typically available [8] . New high-throughput technologies have been used to identify large numbers of binding sites for particular TFs [16]–[18] , but there remains a need for methods that predict TF binding sites given a small number of positive examples . Such methods can be used , for example , to complement analysis of high-throughput data . Binding sites detected by high-throughput in vitro methods , such as protein-binding microarrays [16] , can be compared with predicted binding sites to prioritize studies aimed at confirming the importance of sites in regulating gene expression in vivo . The fine three-dimensional ( 3D ) structure of DNA is sequence dependent and TF-DNA interactions depend on various physicochemical parameters , such as contacts between nucleotides and amino acid residues and base pair geometry [19] . These parameters are not accounted for by conventional methods for predicting TF binding sites , which rely on sequence information alone . Letter representations of DNA sequences do not capture the biophysics underlying TF-DNA interactions . Given that a TF does not read off letters from a DNA sequence , but interacts with a particular sequence because of its chemical and structural features , we hypothesized that better predictions of TF binding sites might be generated by explicitly accounting for these features in an algorithm for predicting TF binding sites . The mechanisms by which TFs recognize DNA sequences can be divided into two classes: indirect readout and direct readout [19] . For indirect readout , a TF recognizes a DNA sequence via the conformation of the sequence , which is determined by the local geometry of base pair steps , the distortion flexibility of the DNA sequence , and ( water-mediated ) protein-DNA interactions [20] , [21] . For direct readout , a TF recognizes a DNA sequence through direct contacts between specific bases of the sequence and amino acid residues of the TF [22] , [23] . These two classes of recognition mechanisms are not mutually exclusive . In this study , we introduce a method , SiteSleuth , for predicting TF binding sites on the basis of sequence-dependent structural and chemical features of short DNA sequences . By using molecular dynamics ( MD ) methods to calculate these features , we can map a set of known or potential binding sites for a given TF to vectors of structural and chemical features . We use features of positive and negative examples of TF binding sites to train a support vector machine ( SVM ) to discriminate between true and false binding sites . Negative examples are derived from randomly selected non-coding DNA sequences . Positive examples are taken from RegulonDB [24] , which collects information about TFs in Escherichia coli . Classifiers for E . coli TFs developed through the SiteSleuth approach are evaluated by cross validation , and the classifier for Fis is tested against chromatin immunoprecipitation ( ChIP ) -chip assays of Fis binding sites [17] . Combining ChIP with microarray technology , ChIP-chip assays provide information about DNA-protein binding in vivo on a genome-wide scale [25] . We also evaluate the performance of SiteSleuth against four other computational methods: the method of Berg and von Hippel ( BvH ) [3] , MATRIX SEARCH [5] , Match [7] , and QPMEME [6] . The BvH , MATRIX SEARCH , and Match methods rely on the PWM approach to capture TF preferences for binding sites . The QPMEME method is similar to SiteSleuth in that it employs a learning algorithm . In the case of Fis , we show that SiteSleuth generates significantly fewer estimated false positives and provides higher prediction accuracy than the other computational approaches .
Let us use X = {x1 , … , xN} to represent the set of training data , where xk ( k = 1 , … , N ) is a real-valued n-dimensional feature vector that characterizes the kth training example and n is the number of features considered . The features considered are described below . Given input xk and scalar output yk = {−1 , 1} , which identifies a training example as a positive or negative example of a binding site , classifier training produces an ( n−1 ) -dimensional hyperplane in the space of features that satisfies the equation wTx+d = 0 and a set of linear inequality constraints , each involving a slack variable . The parameters w and d and the slack variables ξk ( k = 1 , … , N ) are found by solving the minimization problem ( 1a ) subject to the following constraints ( 1b ) where C+ and C− are penalty parameters [27] . These parameters are introduced to balance the contributions of negative and positive training examples to the objective function ( Eq . 1a ) , as we typically have available many more negative examples than positive examples . The penalty parameters are determined for each TF via a grid search over ranges of C− and C+ values as part of a 3-fold cross-validation procedure for each classifier . In 3-fold cross validation , we randomly divide the training set into three subsets of roughly equal size . One subset is then used to test the accuracy of the classifier trained on the remaining two subsets until each subset has been used in testing . We used the F-measure to assess accuracy . The F-measure is the harmonic mean of precision ( p ) and recall ( r ) :Precision is the fraction of predicted binding sites that are actually binding sites and recall is the fraction of actual binding sites predicted to be binding sites:where TP , FP , and FN represent true positives , false positives and false negatives from 3-fold cross validation . To find values of C− and C+ that maximize the F-measure , we first performed a coarse grid search over the following grid points: C− = [2−5 , 2−3 , … , 215] and C+ = [2−5 , 2−3 , … , 215] . We then performed fine grid searches using progressively smaller grid spacing ( 2 , 20 . 5 , 20 . 125 , … ) around the best C− and C+ values found in the coarse grid search . Once trained , a classifier for a TF , taken to recognize binding sites of length L , is used for prediction as follows . The classifier is used to scan an organism's genome for binding sites of length L . Given a feature vector xm for a potential binding site m , we calculate the quantity wTxm+d . The decision function of the classifier is the sign of wTxm+d . Thus , if the sign of this quantity is positive , then site m is predicted to be a TF binding site . Conversely , a negative quantity indicates that m is not a binding site . This step is repeated for all non-coding sequences in the E . coli genome of length L . The length L was chosen for each TF based on information in RegulonDB [24] . Structural and chemical features of short DNA sequences were defined based on the predicted 3D structures of these DNA sequences , which were determined via MD simulations . MD simulations of solvated nucleic acids have been performed for almost three decades [28] , [29] . Simulations of DNA oligomers have been studied systematically and results have been discussed in multiple publications [30]–[32] . Our approach is similar to that used in Refs . [30]–[32] and is described below . Because the available experimental data are incomplete ( i . e . , structures are unavailable for all 4-mers , at least in the Nucleic Acid Database [33] ) and available structures have been determined under various experimental conditions ( e . g . , free or bound to protein ) , we used simulated structures rather than experimentally determined structures for determining structural and chemical features . Predicted structures were obtained for a common condition in a uniform manner . For comparison , we implemented four other computational TF binding site prediction methods: the method of Berg and von Hippel ( BvH ) [3] , Match [7] , MATRIX SEARCH [5] , and QPMEME [6] . These methods were implemented as described in the cited papers and , for the 54 TFs studied , a list of binding sites predicted by each method can be found online at http://cellsignaling . lanl . gov/EcoliTFs/SiteSleuth/ . For completeness , each method is briefly presented below . To discuss these methods we will need to first introduce a few quantities . For a set of N DNA binding sites of a particular TF , the length of each binding site is denoted by L . The value of L is set equal to the length of binding sites reported in RegulonDB for a given TF . In the case of Fis , we set L = 21 . We define to be the number of times base b appears in the jth position in the sequences of the binding sites , and to be the corresponding frequency . We denote as the overall background frequency of base b . We use S to denote a potential TF binding site of length L and we use Sj ( j = 1 , … , L ) to denote the jth base of sequence S . For the BvH method , we denoted the number of occurrences of the most common base in position j of the set of binding sites by . Using a training set of N binding sites , the BvH method calculates the score of each binding site as the summation over every position of the log-odds score of observing a base of S versus the most frequent base in the corresponding position of the sequence . Thus , the score is given byA pseudocount of 0 . 5 is used in the formula [3] . A cutoff threshold is defined as the mean score of the N positive training examples . To evaluate whether a new sequence S is a binding site , the score of S is calculated based on the above formula and compared with the cutoff threshold . If the score of sequence S is greater than the cutoff threshold , it is predicted to be a binding site . For the Match method , a set of N training examples is used to define an information vector , which describes the conservation of the position j in a binding site from the training set:The information vector is used to evaluate whether a new sequence S is a binding site or not by calculating a score defined asand min and max are calculated using the lowest and highest nucleotide frequency in each position , respectively . A cutoff threshold is defined as the mean score of the N positive training examples . If the score for a new sequence S is larger than the cutoff threshold , S is predicted to be a binding site . Using a set of N binding sites as training examples , the MATRIX SEARCH method calculates the score of each binding site S as the summation over every position of the log-odds score of observing a base in S versus the overall background frequency of that base in the corresponding position of the sequences . Thus , the score is given byA pseudocount of 0 . 01 is used in the formula [5] . A cutoff threshold is determined as the mean of the N scores calculated from the training data . A new sequence S is predicted to be a binding site if its score is greater than the cutoff threshold . The QPMEME ( Quadratic Programming Method of Energy Matrix Estimation ) method defines a weight for each base b at position j in S . The score for a sequence S is defined asThe weight is estimated via a learning algorithm that only uses positive examples . The learning algorithm minimizes the variance subject to the constraint that the score for each known binding site is less than a predefined cutoff value . Consistent with the Methods section of Djordjevic et al . [6] , we used −1 for the cutoff value in our implementation of QPMEME , which constrains all known binding sites to one side of a hyperplane . Mathematically , the learning algorithm is described byfor every S in the training data set .
To make a preliminary assessment of our hypothesis that we can produce better predictions if we consider the chemical and structural features of sequence-specific DNA , we examined the features of various sequences and found that the same base in the same position in a sequence can have different chemical and structural features depending on its environment . We illustrate this finding in Figure 3 , which shows sequence-specific DNA structures . From the structures , one can see the context-dependent variation in the twist angle between the center two base planes . The center base pair is the same in each structure , but the twist angle for the left structure of Figure 3A is −20 . 4° , whereas the twist angle for the right structure of Figure 3A is −4 . 3° . Figure 3A demonstrates that different local structural features may characterize the same nucleotide at the same position in a sequence . The feature vectors for TGG and AGA are given in Table S2 . Similarly , Figure 3B demonstrates that different nucleotides in the same position may be characterized by the same local structural features . The twist angles of the middle base pairs of the two structures in Figure 3B are the same , even though the base pairs are different . These observations suggested to us that chemical and structural features may capture sequence correlations relevant for TF-DNA interactions that are not apparent from sequence data alone and encouraged us to build classifiers that separate negative and positive examples of TF binding sites based on their positions in chemical and structural feature space . This approach , which we call the SiteSleuth method , combines DNA structure prediction , computational chemistry and machine learning . To demonstrate the reliability of MD simulations for prediction of structural features of DNA oligomers , we calculated the propeller feature using 1 ) available experimental structural data ( obtained from the Nucleic Acid Database [33] ) and 2 ) predicted structures obtained via MD simulations , and we found significant correlation ( about 0 . 8 ) . The results are shown in Figure S2 . As described in the Methods section , binary SiteSleuth classifiers were developed to identify and predict the binding sites of 54 TFs based on TF binding sites documented in RegulonDB . The input to a classifier is a vector of structural and chemical features generated from DNA sequences , each labeled as either a positive or negative example . Negative examples were taken from randomly chosen non-coding sequences of the E . coli genome . The classifiers were then used to scan both strands of non-coding sequences in the E . coli genome from 5′ to 3′ to identify potential TF binding sites . For comparison , we also considered four other computational TF binding site prediction methods: BvH [3] , MATRIX SEARCH [5] , Match [7] , and QPMEME [6] These methods are each briefly described in the Methods section . The accuracy of predictions of each method was evaluated through a 3-fold cross-validation procedure , described in the Methods section . For each method , the mean cross-validation score , V , for the 54 TFs considered are listed in Table S4 and classifier accuracy is summarized in Figure 4 . Recall that V is the fraction of positive examples predicted to be true binding sites in the cross-validation procedure . Figure 4 is a heat map showing the cross-validation score , , produced by each of the five computational methods . Brighter red indicates a higher cross-validation score and black represents . A cross-validation score of indicates perfect prediction , whereas a cross-validation score of zero indicates that the method fails to predict any TF binding sites correctly . Of the 54 TFs studied , SiteSleuth outperforms all the other methods in 28 cases , equals the next best method in 11 cases , and performs more poorly in 15 cases . Based on the number of times a method outperformed all the other methods , SiteSleuth ( 28 ) performed better than QPMEME ( 8 ) , which performed better than MATRIX SEARCH ( 2 ) , which equaled the performance of BvH ( 2 ) , which performed better than Match ( 0 ) . In one case , IcsR , SiteSleuth is the only method for which . The data used to construct Figure 4 are given in Table S4 . Interestingly , Figure 4 reveals that all methods give cross-validation scores of zero for several TFs: CysB , GcvA , OxyR , RcsAB , and Rob . This observation suggests that methods that rely on DNA sequence information , including SiteSleuth , are insufficiently equipped to predict the binding sites for these TFs . Some of these TFs , such as GcvA [42] , may perhaps recognize DNA indirectly via interaction with a second protein that recognizes DNA directly . Another explanation could be that some of these TFs , such as Rob [43] , may be recognizing very short sequences . The total number of TF binding sites predicted by each computational method is given in Table S3 . For most TFs , QPMEME and Match both predict a large number of TF binding sites in the E . coli genome . The BvH and MATRIX SEARCH methods predict fewer binding sites , but still more than the number of predictions generated by SiteSleuth . In Figure 5 , we show the performance of SiteSleuth relative to that of BvH for the TFs with five or more known binding sites . The relative performance ( RP ) score shown in Figure 5 is defined as the number of TF binding sites predicted by BvH divided by the number of TF binding sites predicted by SiteSleuth . This score indicates how many times more TF binding sites are predicted by BvH than by SiteSleuth . For example , BvH predicts 23 times more TF binding sites for MetJ than does SiteSleuth . For reference , the log transformed number of TF binding sites predicted by SiteSleuth is also indicated in Figure 5 and a solid line is drawn at RP = 1 . As can be seen in Figure 5 , 41 TFs have RP>1 and 13 TFs have RP<1 . Thus , there is a large class of TFs for which SiteSleuth predicts fewer binding sites than BvH ( RP>1 ) and , by extension , the other computational methods . From these results alone , it is not clear whether fewer predictions are a result of fewer false positives or more false negatives . To examine this question , we considered ChIP-chip data for Fis binding to DNA [17] , which , as shown in Figure 5 , has RP>1 . Our findings are discussed in the next section . As described in the Methods section , we also generated ROC curves and calculated AUC to compare classifiers . For each of the five computational methods and for TFs in RegulonDB with 20 or more known binding sites , the AUC values are tabulated in Table S6 . We find that SiteSleuth had the largest AUC for 60% of the TFs tested , BvH had the largest AUC for 25% of the TFs , MATRIX SEARCH had the largest AUC for 10% of the TFs tested , QPMEME had the largest AUC for 5% of the TFs tested , and Match had the largest AUC for 0% of the TFs tested . ChIP-chip assays have identified 894 DNA sequences that bind Fis in E . coli [17] , which we used to validate the Fis binding sites predicted by each method . Looking at SiteSleuth results for Fis , SiteSleuth predicted 129 , 150 binding sites for Fis from a positive training set of 133 binding sites published in RegulonDB ( Table S3 ) , the second largest training set available for the 54 TFs we studied . The relative performance of SiteSleuth for Fis binding site prediction is close to one for three of the other methods under consideration ( RPBvH = 1 . 56 , RPMatch = 2 . 03 , RPMATRIX SEARCH = 1 . 55 , and RPQPMEME = 11 . 67 ) . SiteSleuth's cross-validation score for Fis ( V = 0 . 33 ) is low ( Table S4 ) . The availability of empirical data on Fis binding , including a larger number of known binding sites in RegulonDB for training , and the indirect recognition mechanisms of Fis binding to DNA [33] suggested that Fis may provide a good example to test whether SiteSleuth , which accounts for DNA structure , performs better than the other methods , despite its low cross-validation score . Predictions of Fis binding sites from each computational method are compared to experimentally identified DNA sequences that bind Fis in E . coli in ChIP-chip assays [17] . We assume that the sequences found in this study contain , to a first approximation , the complete set of Fis binding sites . For each method , the approximate number of false positives was determined by subtracting the number of predictions that matched experimentally defined Fis binding sequences from the total number of predictions made by the method . Figure 6 shows the number of false positives generated by each computational method ( black bars ) . As can be seen , the QPMEME method produced more than 1 . 5 million estimated false positives . Match generated approximately 261 , 000 false positives and BvH and MATRIX SEARCH both generated roughly 200 , 000 false positives . SiteSleuth produced the fewest false positives , over 70 , 000 fewer than the next best method , a reduction of 35% in the estimated false positive rate . In absolute terms , QPMEME predicted a binding site within 889 of the 894 experimentally defined Fis binding sequences ( 99 . 44% ) . However , the predictions are not practically useful , since they are hidden within over 1 . 5 million estimated false positive results . The gray bars in Figure 6 report the percentage of TF binding sites correctly predicted by the five computational methods normalized by the total number of predictions . After normalization , QPMEME was the lowest performer for Fis . The BvH , Match , and MATRIX SEARCH methods gave approximately equivalent results . SiteSleuth outperformed these methods , showing a 41% improvement over MATRIX SEARCH , the next best method .
We postulated that a better TF binding site prediction method could be developed on the basis of chemical and structural features , instead of letter sequences . To test this hypothesis , we developed the SiteSleuth method , in which potential TF binding sites are associated with DNA sequence-specific structural and chemical features . These features are then used to build classification models for and to predict TF binding sites . Compared to the other computational methods we tested , including the three methods that use a PWM representation of TF binding sites ( BvH , Match , and MATRIX SEARCH ) , our method provides a higher cross-validation accuracy . For 72% of the TFs studied , SiteSleuth cross-validation accuracy is as high as or higher than any other method ( Table S4 ) . SiteSleuth also generates 35% fewer estimated false positive results ( Figure 6 ) , and gives more accurate predictions ( 41% improvement over the next best method ) for TF binding sites ( Figure 6 ) . In addition , the four other methods considered here each rely on the additivity assumption , which states that each nucleotide in a DNA binding site contributes to binding affinity in an independent fashion . In the study of Benos et al . [44] , the additivity assumption was tested . In general , the additivity assumption holds rather well as shown by ddG measurements of mutated DNA sites in several protein-DNA complexes [44] . However , it was shown that additivity is a poor assumption for some cases [44] . SiteSleuth does not rely on the additivity assumption , which may partially explain its better performance . It must be noted that none of the methods for predicting TF binding sites considered here can be deemed reliable when used alone . In Figure 6 , although SiteSleuth indeed produces the highest fraction of correct predictions , the fraction of correct predictions is still small at 0 . 4% . Nonetheless , SiteSleuth constitutes an advance over existing methods and the approach warrants further investigation . The chemical and structural features we have considered are crude and additional determinants of specificity and other biologically relevant features , such as amino acid side chain interaction energy with DNA , could be incorporated into the SiteSleuth approach in the future . It may also be possible to incorporate experimental measurements of short DNA sequence properties into the SiteSleuth framework . A mechanistic understanding of TF binding to DNA could guide the design of novel model features . For example , a recent study of Fis showed that the shape of the DNA minor groove affects Fis-DNA binding [45] . This property is hard to capture using only DNA letter sequences , but could be captured by defining a new feature in SiteSleuth based on the available structural data . Presently , the features defined in SiteSleuth are unable to capture the effects of the minor groove on Fis binding , which may account for SiteSleuth's poor performance in absolute terms . The QPMEME method is similar to the SVM-based approach of SiteSleuth . Both methods involve a quadratic programming minimization procedure with linear inequality constraints . QPMEME maps sequences of L bases into 4L multidimensional spaces with energy terms for each dimension and constructs a hyperplane such that all positive examples are located on one side of the plane . This quadratic optimization procedure defines a separating hyperplane by minimizing the variance of energies in an energy matrix so as to minimize the number of random sequences lying on the side of the plane that contains the positive examples . In contrast , the separating hyperplane of an SVM divides true binding sites from nonbinding sites with maximum margin . The distinction between random sequences , considered in QPMEME , and negative examples , considered in SiteSleuth , is important because sequences do not appear with equal probability in the E . coli genome , as is shown in Figure S1 . SiteSleuth used negative examples directly sampled from non-coding regions of the E . coli genome . In the report of Djordjevic et al . [6] , the QPMEME method is applied to non-ORF regions of the E . coli genome to predict binding sites for 34 TFs , including Fis . For Fis , Table 1 of Ref . [6] indicates that QPMEME predicts 255 Fis binding sites , compared to the 1 . 5 million found with QPMEME in our hands ( Table S3 ) . To ensure that our implementation was correct , we applied QPMEME using the same training data set used by Djordjevic et al . [6] from DPInteract and were able to reproduce their weight matrix [6] . For Fis , RegulonDB reports 133 binding sites , compared to only 19 reported Fis binding sites in DPInteract . This difference in the size of the training data set ( 19 versus 133 positive examples of Fis binding sites ) may be responsible for the difference in number of predicted binding sites ( 255 vs . 1 . 5 million ) . As can be seen by comparing the common entries in Table 1 of Ref . [6] and in Table S3 , Fis is not an isolated example of QPMEME predicting a larger number of TF binding sites when the number of positive training examples is larger . It is also the case for the TFs ArcA , ArgR , CRP , CytR , DnaA , FadR , FarR , Fnr , FruR , GalR , GlpR , H-NS , IHF , LexA , LRP , MetJ , NagC , NarL , OmpR , SoxS , and TyrR . The QPMEME method may perform poorly for TFs with relatively large numbers of known binding sites because QPMEME requires that all positive examples be located on one side of a hyperplane in the space spanned by an energy matrix [6] ( see Methods section ) . Thus , known binding sites that are outliers in this space may potentially expand the range of sequences considered to be binding sites , such that recall is maximized at the expense of precision . We have not systematically investigated the reasons underlying our observation that QPMEME performs poorly for the TFs identified above when using positive training data from RegulonDB , as such an investigation was beyond the intended scope of our study . In summary , how TFs selectively bind to DNA is one of the least understood aspects of TF-mediated regulation of gene expression . An ability to better predict TF binding sites from small training data sets may advance our understanding of TF-DNA binding , and may reveal important insights into TF binding specificity , regulation and coordination of gene expression , and ultimately into gene function . A long-standing problem has been how to identify new TF binding sites given known binding sites . The accuracy and usefulness of computational methods for genome-wide TF binding site prediction has been limited by the inability to validate , verify , and inform these methods . Only recently has technology matured to the point that we can assay for TF binding sites on a genome-wide scale . This capability should allow us to critically evaluate predictions from computational methods and to develop methods that are more predictive than those currently available . Toward this end , the work presented here provides a starting point for future investigations of how TF binding site prediction can be improved by considering the physical and chemical aspects of TF-DNA binding .
|
An important step in characterizing the genetic regulatory network of a cell is to identify the DNA binding sites recognized by each transcription factor ( TF ) protein encoded in the genome . Current computational approaches to TF binding site prediction rely exclusively on DNA sequence analysis . In this manuscript , we present a novel method called SiteSleuth , in which classifiers are trained to discriminate between true and false binding sites based on the sequence-specific chemical and structural features of DNA . According to cross-validation analysis and a comparison of computational predictions against ChIP-chip data available for the TF Fis , SiteSleuth predicts fewer estimated false positives than any of four other methods considered . A better understanding of gene regulation , which plays a central role in cellular responses to environmental changes , is a key to manipulating cellular behavior for a variety of useful purposes , as in metabolic engineering applications .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"molecular",
"biology/bioinformatics",
"mathematics/statistics",
"computational",
"biology/sequence",
"motif",
"analysis",
"computational",
"biology/transcriptional",
"regulation"
] |
2010
|
Using Sequence-Specific Chemical and Structural Properties of DNA to Predict Transcription Factor Binding Sites
|
Genome size and complexity vary tremendously among eukaryotic species and their organelles . Comparisons across deeply divergent eukaryotic lineages have suggested that variation in mutation rates may explain this diversity , with increased mutational burdens favoring reduced genome size and complexity . The discovery that mitochondrial mutation rates can differ by orders of magnitude among closely related angiosperm species presents a unique opportunity to test this hypothesis . We sequenced the mitochondrial genomes from two species in the angiosperm genus Silene with recent and dramatic accelerations in their mitochondrial mutation rates . Contrary to theoretical predictions , these genomes have experienced a massive proliferation of noncoding content . At 6 . 7 and 11 . 3 Mb , they are by far the largest known mitochondrial genomes , larger than most bacterial genomes and even some nuclear genomes . In contrast , two slowly evolving Silene mitochondrial genomes are smaller than average for angiosperms . Consequently , this genus captures approximately 98% of known variation in organelle genome size . The expanded genomes reveal several architectural changes , including the evolution of complex multichromosomal structures ( with 59 and 128 circular-mapping chromosomes , ranging in size from 44 to 192 kb ) . They also exhibit a substantial reduction in recombination and gene conversion activity as measured by the relative frequency of alternative genome conformations and the level of sequence divergence between repeat copies . The evolution of mutation rate , genome size , and chromosome structure can therefore be extremely rapid and interrelated in ways not predicted by current evolutionary theories . Our results raise the hypothesis that changes in recombinational processes , including gene conversion , may be a central force driving the evolution of both mutation rate and genome structure .
Explaining the origins of variation in genome size and complexity has become the defining challenge for the field of molecular evolution in the genomic era . Historically , numerous evolutionary models have been developed , involving mechanisms such as insertion and deletion ( indel ) bias [1] , [2] , selfish element proliferation [3] , [4] , and natural selection on cell size [5] , replication rate [6] , and evolvability [7] . In recent years , a body of theory known as the mutational burden hypothesis ( MBH ) has emerged as a potentially unifying explanatory framework rooted in the principles of population genetics and the basic evolutionary processes of mutation and genetic drift [8] , [9] . The MBH posits that noncoding elements are generally deleterious but proliferate nonadaptively when small effective population sizes reduce the effectiveness of selection relative to genetic drift , offering an explanation for why noncoding sequences are so abundant in large multicellular eukaryotes . This hypothesis is based on the idea that noncoding elements impose a selective cost associated with the increased chance of mutations disrupting an essential genome function ( e . g . , alteration of a conserved sequence required for intron splicing ) or generating a novel deleterious feature ( e . g . , an improper transcription-factor binding site in an intergenic region ) . The MBH has potentially sweeping explanatory power , but some of its tenets are controversial [10]–[13] , and its generality as a mechanism of genome evolution remains uncertain [14]–[21] . Mitochondrial genomes display striking diversity in size and complexity [22] , [23] , reflecting patterns of variation in genome architecture observed more broadly across the tree of life [9] , [24] . For example , in contrast to the small ( typically 14–20 kb ) and streamlined genomes found in most animal mitochondria [25] , seed plant mitochondrial genomes are very large ( 200–2 , 900 kb ) , containing introns and abundant intergenic sequences [26]–[28] . Plant mitochondrial genomes are also typically characterized by extremely low point mutation rates , further distinguishing them from their fast-evolving animal counterparts [29]–[31] . The observed disparity in mitochondrial mutation rates across eukaryotes motivated the hypothesis that mutation rates are a major determinant of variation in organelle genome architecture [32] . This argument is a direct extension of the MBH and is based on the premise that the probability of mutational disruption of noncoding elements ( which is equivalent to the selective cost associated with maintaining those elements ) is directly proportional to the mutation rate . Therefore , genomes with elevated mutation rates are predicted to experience more intense selection for genomic reduction [32] . The discovery that some angiosperms have greatly accelerated mitochondrial mutation rates , sometimes orders of magnitude greater than closely related species [33]–[35] , presents an opportunity to test the prediction that high mutation rate environments select for reduced and streamlined genomes . In particular , several species in the genus Silene ( Caryophyllaceae ) have experienced dramatic increases in mitochondrial mutation rates within just the last 5–10 Myr , while other members of this genus have maintained their ancestrally low rates [35]–[37] . We compared complete mitochondrial genome sequences from four Silene species with very different mutation rates and found that accelerated mutation rates have indeed been associated with dramatic changes in genome size and complexity . However , the direction of these changes is not always consistent with the predictions from existing theory . We discuss the implications of the unprecedented mitochondrial genome diversity found within Silene and possible alternative explanations for the rapid genome evolution in this genus .
Sequencing of purified mitochondrial DNA ( mtDNA ) from three Silene species generated complete genome assemblies for S . noctiflora and S . vulgaris and a high quality draft assembly for S . conica . We also included the previously published mitochondrial genome of S . latifolia in our analyses [38] . The genomic data extend previous results [35]–[37] by showing that S . noctiflora and S . conica have experienced massive accelerations in nucleotide substitution rates ( Figure 1 ) across all protein genes ( Figure 2 ) with correlated increases in the frequency of both insertions and deletions ( Figure 3 ) . Contrary to the prediction of genomic streamlining in response to high mutation rate , the fast-evolving mitochondrial genomes of S . noctiflora and S . conica have experienced unprecedented expansions , resulting in sizes of 6 . 7 Mb and 11 . 3 Mb , respectively . In contrast , the more typical slowly evolving mitochondrial genomes of S . vulgaris ( 0 . 43 Mb ) and particularly S . latifolia ( 0 . 25 Mb ) are on the lower end of the angiosperm size range . Thus , Silene mitochondrial genomes have diverged more than 40-fold in size in just the past few million years . The genomic expansion in S . noctiflora and S . conica does not reflect detectable increases in gene or intron content . Although these genomes contain duplicate copies of some genes ( particularly rRNA genes; Table S1 ) , they possess fewer unique genes than other angiosperm mitochondrial genomes ( Figures 1 and 4 ) . Notably , the S . conica and S . noctiflora mitochondrial genomes contain only two or three identifiable tRNA genes , which is far fewer than most angiosperms and even less than the already reduced tRNA gene content of S . latifolia and S . vulgaris ( Figures 1 and 4 ) [38] . The four Silene genomes have nearly identical sets of introns ( Table 1 ) . With the exception of additional intron copies associated with gene duplications , there were no intron gains among the four Silene species and only one observed intron loss ( the third intron of nad4 in S . noctiflora ) . Interestingly , in contrast to the overall pattern of genome expansion in S . noctiflora and S . conica , average intron lengths in the expanded S . noctiflora and S . conica genomes are actually ∼10%–15% shorter than in their congeners ( Figure S1 ) . Intergenic sequences account for 99% of the bloated mitochondrial genomes in S . noctiflora and S . conica . As in other vascular plants [28] , [39] , the intergenic regions of all four Silene mitochondrial genomes contain sequences of both nuclear and plastid ( chloroplast ) origin . Although the expanded mitochondrial genomes of S . noctiflora and S . conica contain more of this “promiscuous” DNA than their smaller Silene counterparts ( Table 1 ) , contributions from these sources do not scale proportionally with the increases in genome size and constitute less than 1% of the intergenic content in both species ( Table 1 ) . A larger fraction of the intergenic regions in each of these two genomes exhibit similarity to sequences in other plant mitochondrial genomes ( Table 1 ) , but most of this sequence ( >650 kb ) is only shared between S . noctiflora and S . conica and not with any other angiosperms . Overall , >85% of the voluminous intergenic sequence in these two species lacks detectable homology with any of the nuclear , plastid , or mitochondrial sequences available in the GenBank nr/nt database . Repeated sequences constitute a variable and often large component of seed plant mitochondrial genomes [40] , and Silene species are noteworthy in both respects ( Figures 5 , S2 , and S3; Table 1 ) . The S . conica mitochondrial genome contains a remarkable 4 . 6 Mb of dispersed repeats , which is more than any other sequenced plant mitochondrial genome in both absolute and percentage ( 40 . 8% ) terms [40] . The largest repeats are >80 kb in size , but the bulk of the repetitive content consists of an enormous number of small , imperfect , and often partially overlapping repeats ( Figures 5 , S2 , and S3 ) . In contrast , repeat sequences make up just 6 . 7%–18 . 8% of the other three Silene mitochondrial genomes . Silene noctiflora and S . conica have also evolved extraordinary mitochondrial genome structures . Although the relationship between genome maps and in vivo physical structure remains uncertain for angiosperm mtDNAs [41] , the entire sequence content of the genome typically can be mapped as a single “master circle , ” which can be subdivided into a collection of “subgenomic circles” that arise via high-frequency recombination between large direct repeats ( Figure S4A ) [42] , [43] . This model applies to S . latifolia [38] , whereas the S . vulgaris genome assembles into four circular-mapping chromosomes , with the largest ( 394 kb ) comprising most ( 92% ) of the genome and containing numerous repeats inferred to undergo active recombination on the basis of their association with alternative rearranged genome conformations ( Figure S4 ) . Two of the three smaller mitochondrial chromosomes in S . vulgaris share recombinationally active repeats with the large chromosome , but the majority of sequencing reads support the smaller subgenomic conformations ( see Materials and Methods and Figure S4 ) . In contrast , the smallest of the four S . vulgaris chromosomes appears to be almost completely autonomous . It does not share any repeats longer than 100 bp with the rest of the genome , and in the case of all shorter repeats shared between the smallest chromosome and the main chromosome , >99 . 5% of sequencing read-pairs support the smaller subgenomic conformation . While the presence of this small chromosome is itself unusual for plant mtDNAs , far more extreme are the S . noctiflora and S . conica mitochondrial genomes , each of which assembled into dozens of mostly autonomous and relatively small , circular-mapping chromosomes . The S . noctiflora mitochondrial genome consists of 59 circular-mapping chromosomes ranging from 66 to 192 kb in size ( Table S2 ) . Many of these do not share any large ( >1 kb ) repeats with other chromosomes . Even when S . noctiflora chromosomes do share large repeats ( up to 6 . 3 kb ) , the clear majority of paired-end sequencing reads ( >90% in all cases ) support the conformation consisting of two smaller circles rather than a single combined circle . Although the extremely repetitive nature of the S . conica mitochondrial genome precluded complete genome assembly , its structural organization is similar to that of S . noctiflora . The vast majority ( 98 . 2% ) of sequence content assembled into 128 circular-mapping chromosomes ranging from 44 to 163 kb in size ( Table S2 ) . Most of these chromosomes share only short repeats with other parts of the genome . The number of sequencing reads that cover a given position in a shotgun genome assembly ( i . e . , the read depth ) can be used to estimate the relative abundance of different sequences . The difference in average read depth between the chromosomes with the highest and lowest coverage was only 1 . 7-fold in S . noctiflora and only 3 . 1-fold in S . conica ( after excluding repetitive regions ) , indicating that the abundance of the numerous chromosomes was relatively even in both genomes . The different chromosomes also exhibited a high degree of similarity in GC content within each genome ( Table S2 ) . Assembly of repetitive genomes is inherently complicated , and this is particularly relevant to the identification of genomic subcircles because tandem duplications within a larger chromosome can misassemble as subcircles . However , such assembly errors leave clear signatures , including dramatic variation in read depth and conflicting read-pairs associated with the boundary between tandem repeats and flanking regions . The absence of such patterns in our dataset indicates that the assembled circles are not an artifact of tandem repeats within larger chromosomes . Nevertheless , it is possible , particularly in the draft assembly of S . conica mitochondrial genome , that some repeat pairs have been “collapsed” into single sequences , leaving open the possibility that the reported 11 . 3 Mb genome size for S . conica is a slight underestimate . Sequencing of the S . latifolia mitochondrial genome showed that it contains a six-copy 1 . 4-kb repeat that is highly recombinationally active with physical cross-overs between repeat copies generating a suite of rearranged genome conformations [38] . Southern blot analysis confirmed that the many alternative genome conformations occur in roughly equivalent frequencies in S . latifolia [38] . Paired-end sequencing reads can also be used to quantify the relative abundance of alternative genome conformations ( see Materials and Methods and Figure S4 ) , and our 454 data suggest a comparably high level of repeat-mediated recombinational activity for the largest repeats in the S . vulgaris mitochondrial genome ( Figure 6A ) . The relative frequency of recombinant genome conformations increases with repeat size , and all surveyed repeats longer than 100 bp exhibit evidence of a history of recombination . The two largest surveyed pairs of repeated sequences ( 0 . 9 and 3 . 0 kb ) in the S . vulgaris genome each appear to be at or near a 50∶50 level of alternative genome conformations ( Figure 6A ) . The rapidly evolving mitochondrial genomes of S . noctiflora and S . conica exhibit reduced frequencies of recombinant genome conformations compared to other Silene genomes ( Figure 6B ) and all other angiosperm mitochondrial repeats for which recombinational activity has been assessed . Even the largest repeats in the S . noctiflora genome ( up to 6 . 3 kb ) are associated with only a small minority of recombinant products ( Figure 6B ) . The largest repeats in the S . conica genome ( up to 87 kb ) far exceed our paired-end library span and therefore cannot be analyzed for recombinational activity , but analysis of the shorter repeats suggests that the genome has experienced a similar shift in the relationship between repeat length and the frequency of recombinant products ( Figure 6B ) . Recombinational activity ( including gene conversion ) is expected to homogenize copies of repeated sequences throughout the genome . Therefore , the dramatic increase in the proportion of divergent pairs of repeated sequences within the mitochondrial genomes of S . noctiflora and S . conica ( Figures 7 and S5 ) is consistent with a reduction in recombinational activity in these species , though the existence of divergent repeats could also result from the increased mutation rate in these species or a reduced probability of gene conversion events between physically disparate repeat copies in expanded genomes . The coexistence of maternally and paternally derived mitochondrial genomes in a heteroplasmic state within the same individual or maternal family would introduce complications for genome sequencing and assembly . Therefore , we looked for evidence of heteroplasmy and nonmaternal inheritance in the families used in this study . S . vulgaris has been the subject of extensive investigation into the patterns of mitochondrial genome inheritance [44]–[47] . These studies have found that mtDNA transmission is predominantly maternal in S . vulgaris , with a low frequency of biparental inheritance or paternal “leakage . ” Because of this evidence , the S . vulgaris family used for genome sequencing was chosen , in part , because the maternal source plant had previously been screened with two highly polymorphic mitochondrial markers and revealed no evidence of heteroplasmy [46] . Although similarly intensive investigations of mtDNA inheritance have not been performed in other Silene species , we found evidence of maternal transmission in S . latifolia , S . noctiflora , and S . conica . An analysis of cleaved amplified polymorphic sequences ( CAPS ) showed that all progeny ( 16–48 per species ) from controlled greenhouse crosses inherited the maternal variant of a SNP . Mitochondrial inheritance therefore appears to be at least predominantly maternal in all four Silene species , making it unlikely that genome assembly complications arising from biparental inheritance and heteroplasmy can explain the observed differences in mitochondrial genome size and complexity among Silene species . S . noctiflora and S . conica do not show the proportional increases in mitochondrial nucleotide diversity that would be expected on the basis of their accelerated mutation rates ( even after accounting for the approximately 2-fold differences in generation times across the four Silene species [48] ) , suggesting a recent history of lower effective population size ( Ne ) than their congeners and/or a recent reversion to lower mitochondrial mutation rates as observed in other accelerated angiosperm lineages [33] , [34] . In S . conica , there is less than a 10-fold increase in mitochondrial synonymous nucleotide diversity relative to the more slowly evolving Silene species , and S . noctiflora exhibits no sequence variation whatsoever across our sample of mitochondrial , plastid , and nuclear loci ( Table S3 ) ( see also [49] ) .
The dramatic expansion of intergenic content in the mtDNA of S . noctiflora and S . conica has resulted in mitochondrial genomes that are larger than most bacterial genomes ( Figure 8 ) and even some nuclear genomes [50] . These enormous genomes add to the long-standing mystery regarding the origins of intergenic sequences in plant mtDNA [28] . It is possible that a significant portion of this intergenic content is derived from the nuclear genome , for which sequence data are still limited in Silene . However , by comparing the mitochondrial genomes against a large set of cDNA sequences derived from a recent transcriptome project in S . vulgaris [51] , we detected similarity for only a trivial amount ( <0 . 1% ) of the otherwise uncharacterized mitochondrial sequence in S . noctiflora and S . conica . Therefore , if nuclear DNA is a major contributor to the expanded mitochondrial intergenic regions in these species , it is most likely drawn from the vast repetitive and noncoding fractions of the nuclear genome . That the origin of only a small fraction of the intergenic sequences in S . noctiflora and S . conica can be identified may reflect the rapid rates of sequence and structural divergence in these mitochondrial genomes . In other plant mitochondrial genomes , the proliferation of “selfish” DNA may have contributed to expansions in intergenic regions . For example , the mtDNA of the gymnosperm Cycas contains numerous copies of repetitive elements known as Bpu sequences [52] , and the expanded mitochondrial and plastid genomes of the green alga Volvox share an apparently self-replicating element with the nucleus [53] . The finding of expanded intergenic sequence in S . noctiflora and S . conica mtDNA raises the question of whether some form of selfish element has been involved . This appears possible in S . conica , given the highly repetitive nature of its mitochondrial genome ( Figures 5 , S2 , and S3; Table 1 ) . However , we did not find evidence for any specific sequence or set of sequences that dominate the repetitive content in S . conica . There is even less evidence for a role of mobile , self-replicating elements in S . noctiflora mtDNA given the small amount of repeated sequence in this genome . Interestingly , S . noctiflora harbors a relatively modest proportion of repetitive sequence compared to many other angiosperms' mtDNAs , including the much smaller S . vulgaris genome ( Figures 5 and S2; Table 1 ) , indicating that there is no strict relationship between repetitive content and genome size . It is noteworthy that S . noctiflora and S . conica share a large amount of intergenic sequences with each other ( 659 kb and 760 kb , respectively ) that show little or no homology with any available sequences in the GenBank nr/nt database including all other sequenced plant mitochondrial genomes . These shared intergenic sequences may be the remnants of an ancestral genomic expansion that preceded the divergence of S . noctiflora and S . conica , suggesting a possible sister relationship between these two lineages , an issue that is currently unresolved by molecular phylogeny [37] , [54] . If so , this could indicate that the atypical mitochondrial genome size , structure , and substitution rates in S . noctiflora and S . conica represent a single set of evolutionary changes rather than phylogenetically independent events . However , we cannot rule out the possibility that the shared sequences are the result of parallel acquisitions from similar sources , such as the nuclear genomes in each species . Generating sequence data from other genomic compartments , particularly from a large number of unlinked nuclear loci , should provide better insight into the phylogenetic history of these Silene species . Although the highly multichromosomal genome structures observed in S . noctiflora and S . conica are novel for plant mitochondria , various forms of multicircular organelle genomes have evolved independently in diverse eukaryotic lineages , including in the mitochondria of kinetoplastids [55] , diplonemids [56] , chytrid fungi [57] , and a number of atypical metazoans [58]–[61] , as well as in dinoflagellate plastids [62] . In addition , the recent analysis of the cucumber mitochondrial genome showed that a small fraction of that genome can be mapped to two circular chromosomes that appear to be independent from the main chromosome [63] . It should be noted that the maps generated from the assembly of DNA sequence data do not necessarily reflect the structure of the genome in vivo . In particular , linear concatamers and overlapping linear fragments can assemble as circular maps [64] . Efforts to directly observe the molecular structure of angiosperm mitochondrial genomes have identified a complex mixture of linear , circular , and branched molecules [65] , [66] , indicating that the circular maps produced by genome projects may be abstractions or oversimplifications . Although on the basis of our current data we cannot distinguish between the various structural alternatives capable of producing circular chromosome maps , the sequence assemblies do support the intriguing finding that many of these chromosomes are structurally autonomous , lacking the large , recombinationally active repeats that are characteristic of most angiosperm mitochondrial genomes . The existence of multichromosomal mitochondrial genomes in Silene raises fundamental questions about the nature of replication and inheritance of these genomes . Notably , we did not detect a single intact gene in many chromosomes , including the smallest chromosome in S . vulgaris , 20 of the 59 chromosomes in S . noctiflora , and 86 of the 128 chromosomes in S . conica ( note that these totals do not include chromosomes in S . noctiflora and S . conica that only contain partial gene fragments that require trans-splicing with transcripts originating from other chromosomes to generate complete coding sequences ) . Therefore , the functional significance ( if any ) of these “empty” chromosomes and the evolutionary forces that maintain their presence and abundance within the mitochondrion are unclear . While it is possible that these chromosomes contain unidentified genes or noncoding elements that are functionally important and therefore conserved by selection , they may also replicate and proliferate in a nonadaptive or even selfish fashion . Our analysis was based on mtDNA extracted from predominantly vegetative tissue pooled across multiple individuals from a single maternal family . Therefore , we do not know whether any of the observed structural variation in mtDNA is partitioned within our pooled sample and , if so , at what level it is partitioned ( i . e . , among individuals , tissue types , cells , or even individual mitochondria ) . In this light , it would be particularly informative to conduct an analysis of mitochondrial genome sequence and structure in meristematic tissue to compare with our results from vegetative tissue . Any differences between these tissue types would be of interest because the mtDNA in meristematic tissue should better represent the inherited form of the genome . The co-occurrence of mutational acceleration and genome expansion in the mitochondria of S . noctiflora and S . conica runs counter to patterns in other eukaryotic mitochondrial genomes ( e . g . , plants versus animals ) . Although we cannot determine the relative timing of these changes , their co-occurrence in these lineages is at odds with the hypothesis that reduced mutation rates are a major cause of mitochondrial genome expansion in plants [32] . An alternative possibility that would be consistent with the MBH is that these species have a small Ne , which has reduced the efficacy of selection against the proliferation of noncoding elements even if the intensity of that selection has increased with higher mutation rates . There is some evidence to support this possibility , particularly in S . noctiflora , which appears to have a very low Ne based on the striking lack of polymorphism in genes from all three genomes ( Table S3 ) [49] . However , the finding of high levels of mitochondrial polymorphism in S . conica ( Table S3 ) is contrary to the predictions of the MBH . Some caution is warranted in interpreting the nucleotide diversity data because standing levels of polymorphism are very sensitive to recent bottlenecks and do not necessarily represent the long-term average Ne over the entire history of a species or lineage . One alternative proxy for Ne and the relative strength of genetic drift is the ratio of nonsynonymous to synonymous substitutions ( dN/dS ) , with higher ratios indicating a reduced efficacy of selection in purging deleterious changes in amino acid sequence [12] . Based on this alternative measure , there is no indication of a long-term decrease in Ne in either S . noctiflora or S . conica since their divergence from the other Silene species ( Table 1 ) . Therefore , with respect to both mutation rate and Ne , the changes in mitochondrial genome size within Silene appear to be inconsistent with any straightforward interpretation of the MBH . In contrast to the differences in overall genome size in Silene mitochondria , some of the observed changes in these genomes are consistent with predictions of the MBH . Most notably , average intron lengths have decreased in the species with elevated mutation rates , and the only example of an intron loss was observed in a high-rate species . These results could indicate that the consequences of mutational burden vary substantially within a genome . For example , the contrasting patterns observed in introns versus intergenic regions within these lineages might suggest that the burden associated with disruptive mutations in functional noncoding elements such as introns is of far greater evolutionary importance than that associated with gain-of-function mutations creating novel deleterious elements in largely nonfunctional intergenic regions . The inability of existing theory to fully account for the extreme patterns of divergence in Silene mitochondrial genomes points to a valuable opportunity to expand our understanding of the evolutionary forces that shape genomic complexity . Although this study was restricted to a small number of species from a single genus , it captured enormous variation in genome architecture ( e . g . , approximately 98% of the known range of organelle genome sizes ) , indicating that profound and perhaps novel evolutionary mechanisms are acting to shape mitochondrial genome size and complexity in Silene . The observed differences among Silene species in the frequency of recombinant genome conformations raise the possibility that recombination could be a key factor underlying the extreme patterns of mitochondrial genome evolution in S . noctiflora and S . conica . The mitochondrial genomes in these species differ from those of other angiosperms in numerous respects , including rates of point mutations and indels , presence of duplicated and divergent gene copies , frequency of RNA editing , genome size , and structural organization ( Table 1 ) . Many , perhaps all , of these traits are likely affected by the related processes of intragenomic recombination and gene conversion . Recombinational processes play an important role in plant mitochondrial genome sequence and structural evolution [43] , [67] . In addition , recombination between repeated sequences ( including very short repeats ) has been shown to be an important mechanism for sequence deletion in plant nuclear genomes [68] . Therefore , changes in recombinational activity are expected to affect the evolution of genome size . However , recombinational processes can also have opposing effects on genome size via sequence duplication or integration of new content , so that the relationship between recombination and genome size is likely to be a complex one . Recombination and gene conversion mechanisms have also been implicated in the evolution of other elements of genome architecture . For example , retroprocessing events involving cDNA intermediates are likely responsible for the loss of introns and RNA editing sites [34] , [69] , [70] . Recombination and gene conversion are key components of DNA repair pathways . Notably , gene conversion mechanisms that are biased against new mutations have been proposed to slow the effective or observed mutation rate in multicopy genomes [71] , [72] . Our findings raise the possibility that template-based recombinational repair and biased gene conversion are important factors underlying the typically low rates of nucleotide substitution in plant mitochondrial genomes and that these mechanisms have been altered or disrupted in fast-evolving species such as S . noctiflora and S . conica . The associated increase in the rate of mitochondrial indels in these species ( Figure 3 ) suggests that alterations in replication and repair machinery can have correlated effects on both point mutations and structural changes , which is consistent with the correlation between rates of mitochondrial sequence and structural evolution observed in other lineages [73]–[76] . Our findings highlight the need to characterize Silene nuclear gene families involved in recombination and other aspects of organelle genome maintenance . Unraveling the process of sequence gain and turnover in these rapidly evolving mitochondrial genomes should provide insight into the evolutionary forces underlying the tremendous variation in size and complexity of eukaryotic genomes .
The genus Silene ( Caryophyllaceae ) consists of approximately 700 predominantly herbaceous species of flowering plants [77] , many of which are used as models in ecology and evolution [78] . S . noctiflora L . and S . conica L . both have annual life histories [79] , and they are largely hermaphroditic but produce a low frequency of pistillate ( female ) flowers and can therefore be characterized as gynomonoecious [80]–[82] ( DBS , personal observation ) . S . latifolia Poir . and S . vulgaris ( Moench ) Garcke are short-lived perennials with an average generation time of approximately 2 y [48] that maintain dioecious and gynodioecious breeding systems , respectively [79] , [80] . Details of the Silene latifolia mitochondrial genome project were described previously [38] . For each of the other three species , approximately 200 g of tissue was collected from multiple individuals of a single maternal family . The maternal lineages were derived from seeds originally collected in Abruzzo , Italy ( S . conica ) , Eggleston , VA , US ( S . noctiflora ) , or Stuarts Draft , VA , US ( S . vulgaris ) . Voucher specimens from each of these maternal lineages have been deposited to Massey Herbarium at Virginia Polytechnic and State University: S . conica ( L Bergner 003 ) , S . noctiflora ( D Sloan 003 ) , S . vulgaris ( L Bergner 007 ) . All aboveground tissue was used for S . vulgaris , including leaves , stems , and flowers , while only leaf tissue was collected for S . noctiflora and S . conica . Mitochondrial DNA was purified from mitochondria from harvested tissue using established protocols based on differential centrifugation , treatment with DNase I , and then either CsCl gradients or phenol∶chloroform extraction [83] , [84] . Restriction digests with MspI and HpaII enzymes , which share identical recognition sequences but differ in methylation sensitivity , were performed to confirm the absence of significant nuclear contamination from the purified mtDNA samples prior to sequencing . For each of the species , 3-kb paired-end libraries were prepared following standard protocols for sequencing on a Roche 454 GS-FLX platform with Titanium reagents . Additional libraries were prepared ( also following standard Roche protocols ) for the larger S . noctiflora and S . conica mitochondrial genomes , including shotgun libraries for both species and a 12-kb paired-end library for S . noctiflora . The latter was constructed following the standard 8-kb protocol , but the larger 12-kb average fragment size range was selected on the basis of the size distribution of the DNA sample after shearing . Each library was run on a single quarter-plate region except for the S . conica shotgun library and the S . noctiflora 12-kb paired-end library , which were each run on two quarter-plate regions . The shotgun library for S . noctiflora was constructed and sequenced by the Genome Center at Washington University in St . Louis ( MO , US ) . All other 454 library construction and sequencing was performed at the Genomics Core Facility in the University of Virginia's Department of Biology . To generate sufficient starting material for Illumina library construction , mtDNA samples were amplified with GenomiPhi V2 ( GE Healthcare ) . Paired-end sequencing libraries were generated and tagged with multiplex barcodes using the NEBNext DNA Sample Prep Reagent set 1 ( New England Biolabs ) in accordance with protocols developed by the University of California Davis Genome Center . In brief , DNA samples were sonicated to a peak fragment size of between 300 and 600 bp . DNA fragments were then end polished and ligated to adaptors carrying a unique 6-bp barcode . The resulting samples were gel-purified and amplified with 14 PCR cycles using paired-end library primers . The three libraries were included in a larger sample pool and sequenced in a single lane of a 2×85 bp paired-end run on an Illumina GAII . Sequencing was performed at the Biomolecular Research Facility in the University of Virginia's School of Medicine . Each quarter-plate 454 run produced between 32 and 104 Mb of sequence . The total sequencing yield was 270 , 210 , and 51 Mb for the S . noctiflora , S . conica , and S . vulgaris mtDNA samples , respectively . However , not all sequence data were used in primary genome assembly . For S . noctiflora , only the shotgun and 3-kb paired-end data were analyzed in the initial assembly process . The 12-kb paired-end data were only used to resolve structures associated with large ( >3 kb ) repeats and to quantify the frequency of alternative genome conformations resulting from recombination among repeat copies ( see below ) . For the smaller , S . vulgaris mitochondrial genome , a single quarter-plate run produced very high coverage ( >80× ) . Preliminary analyses suggested use of the entire dataset increased fragmentation in the assembly . Therefore , a random set of sequence reads totaling 25 Mb was selected for initial assembly . The full S . vulgaris dataset was used for subsequent quantification of alternative genome conformations . For each genome , the 454 sequence reads were assembled with Roche's GS de novo Assembler v2 . 3 ( “Newbler” ) using default settings . The resulting assemblies produced average read depths of 20× , 25× , and 42× for the S . conica , S . noctiflora , and S . vulgaris mitochondrial genomes , respectively . Although the assemblies contained few , if any , gaps or low-coverage regions , they were highly fragmented because of the repetitive and recombinational nature of these genomes ( Figures 5 and 6 ) . The assemblies also contained contigs from contaminating nuclear , plastid , and viral DNA . True mitochondrial contigs were distinguished on the basis of read depth and connectivity to other contigs in the assembly , which was inferred from two types of data: ( 1 ) paired-end reads that mapped to two different contigs and ( 2 ) single reads that were split by the assembler and assigned to the ends of two different contigs . On the basis of these data , contigs were organized into “subgenomes , ” each of which represented either a closed circular assembly or a single-copy assembly flanked on either side by recombinationally active repeats . Each of these subgenomic contig groups was then reassembled using a custom set of Perl and BASH scripts that identified all sequencing reads uniquely associated with the corresponding contigs and ran a new assembly using only those reads . The resulting subgenomic assemblies were then manually edited and combined as necessary with the aid of Consed v17 . 0 [85] . The largest repeats in both the S . conica and S . vulgaris mitochondrial genomes exceed the 3-kb span size of their respective paired-end libraries . Therefore , the relationships between the single-copy regions flanking these large repeats are ambiguous . These ambiguities were tentatively resolved on the basis of the pattern observed in smaller repeats within each genome ( Figure 6 ) . On the basis of the high level of recombinational activity among smaller repeats in S . vulgaris , we assumed that large repeats also have high recombinational activity . Therefore , we assembled the majority of the S . vulgaris genome content into a single chromosome , analogous to the “master circle” typically reported for plant mitochondrial genomes . This large chromosome contains numerous recombinationally active repeats , and , as discussed previously [38] , the arrangement of repeats and single-copy regions within this chromosome should be considered only one of many possible alternative representations . We also identified three small circular-mapping structures that were not included in the main assembly . One of these circles ( Chromosome 4 ) shows almost no evidence of recombinational activity with the rest of the genome , while the other two do share repeats that appear to recombine frequently with the main chromosome . However , in both of these cases , the repeats are small ( <500 bp ) , and the clear majority of reads support the closed circle conformations over a single combined circle . For convenience , we refer to these three circles as chromosomes , but their small size and ( in the case of Chromosomes 2 and 3 ) substantial degree of recombinational activity with the rest of the genome distinguish them from the chromosomal structure that characterizes the S . noctiflora and S . conica mitochondrial genomes . In contrast to S . vulgaris , the bulk of the S . noctiflora and S . conica mitochondrial genomes map to discrete circular chromosomes that exhibit little or no recombinational activity with the rest of the genome . In both species , repeats show much less evidence of recombination than repeats of similar size in S . latifolia and S . vulgaris ( Figure 6 ) . Moreover , in cases of recombinationally active repeats , the clear majority of paired-end reads ( >90% in all cases in S . noctiflora and the vast majority of cases in S . conica; Figure 6 ) support minimally sized circular conformations rather than larger combined circles . Therefore , for assembly ambiguities associated with repeats exceeding the 3-kb paired-end library span in S . conica , it was assumed that minimally sized circles predominate over larger combined conformations . To correct base-calling errors including insertion and deletion errors known to be associated with long single-nucleotide repeats ( i . e . , homopolymers ) in 454 sequence data , we mapped Illumina sequence data onto the completed mitochondrial genome assemblies for each species . After removal of multiplex barcodes and quality trimming , Illumina sequencing yielded average read lengths between 53 and 69 bp with a total of 398 , 326 , and 168 Mb of sequence data for S . noctiflora , S . conica , and S . vulgaris , respectively . Paired-end read mapping was performed with SOAP v2 . 20 [86] with the following parameters: m 100 , x 900 , g 3 , r 2 . A set of custom Perl scripts were used to call SOAP , parse the resulting output , and modify the genome sequence on the basis of well-supported sequence conflicts . These scripts were run recursively until additional iterations did not produce any further improvement to the sequence . For both S . vulgaris and S . noctiflora , Illumina mapping provided high-depth ( >10× ) coverage for essentially the entire genome ( >99 . 9% ) . This process identified 55 sequence corrections in S . vulgaris and 1 , 734 corrections in S . noctiflora , the vast majority of which were associated with homopolymer runs . In contrast , because of the larger size and repetitive complexity of the S . conica mitochondrial genome , more than 10% of the sequence had coverage levels below 10× . Furthermore , the recursive mapping approach described above failed to converge for numerous regions in the genome , indicating low confidence in many of the sequence corrections indicated by the Illumina data . To avoid incorporating false sequence changes , we did not use the Illumina data to perform genome-wide corrections in S . conica . Consequently , the reported genome sequence likely contains some errors associated with homopolymer runs . We did , however , use the Illumina data to verify basecalls in S . conica coding genes and introns , including cases of frameshift mutations . The annotation of protein , rRNA , and tRNA genes was performed using a combination of local BLAST [87] and tRNAscan [88] as described previously [20] . Annotated genome sequences were deposited in GenBank ( Table S2 ) . To identify sequence of plastid origin in the Silene mitochondrial genomes , each genome was searched against a database of seed plant plastid genomes , using NCBI-BLASTN ( v2 . 2 . 24+ ) with the following parameter settings: dust no , gapopen 8 , gapextend 6 , penalty -4 , reward 5 , word_size 7 . Only hits with a raw score of at least 250 were considered . These hits were subsequently filtered to exclude matches involving mitochondrial protein and rRNA genes known to have ancient plastid homologs ( e . g . , mitochondrial atp1 and plastid atpA [89] ) . We also excluded hits with very high AT contents ( >72% ) , because we found these to be almost exclusively false positives resulting from the use of sensitive BLAST parameters . To identify intergenic sequence conserved in other plant mitochondrial genomes , all intergenic regions ( excluding those of plastid origin ) were searched against a database of all sequenced seed plant mitochondrial genomes using NCBI-BLASTN ( v2 . 2 . 24+ ) and the following search parameters: task blastn , dust no , gapopen 5 , gapextend 2 , reward 2 , penalty -3 , word_size 9 . All hits with a raw score of at least 70 were considered homologous . Note that we included all sequences from “empty” chromosomes in the intergenic category even though such sequences are not technically bounded by genes on either side . To identify additional conserved sequences ( particularly ones of nuclear origin ) , the remaining intergenic regions ( i . e . , excluding annotated genes , plastid-derived sequence , and regions conserved with other plant mitochondrial genomes ) were searched against the GenBank nr and nt databases ( release date 12/15/2010 ) using NCBI-BLASTX and BLASTN ( v2 . 2 . 24+ ) . Default settings were used for BLASTX , whereas the BLASTN search parameters were as follows: dust yes , gapopen 5 , gapextend 2 , reward 2 , penalty -3 , word_size 9 . All BLASTX hits with a raw score of at least 140 and all BLASTN hits with a raw score of 70 or above were considered homologous . Searches with these same parameters were also conducted against a set of assembled cDNA sequences from a recent S . vulgaris transcriptome project [51] . Tandem repeats in each Silene mitochondrial genome were identified with Tandem Repeat Finder v4 . 04 [90] , but these represented a negligible fraction of total repeat content in each genome and are not reported separately . Dispersed repeats were identified by searching each genome against itself with NCBI-BLASTN ( v2 . 2 . 24+ ) using default parameter settings . All hits with a raw score of at least 30 were considered repeats . The shortest possible sequence that can satisfy this criterion is a perfect 30-bp repeat , but longer sequences with less than 100% sequence identity can also be identified by this method . Finally , Vmatch ( http://www . vmatch . de ) was used to precisely define the boundaries of all repeats with 100% sequence identity . We used paired-end reads from 454 sequencing to quantify the relative abundance of alternative genome conformations associated with repeat-mediated recombination ( Figure S4 ) . In the absence of any recombination or alternative genome conformations , 454 read pairs should map to positions in the genome that are consistent with the size span of the sequencing library ( ∼3 or 12 kb in this case ) . However , the presence of genomic rearrangements will result in read pairs that are inconsistent with the reported genome conformation ( Figure S4 ) . Therefore , for each pair of repeated sequences in a genome , we quantified the number of 454 read pairs that are inconsistent with the reported genome assembly but are consistent with either of the predicted products of recombination between the repeats . This number was then compared against the total number of consistent read pairs in the genome that span one of the two repeat copies to determine the relative abundance of the recombinant products . To perform this analysis , 454 paired-end reads were mapped on the corresponding genome sequence using Roche's GS Reference Mapper v2 . 3 software with default parameters . For S . noctiflora , only reads from the 12-kb paired-end library were used . The resulting output was filtered to exclude duplicate read pairs with identical start positions for both the left and right sequences , as these were assumed to have been generated by the PCR amplification step in paired-end library construction , making them nonindependent data points . Inspection of the mapping output suggested that the analysis was too stringent in identifying consistent read pairs . Therefore , any “inconsistent” read pairs that mapped in a proper orientation within a distance of 4–16 kb for a 12-kb library or 1–6 kb for a 3-kb library were reclassified as consistent . These size ranges were determined on the basis of manual inspection of the distribution of mapping spans . Identified repeats within each genome ( see above ) were filtered on the basis of multiple criteria prior to inclusion in recombination analyses . First , only repeats of at least 50 bp in length and at least 95% sequence identity were considered . Additional repeat pairs were excluded because their proximity to each other or to other repeats would have led to ambiguity in the interpretation of paired-end mapping results . Specifically , repeats were excluded if the two copies were separated by less than the maximum library span or if there was a “correlated” pair of larger repeats within the maximum library span of each repeat copy . Finally , for S . conica and S . vulgaris ( for which only 3-kb paired-end libraries were available ) , repeat pairs were excluded if one of the repeat copies was within 100 bp of the start of any other repeat >500 bp in size . These cases were excluded because the presence of adjoining repeats would preclude unambiguous mapping of reads to the flanking sequence . Because of the limited physical coverage and short ( 3 kb ) span length in the S . conica paired-end data , there are many repeat pairs ( particularly large repeats ) in this genome that passed the aforementioned criteria , but have an insufficient number of read pairs to precisely measure the relative frequency of alternative genome conformations . Therefore , frequencies are only reported for repeat pairs that have at least five consistent read pairs spanning each copy . Finally , because of the enormous number of small repeats in the S . conica mitochondrial genome ( Figure 5 ) , only a random sample of 5% of repeat pairs shorter than 200 bp was included . To validate our methodological approach , we ran a set of control analyses that used the same set of repeats except that we reversed the coordinates for one of the copies . Therefore , these analyses assessed rearrangements associated with the same genomic regions but would only detect alternative genome conformations if recombination occurred between two homologous sequences lined up in opposite orientations . The frequency of alternative genome conformations was at or near zero for every one of these control analyses ( Figure S6 ) . This suggests that baseline level of genome rearrangement and chimeric artifacts is very low in our dataset and that the alternate genome conformations detected by these methods are the genuine result of repeat-mediated recombination . In addition , the differences in assembly methods across species ( see above ) should have no effect on the reported estimates of recombinational activity because these differences only pertain to large repeats exceeding the span of our paired-end libraries , which were not assayed for recombination . Previous analyses based on individual genes have identified massive variation in mitochondrial substitution rates among genes and species within the genus Silene [35]–[37] , [91] . To assess these patterns at a genome-wide scale , all protein genes were aligned with MUSCLE v3 . 7 [92] and levels of synonymous ( dS ) and nonsynonymous ( dN ) divergence were estimated using PAML v4 . 4 [93] as described previously [37] . Analyses were run both on individual genes and on a concatenated dataset of all shared protein genes . Most analyses included six species ( Arabidopsis thaliana , Beta vulgaris , and all four Silene species ) , but a larger dataset of sequenced seed plant mitochondrial genomes was also analyzed . In all cases , the phylogenetic relationships among the four Silene species were left unresolved ( i . e . , as a four-way polytomy ) , reflecting the apparently rapid radiation of these four lineages [37] , [54] . Because substitutions at RNA editing sites can artificially inflate estimates of dN [94] , we excluded all codons that were found to be edited based on genome-wide datasets from four species [70] , [95] , [96] . To estimate absolute rates of nucleotide substitution in these genomes , dN and dS values were divided by an approximate divergence time of 6 Myr [35] , [37] , [97] . However , these estimates should be considered only rough approximations because of the uncertainty in divergence time [37] and the potential bias associated with recent polymorphisms [98] , [99] . To determine the frequency and size distribution of indels , all protein genes ( including cis-spliced introns ) from the four Silene species and the outgroup B . vulgaris were aligned with MUSCLE v3 . 7 and adjusted manually . Unalignable regions at the 5′ and 3′ ends of genes were excluded . The resulting alignments were analyzed to identify all indels that were unique to a single species and did not overlap with any other indels . A genome-wide analysis of C-to-U RNA editing sites by cDNA sequencing has been reported previously for S . latifolia and S . noctiflora [70] . To estimate the frequency of RNA editing in S . vulgaris and S . conica , protein gene sequences were analyzed with a predictive algorithm ( PREP-mt ) [100] . Control analyses using Silene sequences with known editing sites suggested that different stringency settings ( C-values ) are appropriate for species with different rates of sequence evolution . Specifically , the S . conica data were analyzed with C = 0 . 8 and the S . vulgaris data were analyzed with C = 0 . 7 . PREP-mt does not identify synonymous editing sites , so the reported totals were increased by 10% to approximate the contribution of synonymous edits on the basis of observed rates in other Silene genomes [70] . All intact protein genes were included as well as the following putative pseudogenes: rps13 ( S . latifolia ) , rps3 ( S . conica , S . latifolia , and S . noctiflora ) , and ccmFc ( S . conica ) . For genes with duplicates within the genome , only a single gene copy was included . To estimate levels of sequence variation within each of the four Silene species in this study , we PCR amplified and Sanger sequenced a sample of five mitochondrial loci as well as a single plastid and nuclear locus for multiple , geographically dispersed populations . Sequencing methods , source populations , and polymorphism data for S . vulgaris and S . latifolia were reported previously [36] , [91] . Source populations for S . noctiflora and S . conica are summarized in Table S4 . A single individual was sampled from each population . Sequence data from each species were analyzed with DnaSP v5 [101] to calculate nucleotide diversity and the number of segregating sites for each locus . Maximum likelihood estimates of Watterson's Θ and corresponding 95% confidence intervals were calculated as described previously [91] . For the nuclear X4/XY4 locus , a single haplotype was randomly selected from each individual for calculation of polymorphism data . Only X-linked copies were included for S . latifolia males . Haplotypes were inferred from diploid sequence data using the program PHASE v2 . 1 [102] . Novel sequences were deposited in GenBank ( accessions JF722621–JF722652 ) . We performed a set of greenhouse crosses to test for maternal transmission of mtDNA in S . latifolia , S . noctiflora , and S . conica ( S . vulgaris was not included because it has already been the subject of numerous studies examining mitochondrial genome inheritance and heteroplasmy [44]–[47] ) . Each cross involved an individual from the maternal family used for mitochondrial genome sequencing and an individual from another family in that species known to differ in mtDNA haplotype . For each species , a single pair of reciprocal crosses was performed , and a SNP was used to design a CAPS marker capable of distinguishing the two parental genomes ( Table S5 ) [103] . For each pair of crosses , 16 to 48 progeny were analyzed with the corresponding CAPS marker .
|
A fundamental challenge in evolutionary biology is to explain why organisms exhibit dramatic variation in genome size and complexity . One hypothesis predicts that high rates of mutation in DNA sequence create selection against large and complex genomes , which are more susceptible to mutational disruption . Species of flowering plants in the genus Silene vary by approximately 100-fold in the rates of mutation in their mitochondrial DNA , providing an excellent opportunity to test the predicted effects of high mutation rates on genome evolution . Contrary to expectation , Silene species with elevated mutation rates have experienced dramatic expansions in mitochondrial genome size compared to their slowly evolving relatives , resulting in the largest known mitochondrial genomes . In addition to the increases in size and mutation rate , these genomes also reveal a history of rapid change in genome structure . They have been fragmented into dozens of chromosomes and appear to have experienced major reductions in recombination activity . All of these changes have occurred in just the past few million years . This mitochondrial genome diversity within the genus Silene provides a striking example of rapid genomic change and raises new hypotheses regarding the relationship between mutation rate and genome evolution .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biology",
"genomics",
"evolutionary",
"biology",
"genetics",
"and",
"genomics"
] |
2012
|
Rapid Evolution of Enormous, Multichromosomal Genomes in Flowering Plant Mitochondria with Exceptionally High Mutation Rates
|
This paper demonstrates a previously unreported property of deoxyribonucleic acid—the ability of dye-labeled , solid-state DNA dried onto a surface to detect odors delivered in the vapor phase by changes in fluorescence . This property is useful for engineering systems to detect volatiles and provides a way for artificial sensors to emulate the way cross-reactive olfactory receptors respond to and encode single odorous compounds and mixtures . Recent studies show that the vertebrate olfactory receptor repertoire arises from an unusually large gene family and that the receptor types that have been tested so far show variable breadths of response . In designing biomimetic artificial noses , the challenge has been to generate a similarly large sensor repertoire that can be manufactured with exact chemical precision and reproducibility and that has the requisite combinatorial complexity to detect odors in the real world . Here we describe an approach for generating and screening large , diverse libraries of defined sensors using single-stranded , fluorescent dye–labeled DNA that has been dried onto a substrate and pulsed with brief exposures to different odors . These new solid-state DNA-based sensors are sensitive and show differential , sequence-dependent responses . Furthermore , we show that large DNA-based sensor libraries can be rapidly screened for odor response diversity using standard high-throughput microarray methods . These observations describe new properties of DNA and provide a generalized approach for producing explicitly tailored sensor arrays that can be rationally chosen for the detection of target volatiles with different chemical structures that include biologically derived odors , toxic chemicals , and explosives .
Odor sensor arrays composed of materials that are cross-reactive have advantages over narrowly tuned receptor systems . These include combinatorial responses to sets of compounds that exceed the number of receptor or sensor types , tolerance to partial system failure , and an ability to be flexibly trained , all of which are adaptive attributes that have emerged in biological systems over evolutionary time . In building sensors for artificial noses , there are generally two strategies that are used to develop a diversity of detectors that is reminiscent of biological receptors and thus exploits the advantages of cross-reactive arrays . First is the explicit synthesis of polymers designed to interact with defined chemical properties of a target volatile compound ( for example , fluorescent conjugated polymers designed to detect nitroaromatic explosives ) [1–3]—these sensors are commonly designed to be highly specific . Second is the use of available off-the-shelf polymers to test for responses to target vapor phase compounds without prior knowledge of how their chemical properties generate different odor sensations . This second method has been used to make cross-reactive , polymer-based sensors in a number of array-based electronic nose devices , including the one we have developed [4–14] . Both methods have been effective for producing useful devices; however , the sensors tend to be labor intensive to build , yielding relatively small numbers of detector candidates . A more ideal class of sensing molecules would be one which , in addition to responding to volatiles , has the following properties: ( 1 ) has a chemistry that provides a large combinatorial complexity of structure , ( 2 ) has a molecular structure amenable to being replicated exactly in order to make large amounts of identical material , and ( 3 ) provides the opportunity to screen for sensors that can thereby be tailored to respond to many different volatile chemicals . Sensors in artificial noses have not yet come close to achieving the sensor diversity and complexity found in biological olfactory systems [15] . As an alternative to the classes of synthetic polymers usually used for electronic nose sensors , we hypothesized that combinations of fluorescent dyes and a bio-polymer ( namely , DNA ) could be used as a sensing material . DNA is an attractive candidate for such use because of its stability and its potential for tremendous combinatorial complexity . Furthermore , once appropriate molecular sequences are identified , large numbers of identical DNA-based sensing molecules can be easily made using standard methods of synthesis . As described below , we have tested both double-stranded ( dsDNA ) and single-stranded DNA ( ssDNA ) using two methods of staining with fluorophores , and we report here the proof of concept demonstrating that ssDNA shows sequence-specific responses to a variety of volatile compounds and that libraries of these materials can be efficiently screened for odor responses . Portions of this work have been previously presented in abstract form [16–18] .
In an initial test of this hypothesis , we constructed sensors of dsDNA using a standard 2 . 9-kb pBlueScriptSK plasmid mixed with the intercalating DNA dye YO-PRO ( Molecular Probes ) , dried onto a polyethylene substrate material , and tested in our electronic nose device [13 , 14] ( see Materials and Methods ) . YO-PRO alone showed no changes in fluorescence over the time course of brief ( 1 . 6 s ) odor sniffs ( Figure 1A ) applied by negative pressure in our artificial nose device . In contrast , sensors made from plasmid mixed with YO-PRO and dried onto the substrate produced large and rapid decreases in fluorescence upon exposure to propionic acid , with smaller or no changes when exposed to water , methanol , and triethylamine ( Figure 1B ) . These odor responses were relatively stable over repeated trials ( see Figure 2C for ssDNA ) . Responses from sensors made from dsDNA differing significantly in primary sequence ( unpublished data ) , however , were qualitatively similar to those from sensors made from the pBlueScriptSK DNA ( Figure 1B ) . We also constructed hairpin 33mer sequences that were much smaller than the plasmids described above and of about the size of the ssDNA described below , also stained with YO-PRO . These constructs of complementary G-C or A-T pairs ( see sequences DS001 and DS002 in Table S1 ) , which were designed to hybridize over a distance of 15 base pairs , all responded similarly , giving sequence-independent responses to the same odor set tested on the plasmid ( unpublished data ) . These tests showed that , at least for these dsDNA constructs , sequence did not govern odor response . Because different dsDNA sequences did not show odor responses that were modified by changes in sequence , we then tested sensors made from short , ssDNA oligomers stained with the fluorescent dye OliGreen ( Molecular Probes ) . OliGreen dye alone showed a decrease in fluorescence upon exposure to propionic acid , but little change to the other odors tested ( Figure 2A ) . In contrast , SEQ01 ( 22 bases long; see sequence descriptions for SEQ01–SEQ30 in Table S1 ) stained with OliGreen and dried onto a polyethylene substrate had a markedly different odor response profile ( Figure 2B ) from OliGreen alone and from those using dsDNA and YO-PRO . This sensor sequence showed an increase in fluorescence in response to propionic acid and methanol ( to which OliGreen alone did not respond ) , with relatively little change to the other odors tested . Furthermore , as shown in Figure 2C , the propionic acid signals from this sensing material returned to baseline levels between successive odor applications delivered as often as every 10 s , indicating that the odor–sensor interaction was reversible with a short time constant . In these tests , OliGreen and YO-PRO were applied to the DNA in solution , prior to drying onto a substrate for testing . In applying the dyes in this way , there is little control over how and where the fluorophore binds to the DNA . To define better the dye–nucleotide interaction explicitly , we generated labeled oligonucleotides ( 20–24 bases long ) by covalently attaching the fluorescent dye Cy3 ( Amersham Biosciences ) to the 5′ end during synthesis . We then adapted microarray techniques to screen these potential sensors for odor responses . An odor test chamber was constructed having the dimensions of a microscope slide to allow its use in a standard microarray scanner for examining the vapor responses from libraries of sensor sequences . Cy3-labeled oligonucleotides were spotted ( ∼50 μm diameter ) onto cover slips , which were then mounted with the spots facing the interior volume of the test chamber ( see Materials and Methods ) . Vapor-phase test odors were then injected into the chamber prior to scanning . To measure odor responses using this method , we first tested control arrays in which the same DNA-Cy3 construct ( SEQ02 ) was spotted at all locations . The responses of 30 replicates of the same SEQ02 construct ( rows ) to saturated vapors of eight odors ( columns ) are shown in Figure 3A ( increases in fluorescence over baseline indicated by graded red colors and decreases indicated by graded blue colors ) . The responses of the replicated spots in this control array were highly correlated and therefore considered to be essentially identical . Pearson correlation coefficients calculated between pairs of sensors were all ≥0 . 90 ( see Figure 3A , legend ) . This high degree of correlation is also represented by the compact cluster analysis dendrogram shown to the left of the data matrix in Figure 3B ( see [19] for description of Pearson correlation coefficients and cluster analysis that are the standard methods applied to microarray data ) . In contrast to the correlated responses from spots having the same sequence , odor response data from 29 different DNA-Cy3 sequences ( Table S1 ) showed dramatic response differences ( Figure 3B ) . Using a conservative correlation coefficient threshold of 0 . 90 ( the minimum pairwise correlation from analysis of the SEQ02 control array ) , cluster analysis of these data indicates there are at least 10 discriminable sensor types in this small sensor library . Three clusters had Pearson correlation coefficients >0 . 90 ( clusters to the right of the dashed line in Figure 3B ) with others having much lower correlations . The response variety obtained with this small array is clearly sequence-dependent and strongly suggests that large numbers of different sensor types are likely to be found when larger numbers of DNA-Cy3 sensor candidates are tested . It is important to note that the sensors in this sample respond to a variety of compounds with different chemical structures and that all of the odor response profiles , with the possible exception of SEQ06 , were distinctly different from that of Cy3 dye alone ( Figure 3B , top row ) . In addition to response diversity , the sensitivities of a number of DNA-Cy3 sensors were also tested in our electronic nose device ( similar to the tests shown in Figures 1 and 2 ) using an air-dilution olfactometer to deliver ranges of controlled odor concentrations . The dynamics of the responses of individual DNA-Cy3 sensors to sniffs of odorant were similar to those shown in Figure 2 for ssDNA stained with OliGreen . The concentration-response functions of different sequences for a small odorant test set summarize sensor concentration dependence as well as response diversity . For example , SEQ02 ( Figure 4A ) showed responses to propionic acid and triethylamine at lowest concentrations of 4 and 75 ppm , and to methanol , 2 , 4-dinitrotoluene ( DNT; found in the vapor signature of TNT-containing landmines ) [20–22] , and dimethyl methylphosphonate ( DMMP; precursor to Sarin nerve gas ) at ∼33 , 900 ppm , ∼6 ppb , and ∼30 ppm , respectively . In contrast , SEQ03 ( Figure 4B ) responded to triethylamine at ∼75 ppm , to DMMP at ∼30 ppm , and showed no response to propionic acid , methanol , or DNT . The DNT response of SEQ02 ( 2 × 10−2 of DNT saturation is ∼6 ppb , or 2 . 3 × 10−10 M; see [23] for vapor pressure ) indicates that these sensors are capable of detecting certain compounds with low vapor pressures at low concentrations . With the exception of polymers that are specifically synthesized for detecting nitroaromatic compounds [1 , 3] , the DNA-based sensors described here are the only fluorescent polymeric sensor materials of which we are aware that show significant responses to DNT . It should be noted that the responses used to plot Figure 4 consist of both positive- and negative-going changes in fluorescence , but the signs of these responses are not represented in these graphs , which are intended only to show the gradation of response over concentration ranges . It is also important to note that the responses of those sequences ( e . g . , SEQ02 and SEQ03 ) that have been tested both in the array scanner and in our electronic nose device show some differences , which can be due to the following: ( 1 ) the substrates onto which they were deposited ( glass versus silkscreen ) ; ( 2 ) the concentration of odor used for testing ( saturated versus graded series with maximum concentration of 1/10 dilution ) ; ( 3 ) the duration of odor exposure ( an extended time of minutes versus 1 . 6 s ) ; and ( 4 ) the buffers used for deposition . We know from other experiments that each of these variables influences response .
The solid-state , DNA-based vapor sensors described here have two key properties that are important for use in electronic noses: ( 1 ) diverse and broad odor response profiles , ( Figure 3B ) and ( 2 ) high sensitivity ( Figure 4 ) . These response properties , combined with the intrinsically combinatorial nature of DNA , indicate that using DNA in this way provides a major new approach toward designing multitudes of sensors for detecting and identifying volatile compounds . For example , using 21mers yields 421 or more than a trillion different sequences . Further , these data describe a new property of dried , solid-state ssDNA—namely the ability to interact with vapor phase substances . The mechanisms underlying these responses are likely to be complex and we are currently in the process of trying to elucidate them . Briefly , the data we have so far indicate that fluorescence changes related to odor responses arise from minute changes in the three-dimensional structure of the dried ssDNA within the hydration shell that envelopes this hydroscopic molecule at normal relative humidities ranging from about 30%–70% [17 , 18] . It is clear that these solid-state , DNA-based odor sensors are distinctly different from other nucleotide-based sensing materials , such as aptamers . Although aptamers have been used to detect some kinds of small , non-nucleotide ligands , these interactions have all been carried out in aqueous solution [24–26] , not with the oligomers dried onto surfaces , and they have not been used to detect vapor-phase compounds as described here . Furthermore , solid-state , DNA-based sensing materials for cross-reactive sensor arrays can be explicitly selected for nuanced differences in specificity and response breadth , whereas aptamers are selected for optimal “monospecificity” . Analysis of sensor array and olfactory receptor properties that we have carried out using information theory [27 , 28] has shown that the breadth of cross-reactivity and the number of different sensor ( or receptor ) types are crucial parameters for describing how odor discrimination is carried out in both biological and electronic olfactory systems . Using the methods detailed here , sensors that are required to have a particular response breadth and sensitivity for a defined odor detection problem can be explicitly chosen from DNA libraries . Using standard microarray equipment and the odor test chamber described above , it is possible to rapidly screen thousands or hundreds of thousands of DNA-based sensors . A number of methods are available for generating large sensor libraries , including direct synthesis ( e . g . , lithographic [29] and ink-jet [30] methods ) and PCR amplification of random libraries . It is therefore feasible to combine large-scale , DNA-based sensor library generation with high-throughput screening . This provides a means of selecting sensors appropriate for detecting volatile compounds related to many real-world problems in security , industrial quality control , environmental monitoring , and medical diagnosis [31–33] . Based on our early results published in abstract form [16] , there have been two reports of DNA coated onto single-walled , carbon nanotube field effect transistors [34 , 35] in which sequences from our abstract [16] were used to detect vapor phase compounds . This method illustrates another way of potentially reading out odor-related responses from DNA .
Our electronic nose [13 , 14] uses an array of optically based chemical ( polymer ) sensors that change their fluorescence intensity upon exposure to brief pulses of vapor-phase compounds . The device contains 16 sensors that can be illuminated and observed at 16 different excitation and emission wavelengths . The sensors and the optical elements for illuminating and monitoring the sensors are placed along a narrow chamber through which target odors are drawn . Excitation light is produced by filtered LEDs providing bandpass wavelengths appropriate for the sensors being used ( 450 nm , bandwidth ( BW ) 40 for OliGreen and YO-PRO; 540 nm , BW 40 for Cy3 ) . The fluorescent sensors emit light at longer wavelengths which passes through filters ( 550 nm , BW 70 for OliGreen and YO-PRO; 600 nm , BW 10 for Cy3 ) and is monitored by photodiodes . Thus , there is an LED/bandpass filter/sensor/bandpass filter/photodiode set for each sensor channel . Electric current produced by each photodiode is converted to voltage , amplified , and digitized ( every 100 ms ) over the time of the sniff duration ( 1–2 s ) at nominal 20-bit resolution . The digital signals are recorded and processed by an embedded microprocessor . User control of the device is via a touch screen panel . The current device is self-contained , hand-held , weighs about 1 . 5 kg , and is about the size of a facial tissue box . DNA was diluted to the desired concentration ( 0 . 2–50 ng/μl ) in TE ( 10 mM Tris base , 0 . 5 mM EDTA , pH 8 . 0 ) or Tris-NaCl ( 2 mM Tris base , 11 mM NaCl , pH 8 . 0 ) . Twenty μl of dilute DNA was mixed with 1 μl stock solutions of YO-PRO ( 1:40 in Tris-NaCl ) or OliGreen ( 1:1 in Tris-NaCl ) and incubated at room temperature for 5 min . Dye-only controls were made of 1 μl dye stock in 20 μl TE or Tris-NaCl . DNA/dye mixtures were applied to a substrate of acid-washed 16xx polyethylene silkscreen ( 10 mm × 12 mm ) and allowed to dry for 25 min . Each sensor was rinsed in 70% ethanol for 5 min , allowed to dry , and then attached to supports on glass cover slips for testing in the electronic nose device . The dried DNA material adhered to the silkscreen mesh without occluding the openings . The exact thickness is unknown , but based on superficial appearance , it simply forms a thin film stuck to the strand supports that make up the silkscreen mesh . An air-dilution olfactometer of standard design and modeled after a system used in dog studies [36] was used to deliver controlled dilutions of odors to the electronic nose device . Odor dilutions were determined by the ratio of flow rates through the clean air and saturated odor channels . Total flow rate was 10 l/min . Twenty-nine ssDNA sequences labeled with Cy3 on the 5′ end were synthesized using standard phosphoramidite chemistry . The constructs were reconstituted and diluted into buffer ( 10 mM Tris , 50 mM NaCl , pH 8 . 0 ) at a concentration of 4 μM . Sensor constructs were then spotted onto clean 22 × 60 mm cover slips using a BioRobotics MicroGrid II . A chamber was constructed for testing sensor array responses to odors in a Packard ScanArray 4000 microarray scanner by milling a stainless steel blank to have the outer dimensions of a standard microscope slide ( 1 mm thick × 25 mm wide × 75 mm long ) , which the scanner will accept . A rectangular hole slightly smaller ( 20 mm × 57 mm ) than a 22 × 60 mm cover slip was cut through the center of the blank , and shoulders were cut around the hole on each side to accommodate two 22 × 60 cover slips , which , when placed on the shoulders , created an interior volume of approximately 1 cc . Three edge holes were drilled from one end of the blank into the closed volume of the chamber formed by the cover slips in order to inject odors via a 22 gauge hypodermic needle . A blank cover slip was taped into the bottom shoulder , and a sensor array cover slip was taped into the top with the DNA spots facing the interior of the chamber . For odor testing , 30-cc glass syringes containing KimWipes saturated with different chemical compounds ( or containing crystalline solid , as in the case of DNT ) were used to inject odor vapor into the test chamber immediately before scanning . Ten cc of vapor was injected into each of the three edge holes and allowed to escape through the remaining holes , leading to a 30-fold exchange of chamber air . After an odor test , the chamber was opened for 15 min to allow the odor to escape . Clean humidified air was injected before each odor test to maintain constant chamber humidity . For the data analysis shown in Figure 3 , fluorescence changes are expressed relative to a clean air control presented in the same manner as an odor test . We used epifluorescence microscopy and video imaging to confirm that fluorescence changes in sensor spots after odor injection into the chamber were stable for at least two minutes , which is the time required for scanning the full sensor array . For the data shown in Figure 3 , sensor spot fluorescence was quantified using Dapple ( http://www . cs . wustl . edu/~jbuhler/research/dapple ) . Data were then analyzed using Cluster 3 . 0 ( http://bonsai . ims . u-tokyo . ac . jp/~mdehoon/software/cluster/software . htm ) and visualized using Java TreeView ( http://jtreeview . sourceforge . net ) .
|
Biological systems can provide engineering guidance on how evolution has solved particular problems . In the context of detecting chemicals in either the aqueous or vapor phase , two general biological approaches have emerged . The first relies on individual highly specific single receptors ( sensors ) , each tuned to detect a single molecular species—examples include the receptors that mediate pheromone detection in insects or those that function in neurotransmission . Specificity is achieved by narrow band design . The second approach is implemented by arrays of receptors with relatively broad responses . In this case , specificity emerges from a constellation of receptor types that recognizes the molecule of interest—the canonical example here is the olfactory receptors in the main olfactory system of vertebrates . Specificity is achieved by a “one chemical–many broadly responsive detectors” paradigm . While trying to mimic the enormous odor coding ability of biological olfaction in an “artificial nose , ” we searched for molecules with the requisite combinatorial capacity to serve as odor detectors . Here we show that single-stranded DNA molecules tagged with a fluorescent reporter and deposited onto solid surfaces can respond to vapor phase odor pulses in a sequence-selective manner . These findings demonstrate new properties of nucleotide molecules that can be exploited in engineered odor detection devices . In addition , this broadband responsivity to small molecules should be explored as a functional aspect of DNA ( and RNA ) as they exist in the normal cellular milieu .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biotechnology",
"chemistry",
"neuroscience",
"molecular",
"biology"
] |
2008
|
Solid-State, Dye-Labeled DNA Detects Volatile Compounds in the Vapor Phase
|
Coordinated cardiomyocyte growth , differentiation , and morphogenesis are essential for heart formation . We demonstrate that the bHLH transcription factors Hand1 and Hand2 play critical regulatory roles for left ventricle ( LV ) cardiomyocyte proliferation and morphogenesis . Using an LV-specific Cre allele ( Hand1LV-Cre ) , we ablate Hand1-lineage cardiomyocytes , revealing that DTA-mediated cardiomyocyte death results in a hypoplastic LV by E10 . 5 . Once Hand1-linage cells are removed from the LV , and Hand1 expression is switched off , embryonic hearts recover by E16 . 5 . In contrast , conditional LV loss-of-function of both Hand1 and Hand2 results in aberrant trabeculation and thickened compact zone myocardium resulting from enhanced proliferation and a breakdown of compact zone/trabecular/ventricular septal identity . Surviving Hand1;Hand2 mutants display diminished cardiac function that is rescued by concurrent ablation of Hand-null cardiomyocytes . Collectively , we conclude that , within a mixed cardiomyocyte population , removal of defective myocardium and replacement with healthy endogenous cardiomyocytes may provide an effective strategy for cardiac repair .
The left ventricle ( LV ) of the heart drives systemic circulation . Because the LV must be large enough to support adequate cardiac output but not hypertrophic or hypoplastic , such that it obstructs blood flow , congenital heart defects ( CHDs ) and acquired diseases that impact LV morphology or cell number present a significant cause of morbidity and mortality in the human population [1] . These CHDs include left ventricular noncompaction or hypertrabeculation ( LVNC; OMIM: 604169 ) , which is a cardiomyopathy characterized by prominent trabeculations occluding the ventricular lumen and associated with a high risk of heart failure and sudden death [2 , 3] and hypoplastic left heart syndrome ( HLHS; OMIM: 614435 ) , which presents an underdeveloped LV unable to sustain sufficient blood flow [4 , 5] . From the perspective of adult cardiac disease , especially cardiomyopathies , it is of particular importance to establish the capacity of proliferating cardiomyocytes to replace dead or dysfunctional cardiomyocytes [1] . Critical breakthroughs in patient outcomes demand a better understanding of the etiology of cardiac growth , differentiation and morphogenic patterning . The embryonic heart forms from two distinct cardiomyocyte progenitor populations , termed the primary ( PHF ) and secondary ( SHF ) heart fields , which give rise to the LV and right ventricle ( RV ) , respectively [6] . The bHLH transcription factor Hand1 is predominantly expressed within the LV myocardium , with minimal expression in SHF derivatives [7] . Interestingly , HAND1 mutations have been identified in HLHS patients [8]; however , cardiac-specific Hand1 ablation in mice does not recapitulate HLHS [9] . Previous studies suggest that the related bHLH transcription factor Hand2 can functionally cooperate with Hand1 during murine LV development [9] . Although Cre-loxP technology has been used to great effect in mice to genetically model SHF myocardium defects , the genetic tools to specifically interrogate genetic loss-of-function in LV cardiomyocytes have , thus far , not been available . We have isolated a conserved 744 bp enhancer 5’ to the Hand1 transcription start site that is sufficient to drive reporter and Cre recombinase gene expression specifically within the LV . Hand1LV-Cre recapitulates endogenous Hand1 expression within an estimated 80–90% of LV cardiomyocytes between embryonic stages ( E ) E8 . 5-E13 . 5 . Ablation of the Hand1-lineage LV myocardium results in a markedly smaller LV by E10 . 5; however , LV chamber size and adult cardiac function is ultimately rescued via proliferation of non-Hand1-lineage LV cardiomyocytes . In contrast to this LV cell ablation model , Hand1;Hand2 loss-of-function within the LV results in increased cardiomyocyte proliferation resulting in morphogenic defects that lead to the occlusion of the LV lumen . LV cardiomyocytes invade the LV chamber and show mis-expression of compact zone , trabeculae , and intraventricular septum ( IVS ) restricted genes . We reveal a cooperative Hand factor function that is required to morphogenetically specify subpopulations of ventricular myocardium , and to thereby regulate cardiomyocyte growth . Functional analysis of surviving Hand1;Hand2 double conditional knockouts ( CKOs ) reveals impaired LV function that can be rescued by ablating the mutant Hand1 LV-lineage cells . Taken together , these data suggest that , due to its inherent proliferative capacity , the developing LV tolerates myocardial cell death , whereas developmentally abnormal cardiomyocytes adversely impact cardiac function .
We have identified a cis-regulatory element ( s ) that drives gene expression specifically within the LV , and not within other cardiac tissues ( Vincentz and Firulli , manuscript in preparation ) . We used this Hand1 enhancer to make an LV-specific Cre driver . Lineage analysis using the Hand1eGFPCre knock-in allele revealed that Hand1-lineage cells are restricted to the LV myocardium and to a ring of SHF-derived myocardium occupying the OFT , termed the myocardial cuff [7 , 10] . Using the 2 . 7kb Hand1 basal promoter to provide the eGFPCre with a transcriptional start site , we cloned the Hand1 LV-enhancer 5’ ( Fig 1A ) and generated several F0 transgenic lines . In two of these lines , Cre activity , as assessed via the R26RlacZ reporter allele , was detectable within the forming LV at E9 . 0 ( Fig 1E; white arrow ) specifically marking only the Hand1 LV cardiomyocyte-lineage ( Fig 1E , 1G , 1I , 1K and 1M ) . Hand1-lineage epicardium ( black arrowhead; Fig 1H and 1I ) and OFT tissues ( Fig 1J and 1K; black arrows ) , detectable from the Hand1eGFPCre allele , were not observed in the Hand1LV-Cre transgenic . Immunohistochemistry using β-galactosidase to mark Cre-lineage cells and Mlc2v to mark ventricular cardiomyocytes showed total co-localization of these two markers ( S1A , S1C , S1E and S1G Fig ) , whereas expression of β-galactosidase and PECAM , an endothelial/endocardial marker , appears mutually exclusive ( S1A , S1C , S1E and S1G Fig ) . Together , these data validate that this Hand1LV-Cre driver recombines specifically within LV cardiomyocytes . To address the importance of the Hand1LV lineage during cardiac development , we utilized the conditionally active Rosa26 ( R26R ) Diphtheria Toxin A chain ( R26RDTA ) allele to conditionally ablate Hand1-expressing LV cardiomyocytes . Analysis of cell death via TUNEL on E9 . 5 sections revealed that the LVs of Hand1LV-Cre; R26RlacZ/+ control embryos displayed few apoptotic cells ( Fig 2A ) , whereas Hand1LV-Cre; R26RlacZ/DTA LVs contained scattered TUNEL-positive cells ( Fig 2B , quantified in Fig 2C ) . However , given the robust LV expression of the Hand1LV-Cre , fewer TUNEL-positive cells than would be predicted were detected in E9 . 5 R26RDTA-ablated LVs . Additional TUNEL staining , ( Fig 2D ) and whole mount lysotracker staining at E10 . 5 revealed markedly increased cell death within the LV of Hand1LV-Cre; R26RlacZ/DTA embryos when compared to controls ( white arrows Fig 2E and 2F ) . At both time points , no difference in cell death was observed in the RV myocardium ( Fig 2A–2D , 2G and 2H ) , and cell death in the pharyngeal arches served as a positive control ( white arrowheads Fig 2A , 2B and 2E–2H ) . We then bred the R26RlacZ reporter onto the R26RDTA allele to monitor the loss of Hand1-expressing LV cardiomyocytes in our ablation model . At E9 . 5 , the Hand1-LV lineage is largely absent from the heart , but the LV is not grossly hypoplastic ( Fig 2J ) . By E10 . 5 , the lack of Hand1-LV lineage cells results in a markedly hypoplastic LV ( Fig 2L ) . At E12 . 5 , the LVs of Hand1LV-Cre; R26RlacZ/DTA embryos were visibly smaller than Hand1LV-Cre; R26RlacZ/+ controls ( Fig 2M and 2N ) and showed few X-gal-stained cells when compared to littermates that do not carry the DTA allele . Interestingly , these mid-gestation LVs , from which the Hand1-lineage has been ablated , do not display perturbed expression of the LV markers Nppa and Gja5 ( S2 Fig ) . Thus , ablation of Hand1-lineage LV myocardium between E8 . 5 , when the Hand1LV enhancer is first upregulated , and E12 . 5 , when the Hand1LV enhancer begins downregulation , results in a hypoplastic LV . To follow the biological impact of Hand1-LV lineage ablation , we assayed the phenotypes of Hand1LV-Cre; R26RlacZ/DTA and control embryos at E16 . 5 . To our surprise , Hand1LV-Cre; R26RlacZ/DTA hearts were indistinguishable in size from controls that lacked the DTA allele ( Fig 2O and 2P; arrows ) . X-gal staining confirmed that embryos that contain the DTA allele exhibit minimal Hand1-lineage cells within their LVs . Although adult cardiomyocytes are refractory to reentering the cell cycle [1 , 11–13] , embryonic and neonatal cardiomyocytes retain proliferative potential [14–16] . To confirm that the restoration of LV size was simply due to an enhanced cardiomyocyte proliferation from non-Hand1-lineage LV myocardium , phospho-Histone-H3 immunostaining was carried out at E12 . 5 and E14 . 5 ( Fig 2Q–2U ) . Increased proliferation was observed within the LVs of Hand1LV-Cre; R26RlacZ/DTA hearts , compared to controls , at E14 . 5 , but not at E12 . 5 ( Fig 2U ) . Hand1LV-Cre; R26RlacZ/DTA mice were born at Mendelian ratios ( S3A Fig ) and were viable and fertile . These hearts lacked gross structural cardiac abnormalities , exhibiting only a subtle rounding of the free LV wall ( S3B and S3C Fig ) . These hearts exhibited cardiac function , as assayed by ejection fraction ( EF ) , fractional shortening ( FS ) , and other echocardiographic parameters , that was indistinguishable from Cre-negative controls ( S3D–S3O Fig ) , and within the normal range for adult mice under isoflurane anesthesia [17] . Taken together , these data demonstrate that embryonic ablation of the Hand1-lineage is not sufficient to permanently disrupt LV development . As we have established that the Hand1-LV myocardial lineage is not required for cardiogenesis , we next investigated whether the expression of Hand1 and Hand2 within the embryonic LV is required for normal heart development . Although early embryonic expression analysis shows that the majority of Hand2 expression within the heart is restricted to the endocardium , epicardium , and SHF derived myocardium [10 , 18 , 19] , at later embryonic stages , Hand2 mRNA becomes detectable within E11 . 5 LV myocardium in a pattern overlapping with Hand1 ( S4C and S4D Fig ) . We subsequently generated compound heterozygous Hand1LV-Cre;Hand1fx/+;Hand2fx/+ male mice and crossed them to Hand1fx/fx;Hand2fx/fx females to generate Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx offspring ( Fig 3 ) . Hand1LV-Cre;Hand1fx/+;Hand2fx/+ ( Fig 3A–3D ) embryos undergo normal cardiac development and are indistinguishable from wild type controls . Similarly , embryos that delete Hand2 from the LV but retain a single copy of Hand1 ( Hand1LV-Cre;Hand1fx/+;Hand2fx/fx ) also exhibit largely normal cardiac development ( Fig 3E–3H ) . Interestingly , consistent with previous studies [9] , deletion of Hand1 from the LV ( Hand1LV-Cre;Hand1fx/fx ) results in a morphologically disorganized LV , wherein Hand1-lineage cells appear to localize more to trabeculations compared to the compact zone ( Fig 3I–3L ) . At E17 . 5 , abnormal cardiomyocytes localize within the LV lumen ( Fig 3L ) and ventricular septal defects are also observed ( S5H Fig ) . Deletion of Hand1 from the LV leaving a single copy of Hand2 ( Hand1LV-Cre;Hand1fx/fx;Hand2fx/+ ) results in a similar phenotype ( Fig 3M–3P , S5J–S5L Fig ) . Complete deletion of Hand factors from the LV ( Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx ) caused a pronounced internalization of the Hand1-lineage cells where X-gal staining in whole mount is noticeably opaque and in section reveals Hand1-lineage cells occluding the LV lumen ( Fig 3Q–3S ) . Histological analysis at E17 . 5 revealed severely occluded LV lumen and poorly formed IVS ( Fig 3T , S5M–S5O Fig ) , as well as a double outlet right ventricle ( black arrowheads , S5N Fig ) , and hyperplastic mitral valves ( S5O Fig ) , although the aorta and OFT valves are phenotypically indistinguishable from controls ( S5M Fig ) . These phenotypes are summarized in Table 1 . These data suggest that a Hand factor loss-of-function within the LV myocardium alters chamber morphology via a Hand gene dosage dependent mechanism , in which Hand1-lineage cells are found less frequently in the compact zone at the expense of increased cells within the LV lumen representing trabecular and papillary muscle cardiomyocytes . Despite the morphological defects observed in Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx embryos , Hand loss-of-function mice survive perinatally at approximately 50% of the expected Mendelian ratio ( Table 2 ) . Indeed , although Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx pups are underrepresented , this underrepresentation is not statistically significant ( Table 2 ) . Echocardiographic analysis of P56 Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx survivors ( S6 Fig ) revealed that systolic function ( fractional shortening and ejection fraction ) is significantly compromised in Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx mice ( Fig 4 ) , whereas other measures of LV morphology and function , such as chamber dimensions and wall thickness , were not significantly altered ( S7 Fig ) . We next sought to characterize the etiology of the luminal cardiomyocyte overgrowth in Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx embryos . To this end , we performed marker analyses to characterize distinct subpopulations of cardiomyocytes in the heart . The following expression studies revealed no appreciable difference between ( - ) ;Hand1fx/fx;Hand2fx/fx , Hand1LV-Cre ( + ) ; Hand1fx/+;Hand2fx/+ , and Hand1LV-Cre ( + ) ;Hand1fx/+;Hand2fx/ fx embryos . For ease of presentation , Hand1LV-Cre ( + ) ; Hand1fx/+;Hand2fx/+ embryos are presented as controls . Tbx20 and Hey2 expression marks compact and IVS myocardium [20 , 21] . Section in situ hybridization of E11 . 5 hearts showed that Tbx20 and Hey2 are both expressed throughout the presumptive compact myocardium and excluded from the trabecular myocardium in control hearts ( Fig 5A and 5B ) . This sharp delineation between compact and trabecular myocardium is lost in the LVs of Hand1LV-Cre ( + ) ; Hand1fx/fx;Hand2fx/+ hearts ( Fig 5E and 5F; asterisks ) . In Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx hearts , most of the trabeculae within the LV ectopically express compact myocardium markers , whereas the RV trabeculae do not ( Fig 5I and 5J ) . Conversely , the trabecular markers Bmp10 and Nppb show robust expression throughout the trabeculae , and expression is largely excluded from the compact myocardium in control hearts ( Fig 5C and 5D ) . In Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx hearts , Bmp10 and Nppb expression is downregulated within the LV trabeculae when compared to RV trabeculae ( Fig 5K and 5L; asterisks ) . We conclude that Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx hearts display ectopic compact myocardium marker expression and reduced trabecular marker expression within the LV . By Carnegie Stage 16 in the human embryo ( equivalent to E11 . 5 in mice ) proliferation in the ventricular trabeculae has declined [22] . We next sought to correlate differences in the proliferative capacity of compact and trabecular myocardium with the overgrowth seen in Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx LVs . Mki67 marks cells actively undergoing cell cycle , but is excluded from cells in G0 [23] . Mki67 immunohistochemistry revealed that , in control embryos , nuclear Mki67 is robustly detected within compact myocardium and the IVS , but is largely excluded from trabecular myocardium ( Fig 5M–5P ) . In contrast , Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx trabecular cardiomyocytes show robust Mki67 staining ( Fig 5Q–5T; arrowheads ) . These findings indicate that cardiomyocyte proliferation is dysregulated within the LV of Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx embryos . Previous studies have reported that ectopic HAND1 expression throughout the heart disrupts IVS formation [24] . We reasoned that the abnormal proliferation and marker expression seen in Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx hearts may reflect aberrant ventricular septogenesis . Section in situ hybridization of E11 . 5 hearts showed that expression of the chemokine Cxcl12 is strong in the free walls of the ventricles , but is largely excluded from the IVS ( Fig 6A ) . In both Hand1LV-Cre;Hand1fx/fx;Hand2fx/+ . and Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx hearts , LV Cxcl12 expression is excluded specifically from the trabecular myocardium , in addition to the left side of the ventricular septum ( Fig 6D and 6G ) . The secreted Wnt inhibitor Dkk3 and the transcription factor Irx2 are both markers of the IVS [25] . Expression of both Dkk3 ( Fig 6H; black arrowhead ) and Irx2 ( Fig 6I; black arrow ) expands toward the atrioventricular canal in Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx hearts . Together , these data indicate that a loss of both Hand1 and Hand2 function within the LV causes an expansion of the IVS gene expression program . Given that the embryonic heart can recover from DTA-mediated ablation of Hand1-lineage cardiomyocytes , we tested whether ablation of Hand1;Hand2-null cardiomyocytes can rescue the phenotypes associated with their loss-of-function . The R26RDTA allele was bred onto a Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx background . E14 . 5 section in situ hybridization of control hearts ( Fig 7A–7C ) showed expected expression patterns for Bmp10 , Irx2 , and Dkk3 within the trabeculae and IVS respectively . As expected Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx hearts showed reduced Bmp10 LV expression along with expanded Dkk3- and Irx2 LV cardiomyocyte expression ( Fig 7E–7G ) . These changes in gene expression are reversed in Hand1fx/fx;Hand2fx/fx;Hand1LV-Cre ( + ) ;R26RDTA embryos ( Fig 7I–7K ) . Morphological examination of X-gal-stained bisected P56 hearts revealed heart morphology indistinguishable from controls ( Fig 7D , S8A Fig ) , in contrast to the overgrowth of Hand1-lineage cardiomyocytes that occlude the lumen of Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx hearts ( Fig 7H , X-gal stained myocardium; S8B Fig ) . Occluded LV lumens are absent in Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx;R26RDTA embryos ( Fig 7L , S8C and S8D Fig ) . Lineage tracing reveals that the majority of lacZ-positive cells are ablated in these hearts; however , consistent with DTA-ablation embryos ( Fig 2P ) , small populations of lacZ-positive cells were sometimes detectable within these hearts ( S8D Fig ) , which we interpret as cells that have recombined only the lacZ reporter , but not the DTA allele . Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx;R26RDTA pups survive at Mendelian ratios ( Table 3 ) and display restored ejection fraction ( Fig 7M ) and fractional shortening ( Fig 7N ) . Again , other measures of cardiac function were not significantly altered ( S8E–S8J Fig ) . Together , these data demonstrate that ablation of mutant cardiomyocytes from the developing LV restores cardiac function .
CHDs that alter the LV exhibit poor clinical outcomes [5 , 26] . The inability to interrogate LV gene expression in isolation has limited the ability to understand the molecular mechanisms that drive LV morphogenesis . This study reports the generation of a novel LV-restricted Cre driver line that allows for such focused interrogation . First , we ablated Hand1-lineage cardiomyocytes within an E9 . 0-E13 . 5 developmental window . As expected , LV size was greatly reduced by E10 . 5 , and this reduced size remains clear at E14 . 5 ( Fig 2 ) . Hand1 cardiac expression is downregulated by E13 . 5 [27] . We observed a significant increase in LV cardiomyocyte cell proliferation at E14 . 5 ( Fig 2U ) . By E16 . 5 , LV size is indistinguishable from control hearts ( Fig 2O and 2P ) . These findings support published data showing that embryonic and up to 7-day postnatal cardiomyocytes retain regenerative potential [14–16] , and further demonstrate that the Hand1-lineage can be ablated from the developing LV , and the heart can nonetheless undergo full regenerative repair . It is also clear that these replacement cardiomyocytes do not express Hand1—if they did , Cre would also be expressed , thereby activating DTA expression and killing the replacement cell . That said , conclusive identification of the origin ( s ) of the cells that replace the ablated Hand1-lieage cells will require further study employing , for example , dual lineage tracing systems . A candidate for this progenitor population is the SHF . One of the most well studied markers of the SHF is the Isl1-Cre [28] . Although it is excluded from the majority of LV cardiomyocytes , Isl1-Cre-lineage cells do appear in the LV [28–31] , and it is possible that this relatively minor population expands in the absence of Hand1-lineage cells . Indeed , this would explain why early stage DTA-ablation embryos , despite almost completely lacking Cre-lineage cells , are grossly similar to non-ablated littermates ( Fig 2I and 2J ) . If this were the case , it would indicate that SHF-derived cardiomyocytes are sufficient to replace PHF-derived cardiomyocytes . Regardless of potential contributions from other cardiomyocyte lineages , including the SHF , these observations indicate that Hand1-lineage cells are not required for cardiac development , and that , through elevated proliferation , a Hand1-negative fated cardiomyocyte population within the developing heart is sufficient to generate a functional LV . In contrast , Hand1 and Hand2 are required for normal LV morphogenesis . By E11 . 5 , both Hand genes are expressed within the LV myocardium ( S4 Fig ) . Loss of Hand1 , and , more severely , Hand1 and Hand2 , leads to abnormal LV morphology characterized by expanded compact and IVS marker expression ( Tbx20 , Hey2 , Dkk3 , Irx2; Figs . 5 and 6 ) and correspondingly reduced trabecular marker expression . An increase in Mki67-positive nuclei ( Fig 5M–5T ) within the LV trabecular zone indicates that aberrant cell proliferation causes the observed LV hyperplasia . Taking these observations together , we conclude that the expanding cardiomyocyte population observed in Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx mutants is derived from myocardium that has ectopically activated the IVS gene expression program . The Hand1-lineage marks very little of the IVS [7] . Proliferation in Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx mutant hearts is not restricted to this small domain of Hand1-expressing IVS cells on the inner wall near the cardiac apex . This suggests that some Hand1-lineage compact zone myocardium adopts an IVS cell-like fate . The observation that the Hand1-lineage within Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx mutant hearts is largely excluded from the LV compact zone , and is instead localized to the inner curvature supports this finding . These results are especially interesting in light of recent studies , which propose that LVNC likely results from abnormal growth of compact myocardium [32] . It would be informative to test whether IVS markers also expand into luminal domains in such models . In addition to LVNC , these Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx mutants share certain phenotypic similarities with HLHS patients; however , they lack the aortic valve phenotypes characteristic of HLHS ( S5M Fig ) . A recently published mouse model of HLHS [33] posits a digenic etiology of HLHS , in which dysfunction of one gene disrupts cardiomyocyte proliferation and differentiation to cause LV hypoplasia , while a second mutation causes aortic valve abnormalities . Importantly , these findings provide evidence that the aortic phenotypes are not secondary to the LV phenotypes , and therefore , the lack of Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx aortic defects is not surprising . As mentioned , a subset of HLHS patients displays HAND1 mutations [8] . Hand1 is expressed in the OFT [27] as well as the LV , and as such could have pleiotropic functions in both aortic valve and LV development; however , ablation of Hand1 in the neural crest progenitors of the aortic valves is not pathogenic [34] . Given that this study reports the unexpected finding that Hand1 and Hand2 have overlapping roles in LV development , it would be of interest to reevaluate Hand1 function in the OFT considering potential functional redundancy with Hand2 . Increased cardiomyocyte proliferation can negatively impact specification [14 , 15] . The Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx LV hyperplasia occludes so much of the LV lumen that it results , in severe cases , in a single ventricle phenotype ( Fig 3T , S5N Fig ) . In spite of the significant LV hyperplasia , 50% of Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx mutants survive to birth and live to adulthood . These surviving individuals exhibit compromised systolic function ( Fig 4 ) . It is clear that both Hand1 and Hand2 contribute to this phenotype , and that Hand1 plays a more significant role . Hand1LV-Cre;Hand1fx/+;Hand2fx/fx mutants display no observable cardiac phenotypes , whereas Hand1LV-Cre;Hand1fx/fx;Hand2+/+ and Hand1LV-Cre;Hand1fx/x;Hand2fx/+ mutants show both morphological and molecular changes in gene expression . Given that Hand2 LV expression is not detectable until E11 . 5 , and its LV expression is not dependent upon Hand1 , Hand1 mutants are less severe likely due to the later expression of Hand2 . Finally , we observe a functional rescue of Hand1LV-Cre;Hand1fx/fx;Hand2fx/fx mutants when mutant cells are ablated via co-activation of DTA expression ( Fig 7 ) . Given that heart development does not require the Hand1-lineage ( Fig 2 ) this may not be surprising . Nevertheless , this observation suggests that removal of molecularly abnormal cardiomyocytes from a developing heart could be beneficial if molecularly normal cells are present and are still proliferative .
The Indiana University Transgenic and Knock-Out Mouse Core generated the Hand1LV-Cre transgenic mouse line on a C3HeB/FeJ background . Genotyping of the Hand1tm2Eno , Hand2tm1Cse , Gt ( ROSA ) 26Sortm1 ( DTA ) Jpmb , and Gt ( ROSA ) 26Sortm1Sor alleles has previously been described [9 , 10 , 35 , 36] . These mice were maintained on a mixed C57Bl/6;129S background . Embryos were not selected for sex , and were evaluated blindly for all analyses . Mice and other reagents are available from the authors upon request . To generate the Hand1LV-Cre transgene , the genomic sequence corresponding to chr11:57660605–57661348 ( in the mm9 assembly ) was cloned 5’ to a modified eGFP-Cre vector driven by the basal Hand1 2 . 7kb promoter . Mice were genotyped for the Cre allele either via Southern blot or PCR using the primers Cre ( F ) 5’-CGTACTGACGGTGGGAGAAT-3’ and Cre ( R ) 5’-TGCATGATCTCCGGTATTGA-3’ , with the internal controls Smad4 ( F ) 5’-TAAGAGCCACAGGGTCAAGC-3’ and Smad4 ( R ) 5’-TTCCAGGAAAAACAGGGCTA-3’ . X-gal staining was performed as previously described [7 , 19 , 37 , 38] . Cell death analysis on control and mutant embryos was performed as described [39] . Lysotracker ( Life Technologies ) was incubated with embryos as per the manufacturer's instructions . Embryos were imaged in a well slide on a Leica DM5000 B compound florescent microscope . TUNEL analyses were performed upon sectioned embryos using the ApopTag Plus Fluorescein in situ Apoptosis detection kit ( S7111 Chemicon International ) as per manufacturer’s instructions . TUNEL-positive cells occupying the free wall of the LV and RV ( including the myocardial outflow tract ) were counted every other section . Significance was determined by student’s t-test . Immunohistochemistry was performed as previously described [19 , 40–42] . β-galactosidase ( Aves BGL-1010; 1:200 dilution ) , Mlc2v ( Synaptic Systems 310 111; 1:500 ) , PECAM ( BD Pharmingen 550274; 1:200 ) , Phospho-Histone H3 ( Abcam 4797; 1:500 ) , and Mki67 ( Dako; 1:500 ) antibodies were used with biotinylated secondary antibodies and streptavidin-conjugated DyLight 488 or 594 fluorophores ( ThermoFisher ) . Images were collected on a Leica DM5000 B microscope and Leica Application Suite software . Cell proliferation was assayed via counting of phospho-histone H3 positive nuclei . Left and right ventricles from E12 . 5 and E14 . 5 immunostained images were manually isolated in Adobe Photoshop , and nuclei were manually counted using Image J software . For E12 . 5 , n≥9 sections , and for E14 . 5 , n≥10 sections per heart were counted . Total samples analyzed were as follows: E12 . 5 , control embryos ( Hand1LV-Cre ( - ) ; R26RlacZ/DTA ) –n = 2 , RV , 35057 cells , total , LV , 35250 cells , total; LV-ablated embryos ( Hand1LV-Cre ( + ) ; R26RlacZ/DTA ) –n = 5 , RV , 81471 cells , total , LV , 67618 cells , total; E14 . 5 , control embryos ( Hand1LV-Cre ( - ) ; R26RlacZ/DTA or Hand1LV-Cre ( + ) ; R26RlacZ/+ ) –n = 3 , RV , 76254 cells , total , LV , 85121 cells , total; LV-ablated embryos–n = 4 , RV , 167897 cells , total , LV , 80894 cells , total . Mice were lightly anesthetized with mixture of 1% to 1 . 5% isoflurane and 100% oxygen while supine on a heated platform . The heart rate were stabilized at 400 to 500 beats per minute before image recording . Images were obtained with a high resolution Micro-Ultrasound system ( Vevo 2100 , VisualSonics Inc , Toronto , Canada ) equipped with a 40-MHz mechanical scan probe . Two-dimensional images were recorded in the parasternal long and short-axis to guide M-mode recordings in the mid-ventricular level . LV systolic function was computed from M-mode measurement according to the recommendations of American Society of Echocardiography committee [43] . Animal work was approved by the Indiana University School of Medicine Animal Care and use committee ( IACUC ) via protocol 10809 Issued to ABF .
|
The left ventricle of the heart drives blood flow throughout the body . Impaired left ventricle function , associated either with heart failure or with certain , severe cardiac birth defects , constitutes a significant cause of mortality . Understanding how heart muscle grows is vital to developing improved treatments for these diseases . Unfortunately , genetic tools necessary to study the left ventricle have been lacking . Here we engineer the first mouse line to enable specific genetic study of the left ventricle . We show that , unlike in the adult heart , the embryonic left ventricle is remarkably tolerant of cell death , as remaining cells have the capacity to proliferate and to restore heart function . Conversely , disruption of two related genes , Hand1 and Hand2 , within the left ventricle causes cells to assume the wrong identity , and to consequently overgrow and impair cardiac function . Ablation of these mutant cells rescues heart function . We conclude that selective removal of defective heart muscle and replacement with healthy cells may provide an effective therapy to treat heart failure .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"cell",
"death",
"medicine",
"and",
"health",
"sciences",
"cardiovascular",
"physiology",
"in",
"situ",
"hybridization",
"molecular",
"probe",
"techniques",
"cardiovascular",
"anatomy",
"cardiac",
"ventricles",
"cell",
"processes",
"developmental",
"biology",
"molecular",
"biology",
"techniques",
"embryos",
"research",
"and",
"analysis",
"methods",
"myocardium",
"embryology",
"probe",
"hybridization",
"gene",
"expression",
"molecular",
"biology",
"anatomy",
"cell",
"biology",
"physiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"heart"
] |
2017
|
Hand factor ablation causes defective left ventricular chamber development and compromised adult cardiac function
|
We examined relationships between schistosome infection , HIV transmission or acquisition , and all-cause death . We retrospectively tested baseline sera from a heterosexual HIV-discordant couple cohort in Lusaka , Zambia with follow-up from 1994–2012 in a nested case-control design . Schistosome-specific antibody levels were measured by ELISA . Associations between baseline antibody response to schistosome antigens and incident HIV transmission , acquisition , and all-cause death stratified by gender and HIV status were assessed . In a subset of HIV- women and HIV+ men , we performed immunoblots to evaluate associations between Schistosoma haematobium or Schistosoma mansoni infection history and HIV incidence . Of 2 , 145 individuals , 59% had positive baseline schistosome-specific antibody responses . In HIV+ women and men , baseline schistosome-specific antibodies were associated with HIV transmission to partners ( adjusted hazard ratio [aHR] = 1 . 8 , p<0 . 005 and aHR = 1 . 4 , p<0 . 05 , respectively ) and death in HIV+ women ( aHR = 2 . 2 , p<0 . 001 ) . In 250 HIV- women , presence of S . haematobium-specific antibodies was associated with increased risk of HIV acquisition ( aHR = 1 . 4 , p<0 . 05 ) . Schistosome infections were associated with increased transmission of HIV from both sexes , acquisition of HIV in women , and increased progression to death in HIV+ women . Establishing effective prevention and treatment strategies for schistosomiasis , including in urban adults , may reduce HIV incidence and death in HIV+ persons living in endemic areas .
Of the more than 200 million persons who have schistosomiasis worldwide , more than 90% live in Africa , and the disease causes tens to hundreds of thousands of deaths annually [1] . Of the parasitic diseases , the health impact of schistosomiasis is second only to malaria [2] . Common species in sub-Saharan Africa are Schistosoma haematobium and Schistosoma mansoni , which cause urogenital and intestinal schistosomiasis , respectively [2] . Infection is generally established during childhood and is more common among those living in rural areas with frequent freshwater contact [1] . However , urbanization trends have increasingly brought schistosomiasis to municipal areas including Lusaka , the capital of Zambia [3 , 4] . Like many of the neglected tropical diseases , preventive chemotherapy using mass drug administration is the primary strategy for controlling schistosomiasis . Though praziquantel treatment is inexpensive , effective , and safe [5] , only 14% of adults and 54% of school-aged children estimated to need preventive chemotherapy in 2016 received treatment [6] . Schistosome eggs not passed in urine or stool become lodged in tissues and induce granulomas that cause the pathology associated with schistosomiasis which persists after the death of the egg . For example , in female genital schistosomiasis ( FGS ) , deposition of eggs can occur in the cervix , vagina , and/or vulva putting women at risk for genital epithelial bleeding , sandy patches on the cervix and throughout the genital tract [7 , 8] , and vaginal inflammation at the histopathological level [9 , 10] . These lesions and inflammation can persist long after eggs are deposited , irrespective of treatment [11 , 12] . Like other infections that generate a local immune response and/or cause genital lesions ( e . g . , ulcerative and non-ulcerative sexually transmitted infections ( STI ) caused by syphilis , herpes simplex virus ( HSV ) , trichomonas , gonorrhea , and chlamydia ) , schistosomiasis may increase the risk of HIV infection [13 , 14] . Cross-sectional studies show associations between urogenital schistosomiasis and HIV prevalence [15] . Though relatively few longitudinal studies have been published , the literature supports the hypothesis that urogenital schistosomiasis is a risk factor for HIV acquisition in HIV- persons and a risk factor for HIV transmission and disease progression in those co-infected with HIV [16–19] . Because it may increase risk of HIV transmission or acquisition , treating schistosomiasis could be a highly cost-effective HIV prevention strategy , and the World Health Organization has called for more studies examining schistosomiasis in HIV endemic countries [15 , 20–22] . In Zambia , a country with high prevalence of both HIV ( 13% , [23] ) and schistosomiasis ( 5–40% [3 , 24 , 25] ) , we retrospectively analyzed data from urban adults enrolled in a longitudinal HIV discordant couple cohort to test the hypothesis that there is a relationship between a person having schistosome-specific antibodies ( reflecting either active or previous infection , with potential for residual sequelae ) and transmitting HIV , acquiring HIV , and all-cause death . We also describe the effect of infecting schistosome species on HIV acquisition among a sub-set of female HIV- and male HIV+ cohort participants .
The Emory University Office for Human Research Protections-registered Institutional Review Board ( IRB ) and the Zambian Ethics Committee approved this study . The University of Zambia Biomedical Research Ethics Committee ( IORG0000774 ) is registered with the US Office of Human Research Protection ( IRB00001131 ) . Written informed consent was obtained from participants , all of whom were adults . The Centers for Disease Control and Prevention ( CDC ) also reviewed the protocol; CDC investigators were not considered to be engaged with study participants as they had no direct contact with them and no access to identifying information . Heterosexual HIV discordant couples ( M+F- and M-F+ ) were enrolled in an open cohort with longitudinal follow-up every three months in Lusaka , Zambia between 1994 and 2012 . Participants were identified through couples’ voluntary HIV counseling and testing ( CVCT ) . Study recruitment [26] , enrollment , retention and attrition [27] , HIV testing and counseling procedures [28 , 29] , and cohort demographics [30] have been previously published . Briefly , CVCT included group educational sessions , rapid HIV antibody testing , and joint post-test couples’ counseling . HIV serodiscordant heterosexual couples who voluntarily enrolled in the open cohort were provided with free outpatient care including STI testing/treatment and family planning . Couples were censored upon antiretroviral treatment initiation , death of either partner , or relationship dissolution . Data on demographics ( including age , years cohabiting , monthly income , and literacy in Nyanja , the most commonly spoken language in Lusaka Province ) , family planning , and clinical characteristics ( including pregnancy , baseline HIV stage and viral load of HIV+ partners , male partner circumcision status , and STI history ) were collected . Past year STI history included self-reported gonorrhea , chlamydia , trichomonas , syphilis , and HSV-2 diagnoses . Viral load was not collected before 1999 . Genital abnormalities assessed included discharge or inflammation on visual genital exam ( including speculum exam for women ) . Trichomonas , bacterial vaginosis ( BV ) , and candida infections in women were detected by microscopy of vaginal wet mount swabs ( and additional whiff test for BV ) ; genital ulceration ( including cervical/vaginal erosion or friability ) ; and syphilis diagnosed via rapid plasma regain ( RPR ) ( BD Macro-Vue , Becton-Dickinson Europe ) , with Treponema pallidum hemagglutination assay confirmation when available [31] . HSV-2 infections were diagnosed by serology testing with the highly sensitive test Focus Diagnostics HerpeSelect 2 ELISA IgG , [32] , with repeat testing for indeterminate results . HIV incidence was measured via testing of HIV- partners every 1–3 months using rapid antibody tests [29] . When available , plasma from the last antibody negative sample was tested by p24 enzyme-linked immunosorbent assays ( ELISA ) and RNA polymerase chain reaction ( PCR ) . Based on available data , date of infection was defined as the minimum of: the midpoint between the last negative and first positive antibody test; two weeks prior to a first antigen positive test; or two weeks prior to a first viral load positive/antibody negative test [33] . Date of death was reported by study partners or other family members; >90% of deaths among HIV+ persons were HIV-related [34] and thus death is a proxy for HIV disease progression . Outcomes of interest were relationships between schistosome-specific antibody positivity and time-to incident: HIV acquisition ( incident HIV infection in a previously HIV- partner ) , HIV transmission ( onward HIV transmission from an HIV+ index partner ) , and all-cause death . Our analysis is limited to HIV infections genetically linked to the HIV+ study partner determined via comparison of PCR-amplified conserved nucleotide sequences ( gag , gp120 , gp41 , long terminal repeat regions ) between partners [35] . Blood plasma and serum samples were collected at enrollment from all couples and stored in a repository at Emory University . In 2010 , we retrospectively tested plasma samples from individuals enrolled between 1994 and 2009 for antibodies to schistosome soluble worm antigen preparation ( SWAP ) using a previously described ELISA [36] . All seroconvertors with available samples were included along with a random sample of non seroconvertors in this nested case-control deisgn . ELISA data from each complete discordant pair were not available for some couples . To ensure consistency between plates , a standard 1:3 serial dilution curve was prepared and included on each plate . A 4-parameter curve fitting model was used to assign units based on the standard curve to each unknown sera . The positive cutoff value ( 25 units ) was set at three standard deviations above the average anti-SWAP IgG in serum from egg negative controls from the US and Europe [36] . A positive schistosomiasis result was defined as having a positive SWAP antibody response . A coded list of individuals positive for schistosome-specific antibody was sent to the Director of the Lusaka research site ( author WK ) , and those individuals were offered free praziquantel treatment . Immunoblot testing using species-specific antigens was used to distinguish between S . haematobium and S . mansoni antibodies in women and men in a random sample of individuals from the nested case-control study where males were HIV positive and females were negative ( M+F- ) at baseline [37] . Descriptive statistics ( counts and percentages ) described the distribution of schistosomiasis ELISA antibody responses stratified by gender , HIV status , and baseline characteristics , with differences evaluated using Chi-square ( or Fisher’s exact ) tests for categorical variables and Student’s t-tests for continuous variables . Unadjusted associations between baseline SWAP ELISA results and outcomes of interest were estimated from Cox survival models; crude hazard ratios ( cHRs ) , 95% confidence intervals ( CIs ) , and two-tailed p-values are reported . Adjusted Cox survival models were created and adjusted hazard ratios ( aHRs ) were calculated adjusting for factors associated ( p<0 . 05 ) with both the exposure and outcome of interest ( the ‘confounding triangle’ method ) . We also applied a second strategy for adjustment by exploring all subsets of potential confounders to look for meaningful ( +/-10% ) differences in adjusted hazard ratios . In the subset of individuals in M+F- couples with species-specific immunoblot results , we similarly ran unadjusted and adjusted analyses assessing the association between either S . haematobium or S . mansoni and HIV incidence and death . In this subset analysis , we considered the potential for confounding by the other species . All analyses were performed with SAS 9 . 4 ( Cary , NC ) .
Of 2 , 145 individuals tested by SWAP ELISA , 59% were positive for anti-schistosome antibodies at baseline ( 25% had ELISA levels >70 units , 34% had 25–70 units ) , and 41% were negative ( <25 units ) . Schistosome-specific antibody levels were higher in men than women ( p<0 . 0001 ) . This difference was driven by a much higher number of males than females with ELISA levels >70 units ( 31% of men vs . 19% of women ) while the frequencies for 25-<50 and 50–70 units were similar for men and women . There were no differences in the distribution of ELISA results when stratifying by sex and HIV status simultaneously ( Table 1 ) . Women with positive schistosome-specific ELISA results were less likely to be pregnant at baseline ( p<0 . 0001 ) . Positive schistosome-specific ELISA results were also associated with signs and symptoms of genital conditions ( non-ulcerative ) of STI or non-STI etiologies ( p<0 . 001 ) and genital ulcers ( p = 0 . 003 ) in HIV+ women . HIV+ women who were positive for schistosome-specific antibodies were more likely to be at an advanced HIV disease stage III-IV ( p = 0 . 001 ) . In addition , women who were positive for schistosome-specific antibodies were more likely to have a male partner that was also schistosome-specific antibody positive ( p<0 . 01 ) . However , a woman’s age , duration of cohabitation , household income , literacy , and viral load were not statistically significantly associated with schistosome-specific ELISA status . Similarly , number of prior pregnancies , baseline HSV-2 antibody status , or baseline RPR results were not associated with schistosome-specific ELISA status ( data not tabled ) . In the subset of women with information on fertility intentions , tribal/linguistic group , or where they lived prior to age 16 ( which was only collected after 2002 ) , there was also no association with schistosome-specific ELISA status ( data not tabled ) . Most data were missing for fertility intentions ( 72% ) , tribal/linguistic group ( 69% ) , or where they lived prior to age 16 ( 81% ) . Unlike in women , presence of schistosome-specific antibodies in HIV+ men were not related to HIV disease stage but viral loads were higher in men with positive schistosome-specific antibody status ( p = 0 . 025 ) . Interestingly , positive schistosome-specific antibody status was associated with an increased likelihood of being circumcised among HIV+ men ( p = 0 . 004 ) and increased literacy ( p = 0 . 029 ) . Among HIV- men , genital conditions ( non-ulcerative ) ( p = 0 . 062 ) and genital ulcer ( p = 0 . 067 ) were not statistically significantly associated with a positive schistosome-specific antibody result . In addition , men who were positive for schistosome-specific antibodies were more likely to have a female partner that was also schistosome-specific antibody positive ( p<0 . 01 ) . Men’s age , duration of cohabitation , household income , literacy , and reported history of STI in the last year were not associated with schistosome-specific antibody responses . Nor was baseline HSV-2 antibody status , baseline RPR results , or ( in a subset of men with the following information which was only collected after 2002 ) fertility intentions , tribal/linguistic group , or where they lived prior to age 16 associated with schistosome-specific antibody status ( data not tabled ) . Most data were missing for fertility intentions ( 75% ) , tribal/linguistic group ( 72% ) , or where they lived prior to age 16 ( 83% ) . We observed N = 70/296 and N = 70/300 linked HIV transmission outcomes for baseline schistosome-specific antibody positive and negative HIV+ women , respectively . We observed N = 92/275 and N = 97/228 linked HIV acquisition outcomes for baseline schistosome-specific antibody positive and negative HIV- women , respectively . In unadjusted analyses , HIV+ women positive for schistosome-specific antibodies at baseline had a shorter time-to-HIV transmission to HIV- male partners ( cHR = 1 . 8 , p = 0 . 002 ) . When multivariate analyses were performed controlling for genital conditions ( non-ulcerative ) and genital ulceration of the woman , positive baseline schistosome-specific antibody in HIV+ women remained associated with shorter time-to-HIV transmission to HIV- male partners ( aHR = 1 . 8 , p = 0 . 002 ) . When applying a different strategy for adjustment ( exploring at all subsets of potential confounders to look for meaningful ( +/-10% ) differences in adjusted hazard ratios ) , we arrived at the same set of confounders as when using the present ( ‘confounding triangle’ ) method . Because of the nested case-control design , proportions of outcomes should not be calculated/interpreted as risk estimates . We observed N = 112/381 and N = 78/218 HIV transmission outcomes for baseline schistosome-specific antibody positive and negative HIV+ men , respectively . We observed N = 93/309 and N = 53/138 HIV acquisition outcomes for baseline schistosome-specific antibody positive and negative HIV- men , respectively . In unadjusted analyses , HIV+ men positive for schistosome-specific antibodies at baseline had a shorter time-to-HIV transmission to HIV- female partners ( cHR = 1 . 4 , p = 0 . 016 ) . When multivariate analyses were performed controlling for men’s viral load , positive baseline schistosome-specific antibody in HIV+ men remained associated with shorter time-to-HIV transmission to HIV- female partners ( aHR = 1 . 4 , p = 0 . 042 ) . When applying a different strategy for adjustment ( exploring at all subsets of potential confounders to look for meaningful ( +/-10% ) differences in adjusted hazard ratios ) , we arrived at the same set of confounders as when using the present ( ‘confounding triangle’ ) method . Because of the nested case-control design , proportions of outcomes should not be calculated/interpreted as risk estimates . We observed N = 65/296 and N = 44/300 death outcomes for baseline schistosome-specific antibody positive and negative HIV+ women , respectively . We observed N = 17/275 and N = 11/228 death outcomes for baseline schistosome-specific antibody positive and negative HIV- women , respectively . In unadjusted analyses , HIV+ women positive for schistosome-specific antibodies at baseline had a shorter time-to-women’s death ( cHR = 2 . 3 , p<0 . 0001 ) . When multivariate analyses were performed controlling for HIV stage of the woman , positive baseline schistosome-specific antibody in HIV+ women remained associated with shorter time-to-women’s death ( aHR = 2 . 2 , p<0 . 001 ) . When applying a different strategy for adjustment ( exploring at all subsets of potential confounders to look for meaningful ( +/-10% ) differences in adjusted hazard ratios ) , we arrived at the same set of confounders as when using the present ( ‘confounding triangle’ ) method . When adjusting for viral load of the HIV+ woman ( as was done in the model for men , below ) , the point estimate is slightly tempered but still significant ( aHR = 1 . 97; 95%CI:1 . 19–3 . 25 , p-value = 0 . 008 ) ( data not tabeled ) . We observed N = 110/381 and N = 56/218 death outcomes for baseline schistosome-specific antibody positive and negative HIV+ men , respectively . We observed N = 21/309 and N = 13/138 death outcomes for baseline schistosome-specific antibody positive and negative HIV- men , respectively . In unadjusted analyses , HIV+ men positive for schistosome-specific antibodies at baseline were associated with decreased time-to-men’s death ( cHR = 1 . 6 , p = 0 . 008 ) . Baseline schistosome-specific antibody levels were not significantly associated with time-to-HIV+ men’s death once adjusting for men’s viral load . However , viral load was not collected before 1999 , thus only 126 of 166 death events were modeled when controlling for viral load . When applying a different strategy for adjustment ( exploring at all subsets of potential confounders to look for meaningful ( +/-10% ) differences in adjusted hazard ratios ) , we arrived at the same set of confounders as when using the present ( ‘confounding triangle’ ) method . Among 250 HIV- women and 239 HIV+ men , we observed N = 63/76 and N = 144/174 HIV acquisition outcomes for S . haematobium positive and negative HIV- women , respectively . We observed N = 40/46 and N = 167/204 HIV acquisition outcomes for S . mansoni positive and negative HIV- women , respectively . We observed N = 64/80 and N = 132/159 HIV transmission outcomes for S . haematobium positive and negative HIV+ men , respectively . We observed N = 44/54 and N = 152/185 HIV transmission outcomes for S . mansoni positive and negative HIV+ men , respectively . In M+F- HIV discordant couples , species-specific schistosome antibody response were evaluated by species-specific immunoblot results . S . haematobium-specific antibodies ( aHR = 1 . 4 , p = 0 . 034 ) significantly increased the risk HIV acquisition in women , while S . mansoni-specific antibodies also increased risk of HIV acquisition , though not significantly ( aHR = 1 . 3 , p = 0 . 12 ) . The schistosome species infecting men did not significantly influence the likelihood of viral transmission to their HIV- partners . Notably , we had very limited power to detect the difference between S . mansoni status and HIV acquisition in HIV- women ( 3% ) or S . mansoni status and HIV transmission from HIV+ men ( 19% ) . When applying a different strategy for adjustment ( exploring at all subsets of potential confounders to look for meaningful ( +/-10% ) differences in adjusted hazard ratios ) , we arrived at the same set of confounders as when using the present ( ‘confounding triangle’ ) method . Because of the nested case-control design , proportions of outcomes should not be calculated/interpreted as risk estimates .
Our findings indicate a high prevalence of antibodies to schistosomes that was associated with onward HIV transmission from HIV+ men and women and earlier death in HIV+ women . S . haematobium specific immunoblot reactivity was associated ( p = 0 . 034 ) with increased HIV acquisition risk in HIV- women . In this study , 59% of participants were positive for schistosome-specific antibodies , indicating an unexpectedly high prevalence of current or past infection in our urban population . Past infection can lead to persistent residual sequelae , which may increase HIV risk irrespective of treatment [11 , 12] . Men had higher ELISA values at baseline than women , possibly explained by the finding that more men than women had lived in a rural area prior to age 16 . Rural areas in Zambia have a higher prevalence of schistosomiasis , and most infections are acquired during youth [39] . The high observed prevalence of schistosome-specific antibodies in this urban population necessitates a shift in thinking of schistosomiasis as only a disease of children and rural areas . More research is needed to examine urbanization and migration in schistosomiasis transmission , especially in endemic countries . For example , individuals in Lusaka without access to effective water delivery systems may rely on both piped and environmental water sources , which might increase their risk of acquiring schistosome infection [25] . Past or baseline schistosome infection in HIV+ partners was significantly associated with onward HIV transmission to HIV- partners . A possible explanation is increased viral load in the HIV+ partners . In co-infected HIV+ men , schistosome egg excretion in semen may be accompanied by increases in lymphocytes , eosinophils [40] , and other cells associated with HIV replication . Furthermore , HIV binding receptors ( CCR5 and CXC4 ) are denser on CD4 T-cell surfaces in men with intestinal schistosomiasis compared to uninfected individuals and those treated with praziquantel [41] . In co-infected HIV+ women , genital schistosomiasis may lead to inflammatory genital lesions that increase transmissibility of HIV [13] . We also observed that S . haematobium ( but not S . mansoni ) infection as determined by immunoblot was associated increased risk of HIV acquisition by HIV- women in discordant relationships . This finding is consistent with the association of urogenital schistosomiasis and FGS and is similar , though of lesser magnitude , to results from studies in rural Zimbabwe [13] and Tanzania [14 , 42] that showed schistosomiasis was associated with a 2–3 fold increased HIV risk in women . However , prior studies did not rely on antibody status indicating past or current infection but rather active infection , limiting direct comparisons with our study . FGS may increase susceptibility to HIV due to cervical lesions reducing the integrity of the genital epithelial barrier [7 , 10 , 13 , 43] or recruitment of CD4+ lymphocytes , macrophages , and Langerhans giant cells [10] to the genital tract , thus increasing the probability of HIV infection [9] . However , most previous research regarding FGS and inflammation has been at the histopathological level without biopsy [9 , 10] . A previous randomized controlled trial detected no effect of active schistosome infection on time-to-a composite indicator of HIV disease progression ( first occurrence of a CD4 count of <350 cells/ml , first reported use of antiretroviral treatment , and non-traumatic death ) [44] . Conversely , a longitudinal study in Tanzania found that individuals with active schistosome infection at the time of their HIV seroconversion had slower HIV disease progression ( as indicated by CD4 count of <350 cells/ml or death ) [45] . The authors suggest that this unexpected finding indicated complicated interactions between long-term HIV immunological changes and schistosome co-infections , and that additional studies are needed . Furthermore , a 2016 Cochrane Review found scant , “low quality” evidence that treating helminth infections has beneficial effects on slowing HIV disease progression [46] . By contrast , our study is the first to investigate schistosome-specific antibody status , reflecting either past or current infection , as a factor associated with mortality in HIV+ individuals . Interestingly , we also observed an association between schistosome-specific antibody responses and increased baseline HIV stage in women and viral load in men . More research will be needed to confirm these findings . Women who were not pregnant at baseline were more likely to be schistosome-specific antibody positive . Previous research , including case reports , ecological studies , descriptive series , and geographical mapping , has indicated an association between S . haematobium infection and decreased fertility in sub-Saharan Africa [47 , 48] . In a cross-sectional interview study in Kenya , documented treatment for childhood urogenital S . haematobium among women was associated with decreased fertility in adulthood [49] , further highlighting the importance of primary prevention for urogenital schistosomiasis and early treatment . We also found that schistosome-specific antibody positive responses were associated with male circumcision ( significant association for HIV+ men , with a trend for HIV- men ) . To explain this finding , we evaluated circumcision by education , occupation , income , tribal/linguistic group , and whether the man lived in an urban versus rural area before the age of 16 . However , none of these factors attenuated the association between circumcision and ELISA results ( data not tabled ) . This finding warrants further exploration . Given the high prevalence of a history of schistosomiasis in our study , the associated risk of HIV transmission and death among HIV+ persons positive for schistosome-specific antibody , and other studies indicating an increased risk of HIV transmission/acquisition due to schistosome infection , our findings underscore the importance of schistosomiasis treatment and prevention . Further , we argue for integration of routine parasitological testing and treatment in HIV programs . Treating schistosomiasis has been proposed as a cost-effective addition to HIV prevention and treatment programs , and may contribute to slowing the spread of HIV while reducing schistosomiasis-associated morbidity [21] . Praziquantel treatment for schistosomiasis is safe , including in pregnant women [5 , 16] , has no reported widespread drug resistance , has only moderate side-effects , and can be dispensed via community-wide mass administration [20] . Additionally , praziquantel may attenuate HIV replication by decreasing systemic inflammation [50] and slow HIV disease progression [19 , 46 , 50 , 51] . HIV+ individuals with delayed schistosomiasis treatment have increased viral loads and lower CD4 T-cell counts compared to those who received early treatment [19] , and antihelminthic drugs may act to reduce viral load and increase CD4 levels [51] . In resource-limited countries such as Zambia , more efforts are needed to train health care providers , including HIV practitioners , to detect schistosomiasis and administer praziquantel . It will also be important for policy makers to consider the cost-effectiveness of new methods to detect FGS morbidity versus long-term programmatic benefits . Our study has limitations . Though ELISA tests can be highly specific and robust laboratory measures of antibodies to schistosome infection [52] and the use of immunoblots to detect schistosomiasis species has been validated for S . mansoni [53 , 54] and S . haematobium [55] , it is possible that some of the participants were misclassified . In a study in Western Kenya , the SWAP ELISA had a sensitivity of 92% and specificity of 57% when compared to fecal egg microscopy , which only detects active infection . The lower specificity of the ELISA is likely associated with identifying people who were previously infected and/or the relatively lower sensitivity of the parasitological method [36] . ELISA cutoff values would have been more compelling if we had known negative controls from Zambian samples . Furthermore , it would have been informative if we had performed sensitive diagnostic procedures , including antigen detection or schistosomiasis PCR , to delineate active infections . Nevertheless , our data support the growing appreciation that the sequelae of schistosomiasis persist beyond the period of active infection . Colposcopy or cervical biopsy to confirm schistosome eggs in urogenital tissue would have enabled definitive diagnosis of FGS although the latter approach has ethical concerns due to creating a genital tract wound in a population of HIV positive or HIV serodiscordant couples . Schistosomal antibodies were only measured at baseline ( along with several other covariates such as viral load ) , due to funding constraints , and time-varying measures would have been informative to explore how titers changed over time . Type-specific immunoblot testing was only done in women and men who were in M+F- partnerships ( primarily to test the hypothesis that female schistosomiasis infection was associated with risk of HIV acquisition in women ) , and it would have been informative to have species-specific immunoblot data for men and women in M-F+ partnerships . We do not know whether the significantly elevated viral loads ( in schistosome-specific antibody positive men ) and trend towards higher viral loads ( in schistosome-specific antibody positive women ) reflect viral loads early in HIV infection or after several years because we do not know time of infection of the HIV+ index partner . The imperfect specificity of the Focus HSV-2 test ( sensitivity of 99 . 5% and specificity of 70 . 2% in HIV+ and HIV- participants in urban Uganda [56] ) is a limitation , and unfortunately due to funding changes , not all RPR results were confirmed by Treponema pallidum hemagglutination assay . It is unknown if these sources of potential bias would lead to unmeasured confounding by HSV-2 or RPR status in our analysis . We also lack information on some potentially important covariates , including exposures to environmental water , poor sanitation , prior use of praziquantel , and past socioeconomic status . Finally , we did not have sufficient sample size to look at antiretroviral treatment initiation outcomes ( another proxy of HIV disease progression ) since this intervention only began in 2007 . Expanded identification and treatment of schistosomiasis are warranted in endemic countries such as Zambia , including in adults in urban areas . In addition to reducing morbidity and mortality associated with schistosomiasis , our findings suggest such efforts might also contribute to prevention of onward HIV transmission and disease progression in HIV+ men and women . As such , the strategy of preventive chemotherapy may have benefits not just for schistosomiasis , but also for HIV . Given the relatively few quantitative studies to date , additional research on schistosomiasis and HIV transmission , acquisition , and disease progression in both men and women , including the potential effects of former infections , are warranted .
|
This study explored the association between schistosome infections ( a disease caused by parasitic flatworms , also known as ‘snail fever’ , which is very common throughout sub-Saharan Africa ) and human immunodeficiency virus ( HIV ) . We found in Lusaka , the capital of Zambia , that schistosome infections were associated with transmission of HIV from adult men and women , and schistosome infections were also associated with increased HIV acquisition in adult women . We additionally found that schistosome infections were associated with death in HIV+ adult women . Since treatment of schistosome infections with praziquantel is inexpensive , effective , and safe , schistosomiasis prevention and treatment strategies may be a cost-effective way to reduce not only the symptoms associated with the infection , but also new cases of HIV and death among HIV+ persons . Though often viewed as an infection of predominantly rural areas and children , this study highlights that schistosomiasis prevention and treatment efforts are also needed in urban areas and among adults .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"schistosoma",
"invertebrates",
"schistosoma",
"mansoni",
"medicine",
"and",
"health",
"sciences",
"enzyme-linked",
"immunoassays",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"viral",
"transmission",
"and",
"infection",
"helminths",
"pathogens",
"immunology",
"tropical",
"diseases",
"microbiology",
"parasitic",
"diseases",
"animals",
"retroviruses",
"viruses",
"immunodeficiency",
"viruses",
"rna",
"viruses",
"neglected",
"tropical",
"diseases",
"antibodies",
"immunologic",
"techniques",
"viral",
"load",
"research",
"and",
"analysis",
"methods",
"immune",
"system",
"proteins",
"proteins",
"medical",
"microbiology",
"hiv",
"schistosoma",
"haematobium",
"microbial",
"pathogens",
"immunoassays",
"biochemistry",
"helminth",
"infections",
"schistosomiasis",
"eukaryota",
"virology",
"viral",
"pathogens",
"physiology",
"biology",
"and",
"life",
"sciences",
"lentivirus",
"organisms"
] |
2018
|
Schistosomiasis is associated with incident HIV transmission and death in Zambia
|
Yersinia enterocolitica ( Ye ) evades the immune system of the host by injection of Yersinia outer proteins ( Yops ) via a type three secretion system into host cells . In this study , a reporter system comprising a YopE-β-lactamase hybrid protein and a fluorescent staining sensitive to β-lactamase cleavage was used to track Yop injection in cell culture and in an experimental Ye mouse infection model . Experiments with GD25 , GD25-β1A , and HeLa cells demonstrated that β1-integrins and RhoGTPases play a role for Yop injection . As demonstrated by infection of splenocyte suspensions in vitro , injection of Yops appears to occur randomly into all types of leukocytes . In contrast , upon infection of mice , Yop injection was detected in 13% of F4/80+ , 11% of CD11c+ , 7% of CD49b+ , 5% of Gr1+ cells , 2 . 3% of CD19+ , and 2 . 6% of CD3+ cells . Taking the different abundance of these cell types in the spleen into account , the highest total number of Yop-injected cells represents B cells , particularly CD19+CD21+CD23+ follicular B cells , followed by neutrophils , dendritic cells , and macrophages , suggesting a distinct cellular tropism of Ye . Yop-injected B cells displayed a significantly increased expression of CD69 compared to non-Yop-injected B cells , indicating activation of these cells by Ye . Infection of IFN-γR ( receptor ) - and TNFRp55-deficient mice resulted in increased numbers of Yop-injected spleen cells for yet unknown reasons . The YopE-β-lactamase hybrid protein reporter system provides new insights into the modulation of host cell and immune responses by Ye Yops .
Yersinia enterocolitica ( Ye ) is an enteropathogenic bacterium that causes gastrointestinal disorders such as enteritis and enterocolitis as well as extraintestinal manifestations such as lymphadenitis , reactive arthritis , and septicemia [1] , [2] . Ye has been demonstrated to multiply extracellularly in host tissue . To accomplish this , Ye needs to evade the host's immune defense . Beside other virulence factors , Ye evolved a type III secretion system ( TTSS ) consisting of an injectisome and effector proteins the latter of which are injected into host cells [3] . The injection of effectors into host cells via a TTSS injectisome is a common strategy of pathogenic bacteria to counteract the host's immune response [4] . The TTSS injectisome is complex ATP-driven protein-export machinery . Built of ring shaped proteins , the basal body is providing a channel through the bacterial membranes and the periplasm or the peptidoglycan wall , respectively . The injectisome is terminating in a needle-like structure that is protruding from the bacterial surface [5] , [6] . Thus , pore-forming proteins enable the injection of the effector proteins through the membrane of host target cells [7] , [8] . The TTSS is crucial for Yersinia virulence [9] . Ye injects at least six effector Yops into host cells . YopP/J is a potent inhibitor of the NF-κB and the MAPK signaling pathways and thus inhibits downstream effects of these pathways such as proinflammatory responses or antigen uptake [10]–[16] . In addition YopP induces apoptosis in macrophages and dendritic cells [17]–[20] . YopE , YopT and YopO affect RhoGTPase functions which leads to actin cytoskeleton disruption and together with YopH , a tyrosine phosphatase which targets different eukaryotic kinases , promote inhibition of phagocytosis [21] . In addition , phosphatase activity of YopH counteracts T cell activation [22] , [23] . YopM is known to interact directly with protein kinase C-like 2 ( PRK2 ) and ribosomal S6 protein kinase 1 ( RSK1 ) , the function of YopM is so far elusive [24] . Despite well defined and profound in vitro effects of Yops , the contribution of some Yops to establish successful infection in a mouse infection model seems to be rather insignificant ( YopP/O/T ) , while others ( YopE/H/M ) are of great importance for virulence of Ye [9] . This highlights the necessity to understand Yersinia pathogenicity in the context of whole organism infection . It also creates the need for a tool to display the sites of interaction between host organism and pathogen . Various approaches have been taken to enable monitoring of type three secretion into eukaryotic cells by Yersinia and other pathogens during infection in cell culture . TTSS effectors have been detected in host cells by immunocytochemistry and western blot [8] , [25] and fusions of effectors with GFP [26] , the adenylate cyclase ( CyA ) [27] , Elk-tag [28] or Cre-recombinase [29] . Thus , yersiniae expressing YopE-GFP could be used to detect fluorescent bacteria in cell culture and mouse infection , but clear Yop injection into host cells could not be shown or quantified [30] . The YopE-CyA fusion was useful to detect Yop injection directly in vitro and was utilized to demonstrate the requirement of YopB and YopD for Yop injection by quantification of cAMP levels in infected cells [27] . Briones et al . used a SopE-Cre reporter system which allowed the visualization of Salmonella Sop injection at least in a low number of cultured cells [29] . All these systems have enabled interesting discoveries , but failed to detect Yop injection on a single cell level and are not suitable for the quantitative detection of targeted cells in mouse infection models . Recently , a new reporter system for monitoring type III secretion of bacterial proteins into host cells has been described to study effector injection of enteric Escherichia coli [31] and Salmonella enterica [32] as well as Yersinia pestis [30] . The reporter systems consisted of translational fusions of a whole or truncated TTSS effector protein with mature E . coli TEM-1 β-lactamase . Infected cells were stained in these studies with the lipophilic CCF2-AM [33] , an esterified form of the CCF2 substrate . After entry into the cells endogenous cytoplasmic esterases rapidly convert CCF2-AM into its negatively charged form CCF2 , which is retained in the cytosol . CCF2 is a fluorescence resonance energy transfer ( FRET ) substrate , which consists of a cephalosporin core linking 7-hydroxycoumarin to fluorescein . Excitation of the coumarin moiety at 409 nm results in a FRET to the fluorescein residue leading to emission of light with a wavelength of 520 nm ( green fluorescence ) . Cleavage of the CCF2 substrate by TEM-1 β-lactamase separates the two fluorescent moieties and interrupts the FRET between them . Excitation of the coumarin residue now leads to the emission of light at a wavelength of 447 nm ( blue fluorescence ) . Marketon et al . [30] used an expression vector which expressed YopM-β-lactamase fusion protein in Y . pestis for infection experiments of mice . Using this system it was shown that Yops were predominantly injected into granulocytes , dendritic cells and macrophages while injection into T and B cells was found to be a rare event . Geddes et al . [32] used translational fusions of bla and effectors of the SPI-1/2 pathogenicity islands as reporters and demonstrated that Salmonella targets preferentially granulocytes in the spleen after mouse infection . To investigate into which immune cells Yops are injected during Y . enterocolitica infection in vivo , a Bla reporter system was established to track Yop injection in infection in cell culture and in an experimental mouse infection model . The established system was validated as a powerful tool to investigate Yop injection in cell culture as well as in a mouse infection model .
To establish a reporter system to detect Yop injection by Ye , the expression vector pMK-Bla was transformed into E40 Δasd , resulting in the strain E40-pBla , secreting a YopE53-β-lactamase fusion protein ( see Materials and Methods as well as Table 1 ) . As a control , a strain also harboring pMK-Bla but deficient in Yop secretion ( ΔYscN-pBla ) was employed . As a further control the plasmid pMK-Ova , which encodes for a translational fusion of YopE53 with the ovalbumin aa 247–355 fragment , was transformed into E40 Δasd resulting in E40-pOva . Expression and secretion of YopE and the fusion proteins YopE53-Bla and YopE53-Ova by the strains E40-pBla , E40-pOva and ΔYscN-pBla was analyzed by immunoblotting . The data depicted in Figure 1A indicate that YopE wild type and hybrid proteins were expressed and secreted into the supernatant except by the ΔYscN-pBla mutant which is deficient for Yop secretion . To confirm injection of YopE and YopE53-Bla , HeLa cells were mock-infected or infected with E40-pBla or ΔYscN-pBla and injection assays were performed to detect YopE and YopE53-Bla by immunoblot using anti-YopE antibodies . YopE injection was detectable after infection with E40-pBla but not after infection with ΔYscN-pBla ( Figure 1B left panel ) . In contrast , a YopE-Bla band was hardly visible . Only long overexposure led to a clearly visible YopE53-Bla band in E40-pBla , but not ΔYscN-pBla infected cells ( Figure 1B right panel ) . This shows that YopE53-Bla is translocated into cells but with a much lower efficacy than the wildtype YopE . To investigate the usability of the system to assay Yop injection into host cells , HeLa cells were infected with the E40-pBla , E40-pOva or ΔYscN-pBla mutant strains and subsequently stained with CCF4-AM . In the presented study CCF4-AM was used instead of CCF2-AM that was utilized in former studies , because according to the manufacturer the FRET is slightly stronger with this dye , resulting in less background staining . Microscopic analysis of the uninfected cells as well as the controls including infection with E40-pOva or ΔYscN-pBla showed only green net fluorescence resulting from intact FRET within CCF4 ( Figure 1C ) . In contrast , after E40-pBla infection , approximately half of the cells showed a pronounced blue net fluorescence resulting from FRET disruption by β-lactamase , suggesting that injection of the YopE53-Bla hybrid protein occurred ( Figure 1C ) . To quantify cell numbers and fluorescence signal intensity , flow cytometry was used . Green fluorescence resulting from CCF4-AM dye substrate uptake , de-esterification and retention in infected HeLa cells was equal to that observed in uninfected cells ( Figure 1D ) . It also reflects the viability of cells because dead cells cannot retain the CCF4-dye ( data not shown ) . After infection of HeLa cells for 1 hour with E40-pBla ( MOI 50 ) , blue fluorescence resulting from the cleavage of CCF4 was detectable in 58 . 9±11% of the viable cells . In contrast , infection of HeLa cells with E40-pOva or ΔYscN-pBla resulted in 0 . 48±0 . 3% or 2 . 5±2 . 4% blue cells , indicating that spontaneous cleavage of the substrate CCF4 after infection does not significantly occur and that Yop secretion is essential to detect Yop injection into host cells . Adhesion of Ye to host cells is mediated by interaction of host cell β1-integrins with Yersinia invasin protein or YadA ( indirectly via collagen ) ; moreover , adhesion of Ye to host cells is a prerequisite for Yop injection into host cells [34] . To investigate whether β1-integrins are essential for Yop injection , infection experiments were performed with the fibroblast cell line GD25 , lacking β1-integrins and the cell line GD25-β1A , overexpressing β1-integrins as indicated in Figure 2A . Infection of GD25 cells with E40-pBla or E40-pOva ( MOI 50 ) for one hour resulted in 0 . 9±0 . 2% or 0 . 12±0 . 06% blue cells while infection of GD25-β1A cells yielded 18 . 2±6% or 0 . 22±0 . 12% blue cells , respectively ( Figure 2B ) . The strong reduction ( 95% ) of Yop injection into GD25 cells compared to GD25-β1A cells was also confirmed by detection of YopE by immunoblots after infection with E40-pBla ( Figure 2C ) . To address whether reduced Yop injection into GD25 cells was due to reduced adhesion of Ye to GD25 compared to GD25-β1A cells , cell adhesion assays were performed . GD25 showed a modest ( 40% ) but significant decrease of the number of adhering yersiniae compared to GD25-β1A cells ( Figure 2D ) . C . difficile toxin B ( TcdB ) , an inhibitor of RhoA , was shown to inhibit Yop injection while the Rac inhibitor NSC23766 had no impact on Yop injection [35] . According to these data we hypothesized that RhoA might be crucial for Yop injection . To confirm the previous data , HeLa cells were pretreated with TcdB or NSC23766 and then infected with E40-pBla and stained with CCF4-AM ( Figure 2E ) . Flow cytometry analysis revealed that the number of blue cells was significantly decreased by TcdB , but not by the Rac1 inhibitor NSC23766 . Additionally , HeLa cells were transfected with control siRNA ( siCo ) or siRNAs specific for the RhoGTPases RhoA , Cdc42 or Rac1 . The cells were subsequently infected with E40-pBla , stained with CCF4-AM and analyzed by flow cytometry ( Figure 2E ) . Inhibition of RhoA and Rac1 , but not Cdc42 , significantly reduced the number of blue cells by 25–30% indicating that both Rac1 and RhoA play a role in Yop injection . To investigate whether Yops are also injected into primary cells , single cell suspensions were prepared from the spleen of C57BL/6 mice and exposed to E40-pBla , E40-Ova or ΔYscN-pBla mutant strains at a MOI of 50 for one hour in vitro . Similar to infection of HeLa cells , flow cytometry analysis showed that CCF4 dye retention ( green fluorescence ) after infection of splenocytes was not significantly affected by infection with the strains used ( data not shown ) . After infection of spleen cells for one hour with E40-pBla ( MOI 50 ) in 61 . 4±11 . 4% of cells a blue fluorescence signal was detectable indicating that 61% of splenocytes had been injected with Yops ( Figure 3A ) . In contrast , blue fluorescence was observed in less than 0 . 3% of splenocytes after infection with E40-pOva or ΔYscN-pBla . The low background of blue cells after infection with E40-pOva indicates that spontaneous cleavage of CCF4 detectable after infection with yersiniae is negligible . The low background staining after infection with ΔYscN-pBla indicates that secretion of YopE-Bla is a prerequisite for the detection of blue cells ( Figure 3A ) and that expression of YopE-Bla inside the bacteria and potential internalization of bacteria do not lead to significant blue fluorescence . To determine the relationship between the number of bacteria per targeted host cell and the occurrence of blue cells ( FRET disruption ) , spleen cell suspensions were infected with E40-pBla ( and E40-pOva as a control ) at various MOI for one hour and then subjected to flow cytometry analysis ( Figure 3B and C ) . The data show that the percentage of blue cells correlates with the infection dose ( MOI ) in a fashion that a strong increase of the number of blue cells is followed by a saturation state . A hyperbolic regression curve can be calculated ( goodness of fit of r2 = 0 . 0977 ) with the formula: % blue cells = 84×MOI/ ( 16 . 9 + MOI ) . Using this regression curve it can be calculated that an MOI of 16 . 9±1 . 9 would result in 50% blue cells . This regression curve also predicts that in cell culture even at high MOI Yop injection appears to be limited for yet unknown reasons . We also investigated whether there might be preferential Yop injection into certain cell types . To define the populations of the spleen injected with YopE53-Bla , splenocytes were infected with E40-pBla and then flow cytometry analysis was used to detect distinct cell surface markers on and β-lactamase activity in the cells . The data depicted in Figure 4A shows the composition of spleen cells and confirmed that B cells ( 57±2% CD19+ ) and T cells ( 30±7% CD3+ ) are the most abundant spleen cell subpopulations . As shown in Figure 4B , the percentage of blue cells in each analyzed spleen cell population ranged between 58±6% to 82±10% but revealed no significant differences ( p>0 . 05 ) indicating that Yop injection occurs into all cells types to a similar degree . Consistently , the composition of blue spleen cells ( Figure 4C ) is closely related to the frequency of the subpopulations . To investigate whether Yop injection can be detected in vivo , desferrioxamine-conditioned C57BL/6 mice were intravenously infected with E40-pBla or E40-pOva; one to three days later splenocyte suspensions were prepared , stained with CCF4-AM and subjected to flow cytometry analysis ( Figure 5A and Figure S1 ) . As indicated in Figure 5A and Figure S1 , after infection with E40-pBla a distinct population of blue cells ( 3 . 1±1 . 1% of total spleen cells ) could be identified which was not present in the spleen of uninfected or E40-pOva ( 0 . 2±0 . 2% ) infected mice . To confirm data obtained by flow cytometry , cell suspensions were subjected to flow cytometry cell sorting and subsequently analyzed by fluorescence microscopy ( data not shown ) . Analysis of β-lactamase activity after infection for different time periods revealed a time-dependent increase of the percentage of blue cells with a maximum between two and three days after infection . The highest bacterial burden in the spleen was found between 24 and 48 hours after infection ( Figure 5B ) . Increased infection dose led to a modest increase of the bacterial load in the spleen two days after infection ( Figure 5C ) . However , an infection with more than 5×103 E40-pBla did not further increase the bacterial burden in the spleen . The number of blue cells increased with the infection dose . In Figure 5D the bacterial load in the spleen and the percentage of blue cells from various experiments is given . The best fit to approximate the correlation of bacterial load and percentage of blue cells was obtained using a sigmoid regression curve ( Figure 5D ) ( goodness of fit: r2 = 0 . 64 , degrees of freedom = 79 , and Sy . x = 0 . 50 ) . Plotting the percentage of blue cells versus the bacterial load in the spleen also revealed that for the detection of blue cells over background ( defined as % blue cells±2-fold standard deviation = 0 . 6% blue cells ) the bacterial load has to be more than log10 CFU 5 . 6 per spleen . To investigate in which cells Yops were injected , C57BL/6 mice were infected with 5×105 E40-pBla or E40-pOva for two days and subsequently cells were stained with CCF4-AM and fluorescence labeled antibodies binding distinct surface markers . Infection of mice with E40-pOva or E40-pBla for two days resulted in a bacterial load of log10 CFU 7 . 4±0 . 2 or log10 CFU 7 . 3±0 . 2 per spleen . Immunostaining and flow cytometry analysis of spleen cells revealed that infection with E40-pBla led to a significant change in the composition of cell populations compared to uninfected ( Figure 5E ) . Thus , the number of Gr-1+ ( granulocytes ) , CD11c+ ( dendritic cells ) and CD49+ ( NK ) cells increased significantly , while the number of CD3+ T cells decreased significantly . Moreover , we found that 13 . 4±5 . 1% of all F4/80+ ( macrophages ) , 7 . 2±2 . 7% of all CD49+ ( NK cells ) , 10 . 9±3 . 8% of all CD11c+ ( dendritic cells ) and 5 . 5±1 . 8% of all Gr-1+ ( granulocytes ) cells displayed blue fluorescence indicating that they had been injected with YopE53-Bla ( Figure 5F ) . In contrast , only 2 . 3±1 . 3% of all CD3+ ( T cells ) and 2 . 6±1 . 1% of all CD19+ ( B cells ) cells displayed blue fluorescence . These data suggest that Yop injection upon Ye infection in this mouse infection model occurred predominantly into cells of the myeloid lineage representing the innate immune system . Analysis of the composition of total spleen cells with blue fluorescence regarding the percentage of each subpopulation of cells revealed that related to the total number of blue cells 43 . 8±11 . 3% were CD19+ , 40 . 8±7 . 6% were Gr-1+ , 26 . 9±3 . 7% were CD11c+ , 23 . 6±8% were F4/80+ , 15 . 4±5 . 3% were CD49b+ , and 10 . 4±3 were CD3+ . These data indicate that injection of Yops occurred into both myeloid and lymphoid cells , most frequently into B cells and granulocytes ( Figure 5G ) . Splenic B cell subpopulations comprise CD19+CD21hiCD23− ( marginal zone B cells , MZ ) , CD19+CD21+CD23+ ( follicular B cells , FO ) , and CD19+CD21lo ( newly formed B cells , NFB ) B cells . To address whether Yop injection occurs in all of these B cell subpopulations , cultured splenocytes were infected in vitro with E40-pBla ( MOI 50 ) for 1 hour and Yop injection was determined after staining using antibodies for B cell markers and CCF4 ( Figure S2 ) . The data show ∼84% of B cells were blue indicating injection of Yops; moreover , the different B cell subpopulations contributed to the total number of CD19+ blue cells according to their different frequencies of MZ ( ∼10% ) , FO ( ∼70% ) and NFB ( ∼20% ) , as injection of Yops occurred with similar efficacy in all B cell subtypes in vitro . To analyze Yop injection into B cells in vivo , desferrioxamine-conditioned C57BL/6 mice were infected with 5×105 E40-pBla and two days later B cells were analyzed ( Figure 6A ) . After infection , the composition of splenic B cell subpopulations changed . Thus , the percentage of CD21+CD23+ ( FO ) ( 43 . 8±4 in infected versus 66%±2 in uninfected mice ) and the percentage of CD21hiCD23− ( MZ ) ( 3 . 6±0 . 9 in infected versus 9 . 2±1 . 4 in uninfected mice ) significantly decreased whereas the percentage of CD21lo ( NFB ) ( 47 . 2±3 . 8 in infected versus 22 . 2±0 . 9 in uninfected mice ) significantly increased . Taking the changes in total cell numbers in the spleen into account , the total number of NFB significantly increased while the total number of FO and MZ remained largely unchanged ( data not shown ) . Analysis of blue ( Yop-injected ) B cells ( ∼2% of total B cells ) revealed that 64 . 4%±1 . 7 of CD19+ blue B cells were CD21+CD23+ ( FO ) B cells indicating that Yersinia targets predominantly follicular B cells ( Figure 6C ) . In blue FO B cells , a distinct CD21hiCD23hi B cell subpopulation ( Figure 6D ) was detected suggesting that injection of Yops into FO is associated with an increased expression of CD21 and CD23 . To address whether injection of Yops may affect activation of B cells , expression of CD69 , an early activation marker of B cells , was determined . In splenocytes infected in vitro with E40-pBla , increased CD69 expression was found in B cells ( mean fluorescence intensity , MFI; 7 . 5 in uninfected versus 29 . 8 in infected cultures ) . Moreover , in Yop-injected B cells ( green+ blue+ ) , expression of CD69 was higher compared with B cells not injected with Yops ( green+ blue− ) ( Figure 7A ) suggesting that in vitro , injection of Yops into B cells is associated with rapid activation of these cells . Upon infection of mice with E40-pBla , ∼2% of B cells were injected with Yops ( blue staining ) ( Figure 7B ) . Expression of CD69 was increased in splenic B cells compared with uninfected mice ( MFI of 26 . 1 in infected versus 12 . 3 in uninfected mice ) . Moreover , in Yop-injected B cells ( green+ blue+ ) , expression of CD69 was dramatically higher ( MFI 76 . 4 ) compared with B cells not injected with Yops ( green+ blue− ) ( MFI 22 . 3 ) . In conclusion , these data suggest that direct interaction of Ye with B cells including injection of Yops leads to activation of B cells as indicated by increased CD69 expression . Previous studies showed that TNFRp55−/− and IFN-γR−/− mice are highly susceptible to infection with Ye [36]–[40] indicating that the pleiotropic effects of IFN-γ and TNF-α are crucial for the defense against yersiniae . To investigate whether such gene defects may modulate Yop injection , desferrioxamine conditioned TNFRp55−/− , IFN-γR−/− , and wild type mice were infected with Ye and two days later , splenocytes suspensions were prepared , stained with antibodies and CCF4-AM , and subjected to flow cytometry analysis . Infection of TNFRp55−/− and IFN-γR−/− mice with E40-pOva as well as E40-pBla resulted in a bacterial load of log10CFU 7 . 4±0 . 1 and 7 . 3±0 . 1 , respectively and was comparable to the bacterial load of infected WT mice ( log10CFU 7 . 4±0 . 1 ) ( Figure 8A ) . After infection of IFN-γR−/− mice no significant changes in the splenic cell subpopulations were observed ( Figure 8C ) compared to infection of WT mice . However , the percentage of all blue cells increased significantly ( IFN-γR−/− 5 . 9±0 . 9% versus WT 3 . 1±1 . 1% ) ( Figure 8B ) . Likewise , the percentage of blue Gr-1+ , F4/80+ , CD11c+ and CD49b+ , but not CD3+ and CD19+ , cells increased significantly ( Figure 8D ) . Taking both , the total number of blue cells as well as changes in the cell composition into account , no significant alterations in the composition of blue cells was found ( Figure 8E ) . These data indicate that IFN-γR deficiency leads to an increased number of cells , in which Yops are injected , but the composition of blue cells is comparable with that of infected C57BL/6 mice . Infection of TNF-Rp55−/− mice led to a significant increase in the percentage of Gr-1+ , F4/80+ and CD11c+ cells , and a significant decrease of C19+ cells compared with WT mice ( Figure 8C ) . The percentage of total blue cells was significantly higher after infection of TNF-Rp55−/− ( 7 . 6±1 . 1% , p<0 . 001 ) compared with infection of WT mice ( 3 . 1±1 . 1% ) ( Figure 8B ) . In line with these findings , all subpopulations ( F4/80 , Gr-1 , CD11c , CD49b , CD19 and CD3 ) showed a significantly increased percentage of blue cells in infected TNFRp55−/− mice ( Figure 8D ) . The composition of blue cells changed leading to a significantly increased number of blue Gr-1+ and F4/80+ cells compared to WT mice ( Figure 8E ) . These data suggest that TNFRp55 deficiency leads to a more frequent injection of Yops into F4/80+ and Gr-1+ cells . To investigate the localization of yersiniae in the spleen after infection of mice , splenic sections were prepared and immunochemistry using anti-Yersinia Hsp60 antibodies was performed ( Figure 8F ) . Histological and immunohistological studies revealed that Ye formed many microcolonies surrounding the lymphoid follicles . Gross differences in the splenic architecture or localization of yersiniae in the spleen of WT or knockout mice were not observed . To further address why in TNFRp55−/− and IFN-γR−/− mice increased numbers of Yop injected spleen cells were found , splenocytes of C57BL/6 , TNFRp55−/− and IFN-γR−/− mice were infected in vitro for one hour with different MOI of E40-pBla and the percentage of blue cells was determined . The frequency of blue cells for the different MOIs was not significantly different between C57BL/6 and TNFRp55−/− or IFN-γR−/− splenocytes ( Figure S4 ) suggesting that splenocytes of wildtype , TNFRp55−/− and IFN-γR−/− mice upon infection with Ye display a comparable frequency of Yop injection .
Injection of Yop effector molecules into host cells by the TTSS is an important strategy of Ye to suppress the host immune response . However , detection of Yop injection in animal infection models has been difficult so far . Several tools have been employed to detect injection of bacterial proteins into host cells , but their usefulness has been limited to certain in vitro models [8] , [25]–[28] . In similar , results published by Briones et al . and own attempts to create a Cre-recombinase based reporter system to monitor Yop injection into host cells produced promising results in cell culture experiments but failed to deliver a valuable tool in mouse infection models [29] . Recently however , Marketon et al . demonstrated that Yop-β-lactamase fusion proteins can be utilized to monitor Yop injection in vivo [30] . In this study we applied this tool to Ye infection and tested whether it might be useful to detect Yop injection in vitro and in vivo . In line with the report by Marketon et al . [30] , the use of control strains such as ΔYscN-pBla and E40-pOva revealed that detection of β-lactamase reporter activity depends on injection of the YopE53-Bla fusion protein . Therefore , internalization of bacteria does not account for reporter activity . Yop injection into host cells was detectable after infection of HeLa epithelial cells and primary C57BL/6 splenocytes in cell culture experiments . In vitro experiments with splenic subpopulations indicated that Yop injection occurs with similar efficacy into macrophages , dendritic cells , granulocytes , NK cells , T cells or B cells in cell culture . We conclude that the different immune cell subpopulations display the same prerequisites for Yop injection . From these data it is conceivable that Yop injection in vivo actually reflects the interaction frequency of bacteria with distinct splenic subpopulations . The usefulness of this reporter system in cell culture experiments could be demonstrated by investigating the importance of β1-integrins for Yop injection . It is widely accepted that β1-integrins play an important role for adhesion to host cells and eventually for internalization of yersiniae via the direct or indirect interaction with invasin and YadA , respectively [34] , [41]–[43] . In line with this , the number of adhering Ye to fibroblasts which lack β1-integrins was reduced by 40% compared to cells expressing β1-integrins . In contrast , Yop injection ( measured as percentage of blue fluorescent cells ) into fibroblasts which lack β1-integrins was dramatically reduced by 95% compared to fibroblasts expressing β1-integrins . These data are in line with a recent report [35] which suggested a central role for β-integrins for Yop injection . Nevertheless , detection of YopE translocation by immunoblots was still detectable in GD25 cells which lack β1-integrins , indicating that other surface molecules beside β1-integrins may also play a role to modulate Yop injection . At this point it is not clear which other surface molecules beside β1-integrins affect Yop injection and whether depending on the cell types in which Yops are injected different surface molecules modulate Yop injection in vitro and in vivo . After Y . pseudotuberculosis infection β1-integrin mediated signal transduction leading to RhoGTPase activation seems to be the prerequisite for initiation of Yop effector injection [35] . By using the Rac1 inhibitor NSC23766 it was excluded that Rac1 plays a role for Yop injection and because both TcdB ( inhibitor of RhoA , Rac1 , Cdc42 , RhoG , Tc10 ) and C3 exotoxin ( RhoA , RhoB , RhoC inhibitor ) inhibited Yop injection , RhoA , RhoB and RhoC were identified as putative candidates for this effect [35] . In contrast , the siRNA experiments presented herein suggest that both Rac1 and RhoA play a role for Yop injection after Ye infection of HeLa cells . Interestingly , in the present study the Rac1 inhibitor NSC23766 also did not inhibit Yop injection . NSC23766 was demonstrated to bind to Rac1 and to prevent binding to and activation of Rac1 by the Rac1-specific guanine nucleotide exchange factors ( GEF ) Tiam1 and Trio . However , it was described that NSC23766 seems not to prevent binding to and activation of Rac1 by the more promiscuous GEF Vav [44] , [45] . As Vav is able to activate both RhoA and Rac1 , but not Cdc42 [44] , [45] , one may speculate that Vav or other GEFs , rather than Tiam1 and Trio , might be involved in Rac1 mediated Yop injection . One has to anticipate that NSC23766 inhibits Yersinia triggered Rac1 activation only partially and in a GEF specific manner which has to be proven in future studies . In addition it will be interesting to find out which specific GEFs are involved in facilitating Yop injection . Taken together this study clearly pinpoints that β1-integrins play a role in facilitating Yop injection . The YopE-Bla reporter system turned out to work highly efficiently when expressed in E40 ( O9 ) strain , but not in WA-314 ( O8 ) strain; thus , 80% of in vitro infected splenocytes displayed blue fluorescence upon infection with E40-pBla strain , but only 26% upon infection with WA-pBla strain ( data not shown ) . Moreover , in C57BL/6 mice infected with E40-pBla 3% of splenocytes displayed blue fluorescence , while in WA-pBla infected mice we could hardly detect any blue cells for yet unknown reasons . We therefore used the E40 strain for all experiments included in this study . The low virulence of serotype O9 strains such as E40 strain or O3 strains in mice is at least partially caused by their disability to synthesize the iron chelating siderophore yersiniabactin due to their lack of the HPI island [46] , [47] . This could be “complemented” by treatment of mice with desferrioxamine which allows to study Ye O9 and Ye O3 infection in mice [48]–[50] . The disadvantage of desferrioxamine , however , is that it might have immunosuppressive effects on phagocytes [51] , [52] . Infection of desferrioxamine-conditioned C57BL/6 mice revealed a correlation between the percentage of blue cells and the bacterial burden in the spleen which can be best described by a sigmoidal dose response curve . Taking the background staining of 0 . 2±0 . 2% into account , a detection limit of 0 . 6% blue cells was estimated using the sigmoidal regression curve . To reach 0 . 6% blue cells in the spleen the bacterial burden would have to be at least 4×105 bacteria representing a MOI of 0 . 01 . The cell culture experiments , however , showed that much higher MOI are required to achieve half-maximal numbers of blue spleen cells . Therefore , higher efficacy of YopE-bla detection might be due to abscess formation with Ye microcolonies which may allow more Ye bacteria to interact with single host cells at abscesses . Alternatively the TTSS machinery might be upregulated during infection in spleen compared to cell culture . Most experiments were carried out at conditions with a bacterial burden of 5×107 bacteria in the spleen resulting in 3 . 1±1 . 1% of blue , Yop-injected cells . Determination of the percentage of blue cells in each subpopulation revealed that Yop injection occurred into Gr1+ ( predominantly granulocytes ) , CD11c+ ( predominantly dendritic cells ) , F4/80+ ( predominantly macrophages ) and CD49b+ cells ( predominantly NK cells ) . In contrast , only in a small percentage of the CD3+ ( T cells ) and CD19+ ( B cells ) cells , Yop injection was detectable . However , due to the high total number of B cells ( >50% of spleen cells ) their contribution to the total number of blue cells is about 40% . Further characterization of B cells revealed that the composition of B cell subpopulations changed after infection leading to an increased number of CD19+CD21lo ( NFB ) while the number of CD19+CD23+CD21+ cells ( FO ) did not change and the number of CD19+CD21hi CD23− ( MZ ) slightly decreased . Thus , after infection NFB appear to either proliferate in or immigrate into the spleen . Yop injection was predominantly detectable in follicular B cells and to a lesser extent in NFB . Naïve follicular B cells reside in the “follicular niche” and may present T-dependent antigens to activated T cells . The follicular niche therefore represents the major site at which recirculating B cells mediate T-dependent immune responses to protein antigens [53] . Interestingly , in contrast to the Ye infection model presented here , injection of the Bla-reporter into B cells was not described in Y . pestis infection [30] . It remains unclear whether these discrepancies are due to different Yersinia species , different mouse strains or differences in the reporter systems . Nevertheless , it is striking that after Y . pestis infection the analysis gate which was used to define the viable cells was rather small indicating that most spleen cells were necrotic or apoptotic under the chosen experimental conditions . It might well be that Yop-injected B cells might have been missed by the analysis due to high cell death rates in Y . pestis infection . In Ye infection , however , interaction of B cells with Ye is prominent . This finding is in line with the analysis of histological sections of the spleen . Thus , E40-pBla microcolonies were observed adjacent to lymphoid follicles ( B cell zones ) . Balada-Llasat and Mecsas [54] reported that Y . pseudotuberculosis may show a tropism to B and T cell zones in lymph nodes . In our studies we found yersiniae always adjacent to lymph follicles in the spleen and the most frequent interaction of yersiniae with host cells was that with B cells , specifically with CD19+CD21+CD23+ B cells which are defined as follicular B cells ( FO ) . FO B cells organize into the primary follicles of B cell zones focused around follicular dendritic cells in the white pulp of the spleen . Thus in our system Ye appear to preferentially colonize in the follicular niche and interact frequently with follicular B cells . Whether Ye are actively migrating to this region or may be trapped by e . g . dendritic cells needs to be further investigated . Infection with Ye was associated with increased CD21 and CD23 expression levels of FO B cells which suggests that B cells may be activated after interaction with yersiniae . Specific activation of Yop injected B cells was also supported by the observation that the early B cell activation marker CD69 was upregulated in blue B cells . From this data we can conclude that Yop-injected B cells were specifically activated by Ye despite injection of Yops . Whether Yop-injected B cells may undergo subsequent cell death or are affected in their ability to produce IgM or to undergo differentiation needs to be investigated in future study . Previous studies revealed that TNF-α and IFN-γ are important for defense against Ye infection [36]–[38] , [40] . Therefore , we were interested whether these cytokines may affect Yop injection . The experimental conditions were chosen in a way that the bacterial burden in the spleen of the various mouse strains was comparable . Infection of IFN-γR- and TNFRp55-deficient mice revealed a higher number of blue cells in the spleen compared to infected WT mice indicating that either more host cells interacted with yersiniae or the threshold which allows detection of Yop injection was lower in knock-out mice . Histological analysis of the Ye microcolonies in WT and knockout mice revealed comparable tissue distribution of Ye and comparable tissue alterations by the infection . In vitro infection of splenocytes from IFN-γR- and TNFRp55 deficient and wildtype mice did not reveal significant differences in terms of Yop injection . Therefore it remains widely unclear why more blue cells are found in IFN-γR- or TNFRp55−/− mice compared to wild type mice . However , we cannot exclude that minor changes in spleen cell subpopulations may partially cause the different numbers in knock-out and wild type mice . In further experiments we tried to detect Yop injection into cells of the PP after orogastric infection . However , we were not able to detect blue PP cells under these conditions ( data not shown ) . One explanation might be that the number of bacteria that establish infection in the PPs is lower than that in spleen [55] . Thus , one may speculate that the number of bacteria attached to a single host cell in PP might be lower than that in spleen which might result in Yop injection below the detection limit of this reporter system . Alternatively or in addition , the number of cells in which Yops were translocated might be too low . The Bla reporter system described herein is so far the only reporter system that allows tracking of Yop injection in Ye infection in vivo on a single cell level in a mouse infection model . Nevertheless , the system needs to be improved in order to accomplish a more sensitive reporter system . The disadvantages of CCF-4 as a substrate for β-lactamase is that the amount of Yops translocated cannot be quantified and that CCF-4 displays limited sensitivity . Moreover , CCF-4 does not allow studying Yop injection in histological sections in situ . Current experiments in our laboratory aim at overcoming these limitations with new compounds . Taken together , the Ye Bla reporter system can be successfully used in cell culture and in mouse infection models for detection of Yop injection on a single cell level , and thus could be a valuable tool for quantitatively screening for genetic factors involved in Yop secretion and injection at the bacterial as well as at the host cell side , or to identify drugs which target Yop secretion or injection , respectively .
Y . enterocolitica E40 strains used in this study are listed in table 1 . All bacterial strains were grown overnight in Luria-Bertani ( LB ) broth at 27°C supplemented with 50 µg/ml meso-diaminopimelic acid , 50 µg/ml nalidixic acid , 50 µg/ml kanamycin , and 400 µM Na-arsenite ( all from Sigma Chemical , St . Louis , MO ) in combinations according to the indicated resistances and supplementation needs ( table 1 ) . A 1∶25 dilution of the bacterial culture was incubated for additional 3 h at 37°C . The bacteria were washed once with phosphate-buffered saline ( PBS; Invitrogen , Karlsruhe , Germany ) and the optical density at 600 nm was determined . All plasmid constructs and bacteria used are listed in table 1 . To generate pMK-Bla , as a starting vector pBME53-yopT [56] , a pACYC184 derivate containing a HindIII-sycE-yopE/sycE promoter-yopE gene fragment encoding the first 53 N-terminal amino acids ( aa ) -BamHI-YopT -SalI construct was used . YopT was replaced by a translational fusion of the coding sequences ( CDS ) of the SV40 nuclear localization signal ( nls ) and the CDS of Cre recombinase resulting in pBME53-Cre . This vector was digested with HindIII and SalI and the resulting sycE-yopE53-cre fragment was inserted into pIV2 [57] . pIV2 was derived from a small cryptic plasmid of an apathogenic Y . enterocolitica strain ( kindly provided by Eckhard Strauch , Federal Institute for Risk Assessment , Berlin ) resulting in pIV2-SycE-YopE53-Cre . To create a balanced system that compels stable retention of the reporter vector in Y . enterocolitica during in vivo infections , the Y . enterocolitica asd gene coding for the L-aspartate-dehydrogenase , an enzyme that is required for the synthesis of the cell wall component L-lysine via the meso-2 , 6-diaminopimelic acid ( DAP ) pathway and therefore essential for bacterial growth and replication was inserted into the pIV2-SycE-YopE53-Cre plasmid as follows: The asd gene and its 5′region were amplified from Y . enterocolitica E40 chromosome by using the oligonucleotides 5′- AGC TTT AGG GCC CAA AAA CAG CAA CAC CGT TGC C-3′ and 5′-AAC TCG AGT TAC AGA AAA TTC GCA GC-3′ . The PCR product was then digested with ApaI and XhoI and inserted into those sites of pIV2-SycE-YopE53-Cre , thus yielding pMK4 . Together with the deletion of the asd gene from its chromosomal locus ( see below ) in the utilized bacteria , this grants stable retention of the reporter plasmid , by complementing the otherwise lethal DAP auxotrophy of the Y . enterocolitica E40 Δasd strain . The β-lactamase ( bla ) gene was amplified from pCR2 . 1 ( Invitrogen , Karlsruhe , Germany ) using the primers 5′-GGA TCC ATG AGT ATT CAA CAT TTC CG-3′ and 5′-GTC GAC AAC TTG GTC TGA CAG TTA CC-3′ . The PCR product was subcloned into pCR-Blunt II and subsequently donated as a BamHI / SalI fragment to pBME53-YopT , thus replacing the sequence coding for YopT and fusing the translation product to the YopE53 domain , yielding the plasmid pBME53-Bla . pBME53-Bla was digested with HindIII-SalI and the resulting HindIII-sycE-yopE53-bla-SalI fragment replaced the HindIII-sycE-yopE53-cre-SalI fragment of pMK4 yielding pMK-Bla . To generate a control vector , the sequence coding for β-lactamase was excised by BamHI / SalI digestion from pBME53-bla and replaced by a BamHI / SalI fragment from pYopE1–138Ova247–355 [58] encoding an internal ovalbumin fragment comprising aa 247–355 , yielding pBME53-ova . From this plasmid , a fragment encoding the YopE53-Ova247–355 fusion protein was yielded by HindIII / SalI digest and used to replace the HindIII / SalI fragment encoding YopE53-Bla in pMK-Bla yielding pMK-Ova . Generation of Y . enterocolitica strains - Y . enterocolitica E40 is a clinical isolate that belongs to the serotype O:9 [27] . To generate Y . enterocolitica E40 Δasd the asd gene in this strain was disrupted by allelic exchange . The pMK3 asd mutator plasmid was constructed as follows: ( i ) the 5′end of the asd gene was amplified from Y . enterocolitica E40 by using the oligonucleotides 5′-GAT CGT CGA CAT GGT CGG CTC AGT A-3′and 5′- CAG TGA ATT CCG GCG TCC AAT CCA ATA-3′ and the 3′end by using the oligonucleotides 5′-GAT CTC TAG ATT CGC AGC ATA CGG C-3′and 5′-GAC TGA ATT CGT GAC TGC GGC CAC T-3′ . The PCR product corresponding to the 5′ end of asd was digested with SalI-EcoRI and that corresponding to the 3′ end with EcoRI-XbaI . The restricted PCR products were then ligated into pBluescript II SK ( + ) ( Stratagene , Cedar Creek , USA ) digested with SalI-XbaI , which originated pMK1 that contains a deleted asd gene ( Δasd ) . The Δasd allele was subcloned as a SalI-XbaI DNA fragment into pMRS101 [59] digested with the same enzymes , yielding pMK2 . The asd mutator plasmid pMK3 was obtained after digestion of pMK2 with NotI and religation of the vector . The strains E40-pBla and E40-pOva have been generated by electroporation of pMK-Bla or pMK-Ova , respectively into Y . enterocolitica E40 Δasd . Thus , the DAP auxotrophy of these strains was supplemented , thereby creating a metabolically balanced system assuring stable plasmid retention . To enable the generation of reporter strains with pYV40 virulence plasmids that had been subjected to mutagenesis previously , pYV40 was cured from Y . enterocolitica E40-pBla as described previously [60] . Screening for arsenite susceptible clones yielded Y . enterocolitica E40 Δasd pYV− pMK-Bla ( pYV− Δasd pBla ) . Y . enterocolitica E40 Δasd yscNΔ169–177 ( ΔYscN-pBla ) was created by electroporation of pMSL41 [27] , [61] into pYV− Δasd pBla . The plasmid pMSL41 has been yielded by deletion of yscN169–177 from pYV40 . Secretion of the YopE53-β-lactamase fusion protein was examined by the preparation of released proteins as described previously [62] . Overnight cultures grown at 27°C were diluted 1∶20 into brain heart infusion broth ( Difco , Heidelberg , Germany ) . After 2 h of incubation at 37°C , Yop secretion was induced by the addition of 5 mM EGTA for Ca2+ sequestration , 15 mM MgCl2 and 0 . 2% glucose . 3 h after induction secreted proteins were precipitated from the culture supernatant with trichloroacetic acid . Protein concentration was determined by Bradford assay ( Bio-Rad , Hercules , USA ) and 15 µg of protein was loaded on a 12% SDS-PAGE . Separated proteins were transferred electrophoretically to an Immobilon-P PVDF membrane ( Millipore , Bedford , USA ) . For immunostaining , polyclonal rabbit anti-YopE was used . Immunoreactive bands were visualized by incubation with peroxidase-conjugated swine anti-rabbit IgG antibody ( 1∶1000 ) ( DAKO , Glostrup , Denmark ) using enhanced chemiluminescence reagents ( ECL , Amersham Biosciences , Freiburg , Germany ) and CL-XPosure Film ( Thermo , Rockford , IL ) . The fibroblast-like GD25 and GD25-ß1A cell lines were a kind gift from R . Fässler ( Max-Planck-Institute for Biochemistry , Martinsried , Germany ) . The murine GD25 cells were derived from the embryonic stem cell clone G201 which is deficient in the ß1-integrin subunit [63] . The stably transformed cell line GD25-ß1A was obtained by electroporation of wild-type integrin ß1A cDNA into GD25 cells [64] . GD25 cells not expressing ß1-integins were cultivated in DMEM ( Gibco ) supplemented with 10% FCS , the GD25-ß1A line expressing ß1-integrins was maintained in DMEM supplemented with 10% FCS and 10 µg/ml puromycin . HeLa cervical epithelial cells ( ATCC CCL-2 . 1 ) were grown in RPMI 1640 ( Biochrom KG , Berlin , Germany ) supplemented with 10% fetal bovine serum ( Sigma Chemical ) , 2 mM L-glutamine ( Biochrom KG ) , penicillin ( 100 U/ml ) , and streptomycin ( 100 µg/ml ) ( Biochrom KG ) in a humidified 5% CO2 atmosphere at 37°C . Suspensions of single spleen cells were prepared from naïve C57BL/6 mice as described below . For infection with a MOI of 100 for microscopic observation and Western Blot analysis or with a MOI of 50 or as indicated for FACS analysis bacteria were spun onto the cells ( 5 minutes , 400 g ) . After 1 h incubation at 37°C , gentamicin ( 100 µg/ml ) was added to terminate infection . To assess the impact of inhibitors of RhoGTPases on Yop injection , cells were treated with 200 ng/ml medium Clostridium difficile toxin B 10463 ( TcdB; kind gift I . Just , Institute for Toxicology , Hannover Medical School , Hannover , Germany ) 2 h prior to infection or 100 µM Rac1 inhibitor ( NSC23766; Calbiochem , San Diego , CA ) 3 h prior to infection . To analyze Yop injection by immunoblotting , cells were grown in 94 mm dishes , and then were left untreated uninfected or were infected with the E40-pBla reporter strain or the ΔYscN-pBla strain ( MOI 50 ) as a control for 90 minutes . Cells were lysed in 200 µl PBS with 1% Triton-X ( Sigma ) . 100 µg cell lysate protein was analyzed by SDS-PAGE and immunoblotting was performed in similar as already described for detection of Yop secretion with the exception that HRP Substrate Plus ( P . J . K . , Kleinbittersdorf , Germany ) was used instead of ECL and in some cases film exposure was extended to 30 min for increased sensitivity . To assay adhesion of the E40-pBla reporter strain to HeLa , GD25 and GD25-ß1A cells , the respective cells were seeded on coverslips in a 24-well plate ( Nunc Life Technologies , Wiesbaden , Germany ) . One hour after infection ( MOI 50 ) , cells were washed with PBS , fixed with 4% PFA and stained with fuchsine . All samples were prepared in duplicates , of which cells and bacteria were counted in 6 representative details . For the knock-down of RhoGTPases , the predesigned and prevalidated siRNA oligonucleotides Hs RhoA-6 , Hs Rac1-6 and Hs Cdc42-7 were purchased from Qiagen ( Hilden , Germany ) . AllStars negative control siRNA ( Qiagen ) was used as a nonsilencing control . 3×104 HeLa cells were transfected with siRNA ( 5 nM ) in a 12-well plate 48 h prior to infection using HiPerFect transfection reagent ( Qiagen ) according to the manufacturer's fast forward protocol . Successful knock-down was confirmed on RNA level by qPCR . Total mRNA of infected HeLa cells was extracted using RNeasy Mini Kit ( Qiagen ) . 1 µg of total mRNA was reverse transcribed to cDNA using QuantiTect Reverse Transcription Kit ( Qiagen ) . Real-time qPCR was performed on a LightCycler 480 ( Roche Diagnostics , Mannheim , Germany ) using QuantiTect SYBR Green PCR Kit and the respective QuantiTect Primer Assays ( Rac1 1 , RhoA 1 , CDC42 2 and GAPDH 2 ) . Relative expression levels were calculated using the ΔΔCT method [65] . Expression levels of the target genes were normalized to glyceraldehyde-3-phosphate dehydrogenase RNA expression . Thus , we could demonstrate at least 75% knock-down by the used siRNAs at mRNA expression level ( data not shown ) . C57BL/6 mice were purchased from Harlan Winkelmann ( Borchen , Germany ) , TNF-Rp55−/− [66] and IFN-γR−/− [67] mice with a genetic C57BL/6 background were bred under specific pathogen-free conditions . Animal experiments were performed according to german law with permission of the Regierungspräsidium Tübingen . The experiments were performed with 6–8-week old female mice . To enable profound infection , 2 . 5 mg desferrioxamine ( Sigma Chemical ) in 200 µl PBS were administered to the mice intravenously one hour before infection . Mice were infected with 5×105 bacteria from frozen stock suspensions in 200 µl PBS iv . into the tail vein . Uninfected control mice were in parallel treated with desferrioxamine . After two days , mice were sacrificed by CO2 asphyxiation and their spleens were surgically removed and placed in ice-cold HBSS ( Ca2+ and Mg2+ free Hanks' balanced salt solution ) ( Biochrom ) supplemented with 2% v/v fetal calf serum ( FCS ) ( Sigma Chemical ) and 10 mM HEPES buffer ( Biochrom ) . Single cells were obtained by forcing the organs with a 5 ml syringe pestle through a 40 µm-pore nylon mesh cell strainer ( Falcon; BD Labware , Franklin Lakes , USA ) . Cell suspensions were washed twice with ice-cold HBSS . Red blood cells were lysed from spleen samples by incubating the cell suspensions for 5 min at room temperature in lysis buffer ( 170 mM Tris , 160 mM NH4Cl , pH 7 . 4 ) followed by two washes in ice-cold HBSS . For the detection of Bla activity by immunofluorescence microscopy , HeLa cells were washed once with PBS and covered with 1× CCF4-AM staining solution supplemented with probenecid , prepared according to the manufacturer's instructions . Cells were incubated 30 min at a dark place at room temperature prior to observation with an Axiovert 200 microscope . Pictures were taken with an Axiocam HRc and Axiovision 4 . 4 Software was used to capture the shots and to produce overlay images ( Carl Zeiss Microimaging , Esslingen , Germany ) . Filter sets for individual observation of coumarin and fluorescein fluorescence respectively were purchased from AHF ( Tuebingen , Germany ) . For the detection of Bla activity by flow cytometry , cells were resuspended in 1× CCF4-AM staining solution supplemented with probenecid , prepared according to the manufacturer's instructions ( Invitrogen , Carlsbad , CA ) and incubated 30 min at a dark place at room temperature prior to FACS analysis . In case the detection of Bla activity was combined with the staining of leukocyte surface markers , 106 cells per staining were resuspended in PBS . To avoid non-specific labeling , FcγII/III receptors were blocked by preincubation with mAb 2 . 4G2 for 20 min at RT . Flow cytometry analysis was performed on a Dako Cyan cytometer using Summit 4 . 3 software ( Dako , Carpinteria , CA ) or on a BD Biosciences ( Heidelberg , Germany ) FACSCanto II using FACSDiva software . From each sample , at least 100 , 000 cells have been analyzed; error bars indicate the standard deviation between samples from different animals . An exemplary description of the steps how analysis was performed is shown in Figure S1 . For immunohistological analysis the tissues were embedded in Tissue-Tek OCT compound ( Nunc , Roskilde , Denmark ) , snap-frozen in liquid nitrogen , and stored at −80°C . Frozen sections were prepared and stained by an immunoperoxidase method using 3 , 3-diaminobenzidine-tetrahydrochloride acid ( DAB; Sigma , Deisenhofen , Germany ) as chromogenic substrate . Nonspecific binding sites were blocked by incubation of the sections with PBS containing 10% fetal calf and 5% normal goat serum . For the yersiniae staining , rabbit anti-Hsp60 antibody was diluted 1∶200 in PBS containing 5% FCS and 5% normal goat serum for 1 h at room temperature . The secondary antibody was peroxidase-conjugated affinity purified F ( ab' ) 2 fragment goat anti rabbit IgG ( Jackson ImmunoResearch; diluted 1∶200 ) . Isotype-matched irrelevant rabbit IgG was used in controls and revealed no staining signal . The sections were counterstained with Mayer's hemalaun , mounted , and assessed microscopically on a BX51 microscope ( Olympus Optical Co , Leinfelden , Germany ) by two independent investigators . Pictures were taken with a DP71 camera and analysed with cellB software ( Olympus ) . Immunostaining of controls was negative for all groups tested . If not otherwise stated , the means and standard deviations ( SD ) of data derived by cell culture experiments are calculated from four independent experiments . The number of mice which were used to calculate means and SD in mouse infection experiments is indicated in the manuscript or figure legends . Statistical analyses were performed using one-way ANOVA analyses with either Dunnett ( comparison of control group with other groups ) or Bonferroni corrections ( comparison of all groups ) as indicated in the figure legends or the text by using Graph Pad Prism software ( GraphPad Software , La Jolla , USA ) . This software was also used to fit non-linear regression models .
|
An important strategy of Yersinia enterocolitica ( Ye ) to suppress the immune defense is to inject bacterial proteins ( Yersinia outer proteins , Yops ) after cell contact directly into host cells , which affects their functions . However , tracking of cells in which Yop injection occurred has only been described for Yersinia pestis thus far . We adapted the described reporter system specifically for the use of infections with Ye and report the usefulness and limitations of this system . Using cell culture experiments , we demonstrated that β1-integrins and the RhoGTPases RhoA and Rac1 are involved in Yop injection . Since cell culture experiments also revealed that Yop injection is detectable in a similar manner into all subpopulations of the spleen , the system can be used to detect interaction of bacteria with host cells in vivo . In a mouse infection model we found that follicular B cells , granulocytes , macrophages , and dendritic cells are the main targets of Yop injection . Interestingly , Yop-injected B cells displayed an increased activation as indicated by increased CD69 expression . In contrast , interaction of bacteria with T cells seems to be rather a rare event . In immunocompromised gene-targeted mice we found increased frequencies of Yop-injected host cells for yet unknown reasons . Taken together , this novel reporter system represents a powerful tool to further study interaction of host cells with Ye .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"microbiology/immunity",
"to",
"infections",
"microbiology/innate",
"immunity",
"immunology/immunity",
"to",
"infections",
"microbiology",
"microbiology/cellular",
"microbiology",
"and",
"pathogenesis"
] |
2009
|
Yersinia enterocolitica Targets Cells of the Innate and Adaptive Immune System by Injection of Yops in a Mouse Infection Model
|
The diversity and geographical distribution of fleas parasitizing small mammals have been poorly investigated on Indian Ocean islands with the exception of Madagascar where endemic plague has stimulated extensive research on these arthropod vectors . In the context of an emerging flea-borne murine typhus outbreak that occurred recently in Reunion Island , we explored fleas' diversity , distribution and host specificity on Reunion Island . Small mammal hosts belonging to five introduced species were trapped from November 2012 to November 2013 along two altitudinal transects , one on the windward eastern and one on the leeward western sides of the island . A total of 960 animals were trapped , and 286 fleas were morphologically and molecularly identified . Four species were reported: ( i ) two cosmopolitan Xenopsylla species which appeared by far as the prominent species , X . cheopis and X . brasiliensis; ( ii ) fewer fleas belonging to Echidnophaga gallinacea and Leptopsylla segnis . Rattus rattus was found to be the most abundant host species in our sample , and also the most parasitized host , predominantly by X . cheopis . A marked decrease in flea abundance was observed during the cool-dry season , which indicates seasonal fluctuation in infestation . Importantly , our data reveal that flea abundance was strongly biased on the island , with 81% of all collected fleas coming from the western dry side and no Xenopsylla flea collected on almost four hundred rodents trapped along the windward humid eastern side . The possible consequences of this sharp spatio-temporal pattern are discussed in terms of flea-borne disease risks in Reunion Island , particularly with regard to plague and the currently emerging murine typhus outbreak .
Reunion is a small oceanic island of volcanic origin located in the Indian Ocean , Southern Hemisphere ( 21°6′S and 55°36′E ) that forms , together with Mauritius and Rodrigues Islands , the Mascarene archipelago . This oceanic island is geographically isolated from continental landmasses and located within one of the 34 recognized world biodiversity hotspots [1] . The island lies therefore in a biogeographic context favourable to species radiation and potentially high endemism . Its dramatic relief has shaped a highly contrasted climate: the mountainous centre ( >3 , 000 meters ) separates a humid windward coast ( scoring some rain world records ) from a dry leeward coast , which lower part consists mainly in savannah . This peculiar situation has led to the evolution of a strong vegetal endemism with a well-described altitudinal succession of vegetal species observed on both windward and leeward coasts [2] . The diversity of terrestrial animals , specifically mammals , is clearly much less prominent: the only endemic mammal species is the insectivorous bat Mormopterus francoimoutoui [3] . Following human colonization which started in the XVIIth century , five small mammal species have been introduced , namely the insectivores Suncus murinus Linnaeus 1766 ( Asiatic house shrew ) and Tenrec eucaudatus Schreber 1778 ( tailless tenrec ) from Madagascar , and the three cosmopolitan murid rodents Rattus rattus Linnaeus 1758 ( black rat ) , Rattus norvegicus Linnaeus 1769 ( brown rat ) and Mus musculus Linnaeus 1758 ( house mouse ) . Tropical countries and especially tropical islands are known at higher risk for the emergence or re-emergence of infectious diseases [4] . Therefore , updated information on zoonotic pathogens and on the diversity and distribution of their arthropod vectors is warranted for a quicker response to outbreaks threats . Fleas ( Order Siphonaptera ) form a unique group of insects comprising 15 families with a total of about 220 genera and some 2 , 500 described species [5] . Five families including 25 genera are ectoparasites of birds , while all other flea species specifically feed on mammals . In Madagascar , located about 800 km west of Reunion Island , flea diversity has been extensively studied , mainly because of their role as vectors of Yersinia pestis , the plague agent , especially in this country that reports most human plague cases worldwide [6] , [7] . Flea diversity is high in Madagascar , with several endemic species together with a few cosmopolitan ones , which host specificity and distribution have been partly described [6] , [7] . Surprisingly , Xenopsylla brasiliensis ( Baker , 1904 ) has never been collected in Madagascar , even though this species is recognized as a main plague vector in Eastern and Southern Africa [8] , [9] and has been collected in Moroni ( Grande Comore ) and notified on two other islands of the Southwestern Indian ocean: Mayotte ( Comoros archipelago ) and Mauritius [10] . By contrast , almost no data are currently available on flea diversity on the other islands of the Southwestern Indian Ocean , including Reunion . The cosmopolitan and/or tropical species possibly present in the region are Pulex irritans Linné , 1758 , Echidnophaga gallinacea Westwood , 1875 , Leptopsylla segnis Schönherr , 1811 , Xenopsylla cheopis Rothschild , 1903 as well as Ctenocephalides spp . Hence , the recent emergence of murine typhus in Reunion Island , where ten autochthonous human confirmed cases were reported between 2011 and 2013 ( Balleydier E . 2014 pers . comm . ) has stimulated the investigation of fleas for vector-assessment of indigenous species for the agent Rickettsia typhi . The objective of our study was therefore to report the diversity and distribution of fleas in Reunion Island with the aim of highlighting patterns of possible epidemiological importance .
Trapping was conducted throughout a one-year period survey ( November 14th 2012–November 16th 2013 ) in different biotopes along two altitudinal transects lying on each side of the island: the eastern transect comprised eight sampling sites and the western transect , seven . In addition , two sites located in the western coast and a few sites in the urban northern part of the island were included in the present survey ( Figure 1 ) . The sampling encompassed the two local seasons , i . e . hot-wet summer from November to June , and cool-dry season from July to October , with twelve out of the twenty sampling sites being sampled twice , i . e . during the two seasons . Trapping was conducted following a standardized protocol: wire cage live traps ( 29 by 18 by 12 cm ) were used for rats trapping ( and accidentally tenrecs ) , and Sherman live traps for mice and shrews . On each sampling site , forty to eighty traps were placed in line approximately 15 meters apart in the afternoon; trapped animals were collected the following morning and brought back to the laboratory for processing . Traps were baited during three consecutive nights using successively within each line cheese , coconut or a mixture of peanut butter and canned sardine oil . This baiting setup ( bait A , bait B , bait C , bait A , … ) was implemented in order to trap most of the prevalent mammal diversity at each sampling site . Traps were left open in the same place during the day , with productive traps being immediately replaced every morning with the same bait over the 3-days trapping session . Animals were sacrificed by cervical dislocation without anaesthesia to avoid bleeding in accordance with guidelines accepted by the scientific community for the handling of wild mammals [11] and the institutional guidelines published by the Centre National de la Recherche Scientifique ( http://www . cnrs . fr/infoslabos/reglementation/euthanasie2 . htm ) . All animal procedures carried out in this study were approved by the French Institutional Ethical Committee “Comité d'éthique du CYROI” ( No . 114 ) . Following sacrifice , each animal was visually examined for 10 minutes and all ectoparasites , including fleas , were manually collected either with a brush soaked in ethanol when insects jumped off the host , or forceps when eye-spotted in the fur . Collected fleas were preserved in 70% ethanol for later morphological and molecular analyses . Fleas were identified at the species level using taxonomic keys provided by Lewis [12] , and Hoogstraal and Traub [13] . A subsample of Xenopsylla spp . fleas were mounted permanently on slides using Euparal medium , following a procedure adapted from Brigham Young University ( http://fleasoftheworld . byu . edu/Systematics/MountingTechniques . aspx ) . The gender , genus , and species were recorded for each flea specimen . Xenopsylla cheopis and X . brasiliensis were mainly differentiated using the occurrence of marginal cones at the basis of the antepygidial bristle in males , and shape of spermatheca on mounted females [7] . Rodents body mass , ear , and back foot lengths , together with tail and body lengths were recorded . Rattus spp . was identified using morphological criteria including the comparison of ( i ) the ratio of tail to body lengths , ( ii ) the ear length and ( iii ) the hind foot length [14] . The morphological diagnosis of Rattus spp . was confirmed by molecular data through sequencing of cytB locus from 15 randomly selected animals morphologically identified as R . rattus or R . norvegicus . Briefly , DNA was prepared from 20 mg of kidney tissue as previously described [15] and used as a template with L14723 and H15915 primers set , following a previously described PCR protocol [14] . For molecular diagnosis of fleas , DNA was prepared as follows: fleas were dried individually and subsequently crushed with a TissueLyser ( Qiagen , Valencia , CA ) using 3 mm tungsten beads and cetyl trimethyl ammonium bromide 2%; DNA was further extracted following a previously described procedure [16] . Both nuclear and mitochondrial loci were sequenced by amplifying 28S ribosomal RNA ( 28S rRNA gene ) and cytochrome oxidase II ( COII ) encoding gene using 28S A/28S rD7b1 and COII F-leu/COII R-lys primer pairs , that produce 1473-bp and 770-bp PCR fragments respectively [17] . Amplicons were sequenced on both strands by Genoscreen ( Lille , France ) using the same PCR primers , and sequences were edited using Geneious Pro [18] . All sequences used in this study were deposited in Genbank and are accessible under accession numbers KJ638526 to KJ638590 . All sequences were automatically aligned using MUSCLE implemented in Geneious Pro version 5 . 3 . 4 [18] . Alignments were constructed separately for the nuclear ( 28S ) and mitochondrial ( COII ) datasets using sequences available in GenBank to complete our dataset . Bayesian analyses were performed to infer phylogenetic relationships between flea species . First , the best-fitting model and associated parameters were selected by jModelTest [19] and phylogenies were constructed by Bayesian inference . Two sets of four MCMCMC ( Metropolis Coupled Markov Chain Monte Carlo ) chains incrementally-heated were run in MrBayes 3 . 1 . 2 [20] for 20 , 000 , 000 generations . Trees and associated model parameters were sampled every 300 generations . The initial 2 , 000 trees were discarded as a conservative “burn-in” and the harmonic mean of the likelihood was calculated by combining the two independent runs . The 50% majority-rule consensus tree was then computed from the sampled trees in the two independent runs under the best model . The data were entered into EPIData 3 . 1 and analyzed with Epi info 6 . 04 statistical software using the chi-squared or Fisher exact tests for observed frequencies . We used a p-value threshold of 0 . 001 . The effect of “habitat” on fleas' diversity was measured at two scales , host and sampling region , by using the flea percentage incidence index ( PII: mammals parasitized by fleas of species A/mammals caught ( % ) ) , the specific flea index ( SFI: number of fleas of species A collected from host species Y/mammals of species Y parasitized by fleas of species A ) and the total flea index ( TFI: total fleas collected/total trapped mammals , i . e . mean number of fleas per trapped mammal ) [21] . The seasonality of flea diversity was tested by comparing PII on animals trapped at each site during the cool-dry versus hot-wet seasons .
A total of 960 small mammals were trapped . They belong to the five introduced small terrestrial mammal species occurring in Reunion Island: 39 mice ( Mus musculus ) , 168 shrews ( Suncus murinus ) and 25 tenrecs ( Tenrec eucaudatus ) , all other specimens being rats ( Rattus rattus: N = 554; R . norvegicus: N = 174 ) ( Table 1 ) . Almost 10% ( 95 ) of trapped mammals were infested with fleas ( Table 1 ) and the TFI ( mean number of fleas per host ) was equal to 0 . 3 when based on all trapped mammals , and equal to 3 when based on parasitized mammals only . Of 288 fleas collected during the survey , 286 could be identified on a morphological basis . They were distributed within three genera and four distinct species , namely Xenopsylla cheopis ( N = 171 ) , Xenopsylla brasiliensis ( N = 63 ) , Leptopsylla segnis ( N = 43 ) and Echidnophaga gallinacea ( N = 9 ) ( Table 2 ) . Rattus rattus was found to be the most parasitized host , predominantly by Xenopsylla spp . ( p<10−3 ) . Only five mice , eight shrews and one tenrec were found parasitized by fleas ( Xenopsylla spp . and L . segnis ) ( Table 2 ) . Rattus rattus was more heavily infested in the western side of the island ( p<10−7 ) whereas R . norvegicus was most infested in the northern part ( p<10−4 ) and mice in the eastern part ( p<10−4 ) . No difference according to the sampling region was found in shrews or tenrecs . When considering Xenospylla spp . , X . cheopis was mainly found on Rattus spp . ( p<10−5 ) with no difference between R . rattus and R . norvegicus but X . brasiliensis was significantly more abundant on R . rattus ( p<10−4 ) than on any other mammal species . The number of flea species per host species ranged from one to four ( Table 2 ) , but most mammals were parasitized by a single flea species although nine R . rattus were found co-infested with two distinct species as follows: X . cheopis+X . brasiliensis ( N = 1 ) , X . cheopis+E . gallinacea ( N = 3 ) , and X . cheopis+L . segnis ( N = 3 ) . Xenopsylla spp . were by far the most common fleas ( 234/286 fleas ) with X . cheopis and X . brasiliensis representing 59% ( 171/286 ) and 22% ( 63/286 ) of all identified fleas , respectively ( Table 2 ) . Xenopsylla cheopis was also the most geographically widespread species , as it was present in all of the fourteen flea-positive sampling sites out of the twenty prospected ones . X . brasiliensis was collected at only two sites throughout the island , both of them being located on the western transect . Noteworthy , X . brasiliensis/R . rattus SFI index was relatively high in one of those 2 sites ( Sans Soucis , SFI = 2 ) . Leptopsylla segnis was collected on mice and both rat species in four elevated sites ( >1 , 000 meters ) , and E . gallinacea was only collected on R . rattus at three distinct sites along the western transect ( Figure 2 ) . Windward and leeward transects displayed dramatically different results , in terms of abundance of fleas and species richness ( Table 3 ) . The PII was significantly lower ( p<10−7 ) in the eastern region compared to the northern and western regions . Indeed , 201 Xenopsylla fleas were collected out of 405 mammals trapped in the western transect while this species was totally absent on the 464 rodents trapped on the eastern transect ( see Tables 1 , 3 ) ; the only two X . cheopis specimens collected in the eastern side were from one tenrec trapped on the top of the eastern transect located in an elevated plateau at the centre of the island ( Table 3; Figure 2 ) . All other fleas collected in the eastern transect were identified as L . segnis ( 21 of 24 collected fleas; Table 3 ) . Lower flea species richness was recorded in animals trapped along the eastern than in the western transect: fleas were absent on six of the nine eastern sampling sites , and on the remaining sites , only seven mammals were found parasitized . The specific flea indexes ( SFI ) were 1 . 47 for X . cheopis/R . norvegicus on the northern sampling sites; 0 . 53 for X . cheopis/R . rattus in the western sites; and 0 . 26 for X . brasiliensis/R . rattus in the western sites ( see Tables 4 and 5 ) . There is no apparent seasonality of flea abundance in the eastern region , which could be explained by the absence or very low abundance of fleas , even during the peak season observed on other parts of the Island . Seasonality is observed in the west , with greater abundance observed during the hot-wet season . Over the fourteen flea-positive sampling sites , seven were sampled during the two seasons . Two sampling sites were flea-positive during both seasons , four were flea-positive only during the hot-wet season and one was found flea-negative during the hot-wet season , and flea-positive during the cool-dry season ( one R . norvegicus and one S . murinus parasitized by one X . cheopis flea each ) , but the difference was not statistically significant ( Table 6 ) . This seasonality was significant for X . brasiliensis on sampling site « Sans soucis » ( p = 0 . 01; RR = 2 . 2 [1 . 1–4 . 3] ) , and for X . cheopis on sampling site « Port est » ( p<10−3; RR = 11 . 7 [1 . 6–86 . 5] ) . Sixty ( 28S ) and seventy ( COII ) sequences were obtained from fleas sampled in Reunion Island . As all sequences of X . cheopis and X . brasiliensis were 100% identical , only a dozen sequences representative of each of those two species were included in the analyses . Few sequences from Genbank were added , including Parapsyllus longicornis used as an extra-group . Since no 28S or COII sequences were available on databases for X . brasiliensis , we sequenced three X . brasiliensis specimens sampled in Tanzania ( KJ638557-59 in COII; KJ638585 , 638589-90 in 28S: collectors Laudisoit A . , Makundi R . , Katakweba A . , S3°58′989″ E35°21′560″ , 1994 m , 10/02/2009 ) . Models selected by jModelTest were GTR+I for 28S phylogeny ( AIC weight = 0 . 62 ) , and GTR+G for COII phylogeny ( AIC weight = 0 . 85 ) . All X . cheopis ( from Reunion Island and two haplotypes from Genbank , 28S sequence ) branched within a single well supported clade , while X . brasiliensis haplotypes fell within two well supported clades , one containing sequences from Tanzanian fleas , the second harboring all haplotypes from Reunion Island ( Figure 3 ) . Both clades formed a well supported monophyletic X . brasiliensis clade distinct from X . cheopis and embedded within Xenopsylla group .
The present investigation provides the first information on flea diversity and distribution on the five introduced small mammal species present on Reunion Island , where no data were available thus far . We describe the presence of three genera composed of four distinct cosmopolitan species , namely X . cheopis , X . brasiliensis , L . segnis and E . gallinacea . Morphological diagnosis of X . cheopis and X . brasiliensis was further confirmed by sequencing of 28S and COII markers: for X . cheopis , fleas sampled in Reunion Island showed 99% and 100% identity with sequences accessible in Genbank ( i . e . EU336145 . 1 and HM188404 . 1 for 28S and COII sequences , respectively ) . As no sequences were currently available for X . brasiliensis on these 2 loci , we generated sequence data using specimens previously sampled in Tanzania and morphologically identified as X . brasiliensis by A . Laudisoit and colleagues . Again , molecular data confirmed X . brasiliensis morphological diagnosis , with 28S and COII sequences obtained from fleas sampled in Reunion Island showing respectively 99% and 94% identity with sequences obtained from Tanzanian fleas . Phylogenetic analysis carried out with both nuclear and cytoplasmic markers provided two well resolved mostly congruent trees , suggesting that no hybridization nor introgression ( two molecular events known to lead to molecular misdiagnosis [22] ) has occurred within our sample . However , the analyses did reveal one incongruency for L . segnis: while 28S-based analysis was coherent with classical taxonomy , COII sequences unexpectedly clusterized L . segnis within Pulicidae . Additional and more informative markers need to be investigated in order to address this incoherence together with other more basic questions such as a previously reported paraphylly of Leptopsyllidae [23] . The absence of molecular data for L . segnis together with the overall scarcity of accessible DNA sequences for other flea species ( including X . brasiliensis , see above ) should stimulate an increased effort towards the release of a proper barcoding tool facilitating the diagnosis of cosmopolitan species . As for X . brasiliensis , nuclear and mitochondrial sequences from Tanzanian specimens formed a cluster separated from Reunion Island sequences , which might indicate an ongoing diversification . However , a proper investigation of eastern African and Indian Ocean X . brasiliensis populations would be required to ascertain any level of genetic structuration . Altogether , our data indicate a low diversity of fleas on small mammals from Reunion Island . In addition , all flea species were cosmopolitan and likely result from the recent introduction of their vertebrate hosts on the island , or from the importation of food stocks with preimaginal stages . This feature is not unexpected considering the low specific richness in mammal hosts , which strikingly contrasts with the neighbouring island of Madagascar where species richness and endemism of both flea [7] and small mammal hosts are high [24] , flea endemism likely resulting from long host-parasite co-evolutionary processes . Host specificity differed between fleas: E . gallinacea was only collected on R . rattus which is likely a spill over host from poultry breeding sites near the concerned sampling sites , i . e . rural areas where R . norvegicus is likely to be less common . Xenopsylla brasiliensis appeared mostly associated with R . rattus ( one flea found on a shrew ) a situation reminiscent to that previously described in the Canary islands [25] . On the contrary there was low host specificity for X . cheopis that was found to most commonly infest Rattus spp . ( 92% ) , but was also found on shrews and tenrecs ( Tables 4 and 5 ) , which is in accordance with previous report from Madagascar [7] . The number of collected specimens from the two other species was too low to conclude about host specificity . This is the first report of Xenopsylla brasiliensis in Reunion Island . This species is native to continental subsaharian Africa where it is the most common plague vector in some areas , often more abundant than X . cheopis [9] . This expanding species has spread to other parts of the world such as Brazil and India [26] . This known plague vector , particularly effective in rural environments , is less tolerant to high temperatures than X . cheopis but is more resistant to drier conditions [21] . These ecological traits are in agreement with X . brasiliensis distribution in Reunion Island , where the species was restricted -in our sample- to a semi-xerophil landscape partly covered with Tamarinus indica and patches of exotic Furcraea foetida and Agave americana on the western side of the island . The heterogeneous distribution of fleas over Reunion Island , with no Xenopsylla flea collected along the windward humid eastern side , might be related to excessive rainfall in this coast . Indeed , temperature , rainfall and relative humidity have direct effects on development and survival of fleas , and a direct effect of rainfall is supposed to occur when high intensity rainfall causes flooding of rodent burrows [27] . Seasonal abundance of fleas that has been largely reported in literature is also driven by climate variables . Warm-moist weather has been described to provide higher flea indices [27] . This is in agreement with the decrease in flea abundance observed during the cool-dry season on the two sampling sites were seasonality was significant . Fleas are of tremendous medical and economic importance as vectors of several diseases including bubonic plague , murine typhus and tularaemia [28] . The discovery of fleas as vectors of Yersina pestis , and later of Rickettsia typhi , the ethiological agent of murine typhus , stimulated flea studies in the early 20th century . Xenopsylla cheopis is now considered as the most important cosmopolitan vector of both Y . pestis and R . typhi , and an important Bartonella spp . carrier , and X . brasiliensis is an efficient plague vector , especially in rural environments . Leptopsylla segnis is a weak vector of Y . pestis according to old standards ( but no recent experimental studies have been performed to establish if the early-phase transmission apply to this species ) and is a dubious vector of R . typhi [29] . Hence , our study showing that Reunion Island hosts several flea species of medical importance warrants better surveillance of potentially emerging flea-borne zoonoses . Among flea-borne diseases , the situation of plague is of major concern for the region . Plague was introduced in Madagascar from India in 1898 and has become endemic in the highlands [30] . Xenopsylla cheopis and the endemic flea Synopsyllus fonquerniei are known as the primary vectors of Y . pestis on Madagascar [31] . In Reunion Island , plague has quite a long history: the disease was likely misdiagnosed as lymphatic filariasis until 1899 when Y . pestis was isolated by André Thiroux and formally identified by Emile Roux [32] . Thus plague was described within the same year in Madagascar , Reunion and Hawaii , but it was considered as introduced in Madagascar [32] and Hawaii [33] where foci were first described in harbors , while André Thirioux described plague as endemic in Reunion [32] . Plague is not a concern anymore in Reunion Island where the last human cases were reported in 1926 [34] . Indeed , an SFI of 0 . 5 to 1 is considered sufficient to maintain plague in a locality and an index ≥1 is reported to represent a potentially dangerous situation with respect to the risk of plague outbreak [8] . Some indexes reported herein ( Tables 4 and 5 ) , specifically the X . cheopis/R . norvegicus SFI measured on the north of the island may be considered of concern and should be monitored systematically . This area is close to the city of Le Port , the only international harbour of Reunion Island , and the most likely entry port for parasitized rodents and/or food . Although the risk of plague introduction from Madagascar is expected to be limited with an SFI index in this area <0 . 5 [7] , the substantial shipping trade between Reunion and Madagascar where plague has already been described in harbours [35] , [36] command a cautious control in order to prevent introduction of rodents from this plague endemic country [28]–[29] . Finally , the role of domestic cats should not be overlooked since Felidea – in contrast to Canidea in general - are sensitive to the disease , can become infected by ingesting infested rodents and develop pulmonary form of the disease , with a risk of direct respiratory transmission of infectious droplets to the people caring for them [37] . Considering other flea-borne diseases , rickettsioses represent an important concern . Interestingly , a retrospective French study ( 2008–2010 ) on travellers returning from Madagascar and Reunion reported two patients who were infected with murine typhus during their trip [38] . More recently , in 2012 and 2013 , several autochthonous human confirmed cases of murine typhus were reported by hospital clinicians from the western and southern parts of the island ( Balleydier E . , pers . comm . ) . The authors were wondering if the heterogeneous distribution of human cases could be related to medical surveillance bias . Although incomplete , since the southern coast of the island wasn't sampled , the distribution of fleas reported herein is at least in part overlaid with that of human cases . This may suggest that the risk of murine typhus in Reunion Island is related to fleas' geographical distribution driven by environmental determinants . The detection of R . typhi in fleas together with the presentation of a more complete Xenopsylla sp . distribution map throughout the island may provide public health agencies with a useful tool for implementing a specific surveillance system for better risk assessment of murine typhus and other emerging flea-borne zoonoses in Reunion Island .
|
Fleas are blood-feeding parasites involved in the transmission of several arthropod borne pathogens . Rat-fleas ( Xenopsylla spp . ) are known vectors of bubonic plague together with other human diseases receiving less attention such as murine typhus . This latter disease was recorded for the first time in 2011 on Reunion Island where seven human cases were further confirmed within the following year . The outbreak motivated a large survey of fleas , as these insects of major veterinary and medical importance have never been investigated on this oceanic island . We collected fleas on almost 1000 small wild mammals trapped on two altitudinal transects along the humid eastern and dry western sides of the island . Our data reveal the presence of four cosmopolitan flea species and shows an astonishing distribution pattern: 81% of all collected fleas were sampled on the western transect while not a single rat-flea was sampled on the eastern humid side of the island . Interestingly , this distribution did at least in part overlay the map of murine typhus human cases . These data stimulate the need for a diagnosis of pathogens in natural flea populations together with a comprehensive distribution map of fleas , allowing a risk assessment of flea-borne diseases in humans .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"taxonomy",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"typhus",
"of",
"rickettsiae",
"animals",
"animal",
"phylogenetics",
"phylogenetics",
"emerging",
"infectious",
"diseases",
"fleas",
"veterinary",
"science",
"infectious",
"diseases",
"veterinary",
"diseases",
"zoonoses",
"epidemiology",
"evolutionary",
"systematics",
"insects",
"arthropoda",
"ecology",
"biology",
"and",
"life",
"sciences",
"evolutionary",
"biology",
"organisms"
] |
2014
|
Fleas of Small Mammals on Reunion Island: Diversity, Distribution and Epidemiological Consequences
|
Cortical activity has distinct features across scales , from the spiking statistics of individual cells to global resting-state networks . We here describe the first full-density multi-area spiking network model of cortex , using macaque visual cortex as a test system . The model represents each area by a microcircuit with area-specific architecture and features layer- and population-resolved connectivity between areas . Simulations reveal a structured asynchronous irregular ground state . In a metastable regime , the network reproduces spiking statistics from electrophysiological recordings and cortico-cortical interaction patterns in fMRI functional connectivity under resting-state conditions . Stable inter-area propagation is supported by cortico-cortical synapses that are moderately strong onto excitatory neurons and stronger onto inhibitory neurons . Causal interactions depend on both cortical structure and the dynamical state of populations . Activity propagates mainly in the feedback direction , similar to experimental results associated with visual imagery and sleep . The model unifies local and large-scale accounts of cortex , and clarifies how the detailed connectivity of cortex shapes its dynamics on multiple scales . Based on our simulations , we hypothesize that in the spontaneous condition the brain operates in a metastable regime where cortico-cortical projections target excitatory and inhibitory populations in a balanced manner that produces substantial inter-area interactions while maintaining global stability .
Cortical activity has distinct but interdependent features on local and global scales , molded by multi-scale connectivity . Data from multiple species including macaque indicate that the ground state of cortex locally features asynchronous irregular spiking with low pairwise correlations [1] and low , layer-specific spike rates [2 , 3] with inhibitory rates exceeding excitatory ones [4–6] , and activity fluctuations on multiple timescales [7] . Globally , resting-state activity has characteristic patterns of inter-area correlations [8 , 9] and propagation [10] . These interactions are layer-specific and distinct between feedback and feedforward directions [11–13] . We present a full-density multi-scale spiking network model in which these features emerge from its detailed structure . Most cortical models concentrate on either the local or the global scale , using two basic approaches . The first approach represents each neuron explicitly in networks ranging from local microcircuits to small numbers of areas [14 , 15] . The second describes large-scale cortical dynamics by simplifying ensemble dynamics to few differential equations . These models predict resting-state oscillations in a metastable regime [16–19] and reproduce the frequency specificity of inter-area interactions [20] . Cortical processing is not restricted to one or few areas , but results from complex multi-area interactions [21 , 22] . Simultaneously , dense within-area connectivity [23 , 24] suggests the importance of local processing , where the population-specific connectivity underlies multidimensional functional properties [25] and supports a set of computational principles that underlie sensory processing across the cortex [26 , 27] . Capturing both aspects requires combining detailed features of local microcircuits with realistic inter-area connectivity . Modeling at cellular resolution enables testing the equivalence with population models instead of assuming it a priori . Two main obstacles of multi-scale simulations are gradually being overcome . First , recent progress in simulation technology enables the efficient use of supercomputers [28] . Second , systematic connectivity data is increasingly available [29 , 30] . However , statistical predictions remain necessary to fully specify large cortical network models . Consequently , few large-scale spiking network models have been simulated to date , and existing ones heavily downscale the number of synapses per neuron [31 , 32] ( but see [33] ) , affecting network dynamics [34] . We here investigate a spiking multi-area network model of macaque visual cortex , covering the scales of single neurons , microcircuits , and cortical areas . The connectivity map , derived in [35] , customizes that of the microcircuit model of [36] to each area based on its architecture and adds layer-specific inter-area connections . Each area is represented by a 1 mm2 microcircuit with the full density of neurons and synapses . A mean-field method [37] refines the connectivity to fulfill the basic dynamical constraint of nonzero and non-saturated activity . By combining simple single-neuron dynamics with complex connectivity , the model enables studying the influence of the connectivity itself on the network dynamics . We first describe the refinement of the connectivity by dynamical constraints , leading to plausible spike rates . Next , we vary cortico-cortical synaptic strengths and find that with increased coupling , connections onto inhibitory neurons must outbalance connections onto excitatory neurons for stability at low rates . The resulting network state reproduces spiking statistics of V1 resting-state data [38] and yields population bursts reflecting a metastable regime [39–42] . Outside this metastable regime , the spiking statistics deviate considerably from the experimental data . Our findings thus extend previous works demonstrating metastability of cortical networks via modeling [16–19] by unifying microscopic and macroscopic descriptions and supporting the hypothesis that plausible spiking statistics require cortex to be poised in a metastable regime . Analyzing the order of activation of the areas reveals that the population bursts propagate mainly in the feedback direction . Subsequently we show that , for intermediate cortico-cortical synaptic strengths , inter-area correlation patterns resemble fMRI functional connectivity [43] . Finally , we observe directional differences in laminar patterns of inter-area communication that reflect both structural relationships and dynamical states . Our work provides a platform for future studies addressing spiking-level functional properties and for the development of analogous models of other cortical regions . Preliminary results have been presented in abstract form [44] .
We performed simulations on the JUQUEEN supercomputer [75] with NEST version 2 . 8 . 0 [76] with optimizations for the use on the supercomputer which were subsequently released in NEST version 2 . 12 . 0 [77] . The simulations use 1024 compute nodes ( corresponding to 1 rack of JUQUEEN ) with 1 MPI process per node and 64 threads per MPI process . A model instance requires about 2GB of working memory on each compute node and takes about 5 minutes for the creation of the network and approximately 12 minutes per 1 s biological time for propagation of the dynamical state . All simulations use a time step of 0 . 1 ms and exact integration for the subthreshold dynamics of the leaky integrate-and-fire neuron model [78] . Simulations are run for 100 . 5 s ( χ = 1 . 9 ) , 50 . 5 s ( χ ∈ [1 . 8 , 2 . 0 , 2 . 1] ) , and 10 . 5 s ( χ ∈ [1 . , 1 . 4 , 1 . 5 , 1 . 6 , 1 . 7 , 1 . 75 , 1 . 8 , 2 . 5] ) biological time discarding the first 500 ms . Spike times of all neurons are recorded , except for the simulations shown in Fig 2A and 2B , where 1000 neurons per population are recorded . The digitized workflow reproducing all results and figures of this work was created in compliance with [79] and is available as Python code from https://github . com/INM-6/multi-area-model . The simulation data presented in this manuscript is available from https://web . gin . g-node . org/maximilian . schmidt/multi-area-model-data . Instantaneous firing rates are determined as spike histograms with bin width 1 ms averaged over the entire population or area . In Figs 3G and 5G we convolve the histograms with Gaussian kernels of optimal width using the method of [80] , implemented in the Elephant package [81] . Spike-train irregularity is quantified for each population by the revised local variation LvR [82] averaged over a subsample of 2000 neurons . The cross-correlation coefficient is computed with bin width 1 ms on single-cell spike histograms of a subsample of 2000 neurons per population with at least one emitted spike per neuron . Both measures are computed on the entire population if it contains fewer than 2000 neurons . To compare the simulated with the experimental power spectrum in Fig 6K , we use the simulated spiking data from 140 neurons ( equal to the number of neurons identified in the experimental data ) , distributed across populations in V1 in proportion to the population sizes . We compute the power spectrogram and power spectral densities using Welch’s method ( signal . welch of the Python SciPy library [83] with a ‘boxcar’ window , segment length of 1024 data points and 1000 overlapping points between segments ) . To make our results as comparable as possible with [38] , we follow these authors and disregard neurons with an average spiking rate < 0 . 56 spikes/s . We employ analytical mean-field theory to predict the stationary population-averaged firing rates of the model . In the diffusion approximation , which is valid for large numbers of sufficiently independent inputs with small synaptic weights , the dynamics of the membrane potential V and synaptic current Is of the leaky integrate-and-fire model neurons used in our model are described by [84] τ m d V d t = - V + I s ( t ) τ s d I s d t = - I s + μ + σ τ m ξ ( t ) , where the input spike trains are replaced by a current fluctuating around the mean μ with variance σ with fluctuations drawn from a random Gaussian process ξ ( t ) with 〈ξ ( t ) 〉 = 0 and 〈ξ ( t ) ξ ( t′ ) 〉 = δ ( t − t′ ) . Going from the single-neuron level to a description at the population level , we define the population-averaged firing rate νi due to the population-specific input μi , σi . The stationary firing rates νi are then given by [84] 1 ν i = τ r + τ m π ∫ V r − μ i σ i + γ τ s τ m Θ − μ i σ i + γ τ s τ m e x 2 ( 1 + erf ( x ) ) d x ≕ 1 / Φ i ( ν ) μ i = τ m ∑ j K i j J i j ν j + τ m K ext J ext ν ext σ i 2 = τ m ∑ j K i j J i j 2 ν j + τ m K ext J ext 2 ν ext , ( 1 ) which holds up to linear order in τ s / τ m and where γ = | ζ ( 1 / 2 ) | / 2 , with ζ denoting the Riemann zeta function [85] . We solve this equation for our high-dimensional network by finding the fixed points of the first-order differential equation [86] ν ˙ ≔ d ν d s = Φ ( ν ) - ν , ( 2 ) for different initial conditions ν0 using the continuous-time dynamics framework of NEST [87] , which uses the exponential Euler algorithm , with step size h = 0 . 1 , where s denotes a dimensionless pseudo-time . To investigate the local stability of the fixed point , we study the evolution of a small perturbation δν around the fixed point ν* to linear order , ν i = ν i * + δ ν ˜ i = Φ i ( ν * + δ ν ) = Φ i ( ν * ) + d Φ i d ν δ ν = ν i * + ∂ Φ i ∂ μ i ∑ j d μ i d ν j δ ν j + ∂ Φ ∂ σ i 2 ∑ j d σ i 2 d ν j δ ν j = ν i * + ∑ j G i j δ ν j ⇒ δ ν ˜ i = ∑ j G i j δ ν j δ ν ˜ = G δ ν . ( 3 ) The perturbation decays to zero if the maximal real value of the eigenvalues of the effective connectivity matrix G , the Jacobian of Φ , is smaller than 1 . To investigate inter-area propagation , we determine the temporal order of spiking based on the location of the extremum of the correlation function for each pair of areas . This measure is chosen to characterize the relative timing of activity fluctuations across areas , as opposed to measures of causal interactions like Granger causality and the directed transfer function [88] . In an analysis of the lag structure of resting-state fMRI , Mitra et al . [10] similarly characterize temporal order using the time-delay matrix derived from the lagged cross-covariance functions . Our method also resembles the assessment of propagation using the relative timing of slow waves in EEG and LFP recordings in different areas [89–91] . In analogy to structural hierarchies based on pairwise connection patterns [92 , 93] , we look for a temporal hierarchy that best reflects the order of activations for all pairs of areas . This overall characterization of temporal order extracts the essence of the more complex picture provided by the pairwise delays . The hierarchy is based on the cross-covariance function computed between area-averaged firing rates and subsequently convolved with Gaussian kernels with σ = 2 ms to obtain smoother curves . We use a wavelet-smoothing algorithm ( signal . find_peaks_cwt of the Python SciPy library [83] with peak width Δ = 5 ms ) to detect extrema for τ ∈ [−100 , 100] and take the location of the extremum with the largest absolute value as the time lag . To order areas hierarchically , we determine the peak locations τAB of the cross-correlation function for each pair of areas A , B . We then define a function for the deviation between the distance of hierarchical levels h ( A ) , h ( B ) and peak locations , f ( A , B ) = h ( A ) - h ( B ) - τ A B . To determine the hierarchical levels , we minimize the sum of f ( A , B ) over all pairs of areas , S = ∑ A , B f ( A , B ) , using the optimize . minimize function of the scipy library [83] with random initial hierarchical levels . We verified that the initial choice of hierarchical levels does not influence the final result . We obtain hierarchical levels on an arbitrary scale , which we normalize to values h ( A ) ∈ [0 , 1]∀A . In the context of our spiking network model we define functional connectivity ( FC ) as the zero-time lag cross-correlation coefficient of the area-averaged synaptic inputs , which we approximate as I A ( t ) = 1 N A ∑ i ∈ A N i | I i ( t ) | = 1 N A ∑ i ∈ A N i ∑ j K i j | J i j | ( ν j * P S C j ) ( t ) , with the normalized post-synaptic current PSCj ( t ) = exp[−t/τs] , * indicating convolution , synaptic time constant τs , the population firing rate νj of source population j , mean indegree Kij , and mean synaptic weight Jij of the connection from j to target population i containing Ni neurons . The population firing rate νj is defined as the spike histogram with bin width 1 ms averaged over the entire population , thus time t is in discrete increments of 1 ms . We compute a BOLD signal from the simulated area-averaged synaptic inputs using the Balloon model [94] , implemented in the neuRosim [95] package of R . Synaptic inputs IA ( t ) drive the responses of cerebral blood flow ( CBF ) f ( t ) and cerebral metabolic rate of oxygen ( CMRO2 ) m ( t ) by linear convolutions f ( t ) = 1 + ( f 1 - 1 ) h ( t - δ t ) * I A ( t ) m ( t ) = 1 + ( m 1 - 1 ) h ( t ) * I A ( t ) with h ( t ) = 1 k τ h ( k - 1 ) ! ( t τ h ) k e - t / τ h τ f = 4 s , τ h = 0 . 242 τ f , f 1 = 1 . 5 , ( m 1 - 1 ) = ( f 1 - 1 ) / 2 , δ t = 1 s . These responses then feed into the Balloon model which is characterized by two dynamical variables q ( t ) , v ( t ) : d q d t = 1 τ MTT [ f ( t ) E ( t ) E 0 - q ( t ) v ( t ) f out ( v , t ) ] d v d t = 1 τ MTT [ f ( t ) - f out ( v , t ) ] with f out ( v , t ) = v 1 / α + τ d v d t E ( t ) = E 0 m ( t ) f ( t ) with τMTT = 3 s , τ = 10 s , α = 0 . 4 , E0 = 0 . 4 . These two variables determine the relative change of the BOLD signal S: Δ S S = V 0 [ a 1 ( 1 - q ) - a 2 ( 1 - v ) ] with a1 = 3 . 4 , a2 = 1V0 = 0 . 03 . The parameters are chosen as in [94] . Clusters in the FC matrices are detected by optimizing the modularity of the weighted , undirected FC graph [96] . We use the function modularity_louvain_und_sign of the Brain Connectivity Toolbox ( BCT; http://www . brain-connectivity-toolbox . net ) with the Q* option , which weights positive weights more strongly than negative weights , as introduced by [97] . The clustering of the structural connectivity is performed with the map equation method [98] , which can handle directed connections but no negative weights . In this clustering algorithm , an agent performs random walks between graph nodes according to their degree of connectivity and a certain probability of jumping to a random network node . We choose the probability for a certain target node to be selected to be proportional to the outdegree of the connection , and p = 0 . 15 as the probability of a random jump . The algorithm detects clusters in the graph by minimizing the length of a binary description of the network using a Huffman code . To investigate causal relations in the network , we compute the conditional Granger causality [99] between pairs of populations . To reduce computational load , we restrict the set of source populations for each target population i to those that form a connection with on average more than 1 synapse per target neuron . A vector autoregressive model ( VAR ) describes the target firing rate νi ( t ) based on the firing rates of other populations with a maximal time lag of 25 ms corresponding to the rounded maximal delay between any two areas in the network . For each source population j , we perform two fits: one using the set of all source populations , yielding VAR{j′} , and one using all source populations except j , yielding VAR{j′|j′ ≠ j} . To determine the causal influence j → i , we test whether the residual variances of the two VARs are significantly different using Levene’s test [100] , which is more robust against non-normally distributed residuals than the F-test . To study dominant paths in the network , we construct the weighted and directed gain matrix G with G i j = K i j | J i j | d μ i d ν j + K i j J i j 2 d σ i 2 d ν j of the network at the population level , where we evaluate the terms d μ i d ν j , d σ i 2 d ν i j at the simulated population-averaged firing rates of the model with χ = 1 . 9 . We denote the eigenvalues of G by λ and define λmax as the eigenvalue with the largest real part . To reflect the near-criticality of the brain , we perform an element-wise division by the real part of λmax: G′ = G/[Re ( λmax ) ] , so that the maximal real part of the eigenvalues λ′ of the resulting matrix G′ is max[Re ( λ′ ) ] = 1 . This scaling modulates the relative strengths of direct and indirect paths: a larger value of max[Re ( λ′ ) ] increases the relative weighting of indirect paths . Subsequently we use the same method as [35] and denote the weight of the edge from population j to i as g i j ′ . The logarithm of the reciprocal of the weight , dij = log ( 1/wij ) , defines the distance between two nodes in the graph so that summing the distances corresponds to a multiplication of the corresponding weights . Next , the Bellman-Ford algorithm [101–103] finds the shortest paths between any two nodes of the graph . This algorithm determines the shortest paths emanating from vertex i on a graph with N vertices in an iterative manner: it initially assigns an infinite path length to all other nodes k of the graph . Then , the algorithm loops through all edges ( j , k ) of the graph , tests if the path length pij plus the distance of the edge djk is smaller than the currently stored path length pik , and , if so , assigns pik ← pij + djk . By repeating the loop over all edges N − 1 times , the algorithm considers paths of increasing length on every iteration and ultimately uncovers the shortest paths between each pair of vertices . In contrast to Dijkstra’s algorithm Bellman-Ford copes with edges with negative distance values . The experimental recordings are described in [38] and are publicly available [104] . The data consist of sorted spike trains from a 64-electrode array implanted into primary visual area V1 of a lightly anesthetized macaque monkey . The array has 8 electrodes , called shanks , with 8 contacts sites per shank , spanning 1 . 4 × 1 . 4 mm horizontally and in depth at 200 μm spacing , covering all cortical layers . For the analysis in Fig 6 , we used the 15 minutes of spontaneous activity , where no visual stimulation was provided to the animal . To obtain single-neuron spike trains , [38] performed super-paramagnetic clustering [105] on the high-frequency component ( 400–5000 Hz ) of the recorded signal . Details of the experimental procedures are given in [38] . In our analysis , we distinguish between low-fluctuation and high-fluctuation phases , with low vs . high activity in the frequency range up to 40 Hz . We defined these phases from the power spectrum |C ( ω ) |2 of the spike histogram for all neurons combined at subsequent intervals of 10 s duration and assign the interval to the low-fluctuation phase if ∫ 0 Hz 40 Hz | C ( ω ) | 2 d ω ≤ θ , with an empirically determined threshold θ = 0 . 8 ⋅ 108 . This leads to 77 intervals being classified as low-fluctuation and 15 intervals as high-fluctuation . Data were acquired from six male macaque monkeys ( 4 Macaca mulatta and 2 Macaca fascicularis ) . All experimental protocols were approved by the Animal Use Subcommittee of the University of Western Ontario Council on Animal Care and in accordance with the guidelines of the Canadian Council on Animal Care . Data acquisition , image preprocessing and a subset of subjects ( 5 of 6 ) were previously described [106] . Briefly , 10 5-min resting-state fMRI scans ( TR: 2 s; voxel size: 1 mm isotropic ) were acquired from each subject under light anesthesia ( 1 . 5% isoflurane ) . Nuisance variables ( six motion parameters as well as the global white matter and CSF signals ) were regressed using the AFNI software package ( afni . nimh . nih . gov/afni ) . The global mean signal was not regressed . The FV91 parcellation was drawn on the F99 macaque standard cortical surface template [47] and transformed to volumetric space with a 2 mm extrusion using the Caret software package ( http://www . nitrc . org/projects/caret ) . The parcellation was applied to the fMRI data and functional connectivity computed as the Pearson correlation coefficients between probabilistically weighted ROI timeseries for each scan [43] . Correlation coefficients were Fisher z-transformed and correlation matrices were averaged within animals and then across animals before transforming back to Pearson coefficients .
From a dynamical systems perspective , we define a state of the network as a set of mean firing rates for all populations . An attractor is a state towards which the network tends to evolve for many different initial conditions . Since the network receives stochastic external input , individual neurons fluctuate around their mean firing rate . An attractor is locally stable if all eigenvalues of the effective connectivity matrix , defined as the Jacobian of the population-level transfer function obtained from mean-field theory ( Eq ( 1 ) ) , have real values < 1 . The global stability of an attractor is assessed by the size of its basin of attraction in phase space . This volume is measured by discretizing the phase space into a grid of initial conditions and defining the global stability of an attractor A as the proportion of initial conditions leading the system to evolve to A . An analysis based on mean-field theory [37] and simulations reveals that across a wide range of configurations of the external input rate νext and the relative inhibitory synaptic strength g , the network possesses a bistable activity landscape with two coexisting locally stable fixed points ( Fig 2 ) . In view of the high dimensionality of the system with 254 populations , the bistability in the mean-field theory is found numerically from a pseudo-time integration that yields the stable fixed points [37] , in which the set of firing rates for the full set of populations consistently converges to one of two possible states for each combination of νext and g . The simulation results are qualitatively consistent with these mean-field results . We identify the stable fixed points based on the fact that , after a short initial simulation phase ( typically ∼100 ms ) and regardless of the initial condition , the network settles in either of these states . The first attractor exhibits asynchronous , irregular activity at moderate firing rates except for populations 5E and 6E , which are nearly silent ( Fig 2A ) , while the second features highly synchronized and regular firing with excessive rates ( Fig 2B ) in almost all populations . Depending on the parameter configuration , either the low-activity fixed point has a sufficiently large basin of attraction for the simulated activity to remain near it , or fluctuations drive the network to the high-activity fixed point . To counter the shortcoming of vanishing infragranular firing rates , we define an additional parameter κ which increases the external drive onto 5E by a factor κ = Kext , 5E/Kext compared to the external drive of the other cell types . Since the rates in population 6E are even lower , we increase the external drive onto 6E linearly with κ such that κ = 1 . 15 results in K6E , ext/Kext = 1 . 5 . However , even a small increase in κ already drives the network into the undesired high-activity fixed point ( Fig 2B ) . The stabilization procedure described by [37] uses mean-field theory to determine the population-averaged firing rates characterizing the fixed points of the system ( cf . Eqs ( 1 ) and ( 2 ) ) . By linearizing the population dynamics around the fixed points , the technique identifies connectivity components that are most critical to the global stability of the fixed points and yields targeted modifications of the connectivity within the margins of uncertainty of the anatomical data . The resulting average relative change in total indegrees ( summed over source populations ) is 11 . 3% . This allows us to increase κ while retaining the global stability of the low-activity fixed point . In the following , we choose κ = 1 . 125 , which gives K6E , ext/Kext = 1 . 417 , and g = 11 , νext = 10 spikes/s , yielding reasonable firing rates in populations 5E and 6E ( Fig 2C ) with sufficient global stability of the low-activity fixed point [37] . The stabilization renders the intrinsic connectivity of the areas more heterogeneous . Cortico-cortical connection densities similarly undergo small changes , but with a notable reduction in the mutual connectivity between areas 46 and FEF . For more details on the connectivity changes , see [37] . In total , the 4 . 13 million neurons are interconnected via 2 . 42 ⋅ 1010 synapses in the stabilized model . The network displays a reasonable ground state of activity with low spiking rates between 0 . 05 and 11 spikes/s ( Fig 3 ) . Inhibitory populations are generally more active than excitatory ones across layers and areas despite the identical intrinsic properties of the two cell types . This behavior , first found and discussed in detail in [36] , is thus caused by the network connectivity which leads to a high excitation-inhibition ratio onto inhibitory cells . Spiking activity is asynchronous irregular across populations . Population activity fluctuates around its stationary point with small amplitude . Pairwise correlations are low throughout the network ( Fig 3E ) . Excitatory neurons are less synchronized than inhibitory cells in the same layer , except for L4 . Spiking irregularity is close to that of a Poisson process across areas and populations ( Fig 3F ) . The only exception is population 6E , which features very low firing rates , so that the measure probably suffers from insufficient spiking data in single cells . To control interactions between areas , we scale cortico-cortical synaptic weights onto excitatory neurons by a factor χ = J cc E / J and provide balance by increasing the weights J cc I onto inhibitory neurons by twice this factor , J cc I = χ I χ J = 2 χ J . For increasing χ , we observe growing fluctuations of the population spiking rates . At χ = 2 and beyond , the network enters a high-activity state at some time point in the simulation , where most populations spike at unrealistically high rates ( Fig 4A ) . Predictions of mean-field theory show that for increasing χ , a growing proportion of initial conditions ( in Eq ( 2 ) ) result in states with increased activity ( Fig 4B ) . We explain this behavior with the global phase space of the model . At any time , there are two stable attractors with basins of attraction divided by the separatrix , a hyperplane in the phase space that contains unstable fixed points . The low-activity fixed point remains locally stable for increasing χ , as determined by the maximal real part of the eigenvalues of the effective connectivity matrix which is below one for all configurations ( Fig 4C ) . At the same time , its global stability , determined by the proportion of initial conditions leading the system to evolve to it , decreases ( Fig 4B and 4D ) . The effect is that fluctuations around the stationary state , which are evident in a stochastic system , let the system approach the separatrix more closely . Close to the unstable fixed points , the dynamics of the system slow down , which causes the rate fluctuations to appear . From χ = 2 , the system is likely to enter the high-activity state within a short amount of simulation time . In the following , we choose χ = 1 . 9 as the parameter configuration where slow fluctuations coincide with a sufficient global stability of the LA fixed point so that the system does not enter the HA fixed point during the simulation . The corresponding activity is irregular with plausible firing rates ( Fig 5A–5C ) . Irregularly occurring population bursts of different lengths up to several seconds ( Fig 5G ) arise from the asynchronous baseline activity ( Fig 5A–5C ) and propagate across the network . The time scales of the population bursts arise from network interactions rather than directly reflecting axonal delays or membrane and synaptic time constants , which only cover a range of 100 ∼ 101 ms . The firing rates differ across areas and layers and are generally low in L2/3 and L6 and higher in L4 and L5 , partly due to the cortico-cortical interactions ( Fig 5D ) . The overall average firing rate is 14 . 6 spikes/s , with the inhibitory populations tending to have higher rates than the excitatory populations . However , the strong participation of L5E neurons in the cortico-cortical interaction bursts causes these to fire more rapidly than L5I neurons . Pairwise correlations are low throughout the network ( Fig 3E ) . Unlike in the model without population bursts , excitatory neurons are more synchronized than inhibitory cells in the same layer , except for L6 . Spiking irregularity is close to that of a Poisson process across areas and populations , with excitatory neurons tending to fire more irregularly than inhibitory cells ( Fig 3F ) . Higher areas exhibit bursty spiking , as illustrated by the raster plot for area FEF ( Fig 5C ) . We compare the simulated spiking activity with experimental data from [38] , who recorded spiking activity in 140 neurons of macaque primary visual cortex in the spontaneous condition . The experimental activity shows activity phases differing in their low-frequency power ( Fig 6A ) . In the early stage of the recording , the population activity exhibits only small fluctuations ( Fig 6B ) , while in later stages , the population activity fluctuates on different time scales up to the order of a second ( Fig 6C ) . We therefore split the recorded data into low-fluctuation and high-fluctuation phases ( see Materials and methods for details ) , distinguished by their power at frequencies up to 40 Hz ( Fig 6E ) . To compare simulated with recorded power spectra , we compute the spike rates of 140 cells in V1 distributed across populations in proportion to the population sizes ( see Materials and methods for details ) . We compare three different simulations with low fluctuations ( χ = 1 , Fig 3 ) , meta-stable dynamics ( χ = 1 . 9 ) and unrealistically high activity ( χ = 2 . 5 ) in Fig 6D . The power spectral densities ( PSD ) of the simulations with χ = 1 , 2 . 5 are flat while for χ = 1 . 9 the PSD clearly reflects the slow oscillations in the spiking activity ( cf . Fig 5 ) . We compare the PSD of this simulation with the experimental results . Overall , the simulated activity in V1 to a good approximation reproduces both the spectrum from the entire recording period and that from the low-fluctuation phase , differing mainly in its increased power between 20 and 40 Hz . The sum of squared deviations ( SSD ) of the logarithmized spectrum from the logarithmized experimental spectrum for the entire recording period is SSD = 44 ( χ = 1 . 9 ) , compared to SSD = 793 for weak cortico-cortical synapses ( χ = 1 ) and SSD = 2180 for strong cortico-cortical synapses ( χ = 2 . 5 ) , showing that this match is unique to the metastable case . For the entire frequency range , the metastable case ( χ = 1 . 9 ) best matches the low-fluctuation phase ( SSD = 42 vs . SSD = 89 for the high-fluctuation phase ) . At frequencies below 3 Hz , the power spectrum of the simulations closely matches that of the high-fluctuation phase ( χ = 1 . 9: SSD = 3 . 2 vs . SSD = 6 . 8 for the low-fluctuation phase ) . The horizontal stripes in Fig 5I and 5J may to some extent be due to the mixing of spike trains from excitatory and inhibitory neurons , as the spike sorting does not distinguish between these . This interpretation is supported by the fact that the simulated activity across all layers and populations of V1 closely reproduces the broad distribution of spike rates across cells ( Fig 5L ) . The model with weak cortico-cortical synapses has an overrepresentation of near-silent neurons , while that with strong cortico-cortical synapses overrepresents neurons with high firing rates , and both settings lead to flatter spectra than observed experimentally . Thus , the close match between simulated and experimental population spectra and firing rate distributions is specific to the metastable state . To investigate inter-area propagation , we determine the temporal order of spiking ( Fig 7A ) based on the correlation between areas . We detect the location of the extremum of the correlation function for each pair of areas ( Fig 7B ) and collect the corresponding time lags in a matrix ( Fig 7C ) . In analogy to structural hierarchies based on pairwise connection patterns [92 , 93] , we look for a temporal hierarchy that best reflects the order of activations for all pairs of areas ( see Materials and methods ) . The result ( Fig 7D ) places parietal and temporal areas at the beginning and early visual as well as frontal areas at the end . The first and second halves of the time series yield qualitatively identical results . Fig 7D visualizes the consistency of the hierarchy with the pairwise lags: positive ( negative ) time lags are placed in the upper ( lower ) triangle of the reordered time lag matrix . To quantify the goodness of the hierarchy , we counted the pairs of areas for which it indicates the reverse order compared to that of the cross-correlation peaks . The number of such violations is 196 out of 496 , well below the 230 ± 12 ( SD ) violations obtained for 100 surrogate matrices , created by shuffling the entries of the original matrix while preserving its antisymmetric character . This indicates that the simulated temporal hierarchy reflects nonrandom patterns . The propagation is mostly in the feedback direction not only in terms of the structural hierarchy , but also spatially: activity starts in parietal regions , and spreads to the temporal and occipital lobes ( Fig 7G ) . However , activity troughs in frontal areas follow peaks in occipital activity and thus appear last . This predominant feedback propagation occurs despite feedforward connections on average constituting a greater proportion of the connections onto the neurons in the network than feedback connections , indicating that the dynamical state to an important extent determines the effective strength of anatomical connections . In particular , the high firing rates of the excitatory populations in layer 5 compared to those in layer 2/3 enhance the influence of feedback compared to feedforward projections , as feedforward and feedback projections arise predominantly from the supragranular and infragranular layers , respectively . We analyze the eigenspectrum of the effective connectivity matrix ( Fig 7E , cf . Fig 4 ) and find that the most critical eigenvector ( whose eigenvalue has the largest real part , marked in red in Fig 7E ) has the largest contributions in the areas at the bottom of the temporal hierarchy . To test whether the local structure of the areas , particularly the increased indegree in higher areas , alone predicts the relative instability in higher areas , we perform another test: We compute the maximum real part of the eigenvalues of the gain matrix in isolated areas with mean-field theory , where we replace inputs from other populations by Poisson input with a global rate of νext = 10 Hz . This does not yield a systematic correlation with the position of the areas in the temporal hierarchy . We conclude that areas at the bottom of the temporal hierarchy are the most unstable areas in the network , i . e . , fluctuations in these areas are less suppressed than in temporal and occipital areas , and that this is not a local effect but rather caused by the global network structure . We compute the area-level functional connectivity ( FC ) based on the synaptic input current to each area , which has been shown to be more comparable to the BOLD fMRI than the spiking output [107] . The FC matrix exhibits a rich structure , similar to experimental resting-state fMRI ( Fig 8A and 8C , see Materials and methods for details ) . In addition , we use the Balloon model of [94] to compute a BOLD signal from the area-averaged synaptic inputs ( see Materials and methods for details ) . The resulting matrix displays a similar structure as the FC matrix based on synaptic input currents . Overall , the values tend to be more extreme ( closer to +1 or −1 ) . There are several possible reasons for this: Since our network model comprises the visual cortex only and does not consider neuromodulation , it is potentially in a more confined state than a larger , neuromodulated model of this type and may therefore be less noisy than real brains . Furthermore , the spatial convergence and divergence of cortico-cortical connections presumably also contribute to the variability while in the model all cortico-cortical connections emanate from and target 1 mm2 of spatially unresolved microcircuitry . Lastly , the experimental data are averaged over six monkeys and inter-individual variability decreases the average absolute FC values due to a spread between both negative and positive values . In particular , the positive and negative FC values considered separately are larger in individual monkeys than in the averaged matrix . In the simulation , frontal areas 46 and FEF are more weakly coupled with the rest of the network compared to the experiment , but the anticorrelation with V1 as also found by [108] is captured by the model ( the corresponding entries in Fig 8A are light blue , see also Figs 8B , 5A and 5C ) . Area MDP sends connections to , but does not receive connections from other areas according to CoCoMac , limiting its functional coupling to the network . The structural connectivity of our model shows higher correlation with the experimental FC ( rPearson = 0 . 34 ) than the binary connectivity matrices from both a previous [109] and the most recent release of CoCoMac ( rPearson = 0 . 20 ) , indicating the importance of taking into account connection weights . For χ = χ I = 1 , areas interact weakly , resulting in low correlation between simulation and experiment ( Fig 8D ) . For increasing weight factor χ ( with χ I = 2 ) , the correlation between simulation and experiment improves . For intermediate cortico-cortical connection strengths , the correlation of simulation vs . experiment exceeds that between the structural connectivity and experimental FC , both for FC based on synaptic currents ( rPearson = 0 . 44 ) and simulated BOLD signal for χ = 1 . 9 ( rPearson = 0 . 38 , red dot in Fig 8D ) indicating the enhanced explanatory power of the dynamical model . To compare the agreement of our simulated FC with the inter-individual variability between the six monkeys used in the experiments , we compute the average correlation across the monkeys to be 0 . 31 . Comparing the simulated FC ( based on synaptic currents ) to the FC of individual monkeys yields an average correlation of rPearson = 0 . 39 ± 0 . 04 , showing that the agreement between the simulated and experimental FC reaches the upper limit determined by the inter-individual variability . From χ = 2 on , the network is increasingly prone to transitions to the highly active state ( cf . Fig 4 ) , causing the correlation to decrease . Thus , the highest correlation occurs in a state in which the model exhibits metastable dynamics and slow rate fluctuations appear . Such dynamical slowing close to instability has previously been demonstrated in models of cortical resting-state dynamics where the individual areas were described by population rate equations [110] or small numbers of spiking neurons [111] . Louvain clustering [112] , an algorithm optimizing the modularity of the weighted , undirected FC graph [96] , yields two modules for both the simulated and the experimental data ( Fig 8E ) . We denote the simulated modules by 1S , 2S and the experimental ones by 1E , 2E and compare these dynamical clusters with the community structure of the structural connectivity . In [35] , we describe six clusters in the connectivity exposed using the map equation method [98] . Applying the same method to the modified connectivity matrix used for the simulations in the present work yields seven clusters that are similar to the clusters of the original connectivity , and reflect known functional groupings ( Fig 8E ) . Three clusters contain dorsal stream areas and one large cluster gathers ventral stream areas . Furthermore , early visual areas and frontal areas each form separate communities . Ventral area VOT is grouped together with early visual area VP . The modules exposed by simulation combine these structural clusters . Cluster 1S contains early visual along with ventral and three dorsal regions . Cluster 2S merges parahippocampal with dorsal but also frontal areas . The experimental module 2E comprises early visual areas , ventral area V4 , and dorsal areas , while 1E consists of all other areas including also eight dorsal areas . This large degree of cross-over is most likely caused by an underestimation of the interactions between the areas in module 1E: ventral stream , frontal , and parahippocampal areas , and a subset of dorsal areas , since these are more strongly correlated in the experimental than in the simulated data . In conclusion , our analysis , summarized in Fig 8 , shows that the interactions between areas in the network resemble resting-state fMRI data and the agreement is highest when the network is in the metastable state at intermediate cortico-cortical connection strength . We investigate the laminar patterns of cortico-cortical interactions by computing conditional Granger causality on pairs of connected populations in different areas . Testing for significance ( p < 0 . 05 ) with Levene’s test [100] yields the pairs of causally influencing populations . We distinguish area pairs into three different categories based on the relation of their structural types . These categories approximately agree with the commonly used terminology of feedforward ( ∼high-to-low-type ) , lateral ( ∼horizontal ) and feedback ( ∼low-to-high-type ) directions . We observe systematic differences between the three different categories ( Fig 9A ) . High-to-low-type interactions preferentially originate in the supragranular layer and target layer 4 , while low-to-high-type interactions are controlled by population 5E and target the infragranular layers . Horizontal interactions share features of both other types . We compare the patterns of dynamical interactions with patterns of shortest paths in the structural connectivity between areas ( Fig 9B ) . High-to-low-type interactions resemble the shortest paths , but the other two types show clear differences to structural paths . Dynamical interactions in the horizontal direction are more similar to low-to-high-type interactions while horizontal shortest paths resemble high-to-low-type paths . The low-to-high-type interactions are dominated by population 5E , while the shortest paths suggest a stronger influence of population 6E . The shortest paths in the structural connectivity differ from the results of [35] because we here consider the modified connectivity after stabilization by the method of [37] as well as synaptic weights with an inter-area scaling factor of χ = 1 . 9 and respect the population-specific firing rates as opposed to [35] , where all populations are assumed to have equal activity levels . The detected shortest paths are nonetheless similar to the ones presented in Figure 8 of [35] . The most obvious difference is that paths onto inhibitory populations are significant in the simulated network ( Fig 9D ) , most likely due to the increased synaptic weight of corticocortical projections onto inhibitory populations . Furthermore , in low-to-high-type connections , paths onto L6 are more relevant . Overall , the number of significant connections is low , with approximately 4% of the cortico-cortical connections leading to significant causal interactions ( Fig 9C ) . Locally , one third of the connections carries causal interactions between the populations of an area . This observation reflects the high degree of local connectivity in cortex . The indegrees of connections with significant interactions are higher compared to all connections , for both cortico-cortical and local connections ( Fig 9E ) . Still , there is no strict dependence of causal interactions on high indegrees , since there are weak but causal connections as well as strong , non-causal connections . One reason is that the activity level of the projecting population plays a major role . The connectivity structure alone predicts for instance that 6E plays a dominant role in low-to-high-type projections , but these connections are actually not significant in the simulations . This is caused by the low activity level of 6E rendering these populations only influential in local interactions . Also the activity of the target populations modulates the effective interactions by determining the susceptibility to inputs due to the nonlinearity of the firing rate response curve [114 , 115] . In this sense , network simulations are more powerful than studies merely considering structural connectivity because they add these dynamical aspects to the cortico-cortical interactions . In conclusion , causal interactions in the network , summarized in Fig 9 , follow laminar patterns that depend on the relative architectural differentiation of areas . This dependence results from a combination of structural connectivity differences and dynamical states of source and target populations .
We present simulations of a multi-scale spiking network model of macaque visual cortex relating cortical connectivity to its dynamics . The connectivity [35] is refined by a mean-field-based stabilization procedure [37] incorporating fundamental activity constraints of non-vanishing , non-saturating spiking rates . The network produces a state with single-cell spiking statistics close to those from recordings in macaque V1 along with large-scale interactions resembling inter-area correlation patterns in macaque fMRI . The model predicts that cortex operates in a metastable regime , and exposes layer-specific , hierarchically organized channels mediating inter-area interactions . Population firing rates are layer-specific , inhibitory rates generally exceeding excitatory ones , in line with experiments [4 , 116 , 117] . Since excitatory and inhibitory neurons are equally parametrized and excitatory neurons receive equal or stronger external stimulation compared to inhibitory ones , we conclude that the connectivity causes these differences . The contribution of faster intrinsic dynamics of inhibitory neurons [118 , 119] merits future investigation . The mean firing rate of the network is not far above the mean spontaneous rate in V1 of alert monkeys [2] . However , common recording methods may miss ( near- ) silent neurons that would lower overall rates [120] . Future work can address this potential discrepancy . With sufficiently strong cortico-cortical connections , fluctuations let the system approach an instability where the dynamics drastically slow down [110 , 111 , 121 , 122] . In this metastable regime , the network closely reproduces the spike rate distribution across V1 neurons in lightly anesthetized macaque [38] . This state features variable-length population bursts mediating inter-area interactions , resembling cortical synchronized fluctuations in both spontaneous and stimulus conditions observed in mouse [123] , rat [124 , 125] , and macaque [7 , 104] . Overall , our simulations to a good approximation match the power spectrum of [38] , matching periods of higher synchrony at low frequencies . Simultaneous recordings of spontaneous activity in V1 and eye movements of a monkey sitting in darkness suggest that higher synchrony reflects drowsiness [126] . The low-frequency fluctuations may represent irregular Up-Down fluctuations produced by a bistable network with transitions between Up and Down states driven by external input to each area rather than arising locally from an adaptation mechanism , consistent with a recent experimental and modeling study [125] . Further work could distinguish such network effects from other sources of low-frequency fluctuations including NMDA and GABAB transmission , neuromodulation , and adaptation . The pattern of simulated interactions resembles fMRI resting-state activity from the anesthetized macaque . The functional connectivity ( FC ) based directly on the synaptic inputs provides similar predictions to that computed using the Balloon model , indicating that the low-frequency fluctuations present already at the level of the spiking activity drive these interactions . The agreement between simulation and experiment peaks at intermediate coupling strength in the metastable regime , potentially related to the slight subcriticality of the resting brain [127] . By combining these large-scale dynamics with plausible single-cell spiking statistics and layer-specific communication , our study extends earlier models exhibiting similar behavior with simpler local circuits [18 , 128] . We compute a BOLD signal based on the Balloon model of [94] from the simulated spiking activity and find that the resulting functional connectivity resembles the experimental data , but that the absolute values of the functional connectivity are larger than for the experiment and the synaptic input currents of the model . The model may underestimate the level of noise and thereby overestimate the functional connectivity due to the restriction to visual cortex and the lack of neuromodulatory effects , as well as the lack of convergence and divergence in cortico-cortical connections beyond the 1 mm2 scale . On the other hand , the experimental functional connectivity may be underestimated due to averaging of the experimental data across monkeys , which leads to a decrease in the absolute values owing to inter-individual variability . There is an ongoing debate in the literature regarding the extent to which fluctuations of functional connectivity during the resting state ( e . g . [129 , 130] ) are significant . Our simulations do not yield slow alternations between the clusters in our FC matrix . There are various possible reasons: experimentally observed FC dynamics in the resting state may be largely due to sampling variability and head motion [131]; the limited duration over which we can simulate and save the corresponding spiking data , which becomes gigabytes or even terabytes in size , is still too short to enable FC dynamics to be observed [132]; or FC dynamics may depend on variations in cognitive state [133] or vigilance [130] . Because we do not simulate the rest of the brain nor neuromodulation , such variations in cognitive and vigilance states are not captured . The model construction starts with the binary connectivity , defines the weighted structural connectivity based on connection densities , and finally yields the functional connectivity from simulations . Comparing with the experimentally observed functional connectivity , we find that each level ( binary structure → weighted structure → dynamical simulation ) adds explanatory power . In relation to fMRI functional connectivity , the role of anesthesia is worth investigating , as anesthetics influence BOLD responses in complex ways , although low doses as used in the experiment described here may have only mild effects , as shown in rat [134] and marmoset [135] . The model predicts that population bursts propagate stably across multiple areas , predominantly in the feedback direction , because parietal areas are more unstable than temporal and occipital areas , as revealed by linear stability analysis . This occurs despite higher relative indegrees of feedforward connections , indicating the importance of the dynamical state for the effective influence of the structural connectivity . Specifically , the comparatively high firing rate of the layer 5 excitatory neurons enhances the influence of feedback compared to feedforward connections , which mainly originate in supragranular layers . The systematic activation of parietal before occipital areas in the model parallels human EEG findings on information flow during visual imagery [136] and top-down slow-wave propagation during sleep in humans [89 , 90] and mice [91] . Our method for determining the order of activations resembles one recently applied to fMRI recordings [10] . It could be extended to distinguish between excitatory and inhibitory interactions like those we observe between V1 and frontal areas or to identify multiple ‘lag threads’ in the network [137] . Granger causality analysis of cortico-cortical interactions reveals layer- and population-specific communication channels that depend on the difference in the architectural types of the areas . Low architectural types correspond to low neuron density and a thin or absent layer 4 ( more limbic areas ) ; high architectural types correspond to high neuron density and a pronounced layer 4 ( more sensory areas ) [50 , 53 , 54 , 138] . In terms of visual processing hierarchies , areas of high architectural type are found at the bottom of such hierarchies while areas of low type constitute the top level of visual hierarchies . Interactions from high to low types , corresponding to feedforward communication , originate in layer 2/3 and target layer 4 , while interactions from high to low types , associated with feedback communication , are predominantly mediated by source neurons in layer 5 and target neurons in layers 6 and 4 . Thus , layer 5 neurons are the dominant source of feedback interactions , in contrast with an equal division between layers 5 and 6 in terms of anatomical connectivity [35] . This distinction is due to the higher activity in layer 5 than in layer 6 which enters the Granger causality analysis on both source and target side , whereas it influences only the target side in the path analysis of the structural connectivity via the gain of the receiving population . These findings differ from existing theories about predictive coding in cortical microcircuits [139 , 140] insofar that feedback signals preferentially reach granular and infragranular neurons rather than supragranular neurons . This suggests a more prominent role of layer 4 because in addition to feedforward signals and intrinsically produced feedback predictions ( via the local L2/3→L5/6→L4 pathway ) , statistical mapping of synapses to target neurons according to dendritic length [51] provides it with direct feedback predictions from higher areas via apical dendrites in upper layers . Incorporating a dual counterstream organization of feedforward and feedback connections [66 , 141] would allow a more refined analysis of these laminar interactions . Comparing these results with the network structure shows that substantial effective communication arises over a range of connection strengths due to the influence of the dynamical state of the populations , but that the weakest connections do not contribute significantly to communication . Our insights open up a new perspective on the significance of weak ties [142] , since our results suggest that they can gain relevance if they link highly active populations but their influence should not be overestimated by binarizing the connectivity . Incorporating local structure beyond population-specific connectivity of point neurons would enable studying how this could strengthen the influence of weak connections [143 , 144] . Our analysis further stresses the dominance of local communication in cortical dynamics . In the model , cortico-cortical indegrees are higher onto excitatory than onto inhibitory neurons , but stronger synapses onto inhibitory than onto excitatory neurons are required to maintain stability . This stabilization mechanism is similar to stabilization by inhibition locally in V1 [145] , but uses a different way to balance excitation and inhibition . Locally , the activity of inhibitory neurons increases in proportion to the activity of excitatory neurons [6] via a feedback mechanism [146–149] . In the corticocortical interactions in our model , the activity of the excitatory and inhibitory neurons is increased proportionally by a common input . In principle , other features could support stable inter-area activity propagation , such as spike-frequency adaptation of excitatory neurons [150 , 151] , short-term synaptic plasticity [152–154] , more detailed local balance [155] , and fine-tuned excitatory and inhibitory pathways [156–162] . However , optogenetic photostimulation experiments in mice provide some support for our prediction: inter-area EPSCs onto parvalbumin-expressing ( PV ) interneurons were found to be stronger than onto pyramidal neurons in mouse visual cortex , and mean EPSCs per pixel were larger onto PV interneurons at least for feedforward connections [163] . This outbalancing of excitation by inhibition resembles the “handshake” mechanism in the microcircuit model of [36] where interlaminar projections provide stability by their inhibitory net effect . In the model , this stabilization is reflected in the slightly larger proportion of significant interactions onto inhibitory than onto excitatory populations . In contrast , [20] model feedback connections onto layer 5/6 as net excitatory and onto layer 2/3 as net inhibitory , to reflect the hypothesis that the latter convey predictions suppressing feedforward activation . However , feedback connections may be facilitatory for stimuli within the classical receptive field and suppressive outside the receptive field [139 , 164–166] . Such details in feedback processing can be integrated and studied in future model versions . Random connectivity below the population level is unlikely to suffice for the brain to perform its computations . Including higher-order structures such as neuron-level motifs [167] and patchy connections would make the connectivity more realistic and enable the network to support specific activity patterns between groups of neurons inside a population [168] , such as synfire chains [169] . For instance , reciprocal connections and multiple synaptic contacts between neurons are overrepresented in cortex compared to random networks [156 , 161 , 162] , and the existence and strength of a synaptic connection between two neurons correlates with the existence of connections with other neurons [156 , 159] . Such cell-specific connectivity is likely to influence firing rate distributions , correlations , and propagation of activity; and theoretical studies have shown that they influence the computational capacity of cortical circuits [15 , 170] . To determine the large-scale connectivity , we included only connections from CoCoMac and [30] . CoCoMac alone contains information on approx . 78% of all pairs of cortical areas in the scheme of [45] and approx . 66% of all pairs of visual areas . Including the data of [30] , the existence or absence of a connection has been established for 76% of pairs of visual areas . It would be interesting to study the influence of additional connections predicted from graph-theoretical analyses [171] . Our model distinguishes only excitatory and inhibitory cells while cortical networks consist of many different cell types [162 , 172] . In particular , integrating different types of inhibitory cells and their detailed projection patterns [173–177] would enrich the model dynamics [178 , 179] . Furthermore , going beyond the simple single-neuron dynamics used in this study would enable one to study effects of intrinsic neuron dynamics , such as active calcium conductances in dendrites [180 , 181] , on the network state . For tractability , the model represents each area as a 1 mm2 patch of cortex . True area sizes vary from ~3 million cells in TH to ∼300 million cells in V1 for a total of ∼8 ⋅ 108 neurons per hemisphere of macaque visual cortex , a model size that with recent advances in simulation technology [28] already fits on the most powerful supercomputers available . Approaching this size would reduce distortions imposed by downscaling [34] and enable a more realistic representation of synaptic convergence . Overall , our model elucidates multi-scale relationships between cortical structure and dynamics , and can serve as a platform for the iterative integration of new experimental data , the creation of hypotheses , and the development of functional models of cortex .
|
The mammalian cortex fulfills its complex tasks by operating on multiple temporal and spatial scales from single cells to entire areas comprising millions of cells . These multi-scale dynamics are supported by specific network structures at all levels of organization . Since models of cortex hitherto tend to concentrate on a single scale , little is known about how cortical structure shapes the multi-scale dynamics of the network . We here present dynamical simulations of a multi-area network model at neuronal and synaptic resolution with population-specific connectivity based on extensive experimental data which accounts for a wide range of dynamical phenomena . Our model elucidates relationships between local and global scales in cortex and provides a platform for future studies of cortical function .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"action",
"potentials",
"medicine",
"and",
"health",
"sciences",
"diagnostic",
"radiology",
"functional",
"magnetic",
"resonance",
"imaging",
"neural",
"networks",
"nervous",
"system",
"membrane",
"potential",
"vertebrates",
"electrophysiology",
"neuroscience",
"animals",
"mammals",
"magnetic",
"resonance",
"imaging",
"primates",
"brain",
"mapping",
"computational",
"neuroscience",
"neuronal",
"dendrites",
"neuroimaging",
"old",
"world",
"monkeys",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"imaging",
"techniques",
"monkeys",
"animal",
"cells",
"macaque",
"cellular",
"neuroscience",
"radiology",
"and",
"imaging",
"eukaryota",
"diagnostic",
"medicine",
"cell",
"biology",
"anatomy",
"synapses",
"physiology",
"neurons",
"single",
"neuron",
"function",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"computational",
"biology",
"amniotes",
"neurophysiology",
"organisms"
] |
2018
|
A multi-scale layer-resolved spiking network model of resting-state dynamics in macaque visual cortical areas
|
While three countries in South Asia decided to eliminate anthroponotic visceral leishmaniasis ( VL ) by 2015 , its control in other regions seems fraught with difficulties . Is there a scope for more effective VL control in the Americas where transmission is zoonotic ? We reviewed the evidence on VL control strategies in Latin America—diagnosis , treatment , veterinary interventions , vector control—with respect to entomological and clinical outcomes . We searched the electronic databases of MEDLINE , LILACS , and the Cochrane Central Register of Controlled Trials , from 1960 to November 2008 and references of selected articles . Intervention trials as well as observational studies that evaluated control strategies of VL in the Americas were included . While the use of rapid diagnostic tests for VL diagnosis seems well established , there is a striking lack of evidence from clinical trials for drug therapy and few well designed intervention studies for control of vectors or canine reservoirs . Elimination of zoonotic VL in the Americas does not seem a realistic goal at this point given the lack of political commitment , gaps in scientific knowledge , and the weakness of case management and surveillance systems . Research priorities and current strategies should be reviewed with the aim of achieving better VL control .
Visceral leishmaniasis ( VL ) in Latin America is a severe systemic disease caused by an intracellular protozoon , Leishmania infantum ( syn . L . chagasi ) . VL is a zoonosis: the domestic dog is the main animal reservoir , while foxes and other wild animals play a role in sylvatic transmission [1]–[5] . The parasite is transmitted by a night-biting sandfly , Lutzomyia longipalpis , a 2 to 3 mm-long insect well adapted to the peri-domestic environment and distributed throughout Latin America [6]–[10] . L . infantum is also transmitted by Lu . cruzi in Brazil [11] and Lu . evansi in Colombia , and Venezuela [12] , [13] . Clinically , VL is characterized by prolonged fever , weight loss , hepatomegaly , splenomegaly , hypergammaglobulinemia and pancytopenia and it is usually fatal if not adequately treated [14] . Not all L . infantum infections lead to overt clinical disease: in Brazil ratios of 8–18 incident asymptomatic infections to 1 incident clinical case were described [15]–[17] . Risk factors for the development of clinical disease are only partially understood . Some studies suggest that the susceptibility to VL could be genetically determined [18]–[22] . Malnutrition places children at higher risk [23] , [24] . Other studies identified being a young male and the presence of animals in the neighborhood [25] , living in houses with a inadequate sewage system and waste collection [26] , and residence in an urban slum or in areas with green vegetation [27] as risk factors . The VL disease burden in Latin America is not exactly known because most countries lack effective surveillance systems [28]–[30] . Brazil declared a total of 50 , 060 clinical VL cases between 1990 and 2006 and this number accounts for 90% of all reported VL cases in the Americas , but is subject to substantial underreporting [29] , [31] . The country reported so far 176 HIV-coinfected VL cases [32] but has a significant number of asymptomatic co-infected individuals [33] , [34] . Whereas VL was initially concentrated in the poor rural areas in the northeast of the country , since the 1980s epidemics have occurred in major cities such as Belo Horizonte , Campo Grande , Natal , and others [35]–[37] . Some of these urban VL outbreaks were attributed to the migration of families from the rural areas to the peri-urban slums after periods of prolonged drought . Whereas the reported VL incidence in the 1980s averaged at 1 , 500 cases per year , this figure increased to an average 3 , 362 per year between 2000 and 2006 [31] . The disease has gradually spread south and eastward and is reported since 1999 from the states of São Paulo and Mato Grosso do Sul [38] . Human VL cases have also been reported from Honduras [39] , Venezuela [40] , Paraguay [41] and Argentina [42] . Sporadic and/or import human or canine cases were described in Chile [43] , Ecuador [44] , Bolivia [45] , Mexico [46] , Costa Rica [47] , and French Guyana [48] . A geographically referenced database providing links to published literature about the spatial distribution of VL can be accessed on http://apps . who . int/tools/geoserver/www/ecomp/index . html ( Accessed on September 19 2009 ) . Control of VL in the Americas has proved challenging . Early diagnosis and treatment is essential for the patient , but has limited impact on transmission if the main animal reservoir or insect vectors are not tackled [49] . Some studies showed a decreased incidence of VL in both dogs and children following serological screening and culling of seropositive dogs [50] , [51] , but this control strategy is increasingly debated [52] . Human VL incidence remained high in Brazil despite intensive application of this strategy in recent years [31] . Lack of impact has been attributed to the low sensitivity of the diagnostic tests , the long delay between diagnosis and culling and the low acceptance of culling by dog owners . Mathematical modeling suggests that vector control and vaccination of dogs would be more efficacious than dog culling [49] . Treatment of infected dogs is not an effective strategy as relapses are frequent , and dogs quickly become infectious again [53] . A controlled trial in a different setting of zoonotic VL ( Iran ) showed how the use of deltamethrin-treated dog collars reduced the risk of infection in dogs ( by 54% ) and in children ( by 43% ) [54] . Another controlled trial in Brazil showed only a modest effect on canine seroconversion rates [55] in spite of the proven effect of deltamethrin-impregnated dog collars on vector density [56] . In the Mediterranean region , where VL is also zoonotic with dogs playing a role as main reservoirs , human cases and canine cases are treated with antiparasitic drugs . In Europe , individual measures to protect dogs from sand fly bites using insecticides are common practices , but no public health surveillance and control interventions such as those applied in Brazil are in place [57] . Recently , the governments of India , Bangladesh and Nepal launched a VL elimination initiative , aiming to reduce the annual incidence of VL to less than 1/10 , 000 population by 2015 [58] . The strategy exploits recent technological developments in diagnosis , drugs and vector control [59] . Though the transmission pattern in this region is totally different , with L . donovani being the causative agent , a different sandfly vector ( P . argentipes ) and -most importantly- anthroponotic instead of zoonotic transmission , we wanted to examine whether there is a scope for VL elimination or at least improved control in the Americas . Given the heterogeneity in causative species , vector and transmission pattern , evidence on VL control tools from one region cannot be readily extrapolated to another . We report a review of the literature on the effectiveness of novel VL control tools and strategies in Latin-America structured around diagnosis of human and canine VL , treatment of human cases and control of the animal reservatoir and arthropod vectors .
As we have stated above , we set out to examine whether the existing control tools allow for elimination of VL in the Americas . The goal of elimination requires diagnostic and therapeutic tools that are very easy to use and can be easily decentralized . The World Health Organization now considers two ‘rapid diagnostic tests’ as appropriate for the diagnosis of VL in control programs: the Direct Agglutination Test ( DAT ) based on whole promastigotes of L . donovani or L . infantum and the rK39-ICT [60]–[62] . As it was not our intention to go into a full review of the available diagnostic tools for VL , we have excluded PCR and serological tests that require substantial laboratory equipment , even though there is extensive experience with the use of IFAT and ELISA tests in the Americas . Moreover , the clinical benefit of antigen-detection and PCR tests still needs to be demonstrated [63] , [64] . We therefore limited our systematic review to DAT and rK39-ICT . The eligibility criteria included: original studies evaluating the DAT or the rK39 immunochromatographic test ( ICT ) ; clinical visceral leishmaniasis diseases in humans as target condition; adequate reference classification; absolute numbers of true-positive , true-negative , false-positive and false-negative observations available or derivable from the data presented . Accuracy measures were summarized as sensitivity and specificity . Clinical trials including uncontrolled and retrospective studies with description of the following characteristics: intervention; case definition; follow-up schedule; therapeutic endpoints; control group; and efficacy measure defined through cure and failure proportions for each treatment . Original studies evaluating any diagnostic test for canine leishmaniasis; Leishmania infection and/or VL disease in domestic dogs as target condition; adequate reference classification; absolute numbers of true-positive , true-negative , false-positive and false-negative observations available or derivable from the data presented . Accuracy measures were summarized as sensitivity and specificity . Field trials of control measures ( canine culling , impregnated dog collars , canine vaccination , insecticide spraying , insecticide treated bednets , environmental management ) evaluating at least one control measure; description of the intervention under analysis; target population , sampling and randomization process; adequate case definitions for asymptomatic infection or VL; definition of outcomes related to humans , dogs or sand flies; at least one effect measure; and at least one point estimation for the magnitude of the expected effect .
A Medline search generated 77 papers , and LILACS 179 . After screening the titles and abstracts of those papers for evaluations of the DAT or rK39 in human VL , we retrieved eight original papers ( Figure 1 and Table 1 ) . We report only descriptive statistics of sensitivity and specificity estimates; without drawing conclusions about differences in these parameters between tests and discuss them in comparison with results of a meta-analysis by Chappuis et al . [65] . Seventy-seven papers were retrieved from Medline/PubMed search and 11 of them were considered relevant . The LILACS database search retrieved 26 papers of which 2 were considered relevant , but 1 was already obtained from the PubMed database ( Figure 2 ) . Finally , 12 papers were included in the review , covering 5 serological tests for canine VL: IFAT , ELISA , dot-ELISA , DAT , and rK39-ICT [66] , [69] , [74]–[83] . IFAT has been the test adopted by the Brazilian Ministry of Health for its dog screening-and-culling campaigns . Published estimates for sensitivity range from 72–100% , for specificity 52–100% ( Table 2 ) . The moderate sensitivity and specificity of this test , the long turn-around time between sample taking and culling , and the complexity of its execution have been invoked as one of the reasons for the low effectiveness of the culling campaign . Several ELISA tests have been evaluated , with assays based on homologous antigens usually showing higher sensitivity . Evans et al ( 1999 ) showed a higher sensitivity of ELISA compared to IFAT and pleaded for a revision of the screening policy [84] . Recently more “user-friendly” diagnostics as the DAT and a canine version of the rK39-ICT were evaluated with good results . For the freeze-dried DAT sensitivity ranged from 85–100% , specificity 89–100% [65] , [76] , [78] and for the rk39-ICT sensitivity ranged from 72–96% , specificity 62–100% [81] , [82] . The main advantage of these rapid tests would be to shorten the delay between diagnosis and culling/treatment . However , the reported estimates of sensitivity in the above studies depend on the type of dogs included in the “true cases” group with higher sensitivity observed in symptomatic than in asymptomatically dogs , and unfortunately , several evaluations failed to include an adequate sample of asymptomatically infected dogs . The sensitivity of the test in asymptomatic dogs is crucial for a control strategy , as those dogs are infectious , and should be targeted by the campaign . Sensitive antigen detection tests as PCR might become a relevant marker of infection in the future with the advantage that they can still be used in vaccinated dogs that will be serologically positive because of the vaccine . However , Quinnell et al ( 2001 ) showed in a longitudinal study of naturally infected dogs how the sensitivity of PCR was high early after infection but declined to 50% thereafter . The sensitivity of serology also varied with time , being lowest at the time of infection but clearly superior thereafter ( 93–100% ) . They concluded that PCR was most useful for detection of active disease , and considered serology as more adequate for the detection of infection [84] . Thirty-nine papers were retrieved from Medline/PubMed search and four of them were considered relevant . The LILACS database search retrieved 42 papers of which 24 were not available from the PubMed database . Three of those 24 studies were considered relevant , one of them , was previously identified through the PubMed search . One paper was identified through specific author's name searching in PubMed . The Cochrane Central Register of Controlled Trials search retrieved 103 trials , three of them were conducted in the Americas but all were also identified through the PubMed and LILACS searches . Finally , seven papers were included for review [85]–[91] . Three papers were excluded from further analysis , one because it was a second publication on the same trial [88] , one for being a retrospective study with heterogeneous therapeutic interventions with meglumine antimoniate and case definition based on clinical findings plus positive serology without description of the methods and test cut-off . A minority of cases was diagnosed through parasite identification [85] , and one paper because it was a case-control study focusing on prognostic factors [87] . The flow for the selection and a summary of the reviewed studies appears in Figure 3 and Table 3 . Dietze et al ( 1993 ) reported an open-label dose-escalating trial with amphotericin B colloidal dispersion ( Amphocil ) in two small groups of patients who showed similar cure rate suggesting that the 7 days was as effective as the 10 days regimen [90] . In 1995 the same authors reported another open-label trial with Amphocil with a shorter regime of 5 days , observing an episode of relapse [91] . Berman et al ( 1998 ) reported the results of an open-label phase II trial with three therapeutic regimens consisting of liposomal amphotericin B 10 , 14 or 20 mg/kg total dose; the reported outcomes were cure , failure and relapse and the follow-up period was of six months . This paper suggested that the lower 10mg/kg total dose was less efficacious than the higher 20mg/kg total dose [86] . Dietze et al ( 2001 ) concluded from an open-label dose-escalating safety and efficacy trial that sitamaquine was not efficacious for the treatment of VL in young adults . Severe adverse events described as renal toxicity lead to trial interruption when using the higher dose of 3 . 25mg/kg/d [89] . Incidence and prevalence estimates of canine VL in the Americas have been reported from several foci [2] , [40] , [92]–[95] , but the specific relationship between canine and human VL cases is not well understood . Transmission in the dog population is mainly due to infected sandfly bites but alternative routes have been proposed such as sexual transmission and other potential insect vectors [96]–[98] . The control of the animal reservoir is complex and frequently involves combined interventions . The Brazilian Control Program recommends a strategy based on canine culling and vector control with insecticide spraying . Insecticide-impregnated collars for dogs and canine vaccination are not currently recommended as public health control measures [99] . One-hundred seventy-two papers were retrieved from Medline/PubMed using the search strategy cited above . The LILACS search was performed using the term visceral leishmaniasis because no document was retrieved when using the PubMed approach . The LILACS search was less specific and 519 documents were retrieved; 514 documents comprised an extensive spectrum of research irrelevant for the purpose of this paper and four of the five relevant papers were already identified through the PubMed search . After reading the titles and the abstracts and hand searching reference lists for related papers , fourteen were selected for full text reading because the main subject was at least one intervention for control VL ( Figure 4 ) [50] , [55] , [100]–[111] . Magalhães et al ( 1980 ) published a retrospective –non controlled- study on the impact of a combined intervention consisting of human VL case treatment , culling of seropositive dogs and insecticide spraying with DDT in 19 municipalities of the Rio Doce Valley , State of Minas Gerais , Brazil reporting the disappearance of human symptomatic cases after 15 years of application of this strategy [100] . Dietze et al ( 1997 ) reported a field trial of dog screening and culling , based on twice-yearly screening with DOT-ELISA . This trial was conducted in three rural valleys , State of Espirito Santo , Brazil , two benefiting from the intervention and one used as control . At 6-months there was a 16% reduction of seroconversion rate in dogs ( 36% in the intervention vs . 52% in the control group ) , but this difference was not significant [101] . Braga et al ( 1998 ) reported the comparison of two strategies of dog screening-and-culling: screening by ELISA was compared to IFAT as routinely recommended by the National Control Program . The main difference consisted in the lag times after blood sampling ( 7 days for ELISA vs . 80 days for IFAT ) . The trial was conducted in a rural area of Northeastern Brazil where 28 communities were systematically allocated to one of the two groups . In the ELISA arm , reduction of canine seroprevalence was higher , probably due faster dog removal plus higher sensitivity of the ELISA test [102] . Ashford et al ( 1998 ) reported a controlled intervention trial of seropositive dog removal in an endemic area of the State of Bahia , Brazil . The intervention area was subjected to screening with FAST-ELISA and removal of seropositive dogs , in the control area no intervention was carried out . A significant reduction of dog seroconversion rate in the intervention area as compared to control was observed , and a significantly lower number of VL cases reported to health facilities in the intervention area [50] . Paranhos-Silva et al ( 1998 ) report a follow- up study of several clusters of seronegative dogs in Jequié , State of Bahia , Brazil . The initial prevalence of infection among 1681 dogs was 23 . 5% . After serological screening every six months for 18 months and removal of the seroconverters , the annual incidence rate of infection was 6 . 55 cases/100dog-years . The migration of dogs between clusters was 2 . 3 cases/100 dog-years . This study is relevant because as highlights the challenges posed by dog migration for any control program dealing with the canine reservoir [103] . Da Silva et al ( 2000 ) reported a phase III vaccine field trial in seronegative dogs screened with IFAT and FML-ELISA and exposed to fucose-mannose-ligand vaccine in three subcutaneous doses at 21 day intervals . Control arm was treated with saline placebo . Endpoints were symptomatic VL or death , seroconversion rates in FML-ELISA and conversion of leishmanin skin test composed of crude L . donovani antigen . Follow-up evaluations were performed at 2 , 7 , 13 and 24 months . A significant difference in the three endpoints was observed during the trial . The overall efficacy to prevent symptomatic VL disease was 75% [104] . Giffoni et al ( 2002 ) reported the effect of application of a 65% permethrin spot-on formulation on canine VL infection and sandfly abundance . A decrease of canine VL prevalence was observed in the intervention area compared with increased prevalence in the control area . No effect was observed on sandfly population [105] . Feliciangeli et al ( 2003 ) described a controlled trial of pyrethroid ( λ-cyhalothrin ) indoor spraying every 5 month and organophosphate ( fenitrothion ) ultra-low volume spatial fogging around the houses twice a month for ten months in one intervention compared to one control area . The main vector captured was Lu . longipalpis . A significant decrease of sandfly abundance was observed , with a residual effect of indoor spraying of 3 months . Main limitation of this study was the specific construction style of the houses: completely cemented , plastered and oil-painted walls and zinc roofs , which lowers its external validity [106] . De Oliveira et al ( 2003 ) reported the evaluation of routine combined control measures of seropositive dog-culling and insecticide spraying during six years . The intensity of the application of control measures correlated with human VL incidence , the coverage of canine surveys , the number of canine surveyed and the number of buildings submitted to insecticide spraying [107] . Reithinger et al ( 2004 ) reported a controlled field trial to evaluate the effectiveness of insecticide impregnated collars to prevent infection detected through serological tests or DNA detection by PCR assay in one intervention compared to one control area . The authors failed to detect a significant difference between groups in the incidence of new infections but they demonstrated a significant reduction of antibody titers in the collar protected dogs . Mathematical modeling using the results obtained in this study suggests that dog collars would be a better alternative than dog culling [55] . Moreira et al ( 2004 ) reported the incidence rates of canine Leishmania infection in a cohort of dogs submitted to an optimized culling strategy consisting of: ( i ) ELISA screening of serum samples; ( ii ) shortening of the time interval from serodiagnosis to removal of dogs; ( iii ) screening a high proportion of the dog population . They demonstrated that the incidence of canine infection remained stable through 2 . 5 years of observation under this strategy but the study had no control arm for comparison . A high replacement rate by susceptible puppies and already infected dogs was observed [108] . Courtenay et al ( 2007 ) reported the barrier effect , the 24-h mortality rate and the human landing rates of Lu . longipalpis in households using deltamethrin-impregnated bednets compared others using untreated bed nets . The study described a 39% increase in barrier capacity of the impregnated bednets , 80% reduction in sandfly landing rates on humans and 98% increase in the 24-h sandfly mortality rates . The study was done under field conditions with a small number of observations during a very short period of exposure to the treated bednets ( three days ) and the residual effect was not measured . However this intervention should be explored further because it could bring an additional benefit in areas where malaria is also endemic [109] . Costa et al ( 2007 ) reported a randomized community intervention trial to compare the effect of four strategies on human VL , as follows: ( i ) spraying houses and animal pens with pyrethroid insecticide; ( ii ) spraying houses and eliminating seropositive dogs; ( iii ) combination of spraying houses and animal pens plus eliminating seropositive dogs; and ( iv ) spraying houses only as the reference comparator . The outcome was evaluated by measuring incidence of seroconversion in humans six months after the application of interventions . The results indicated a positive effect of canine removal on incidence of leishmanial infection in men but surprisingly , the combination of dog culling plus outdoor spraying of peridomestic animal shelters failed to demonstrate any effect . The relevance of this study is that it constitutes the first attempt to measure the effect of combined interventions on human VL incidence [110] . De Souza et al ( 2008 ) reported a randomized community intervention trial to compare the effect of ( i ) pyrethroid insecticide spraying; ( ii ) pyrethroid insecticide spraying plus culling of seropositive dogs with ( iii ) no intervention . The interventions were maintained for two years and outcomes were registered every year , insecticide spraying was performed every 6 months . Although a lower incidence was observed in the groups submitted to interventions and that reduction was more intense after two years , the study failed to detect statistically significant differences [111] . The summarized characteristics and main limitations of these studies are shown in Table 4 .
This review of evidence related to VL control in Latin America revealed that a lot of work remains to be done in order to clarify the dynamics of Leishmania transmission in human , canine and vector populations . The exact burden of disease remains largely unknown . The increasing trend of VL cases observed in Brazil and the spread of transmission to previously not affected areas raise doubts about the impact of ongoing control measures . The determinants of human infection and of symptomatic disease are also poorly understood with the exception of the nutritional status in young children . To diagnose VL in humans the rK39-ICT has clear advantages over the IFAT or ELISA based tests that are widely used in Latin America . The DAT assay has shown similar diagnostic performance but is not as user-friendly as the rK39 . The research priorities in this field should be geared towards diagnostic accuracy studies in large prospective trials ( phase-3 ) and to study diagnostic performance in specific groups such as HIV co-infected patients . Current treatment practice in VL in Latin-America is based on rather weak scientific evidence . It is worrisome that case fatality rates remain high and are even increasing , at least in Brazil . The lack of clinical evidence from the region is very worrying . We retrieved not a single phase-3 randomized controlled trial on VL conducted in the Americas . Nowadays , one phase-2 trial with miltefosine is ongoing and two Brazilian large randomized controlled trials with liposomal amphotericin B , amphotericin B deoxycholate and meglumine antimoniate are expected to initiate recruitment in 2009 . The research priorities include well-designed clinical trials with pentavalent antimonials , amphotericin B deoxycholate and the liposomal formulations , miltefosine and drug combinations . Although the resistance to antimonials observed in India is less relevant in Latin America , drug combinations are attractive because their potential for shortening treatment schemes and reduction of toxicity . Clinical factors associated with treatment failure should be studied to contribute to the development of a prognostic score that allows early interventions to reduce case fatality rates [14] , [87] . Control interventions targeting the dog reservoir for culling/treatment require accurate assays able to detect the asymptomatic infections as well as the symptomatic dogs . Validating such tests is no easy task , as there is no adequate gold-standard for the diagnosis of asymptomatic infection . PCR-assays seem to be very attractive but estimating their accuracy and reproducibility still constitutes a research priority . Moreover , novel screening strategies based on combined , parallel or sequential use of current available tests needs to be validated . Another challenge faced in canine diagnosis is the distinction of positive serology results produced by natural infection from those induced by vaccines . The development and proper validation of tests with capacity to discriminate both phenomena are crucial to avoid interference with concomitant interventions including dog culling and vaccination in the same area . Furthermore , the study of the determinants of dog infectiousness for the sandfly vector is essential to define the best culling strategy [112] , [113] and the determinants of dog susceptibility to infection [114] is crucial for the design of canine vaccine trials . Some of the problems with the design of the community intervention trials we reviewed are related to the lack of accurate diagnostic methods to define the relevant outcomes in the human and canine population . Furthermore , the definition of a control group is challenging because of an obvious ethical dilemma . The heterogeneity of disease transmission within the study area often generated imbalances in the baseline comparisons among groups and the random allocation process is also complex because of the mobility of the human , canine and vector population . Most of the reported community trials used a too limited number of clusters for comparison ( usually a one to one comparison ) . In spite of all those limitations a relevant number of reports could be reviewed in detail , showing no strong evidence for a significant impact on VL transmission for any of the interventions reviewed . Canine culling seems to be the least acceptable intervention at community level for obvious reasons and has low efficiency due to high replacement rate of eliminated dogs with susceptible puppies [103] , [115] , [116] . Vector control interventions are better accepted by the affected populations and mathematical models suggested encouraging efficacy , but they need further study . Better knowledge of vector seasonality and behavior is required for proper timing of these interventions . The current evidence indicates that spatial fogging is useless and that the residual effect of house wall spraying is very short [106] , [117] . Insecticide impregnated collars seem to have a longer residual effect [56] and theoretical advantages over the other methods and should be studied in larger and well-designed controlled trials . The potential emergence of resistance to insecticides should also be considered for the long-term planning of any vector control intervention [118] . Canine and human vaccine development needs to be prioritized . The dog vaccines already registered in Brazil have some protective effect against canine VL but none of them were properly evaluated as control measures against human VL [119] , [120] . Such evaluation is challenging as field trials should include relevant canine endpoints , related to dog infectiousness for the sandfly vector , as well as relevant human endpoints , that include symptomatic and asymptomatic infections in order to obtain precise estimates of the vaccine effect on transmission rates . Human vaccine development is expected to take at least several years to obtain efficacious and safe candidates for clinical trials . Furthermore , the surrogate markers of the desired protective effect are not well understood and the definition of target population for such products will be a matter of intense debate . The role of sylvatic and peridomestic animals such as foxes , marsupials and rodents in some relevant VL transmission scenarios deserves more specific research [6] . Last but not least , in countries such as Brazil , where the government has put the elimination of hunger as a political priority , targeted nutritional support in VL risk areas would be an interesting and probably cost-effective intervention from a societal perspective . Similarly , schemes for the improving of housing and waste management as well as other general measures involving active community participation should be encouraged [121] , [122] . Finally , the strengthening of the surveillance system capacity is essential to avoid the underreporting of human cases [123] and to follow-up the infection behavior in canine population . Strong surveillance will certainly contribute to improve data quality for decision-makers in this complex scenario . The elimination of zoonotic VL in Latin America is not ( yet ) a realistic goal taking into consideration the complexity and diversity of its transmission scenarios , the scientific knowledge gaps and the lack of adequate and properly validated interventions . Many countries perceive the burden of leishmaniasis as negligible; there is not much political support nor funding for VL control . The zoonotic nature of transmission is an additional constraint that limits the impact of the few known effective prevention and control interventions . Nonetheless we believe the improved control of VL is possible if the region builds the political will , develops a more coherent regional control policy , and invests in better case management and epidemiological surveillance systems . The implementation of a focused research agenda to support such control initiative is essential .
|
Visceral leishmaniasis is a vector-borne disease characterized by fever , spleen and liver enlargement , and low blood cell counts . In the Americas VL is zoonotic , with domestic dogs as main animal reservoirs , and is caused by the intracellular parasite Leishmania infantum ( syn . Leishmania chagasi ) . Humans acquire the infection through the bite of an infected sand fly . The disease is potentially lethal if untreated . VL is reported from Mexico to Argentina , with recent trends showing a rapid spread in Brazil . Control measures directed against the canine reservoir and insect vectors have been unsuccessful , and early detection and treatment of human cases remains as the most important strategy to reduce case fatality . Well-designed studies evaluating diagnosis , treatment , and prevention/control interventions are scarce . The available scientific evidence reasonably supports the use of rapid diagnostic tests for the diagnosis of human disease . Properly designed randomized controlled trials following good clinical practices are needed to inform drug policy . Routine control strategies against the canine reservoirs and insect vectors are based on weak and conflicting evidence , and vector control strategies and vaccine development should constitute research priorities .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases/protozoal",
"infections",
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"infectious",
"diseases/epidemiology",
"and",
"control",
"of",
"infectious",
"diseases"
] |
2010
|
Control of Visceral Leishmaniasis in Latin America—A Systematic Review
|
Many bacterial pathogens and symbionts use type III secretion machines to interact with their hosts by injecting bacterial effector proteins into host target cells . A central component of this complex machine is the cytoplasmic sorting platform , which orchestrates the engagement and preparation of type III secreted proteins for their delivery to the needle complex , the substructure of the type III secretion system that mediates their passage through the bacterial envelope . The sorting platform is thought to be a dynamic structure whose components alternate between assembled and disassembled states . However , how this dynamic behavior is controlled is not understood . In S . Typhimurium a core component of the sorting platform is SpaO , which is synthesized in two tandemly translated products , a full length ( SpaOL ) and a short form ( SpaOS ) composed of the C-terminal 101 amino acids . Here we show that in the absence of SpaOS the assembly of the needle substructure of the needle complex , which requires a functional sorting platform , can still occur although with reduced efficiency . Consistent with this observation , in the absence of SpaOS secretion of effectors proteins , which requires a fully assembled injectisome , is only slightly compromised . In the absence of SpaOS we detect a significant number of fully assembled needle complexes that are not associated with fully assembled sorting platforms . We also find that although binding of SpaOL to SpaOS can be detected in the absence of other components of the sorting platform , this interaction is not detected in the context of a fully assembled sorting platform suggesting that SpaOS may not be a core structural component of the sorting platform . Consistent with this observation we find that SpaOS and OrgB , a component of the sorting platform , share the same binding surface on SpaOL . We conclude that SpaOS regulates the assembly of the sorting platform during type III secretion .
Type III protein secretion systems ( T3SSs ) are highly specialized multiprotein molecular machines with the capacity to inject bacterially-encoded proteins into target eukaryotic cells . Encoded by a large variety of gram-negative bacteria , T3SSs are central to the interactions of many pathogens and symbionts with their respective hosts[1–3] . The type III secretion machine is made up of several substructures that come together to form the injectisome[1 , 4–7] . The core component of the injectisome is the needle complex ( NC ) , which is composed of a multi-ring base anchored in the bacterial envelope , and a filament-like extension that protrudes several nanometers from the bacterial surface[4 , 6 , 8 , 9] . The needle filament is traversed by a narrow , ~2 nm channel and is capped at its terminal end by the tip complex , which is thought to be involved in sensing target cells and deploying the translocation pore that mediates the passage of effectors through the target cell plasma membrane[10–15] . The NC is associated to a very large cytoplasmic complex known as the sorting platform , which is responsible for selecting the type III secretion substrates and initiating them into the secretion pathway in the appropriate order[16] . Recent cryo electron tomography ( cryo-ET ) studies in Salmonella Typhimurium and Shigella flexneri have provided a high-resolution view of this substructure of the injectisome[5 , 17] . The sorting platform exhibits a unique cage-like architecture , enclosed by 6 pods that emerge from the NC and converge into a 6-spoke wheel-like structure that caps it at its cytoplasmic side . In the S . Typhimurium T3SS encoded within its pathogenicity island 1 ( SPI-1 ) , the cage-like structure is made up of the OrgA , SpaO , and OrgB proteins , which serve as scaffold to place the associated ATPase InvC in close apposition to the export apparatus[5] . The ATPase plays an essential role in initiating substrates in the secretion pathway by removing their associated chaperones and unfolding the effectors prior to their threading through the narrow secretion channel[18] . Several pieces of evidence indicate that , unlike other substructures of the injectisome , the sorting platform exhibits a dynamic behavior and may undergo cycles of assembly and disassembly[19–21] . Fluorescence-imaging studies have shown that cytoplasmically located subunits of the sorting platform can exchange with a pool located in close proximity to the membrane and thus presumably associated to the NC[19 , 20] . Consistent with these observations , biochemical and super-resolution imaging studies in live bacteria have shown that a substantial proportion of the core components of the sorting platform are not associated to the NC[21] . Although this dynamic behavior is thought to be essential for the function of the sorting platform , very little is known about its regulation . We show here that SpaOS , which is the product of an internally-translated product of spaO[22] , influences the assembly and/or stability of the sorting platform . Therefore , we propose that SpaOS may regulate the dynamic behavior of this injectisomes substructure .
It is often the case that the stability of individual component of multi-protein complexes is affected by the absence of some of the components of such a complex . In addition , insight into the assembly of a multi-protein complex can be often obtained from the observation of the relative stability of its subunits in the presence or absence of one another . We therefore examined the stability of the structural components of the S . Typhimurium SPI-1 T3SS sorting platform SpaO , OrgA , OrgB , and InvC in different S . Typhimurium mutant strains lacking each one of these components . We found that the levels of OrgA , OrgB , and InvC were significantly reduced in the absence of SpaO ( Fig 1A and 1B and Figure A in S1 Text ) . In contrast , SpaO was stable in the absence of OrgA , OrgB , or InvC ( Fig 1C and 1D and Figure A in S1 Text ) . These observations suggest that SpaO serves as the core component of the sorting platform and therefore it may play a more central role in the coordination of its assembly . The internally translated C-terminal polypeptide from the SpaO homologs in Yersinia spp . and Shigella spp . YscQS and Spa33S , respectively , have been shown to be required for type III secretion in these bacteria[23 , 24] . Therefore , it has been proposed that YscQS and Spa33S are essential structural components of the sorting platform . We investigated the contribution of the internally translated fragment of SpaO ( SpaOS ) to the function of the S . Typhimurium SPI-1 T3SS . We constructed a S . Typhimurium mutant strain ( S . Typhimurium spaOGTG ( 203 ) »GCG ) in which the internal translational initiation GTG codon ( codon 203 in the spaO gene ) for SpaOS was changed to the non-initiating GCG codon , which prevented its expression ( Figure B in S1 Text ) . We then examined the function of the SPI-1 T3SS in a Salmonella mutant strain carrying such mutation by biochemical and functional assays . We found that the mutant strain was able to secrete substrates into the culture supernatant although at slightly reduced levels in comparison to wild type ( Fig 2A ) . Furthermore , the mutant was able to enter into cultured epithelial cells , a measure of the function of the SPI-1 T3SS , in a manner indistinguishable from that of wild type indicating that the levels of T3SS-mediated effector secretion , while slightly decreased , were sufficient to mediate a critical function of this system ( Fig 2B ) . These results indicate that , unlike T3SS in other bacteria , SpaOS does not play an essential role in the function of the S . Typhimurium SPI-1 T3SS but rather , it increases the efficiency of the secretion system . It has been previously shown that , similarly to SpaOS , its homolog in the S . Typhimurium SPI-2-encoded T3SS ( SsaQS ) , is not essential for T3SS function[25] . Instead , it has been proposed that SsaQS may help the folding of SsaQL since in the absence SsaQS , the stability of SsaQL is compromised[25] . However , we found that in the absence of SpaOS the levels of SpaOL were not significantly affected . The observation that SpaOL remains stable in the absence of other component of the sorting platform ( see Fig 1C and 1D and Figure A in S1 Text ) prompted us to test whether such stability may depend on SpaOS . We therefore tested the stability of SpaOL in the absence of both , OrgB and SpaOS . We found that the stability of SpaOL is compromised when both SpaOS and OrgB are absent ( Fig 2C ) . These results suggest that SpaOS may serve to stabilize SpaOL prior to its association to other core components of the fully assembled sorting platform . We found that , as previously shown[22] , in vitro SpaOL and SpaOS interact with one another in a 1:2 stoichiometric ratio ( Fig 3A ) . Similar observations have been made with homologs of other T3SSs[23 , 24] . However , the investigation of the interaction of SpaOL with SpaOS in vivo in the presence of all the T3SS components has not been reported . To determine whether SpaOS interacts with SpaOL in vivo we generated a strain that expresses differentially-tagged versions of SpaOL and SpaOS from the Salmonella chromosome ( Figure C in S1 Text ) . The interaction between SpaOL and SpaOS was then probed by affinity purification of SpaOL after bacterial growth under conditions that stimulate the expression of all components of the SPI-1 T3SS . No interaction between SpaOL and SpaOS was detected under these conditions even though the interaction of SpaOL and OrgB was readily detected ( Fig 3B ) . These results suggest that the interaction between SpaOL and SpaOS in vivo is transient and may not be captured in the presence of other components of the SPI-1 T3SS , conditions that lead to fully assembled sorting platforms . These results also suggest that SpaOS may not be a structural component of the sorting platform . To further investigate this hypothesis we compared the in situ structures of the sorting platforms of wild-type S . Typhimurium and that of a mutant unable to produce SpaOS using cryo-ET . We have recently used this approach in combination with bacterial minicells to obtain a high-resolution structure of the entire SPI-1 T3SS injectisome in situ[5] . We found that in the absence of SpaOS , a significant number of injectisomes displayed a needle filament ( Fig 3C ) . This is consistent with the observation that the injectisome is functional in this mutant strain ( Fig 2A and 2B ) , since assembly of the needle filament requires a fully assembled sorting platform . Sub-tomogram classification was utilized to analyze the sorting platforms present in this mutant strain lacking SpaOS . This analysis showed that although there were structures that were indistinguishable from wild type , in the absence of SpaOS there was more heterogeneity in the class averages in comparison to wild type ( Fig 3D ) . Since SpaOS stabilizes SpaOL in the absence of OrgB ( Fig 2C ) , these observations are consistent with a role for SpaOS in the assembly and/or stability of the sorting platform or its components . It has been previously proposed that homologs of SpaOS may serve as structural components of the fully assembled sorting platform[19] . Although the phenotype of the ∆spaOS mutant is inconsistent with this hypothesis , we probed the potential presence of SpaOS within the assembled sorting platform by adding a traceable density that can be imaged by cryo-ET . We have previously used this approach to map the location of SpaOL , OrgA , OrgB and InvC within the assembled sorting platform[5] . We constructed a strain expressing SpaOS tagged at its amino terminus by the fluorescent protein mEos3 . 2 , and examined the sorting platform structures of this strain for the presence of an extra density that could be assigned to SpaOS . Comparison of the sorting platform structures from strains expressing tagged ( Fig 3E , 3F and 3G ) and untagged ( Fig 3H , 3I and 3J ) versions of SpaOS revealed no detectable additional densities ( Fig 3E , 3F and 3G ) . Although the epitope tagging of SpaOS may affect its function in a manner that cannot be detected by our assays , this observation further supports the hypothesis that SpaOS may not be a structural component of the SPI-1 T3SS sorting platform but rather , may be required for its efficient assembly and/or stability . To further characterize the structure and function of SpaOL we carried out a detailed random mutagenesis analysis in an effort to define its functional domains and its protein-interaction network . We used a strategy we have previously described designed to increase the efficiency of the mutagenesis screen[26] . This strategy entails the use of a functional ( able to complement a ∆spaO mutation ) chimeric fusion protein between SpaOL and chloramphenicol acetyltransferase ( CAT ) separated by a flexible linker sequence ( Figure D in S1 Text ) . By imposing the requirement of conferring chloramphenicol resistance , we were able to select against mutations leading to premature termination or gross conformational changes of SpaOL that may prevent the folding of CAT . The spaOL gene was mutagenized by error-prone PCR and the generated mutants were screened as indicated in the Materials and Methods . We identified 223 loss-of-function mutants out of 12 , 350 mutants screened . Forty-seven of the identified mutants expressed full-length protein and therefore were analyzed by nucleotide sequencing to determine the location of the mutation ( s ) . Twenty-two of the 47 mutants had a single nucleotide change , which in some instances was independently identified more than once . The identified mutations were distributed throughout the spaOL coding sequence although clusters of mutations were identified between D45-A74 , L129-L134 , G157-L176 , and L234-G289 ( Figure E in S1 Text ) . Overall , more than 50% of the identified mutations mapped within the amino-terminal half of SpaO , the least characterized of its domains , although mutations were also identified within its carboxy terminal SPOA domains . To examine the effect of the mutations on wild-type spaO ( i . e . competent to produce both SpaOL and SpaOS ) we placed sixteen representative mutants mapping to different domains of SpaO in the wild type chromosome and examined their effect on SpaO function . We found that the majority of the mutants tested ( 11 out of 16 ) behaved like wild type when placed in the context of a chromosomal wild type spaO gene that leads to the synthesis of both SpaOL and SpaOS ( Fig 4A ) . These conditional mutations mapped to the amino-terminal half of SpaOL thus producing wild-type SpaOS . These results indicate that the presence of SpaOS was able to suppress the conditional phenotype exhibited by these SpaOL mutants . In the three cases tested , removal of the internal initiating codon of the conditional mutants resulted in the loss-of-function of SpaO thus confirming that these set of mutations were indeed conditional to the absence of SpaOS ( Fig 4B ) . These results suggest that SpaOS may stabilize SpaOL and that the presence of these conditional mutations in SpaOL may enhance the need for the putative chaperone function of SpaOS . A similar function has been previously proposed for the SpaOS homolog of the SPI-2 T3SS SsaQS[25] . Five of the 16 mutants exhibited a loss-of-function phenotype even in the presence of wild type SpaOS ( Fig 4A ) . Furthermore , all these mutants were complemented by a wild-type copy of SpaO expressed in trans , indicating that they did not exhibit a dominant-negative phenotype ( Fig 4C ) . Three of the identified mutants ( SpaOL234Q , SpaOL276P , and SpaOG289D ) mapped to the SPOA carboxy terminal domains highlighting the importance of these domains in SpaO function . The SPOA domains have been proposed to be implicated in the formation of higher order SpaO structures through homotypic interactions as well as through interaction with OrgB , another structural component of the sorting platform . However , the specific role of these domains in sorting platform assembly remains poorly understood . Our analysis also identified two loss-of-function mutants ( SpaOS51P and SpaOL67P ) that mapped to the amino terminal third of SpaO ( Fig 4A ) . Unlike the carboxy terminus , little information is available about this domain of SpaO since it has been refractory to structural analysis . To gain insight into the function of the amino terminal domain , we further characterized the phenotype of an S . Typhimurium mutant strain expressing SpaOL67P . We found that needle complexes purified from this mutant lack the needle substructure ( Fig 4D ) . Since the sorting platform is required for the secretion of the components necessary for needle assembly ( i . e . PrgI and PrgJ ) , these results indicate that this mutant is either unable to assemble the sorting platform or may assemble a non-functional sorting platform . To distinguish between these two possibilities we examined minicells obtained from the S Typhimurium spaOL67P mutant strain by in situ cryo-ET . We found that all NC observed lacked the sorting platform indicating that SpaOL67P is unable to direct the assembly of the sorting platform ( Fig 4E ) . These results indicate that the N-terminus of SpaO is essential for the assembly of the sorting platform . Our mutagenesis analysis identified specific SpaO residues required for the assembly of the sorting platform . SpaOL is predicted to engage in multiple interactions with itself as well as with other components of the sorting platform such as OrgA and OrgB . Therefore , it is predicted that some of the identified residues may be involved in some of the defined interactions of this critical component of the sorting platform . To define those potential interaction domains we substituted the codons of the identified critical residues ( i . e . SpaOL67 , SpaOL234 , SpaOL276 , and SpaOG289 ) by an amber codon ( AUG ) so that the unnatural photo-cross-linkable amino acid p-benzoyl-L-phenylalanine ( pBpa ) could be incorporated into SpaO in the presence of an orthogonal aminoacyl tRNA synthetase-tRNA pair[27] . The resulting mutant strains were competent for type III secretion exhibiting a secretion profile in culture supernatants that was indistinguishable from that of the wild-type strain ( Fig 5A ) . These results indicate that SpaO containing the unnatural amino acid pBpa can assemble a wild type sorting platform that is competent for type III secretion function . We then grew the resulting mutant strains under conditions that stimulate the expression of the SPI-1 T3SS and exposed them to UV light to promote site-specific crosslinking . Bacterial cell lysates were then run in SDS-PAGE for western-blot analysis to identify cross-linked species . The UV light treatment resulted in the appearance of distinct cross-linking patterns in lysates from several of the mutant strains ( Fig 5B ) . UV-cross-linking of SpaOL67pBpa resulted in the appearance of a ladder of bands whose mobility suggest that , most likely , represent different multimeric forms of SpaO . Western blot analysis of the bands identified only SpaO ( Fig 5C ) , which is consistent with the conclusion that these bands represent crosslinks of SpaO to itself as a consequence of its multimerization . The molecular weight of the cross-linked bands suggests that SpaO may form a tetrameric complex . This is consistent with the predicted stoichiometry of SpaO , 24 copies[21] , and its organization in the assembled 6-pod sorting platform[5] . These results also indicate that the amino terminal domain of SpaO plays a critical role in the formation of high order homotypic structures , which are essential for the assembly of the sorting platform . UV crosslinking of S . Typhimurium SpaOL276pBpa and SpaOG289pBpa cells also resulted in the detection of distinct cross-linked species ( Fig 5B and 5C ) . To identify the cross-linked proteins we probed the cross-linked bacterial cell lysates for the presence other components of the sorting platform functionally tagged with a different ( M45 ) epitope tag . Through this analysis we found that SpaOG289Bpa cross-linked to OrgB indicating that G289 is required for SpaO interaction with OrgB ( Fig 5C ) . We found that the stability of OrgB is compromised in the absence of SpaO ( see Fig 1 ) . Therefore , we reasoned that if OrgB is unable to interact with SpaOG289D its stability should be compromised in the presence of this mutation . Consistent with this hypothesis , the amount of OrgB in cell lysates of a S . Typhimurium spaOG289D mutant was reduced ( Fig 5D ) further supporting the conclusion that OrgB engages in critical interactions with the G289 residue in SpaO . The observed SpaOL276pBpa cross-linked band was smaller in size than that of the SpaOG289pBpa-OrgB cross-link but consistent with a band corresponding to a SpaOL-SpaOS cross-link . To test this hypothesis we expressed M45-epitope-tagged SpaOS in a S . Typhimurium strain expressing SpaOL276pBpa and unable to translate SpaOS . Bacterial cell lysates after exposure to UV light were then analyzed by western blotting . This analysis indicated that the cross-linked band was composed of SpaOL276pBpa-SpaOS ( Fig 5E ) indicating that L276 is essential for the formation of a heterodimer between the SPOA2 domains of SpaOL and SpaOS . Since the N-terminus of OrgB has been shown to interact with a SPOA1-SPOA2 heterodimer of SpaOL making extensive contacts with both SPOA domains[22] , these results indicate that the binding of SpaOL to OrgB and SpaOS is mutually exclusive , which is consistent with our inability to detect an interaction between SpaOL and SpaOS in S . Typhimurium in the presence of other T3SS components . These observations are also consistent with our conclusions that SpaOS may not be a structural component of the sorting platform but that it plays a role in the course of its assembly and/or dynamic behavior during type III secretion .
The sorting platform is an essential substructure of the T3SS injectisome , which is critical for the recruitment and sorting of protein substrates destined to travel the type III secretion pathway[16] . Recent cryo-ET studies have shown , in S . Typhimurium , that the sorting platform is made of a multi-protein cage-like scaffold , integrated by OrgA , SpaO , and OrgB , which serves as support for another component of this substructure , the ATPase InvC , which is involved in the initiation of substrates into the secretion pathway[5] . Unlike other substructures of the T3SS injectisome , the sorting platform is thought to exhibit a behavior that may involve cycles of assembly and disassembly[19 , 20] . However , very little information is available about the mechanisms and functional significance of this dynamic behavior . Furthermore , little is known about the mechanisms of assembly of this critical component of the T3SS machine . In this paper we have used a multidisciplinary approach to gain insight into the role of SpaO in the assembly of the sorting platform . We found that in the absence of SpaO , other components of the sorting platform are destabilized . In contrast , the absence of other sorting platform components did not destabilize SpaO indicating that it must play a central role in its assembly . Consistent with this role , cryo-ET studies have mapped the location of SpaO to the central region of the pods that form the core of the cage-like structure , a position that would allow it to interact with the other components of the sorting platform[5] . In fact , our mutagenesis and cross-linking studies identified critical domains of SpaO involved in interactions with other components of the sorting platform . As previously demonstrated[22] , we found that the carboxy terminal SPOA domains of SpaO mediate its interaction with OrgB , which caps the cage-like structure at its cytoplasmic side serving as a cradle for InvC . In addition , our genetic screen identified critical residues at its poorly characterized amino terminus , which are essential for the oligomerization of SpaO , and presumably , for its interaction with other components of the sorting platform . We hypothesize that this interaction network allows SpaO to nucleate the assembly of the sorting platform . Consistent with this hypothesis , we have shown by cryo-ET in situ analysis that a single amino acid substitution in SpaO that disrupts this interaction network resulted in the complete absence of the sorting platform . spaO and its homologues are unusual in that they are translated as two products , a full length product ( SpaOL ) and a shorter version composed of their carboxy terminal third ( SpaOS ) , the product of an internal translational initiation site[23–25] . The precise role of SpaOS is unclear and the available evidence suggests that its function may differ in different T3SSs . For example , in the case of the Yersinia T3SS , it has been reported that the SpaOS homolog ( YscQS ) is essential for type III secretion[23] . Based on this observation , it has been suggested that YscQS is a structural component of the sorting platform . However , we found that the absence of SpaOS resulted in a rather mild secretion phenotype and no detectable functional phenotype , observations that are inconsistent with a crucial structural role for this component of the sorting platform . Similar observations have been made in the case of the S . Typhimurium SPI-2 T3SS homolog SsaQS[25] . If not a structural component of the sorting platform , what could be the function of SpaOS ? We hypothesize that SpaOS may play a regulatory role during assembly and disassembly of the sorting platform . We based our hypothesis on the following observations . First , the mild phenotype of its absence , which is more consistent with a regulatory rather than a structural role . Second , the observation that in the absence of SpaOS we detect needle complexes that while displaying a needle filament in situ , lack a fully assembled sorting platform . Since assembly of the needle filament requires a fully functional sorting platform , we take this observation to mean that those needle complexes must have been previously associated with a sorting platform but the absence of SpaOS may impede or delay its efficient re-assembly . Third , the stability of SpaOL in the absence of other components of the sorting platform requires SpaOS . This observation suggests that prior to the assembly of the sorting platform ( or during the cycles of assembly and disassembly ) , SpaOS may play a crucial role in chaperoning SpaOL . Fourth , at least in the fraction of in situ structures that we were able to visualize , we detected no obvious differences in the structure of the fully assembled injectisomes from the wild-type and the ∆spaOS mutant strains . Finally , we were not able to detect the presence of SpaOS in fully assembled injectisomes . Taken together , these data suggest a regulatory rather than a core structural role for SpaOS in the assembly of the sorting platform . What this role might be is unclear but stabilizing SpaOL when not associated to other components of the sorting platform must be a central element of this function . We have shown here that the two forms of SpaO , SpaOL and SpaOS , play a central role in the assembly of the S . Typhimurium T3SS sorting platform . The conserved nature of the T3SS components suggests that this function may be maintained in other T3SS .
All the strains used in this study are derivatives of Salmonella enterica Typhimurium SL1344 . Genetic modifications were introduced in SL1344 by allelic exchange using R6K suicide vectors[28] in E . coli ß2163 ∆nic35 as donors of the modified allele[29] . All plasmids in this study were constructed using Gibson assembly [30] . The resulting modified strains were screened by PCR and confirmed by western blot and/or functional assays to behave in a manner indistinguishable from wild type . A complete list of plasmids and strains used in this study can be found in Table S1 . To prevent recombination between the upstream spaOLGTG ( 203 ) >GCG allele and the downstream 3xFlag-tagged spaOS in the strain schematically presented in Figure C in S1 Text the coding sequence of spaOS was changed to: ATGGACTACAAAGACCATGACGGTGATTATAAAGATCATGACATCGATTACAAGGATGACGATGAGACCCTGGATATCCAGCATATTGAAGAGGAGAACAACACGACCGAAACCGCGGAAACCCTGCCGGGCCTGAACCAGTTACCGGTGAAACTGGAATTCGTCTTATATCGCAAAAATGTCACGCTTGCGGAACTTGAAGCGATGGGTCAACAGCAACTCTTGTCGTTACCGACGAACGCAGAGTTAAATGTCGAGATCATGGCCAACGGCGTCCTTTTAGGCAACGGTGAGTTAGTGCAAATGAACGATACGCTGGGTGTCGAAATTCACGAGTGGTTATCGGAAAGCGGGAACGGTGAGTGA The underlined sequence corresponds to the 3xFlag-tag sequence . The changes did not alter the actual amino acid sequence of SpaOS . The new sequence was generated as a synthetic minigene by IDT , Integrated DNA technologies , Inc . Strains were typically maintained in Luria broth ( LB ) . To induce optimal expression of the SPI-1 T3SS , LB was modified to contain 0 . 3 M NaCl and cultures were grown under low aeration to an OD600 of ~0 . 9[31] . When appropriate , further induction of SPI-1 was achieved by expression of hilA[32] , the master regulator of SPI-1 T3SS expression , from an arabinose inducible plasmid . Overnight cultures of the strains of interest were grown in LB with the appropriate antibiotics . The O/N cultures were diluted 1:20 in 10 ml of 0 . 3 M NaCl LB containing the appropriate antibiotics and 0 . 05% arabinose when the hilA plasmid was present . The subcultures were grown under low aeration conditions until an OD600 of ~ 0 . 9 ( 4 to 5 hours ) . Then , 1 ml of the subculture was transferred to a 1 . 5 ml microfuge tube , the cells were pelleted at maximum speed for 1 min , resuspended in 100 μl of SDS-PAGE loading buffer and saved as the whole cell lysate ( WCL ) control at 10x concentration . The rest of the culture was spun down and the supernatant was filtered using a 0 . 45 μm syringe filter to remove remaining bacterial cells . Proteins in the supernatant were recovered by trichloroacetic acid ( TCA ) precipitation , and the protein pellet was resuspended in 100 μl of SDS-PAGE loading buffer , resulting in the sup sample at 100x concentration . All Western blots were imaged using the Odyssey Li-Cor system ( LI-COR Biosciences ) and near-infrared secondary fluorescence antibodies that capture data over the entire linear range in a single image . Quantification of the Western blots was performed using Image Studio Lite software specifically desing by Li-Cor for Western blot quantification . Int 407 embryonic intestinal epithelial cells , seeded in a 24-well plate , were infected in triplicates with the wild-type or ∆spaOS S . Typhimurium strains grown under SPI-1-T3SS expression-inducing conditions with a multiplicity of infection ( m . o . i . ) of 10 in Hank’s balanced salt solution with calcium and magnesium ( Gibco 14025092 ) . Dilutions were plated to determine the exact inoculum used for each strain . Infection was let to proceed for 30 min , after which the cells were washed 3x with 0 . 5 ml of PBS , and incubated for 1 h in the presence of DMEM containing 10% BCS and 50 μg/ml of gentamicin to kill extracellular bacteria . Cells were washed 3x with 0 . 5 ml of PBS , lysed in 0 . 5 ml of PBS + 0 . 05% Na-deoxycholate , and the number of colony forming unites ( c . f . u . ) was determined by dilution plating in LB agar plates containing 100μg/ml of streptomycin . Bacterial invasion was expressed as the percentage of the inoculum surviving the gentamicin treatment . A 50 ml overnight culture of E . coli Lemo21 ( DE3 ) ( New England Biolabs C2528J ) harboring plasmid pSB3775 in LB containing 50 μg/ml of kanamycin and 30 μg/ml of chloramphenicol was diluted in 1 L of LB containing the same antibiotics . The culture was grown at 37°C and 220 rpm to an OD600 of 0 . 6 and then induced with 0 . 5 mM IPTG for 5 hours . Cells were pelleted at 6 , 000 rpm and the pellet resuspended in 10 ml of lysis buffer ( 50 mM NaH2PO4 , 300 mM NaCl , 10 mM imidazole , 1mM Mg2Cl , 2 . 5U of DNAse , and cOmplete™ Protease Inhibitor Cocktail [Sigma 11697498001] ) . Cells were lysed using the One Shot cell disrupter ( Constant Systems Ltd . , Northants , UK ) . After lysis , cellular debris was removed by centrifugation , and the cleared supernatant was transferred to a fresh tube . Two hundred microliters of Ni-NTA agarose resin ( Qiagen 30210 ) were added to the lysate , and after 3 h incubation at 4°C under rocking conditions , the lysate containing the agarose resin was applied to an empty chromatography column ( Bio-Rad 7311550 ) . The beads were washed three times with 10 ml of lysis buffer containing 20 mM imidazole and the bound protein was eluted in five 1 ml aliquots of lysis buffer containing 250 mM imidazole . The protein concentration was estimated using the Bio-Rad Protein Assay ( Bio-Rad 500–0006 ) and fractions containing high amounts of protein were concentrated prior to loading into a Superose 6 10/300 GL ( GE Life Sciences 17517201 ) column using an ÄKTA purifier system ( GE Life Sciences ) . One ml fractions were collected , and 20 μl of each fraction were subsequently analyzed by SDS-PAGE and Coomassie Brilliant Blue R-250 ( ThermoFisher 20278 ) staining . Overnight cultures of S . Typhimurium strains expressing the indicated his-tagged proteins and carrying a plasmid encoding hilA expressed from an arabinose-inducible promoter[32 , 33] were diluted 1:20 in flasks containing 150 ml of LB containing 0 . 3 M NaCl , 100 μg/ml ampicillin , and 0 . 05% arabinose . Cultures were grown at 37°C under 100 rpm shaking ( low aeration ) conditions to an OD600 ~ 0 . 6 , cells were pelleted at 7 , 000 rpm , resuspended in 2 . 5 ml of PBS containing 15 mM imidazole , cOmplete EDTA-free protease inhibitor cocktail ( Sigma 4693159001 ) , and lysed using a One Shot table top homogenizer ( Constant Systems Ltd , Northants , UK ) . Debris was removed by centrifugation and the cleared lysate was transferred to a fresh 2 ml microcentrifuge tube . One hundred μl of Ni-NTA agarose ( Qiagen 30310 ) was added to each sample and the tubes were incubated for 1 hr at 4°C under rocking conditions . After binding , beads were washed 4x with 1 ml of PBS containing 20 mM imidazole . Bound protein was eluted in 100 μl of PBS containing 250 mM imidazole or by boiling the beads in SDS-PAGE running buffer . Samples were collected throughout the procedure and analyzed by western blot . Minicell producing bacterial strains were grown overnight at 37°C in LB containing 0 . 3M NaCl . Fresh cultures were prepared from a 1:100 dilution of the overnight culture and then grown at 37°C to late log phase in the presence of ampicillin ( 200 μg/mL ) and L-arabinose ( 0 . 1% ) to induce the expression of regulatory protein HilA and thus increase the number of injectisomes partitioning to the minicells[34] . To enrich for minicells , the culture was centrifuged at 1 , 000 x g for 5 min to remove bacterial cells , and the supernatant fraction was further centrifuged at 20 , 000 x g for 20 min to collect the minicells . The minicell pellet was resuspended in PBS and mixed with 10 nm colloidal gold particles and deposited onto freshly glow-discharged , holey carbon grids for 1 min . The grids were blotted with filter paper and rapidly frozen in liquid ethane , using a gravity-driven plunger apparatus as described previously[5 , 17] . The frozen-hydrated specimens were imaged with 300kV electron microscopes . The tomographic package SerialEM [35] was utilized to collect single-axis tilt series from -51° to +51° using Polara or Titan electron microscopes equipped with a field emission gun and a direct detection device ( Gatan K2 Summit ) . The tilt series was aligned and reconstructed using IMOD [36] . In total , 2 , 452 tomograms ( 3 , 600 × 3 , 600 × 400 pixels ) were generated for detailed examination of the sorting platform in several mutants ( Table S2 ) . Sub-tomogram analysis was accomplished as described previously [17] to analyze the injectisomes . Briefly , we first visually identified the injectisomes on each minicell . Two coordinates along the needle were used to estimate the initial orientation of each particle assembly . For initial analysis , 4 × 4 × 4 binned sub-tomograms ( 128 × 128 × 128 voxels ) of the intact injectisome were used for alignment and averaging by using the tomographic package I3 [37 , 38] . Then multivariate statistical analysis and hierarchical ascendant classification were used to analyze the needle tip complex [38] . IMOD and UCSF Chimera [39] were used to visualize the sub-tomogram average structures of the T3SS injectisome . The mutagenesis procedure was based on a previously described strategy [26] , which relies on the use of a chimeric fusion protein between the protein to be mutagenized ( SpaOL , spaOGTG203TTG , in this case ) and chloramphenicol acetyltransferase ( CAT ) separated by a flexible linker sequence . By imposing the requirement of chloramphenicol resistance , mutations leading to premature termination or gross conformational changes can be counter-selected . A detailed schematic representation of the plasmid employed for the mutagenesis ( pSB4545 ) can be found in Figure D in S1 Text . This plasmid was able to complement the secretion phenotype of a S . Typhimurium ∆spaO mutant strain ( Figure D in S1 Text ) . The mutagenic PCR was performed as described [40] , but without the addition of MnCl2 to reduce the mutation frequency . SpaOL was amplified under mutagenic conditions using forward ( ATGGACTACAAAGACCATGACGG ) and reverse ( TTCTCTCTAGAAGGCAGGTGTCCCTGCAC ) primers and ligated to plasmid pSB4545 . The ligation was transformed into E . coli selecting for kanamycin , colonies were pooled and plasmid DNA was extracted and electroporated into a ∆spaO S . Typhimurium strain . During the set-up process , we found that the SpaO-CAT fusion protein required time to fold and confer chloramphenicol resistance and that plating the electroporation directly into chloramphenicol-containing plates yielded no colonies . For this reason , the electroporated ∆spaO S . Typhimurium was plated on 30 μg/ml kanamycin LB plates ( to select for the plasmid ) and kanamycin resistant colonies were then replicated onto chloramphenicol-containing plates ( 10 μg/ml ) . Colonies that were chloramphenicol-resistant were then assayed for type III secretion function by a dot-blot assay . Chloramphenicol resistant S . Typhimurium expressing M45 epitope tagged SopB and harboring the mutagenized plasmid were inoculated in a 96-well plate containing 200 μl of 0 . 3M NaCl LB and 10 μg/ml of chloramphenicol per well , and incubated O/N at 37°C with gentle shaking . Using a 96 pin replicator ( Boekel Scientific 140500 ) a small volume of culture was transferred to a nitrocellulose membrane , which was let to dry and then fixed by exposure to chloroform vapors . This procedure did not permeabilize the bacterial cells . The presence ( secretion positive ) or absence of SopB ( secretion negative ) on the bacterial surface was then probed by treating the membranes with a monoclonal antibody to the M45 epitope following standard procedures ( Figure F in S1 Text ) . The clones that showed a loss-of-function phenotype in the dot-blot assay were then analyzed by western blot to examine the expression of the full length SpaO-CAT chimeric protein ( Figure G in S1 Text ) . Clones that showed a loss-of-function phenotype and expressed the full length SpaO-CAT protein were then sequenced to determine the location of the mutation ( s ) ( Figure E in S1 Text ) . Following the outlined screen , 12 , 350 kanamycin-resistant clones were picked and replicated on chloramphenicol plates . Of those clones , 1 , 965 clones ( 17 . 37% ) were chloramphenicol resistant and were then assayed for type III secretion function on dot-blot assays; 223 of the assayed clones were selected as loss-of-function and checked to confirm that they express the full length SpaO-CAT fusion protein . Forty seven clones passed all the requirements and were sequenced . Twenty-two of the 47 clones had a single mutation , some of which were identified independently more than once . Figure E in S1 Text . shows the location of the single mutations identified . A summary of the mutagenesis results can be found in Table S3 . For site-specific in vivo photo-crosslinking , the photoreactive unnatural amino acid pBpa[41] was incorporated into SpaO by replacing the codon of the targeted amino acid with a TAG amber codon . Incorporation of pBpa was accomplished by amber codon suppression provided by the presence of plasmid pSUP encoding an E . coli nonsense suppressor tRNA-tRNA synthetase system that can recognize and incorporate pBpa at the TAG amber codon[27] . Overnight cultures of the strains encoding the different TAG-containing spaO alleles , and carrying plasmids encoding the suppression system and hilA expressed under an arabinose-inducible promoter were diluted 1:20 in 10 ml of 0 . 3M NaCl LB containing 100 μg/ml ampicillin , 10 μg/ml chloramphenicol , 0 . 1% arabinose , and 1mM pBpa and grown at 37°C to an OD600 of ~ 0 . 9 . Bacterial cells were pelleted and resuspended in 5 ml of PBS . Half of the resuspended culture was placed in a 60 mm tissue culture dish and irradiated at 365 nm with a hand-held UV lamp and the other half was left unexposed to the UV lamp . Cells were then pelleted and resuspended in 250 μl of SDS-PAGE loading buffer ( 20x concentration ) and 20 μl of the UV-treated or control samples were loaded onto a SDS-PAGE gel for western blot analysis . In some instance the site-specific in vivo crosslinking experiments were scaled up ( up to 200 ml of culture ) and after crosslinking the crosslinked species were concentrated by anti-flag immunoprecipitation ( see section below ) . After UV-crosslinking as indicated above , bacterial cells were pelleted and resuspended in 3 ml of TBS containing 1mM MgCl2 and cOmplete EDTA-free protease inhibitor cocktail ( Sigma 4693159001 ) . The suspended cells were lysed by sonication ( 5 minutes @ 35% amplitude with cycles of 3 sec ON and 7 sec OFF ) . Debris was removed by centrifugation , the clarified lysate was transferred to a fresh tube and 50 μl of anti-Flag M2 affinity gel ( Sigma A2220 ) were added . Samples were incubated for ~4 hrs at 4°C under rocking conditions , beads were washed 4x with 1 ml of TBS containing 0 . 05% Tween-20 , resuspended in 50 μl of SDS-PAGE running buffer , and boiled to elute all bound proteins . The needle complex purification was carried out by maltose-binding protein ( MBP ) affinity purification as follows . An MBP-tagged PrgH allele was introduced into an S . Typhimurium strain carrying the SpaOL67P mutation . Two liters of 0 . 3M NaCl containing 100 μg/ml of ampicillin and 0 . 1% arabinose were inoculated with the strains of interest and grown for ~10 hrs under gentle ( 100 rpm ) shaking . Cells were recovered by centrifugation , resuspended in 10 ml of lysis buffer ( 200mM Tris pH 7 . 5 , 20% sucrose , 1mM EDTA , 0 . 25mg/ml of lysozyme and cOmplete EDTA-free protease inhibitor cocktail [Sigma 4693159001] ) and incubated on ice for 1 hr . Cells were incubated for 5 min at 37°C and lysed by the addition of 0 . 5% N-Dodecyl- ß-D-maltoside ( DDM ) ( Anatrace D310S ) . Cells were incubated at 37°C for additional 5 to 10 min while monitoring lysis . Cells were then transferred to ice and further incubated for an additional hour . Debris was removed by centrifugation and the clarified lysate was transferred to a fresh tube . Two hundred microliters of amylose resin were added and the suspension was incubated O/N at 4°C under rocking conditions . Beads were then washed 4x with 10 ml of washing buffer ( 20 mM Tris pH 7 . 5 , 100 mM NaCl , 1 mM EDTA ) and finally resuspended in 50 μl of washing buffer containing 20 mM of maltose . After 1 hr incubation on ice with occasional tapping , beads were removed by centrifugation and the needle complex containing supernatant was transferred to a fresh tube . Samples ( 3 . 5 μl ) were directly applied onto glow-discharged grids bearing a continuous carbon film ( EMS CF300-Cu ) . After two minutes , the sample was blotted , then stained with 2% ( w/v ) uranyl acetate . Images were recorded on a FEI Tecnai-12 electron microscope ( LaB6 , 120KV ) equipped with a 4096x4096 pixel Gatan Ultrascan CCD camera .
|
Many pathogenic and symbiotic gram-negative bacteria utilize type III secretion systems to deliver bacterial proteins , known as effectors , directly into the host cell cytosol to promote their survival and the colonization of tissues . Type III secretion systems or injectisomes are large , multiprotein complexes composed of several substructures: the needle complex , a multiring structure with a protruding needle-like appendage anchored in the bacterial envelope; the export apparatus , a set of membrane proteins that form a gate in the inner-membrane for the passage of effector proteins; and the sorting platform , a large cytosolic complex that delivers the effectors to the needle complex in an orderly fashion . In this study , we characterize SpaO , the core component of the Salmonella Typhimurium sorting platform . The spaO gene encodes two simultaneously translated products , a full length protein ( SpaOL ) and a shorter product ( SpaOS ) encompassing the last 101 aa of the full length product . Here we find that in the absence of SpaOS , the sorting platform still forms and functions although slightly less efficiently than in the wild-type situation , and therefore we conclude that SpaOS is most likely not a central structural component of the sorting platform and may play a regulatory role during the cycles of assembly and disassembly that the sorting platform undergoes . In addition , we identify residues critical for the interaction between SpaOL and OrgB and SpaOL and SpaOS and conclude that those interactions might be mutually exclusive further supporting the idea that SpaOS may not be a core structural component of the sorting platform . N-terminal residues in SpaOL are shown to be critical for the formation of the sorting platform . Our findings provide insights into the sorting platform substructure , a highly conserved element in type III secretion systems and may contribute to the development of novel therapeutic avenues to fight infection .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"bacteriology",
"antimicrobials",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"drugs",
"microbiology",
"bacterial",
"diseases",
"physiological",
"processes",
"mutation",
"secretion",
"systems",
"immunoprecipitation",
"antibiotics",
"enterobacteriaceae",
"pharmacology",
"bacteria",
"bacterial",
"pathogens",
"salmonella",
"typhimurium",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"mutant",
"strains",
"microbial",
"physiology",
"medical",
"microbiology",
"microbial",
"pathogens",
"salmonella",
"precipitation",
"techniques",
"mutagenesis",
"bacterial",
"physiology",
"virulence",
"factors",
"phenotypes",
"physiology",
"secretion",
"genetics",
"biology",
"and",
"life",
"sciences",
"microbial",
"control",
"organisms",
"chloramphenicol"
] |
2019
|
Role of SpaO in the assembly of the sorting platform of a Salmonella type III secretion system
|
Staphylococcus aureus exhibits many defenses against host innate immunity , including the ability to replicate in the presence of nitric oxide ( NO· ) . S . aureus NO· resistance is a complex trait and hinges on the ability of this pathogen to metabolically adapt to the presence of NO· . Here , we employed deep sequencing of transposon junctions ( Tn-Seq ) in a library generated in USA300 LAC to define the complete set of genes required for S . aureus NO· resistance . We compared the list of NO·-resistance genes to the set of genes required for LAC to persist within murine skin infections ( SSTIs ) . In total , we identified 168 genes that were essential for full NO· resistance , of which 49 were also required for S . aureus to persist within SSTIs . Many of these NO·-resistance genes were previously demonstrated to be required for growth in the presence of this immune radical . However , newly defined genes , including those encoding SodA , MntABC , RpoZ , proteins involved with Fe-S-cluster repair/homeostasis , UvrABC , thioredoxin-like proteins and the F1F0 ATPase , have not been previously reported to contribute to S . aureus NO· resistance . The most striking finding was that loss of any genes encoding components of the F1F0 ATPase resulted in mutants unable to grow in the presence of NO· or any other condition that inhibits cellular respiration . In addition , these mutants were highly attenuated in murine SSTIs . We show that in S . aureus , the F1F0 ATPase operates in the ATP-hydrolysis mode to extrude protons and contribute to proton-motive force . Loss of efficient proton extrusion in the ΔatpG mutant results in an acidified cytosol . While this acidity is tolerated by respiring cells , enzymes required for fermentation cannot operate efficiently at pH ≤ 7 . 0 and the ΔatpG mutant cannot thrive . Thus , S . aureus NO· resistance requires a mildly alkaline cytosol , a condition that cannot be achieved without an active F1F0 ATPase enzyme complex .
Staphylococcus aureus is a highly invasive human pathogen that is responsible for significant morbidity each year[1] . Treatment of S . aureus infections has become increasingly difficult due to the propensity of S . aureus to quickly evolve antibiotic resistance . Methicillin resistant S . aureus ( MRSA ) and multidrug resistant MRSA clones are now prevalent throughout the world[2] . Although historically a nosocomial pathogen , in recent decades otherwise healthy individuals have begun to contract MRSA outside of hospital settings[3] . These community-acquired MRSA ( CA-MRSA ) strains are typically characterized as hypervirulent and most frequently cause skin and soft tissue infections ( SSTIs ) , although infections often progress to more invasive and systemic disease[4] . Understanding the factors contributing to the persistence and invasiveness of CA-MRSA is of paramount importance to limiting its current spread through both community and hospital settings . A major factor contributing to S . aureus persistence within mammalian hosts is resistance to the antimicrobial radical nitric oxide ( NO· ) , a membrane permeable gas that is produced by activated phagocytes in response to bacterial infection[5] . Production of NO· is important for limiting bacterial proliferation in multiple infection models , so the ability to continue growth in the presence of this broad-spectrum antimicrobial confers S . aureus with a major pathogenic advantage . S . aureus NO· resistance is a highly unique trait , as even closely related Staphylococci such as Staphylococcus epidermidis , S . saprophyticus and S . haemolyticus are sensitive to NO·[6 , 7] . Additionally , NO· resistance is important for S . aureus persistence in murine infection models because deletion of genes required for NO· resistance attenuates S . aureus virulence[6 , 8 , 9] . Exposure to NO· results in a myriad of consequences within a bacterial cell . NO· and its derivatives target metal centers of bacterial enzymes ( heme iron , iron-sulfur clusters , and other transition metal cofactors ) and protein thiols[10 , 11] . At high NO· concentrations , the reversible binding of NO· to cytochrome heme iron results in respiration inhibition[12] . NO· and its derivatives can also cause DNA damage , lipid peroxidation , and nitration of tyrosine residues[13] . S . aureus must therefore employ a diverse array of strategies to cope with these multifaceted effects . We have previously characterized several of the strategies employed by S . aureus to resist the effects of NO· . The induction of a unique L-lactate dehydrogenase enzyme , Ldh1 , in response to NO· exposure is important for the ability of S . aureus to balance redox when respiration is inhibited by NO·[12] . The NO·-mediated activation of the two-component system SrrAB when respiration is limited by NO· results in induction of a flavohemoprotein , Hmp , that detoxifies NO· to nitrate[8] . SrrAB also induces increased expression of two terminal oxidases , Qox and Cyd , which helps to overcome the inhibitory effects of NO· on respiration[14 , 15] . However , with the exception of Ldh1 , these responses to NO· stress are not unique to S . aureus and cannot fully explain its resistance . Bacillus subtilis has both a two-component system homologous to SrrAB and an NO·-inducible Hmp homologue , yet is highly sensitive to this immune radical[16] . Therefore , we are still lacking a significant understanding of what makes the S . aureus response to NO· so unique and effective . Several genome-wide transposon-based screens have been performed to date with the goal of identifying genes important for S . aureus pathogenesis in multiple infection models , including murine bacteremia , skin abscess , osteomyelitis , rabbit endocarditis , and nematode infection[17–21] . Many of these studies used signature-tagged mutagenesis and thus employed a small and highly limited pool of transposon mutants for screening[17 , 18] . Other studies used MSSA strains of S . aureus with significant virulence differences relative to CA-MRSA isolates[19 , 20] . A recent Tn-Seq screen using the MSSA strain HG003 examined Tn-mutant fitness after 24 and 48-hrs of infection in murine SSTIs and identified many genes important for persistence[20] . However , HG003 is not highly virulent in an SSTI model , and abscess bacterial burdens remained constant at around the same level as the inoculum in these experiments . Therefore , significant knowledge remains to be gained regarding genes required for fitness in S . aureus strains that are specifically known for their proclivity to proliferate within SSTIs , such as the CA-MRSA isolates . In the current study , we report the first in vivo Tn-Seq screen to use a relevant CA-MRSA strain of S . aureus , USA300 LAC . We employ a genome-wide Tn-Seq screen with a highly saturated transposon library to broadly identify genes required for NO· resistance in S . aureus . After identifying genes important for NO· resistance in vitro , we then investigated genes important for general fitness in murine SSTIs and identified a subset of genes important for both NO· resistance in vitro and for persistence in murine SSTIs . Most notably , our results highlight a major role for the F1F0 ATPase during both NO· stress and virulence in murine SSTIs . This phenomenon may be generalizable across bacterial species as the F1F0 ATPase was recently reported as being essential to Listeria monocytogenes when cultured anaerobically[22] . Thus , given the non-respiratory nature of inflamed tissue environments ( hypoxia , iron-limitation and high levels of immune radicals ) , the F1F0 ATPase may be required for the full virulence of a variety of pathogenic microorganisms , underscoring the utility of targeted antimicrobials that inhibit this enzyme complex[23] .
Previous genome-wide screening efforts to identify essential genes in S . aureus have used MSSA strains including Newman , RN4220 , RN6390 , SH1000 and HG003[17–21 , 24] . Because our primary interest was to identify genes required for both NO· resistance and persistence in SSTIs , we wanted to select a MRSA strain commonly associated with SSTIs in the current CA-MRSA pandemic . Therefore , we chose to generate a high-density transposon library in S . aureus LAC , belonging to the dominant CA-MRSA PFGE subtype , USA300 . We used the transposon bursa aurealis , a mariner-based transposon that inserts randomly throughout the S . aureus genome into TA dinucleotides[21 , 25] . The Himar1 transposase-containing plasmid pFA545 that has previously been used with bursa aurealis in S . aureus did not successfully result in transposition , so we modified this plasmid by replacing the original transposase allele and its xylose-inducible promoter with a new copy of the transposase allele fused to the constitutive lgt promoter . This plasmid mediated efficient transposition and was used for creation of a saturated S . aureus LAC transposon library comprising 77 , 161 independent Tn insertions ( Fig 1A ) . To verify that constitutive expression of tnp was mediating only one transposition event per genome , we performed a Southern blot with a Tn-specific probe on ClaI digests of DNA from sixteen individually isolated Tn-mutants . We observed single bands of variable sizes for each mutant , indicating that transposition was occurring randomly and only once per genome for each of the 16 mutants ( S1 Fig ) . Moreover , the density of coverage was very high with a median distance between two neighboring Tn insertion sites at 16 bp ( S2 Fig ) . The median distance metric is not subject to the effects of essential DNA segments that significantly elevate the average distance between neighboring insertions . The non-biased nature and high density of the Tn insertions is further reflected by a lack of obvious strand bias ( Fig 1B ) including within the 2 , 132 Tn insertions that were at the identical site in the genome , but inserted into opposing DNA strands ( Fig 1C ) . To calculate the relative fitness contribution of genes we generated representation ( R ) values based on both the density of transposon insertions and the number of reads , where R = ( median read count ) x ( insertion density ) for each gene ( see Materials and Methods ) . For each run , individual R-values were all normalized to the median R-value of the run to correct for variation in read numbers . Thus , the typical R-value distribution is biphasic with a peak at or near 1 for genes with no significant contribution to fitness within a given environment and a peak at 0 indicative of elements that are absolutely essential for survival in a given environment ( Fig 2 ) . More specifically , genes with log-transformed R-values more than three standard deviations ( SD ) below the mean R value were designated as being essential , while those with log-transformed R-values between 2 and 3 SD below the mean were designated as genes required for full fitness . By this calculation , we found 370 genes to be essential/required in our input library . Our findings are consistent with previous reports[20] , with 88% of the essential genes in our dataset being previously designated as essential despite the fact that the two libraries were generated in different strain backgrounds ( HG003 versus LAC ) . Furthermore , comparing essential gene lists between this study and those by Valentino et al . and Chaudhuri et al . finds that the majority of genes ( 288 ) were commonly identified between all three studies ( S3 Fig and S1 Table ) [20 , 24] . R values for all genes in the library under all conditions tested are listed in S2 Table . Given the extensive overlap between our list of essential genes with those previously reported , we chose not to validate our essential genes . Rather , we focused on applying Tn-Seq to identify undiscovered NO·-resistance determinants in S . aureus as well as genes required for persistence within 3-day and 7-day murine skin abscesses . To generate inocula for further selection experiments , we grew library aliquots for 10-hrs overnight in TSB 5mg/ml glucose . We quantified transposon insertions from two different 10-hr overnight cultures to establish the composition of our inoculation culture . We also performed a technical replicate of library preparation for Illumina sequencing on one overnight culture to establish the reproducibility of our method . We performed regression analyses of the insertion densities ( # of actual insertions/ # of possible insertions for each gene ) between replicates and found a high degree of reproducibility between both technical and biological replicates , most with Pearson coefficients of ~0 . 9 ( S4 Fig ) . Before performing Tn-Seq to identify NO· sensitive Tn-mutants , we first validated an assay for fitness selection in the presence of NO· . We mixed cultures of WT LAC and several known NO· sensitive mutants: ΔsrrAB , Δhmp , Δldh1 , and ΔsarA . These mutants were selected because they exhibit a range in severity of NO· sensitivity and are sensitive to NO· via a variety of mechanisms[8 , 12 , 14] . We mixed each mutant with WT in a 1:100 ratio to simulate the fact that in the saturated Tn-library , the majority of bacterial cells will not have fitness defects in the presence of NO· . These mixed cultures were serial passaged aerobically with or without NO· for 24 generations . Over time , NO· sensitive mutants were out-competed by WT in the presence of NO· but not significantly when grown aerobically ( Fig 3 ) with the only exception being ΔsarA . However , the underrepresentation of the ΔsarA mutant passaged for 24 generations under NO· stress was significantly exacerbated when compared to aerobically passaged cultures ( Fig 3 ) . Thus , 24 generations under constant NO· stress is sufficient to select against mutants with varying degrees of NO·-sensitivity . We next serially passaged the Tn-library aerobically either with or without NO· as described above for 24 generations , performing two biological replicates for each condition . We averaged the R-values from the two biological replicates and , similarly to methods outlined above , we determined genes essential for serial passage in aerobic cultures ( genes with R-values ≥ 3 SD below the mean normalized R-value ) . These genes were removed from further analyses since they are required for aerobic growth in the absence of NO· . We then computed the ratios of R-values of remaining genes from NO·-passaged cells to those of aerobically grown cells for each gene . There were a total of 41 genes that were essential ( ≥ 3 SD below mean NO·:aerobic ratio ) for NO· resistance and 127 genes with significant fitness contributions during NO· stress ( NO·:aerobic ratios between 2 and 3 SD below the mean , S3 Table ) . These included many previously characterized NO· resistance determinants such as ldh1 ( SAUSA300_0235 ) , hmp ( SAUSA300_0234 ) , pyk ( SAUSA300_1644 ) , ccpA ( SAUSA300_1682 ) , codY ( SAUSA300_1148 ) , nfu ( SAUSA300_0839 ) , srrAB ( SAUSA300_1442/1 ) and qoxABD ( SAUSA300_0963/2/0 ) [8 , 9 , 12 , 14 , 26 , 27] . However , additional genes not previously known to contribute to NO· resistance in S . aureus were among the essential/required list including genes encoding the F1F0 ATPase ( SAUSA300_2057 through SAUSA300_2064 ) , SodA ( SAUSA300_1513 ) , MntABC ( SAUSA300_0618 , SAUSA300_0619 SAUSA300_0620 ) , UvrA ( SAUSA300_0742 ) , UvrC ( SAUSA300_1045 ) , MprF ( SAUSA300_1255 ) and RpoZ ( SAUSA300_1103 ) . Other genes important for fitness specifically during NO· stress included genes associated with carbohydrate utilization ( ptsI/SAUSA300_0984 and gpmI/SAUSA300_0759 ) , proteases ( clpC/SAUSA300_0510 ) , putatively involved in iron-sulfur cluster homeostasis ( SAUSA300_1248 ) , thioredoxin-like proteins ( SAUSA300_0903 ) , transcriptional regulators ( ctsR/SAUSA300_0507 , sarT/SAUSA300_2437 and the two-component system bceRS/SAUSA300_0645/6 ) and SigB regulation ( rsbU/SAUSA300_2025 and rsbW/SAUSA300_2023 ) . There were an additional 11 genes that were overrepresented by 3 SD in NO·-stressed cultures and 38 genes between 2 SD and 3 SD overrepresented . These mainly consist of hypothetical proteins and genes encoding enzymes involved in nucleotide and amino acid metabolism ( e . g . apt/SAUSA300_1591 , carA/SAUSA300_1095 , aroK/SAUSA300_1499 , aroD/SAUSA300_0787 , gudB/SAUSA300_0861 , thrC/SAUSA300_1227 and leuD/SAUSA300_2013 ) . It is possible that the slowed growth of these mutants ( many were significantly impaired in serial aerobic cultivation ) provides an advantage during NO· stress . To validate the results of Tn-Seq identification of genes required for NO· resistance , we created clean deletion mutants of five genes or operons newly identified as required for growth during NO· stress: atpG ( F1F0 ATPase ) , uvrAB and uvrC ( nucleotide excision repair ) , mntABC ( manganese ABC transporter ) , and mprF ( antimicrobial peptide resistance factor ) . Additionally , we tested mariner Tn-mutants from the Nebraska Library including iron-sulfur cluster assembly genes nfu and SAUSA300_1248 as well as a thioredoxin-like protein , SAUSA300_0903 . With the exception of ΔmprF , all other mutants exhibited longer lag times and slower growth rates than WT in the presence of NO· but not during aerobic growth , confirming their identification as important for NO· resistance ( Fig 4 ) . The lack of NO· sensitivity observed for ΔmprF could be due to the shorter NO· exposure time in the growth curve , or to differences in growth conditions between a 96-well plate ( where growth curves were performed ) and a shaking 5 ml culture in a test tube ( where selection experiments were performed ) . To identify genes important for fitness in murine SSTIs , we infected mice subcutaneously with 107 cfu of the Tn-Library and harvested bacterial DNA from skin abscesses on days 3 and 7 post infection . Abscesses from two mice were pooled at each time point , and two biological replicates were performed ( for a total of four mice at each time point ) . We averaged R-values between biological replicates at each time point and compared them to the R-value of the 10-hr overnight culture as this was our input library for infection . Importantly , the reproducibility between biological replicates of animal infection was very high ( S4 Fig , Pearson coefficients of 0 . 9 and 0 . 7 for days 3 and 7 respectively ) , indicating that consistent selective pressures were encountered by the bacteria infecting different mice and that there are no significant bottlenecks in this infection model up until day 7 . Of the 168 genes required for fitness in the presence of NO· , 22 were also required for fitness in murine SSTIs at day 3 ( i . e . , Tn-insertions in these genes were underrepresented by more than 2 SD ) , and 49 were required at day 7 ( S4 Table ) . These genes included the F1F0 ATPase operon , mntABC , srrAB , ccpA , rpoZ , codY , qoxABCD , sodA , pyk , ctsR and rot . We also found that an additional 144 genes were required for fitness in SSTIs at day 3 that were not required for in vitro NO· resistance . At day 7 there were 213 genes required for fitness that were not involved in in vitro NO· resistance . Many of these genes include those not analyzed in the context of NO· exposure because they exhibited significant growth defects in aerobically passaged cells , including sarA and genes involved in the synthesis of purines , pyrimidines , heme , aromatic amino acids and menaquinone . As expected , genes encoding secreted virulence factors including toxins and proteases were not identified as required for fitness in SSTIs , presumably due to trans-complementation by the rest of the bacteria in the pool . An interesting group of genes required in SSTIs but not for NO· resistance included lgt ( SAUSA300_0744 ) and lspA ( SAUSA300_1089 ) , genes associated with lipoprotein processing and attachment to diacylglycerol[28 , 29] . Related to this , SAUSA300_1741 , a putative lipoprotein , was essential in SSTIs but not required for in vitro growth . Genes of the LytR-CpsA-Psr family ( msrR/SAUSA300_1257 and SAUSA300_0958 ) , which are thought to link wall techoic acids to peptidoglycan[30] , were also required for fitness at both day 3 and day 7 . These data suggest that cell envelope and cell wall modifications play a major role in persistence in SSTIs , although largely unrelated to NO· resistance . Finally , in each of the day 3 and 7 abscess output pools , mutants in each of the Agr quorum sensing genes ( agrB/SAUSA300_1989 , agrC/SAUSA300_1991 , agrA/SAUSA300_1992 ) were consistently and significantly overrepresented . The ΔatpG mutant was selected for further study because of its essentiality in the presence of NO· and in murine SSTIs both at days 3 and 7 ( S3 Table and S4 Table , Fig 4 , S5 Fig ) . Notably , the NO·-specific growth defect of the ΔatpG mutant could be restored by complementing with the entire atpIBEFHAGDC operon cloned on a medium copy vector ( Fig 4 ) . A major characteristic of high-level NO· exposure is the inhibition of respiration[12 , 31] . In addition , S . aureus frequently encounters hypoxia and iron limitation within inflamed tissue , two conditions that would also limit respiration . Therefore , to test whether ΔatpG is specifically sensitive to NO· stress or is sensitive to the general inhibition of respiration , we grew ΔatpG either anaerobically or in the presence of the divalent cation chelator 2’2-dipyridyl to simulate other respiration-limiting conditions encountered within a host . ΔatpG grew poorly compared to WT under both conditions , suggesting that its sensitivity to NO· is likely due to the general inhibition of respiration ( Fig 5A and 5B ) . Importantly , the addition of nitrate , which S . aureus can use as an alternate electron acceptor for anaerobic respiration , rescued anaerobic growth of the ΔatpG mutant . These data indicate that the S . aureus F1F0 ATPase is critical for normal non-respiratory growth and is therefore likely attenuated in SSTIs due to multiple host factors that limit respiration . We next wanted to test whether ΔatpG was more sensitive to general cellular stress , or whether it was specifically sensitive to respiration-limiting conditions . We found that relative to WT , ΔatpG was more sensitive to peroxide and kanamycin , but not to vancomycin , chloramphenicol , or tetracycline ( Fig 5C ) . Thus , the ΔatpG mutant is not sensitive to cellular stress in general , but rather to specific stressors . However , these stressors are not limited to those that inhibit respiration ( See Below ) . In order to thrive and sustain many basic cellular processes , cells must maintain proton motive force ( PMF ) , representing the sum of membrane potential ( ΔΨ ) and pH gradient ( ΔpH ) . PMF is readily generated during bacterial respiration , in which case its energy can be utilized by the F1F0 ATPase to generate ATP . In contrast , many fermenting bacteria rely on the hydrolysis of ATP by the F1F0 ATPase coupled to proton extrusion for maintaining PMF [32] . Because S . aureus lacks identified proton-specific pumps outside its respiratory chain , we predicted that non-respiring S . aureus might similarly require the F1F0 ATPase for production of PMF via ATP hydrolysis and proton extrusion . We first assessed intracellular ATP levels ( BacTiter Glo ) in WT and ΔatpG S . aureus before and after addition of NO· ( Fig 6A ) . Consistent with fermenting S . aureus hydrolyzing ATP to extrude protons , the ΔatpG mutant exhibited drastically higher levels of ATP compared to WT . This difference was apparent even before the addition of NO· , implying that S . aureus generally uses the hydrolysis of ATP to defend PMF under respiratory conditions as well ( Fig 6A ) . Next we determined ΔΨ in WT and ΔatpG S . aureus before and after addition of NO· , reasoning that the inability to extrude protons under fermenting conditions may result in the loss of charge across the membrane . Surprisingly , the ΔatpG mutant exhibited a hyperpolarized membrane compared to WT before and after NO· addition ( Fig 6B ) . This result could be explained by S . aureus compensating for the loss of the proton pumping F1F0 ATPase by extruding alternative cations other than protons . Defending ΔΨ by extruding alternative cations could result in an inability to specifically export protons efficiently leading to the acidification of the cytosol . Indeed , the ΔatpG mutant exhibited significantly reduced intracellular pH compared to WT both before and after NO· addition ( Fig 7A ) . Interestingly , the intracellular pH of NO·-stressed S . aureus elevates above 8 . 0 . In contrast , the ΔatpG mutant cannot increase intracellular pH above 7 . 0 , even during NO·-stress Consistent with the F1F0 ATPase operating in the hydrolysis mode , overexpressing the atpIBEFHAGDC operon when complementing the ΔatpG mutant caused a dramatic increase in intracellular pH that responded to NO· as does WT ( Fig 7A ) . The lower intracellular pH of the ΔatpG mutant cannot solely explain the lack of growth in the presence of NO· since the pH does not significantly differ from aerobically cultured mutants , which grow near WT levels ( Fig 4A ) . Rather , we hypothesize that enzymes specifically required for non-respiratory growth are only active at the elevated pHs attained by WT and are essentially non-functional at the mildly acidic intracellular pH of the ΔatpG mutant . For instance , the optimal pH for the activity of all three lactate dehydrogenases in S . aureus is above 8 . 0 with little measurable activity below 7 . 5[33] . Accordingly , the ΔatpG mutant excretes less lactate per cell than WT when exposed to NO· ( S7 Fig ) . Moreover , while S . aureus is capable of defending intracellular pH under mild acid stress , a threshold can be reached whereby WT cells are no longer able to maintain optimal cytosolic pH ( extracellular pH ≤ 5 . 5 ) . Below pH 5 . 5 , WT cells exhibit a cytosolic pH similar to that of the ΔatpG mutant ( Fig 7A ) . While this reduced intracellular pH does not affect aerobic growth ( S6 Fig ) , it does eliminate the ability of WT S . aureus to resist NO· ( Fig 7B and S6 Fig ) . These data support our hypothesis that S . aureus must be able to maintain an alkaline intracellular pH under non-respiratory conditions to allow key metabolic enzymes to operate at maximal efficiency . The ΔatpG mutant lacks this ability to alkalinize intracellular pH and therefore cannot replicate efficiently without respiration . Given that murine skin abscesses are replete with high NO· levels early on and transition to hypoxia at later time points ( both limiting respiration ) [7] , it is not surprising that we found the F1F0 ATPase operon to be required for fitness at days 3 and 7 in our Tn-Seq experiment ( S3 Table ) . To verify this result in a non-competitive assay , we infected mice subcutaneously with the ΔatpG mutant strain . Indeed , we found the ΔatpG mutant to be highly attenuated relative to WT ( Fig 8 ) . The mutant forms no visible abscess and is severely reduced in viability 3 days post inoculation . This result underscores the essentiality of the F1F0 ATPase for CA-MRSA pathogenesis in the skin and reinforces the notion that S . aureus bacteria persisting in a skin abscess encounter significant respiratory inhibition . Thus , inhibitors targeting the F1F0 ATPase of S . aureus are likely to be more potent in vivo than they are in vitro as the latter is usually tested in aerated cultures[23] .
S . aureus NO· resistance is a complex , multifaceted trait that remains incompletely understood . In the current study , we performed Tn-Seq to broadly screen for genes contributing to fitness during NO· stress . We identified 168 genes specifically required for in vitro NO· resistance , many of which had not been previously associated with the NO· response in S . aureus . We further identified 166 genes specifically required for persistence in murine SSTIs at day 3 and 262 at day 7 , of which 49 may be required due to their role in NO· resistance . Of the genes required under both conditions , the most prominent group encode the S . aureus F1F0 ATPase , which we show to be essential for the non-respiratory growth of S . aureus . Additionally , we validated many previous findings by identifying srrAB , qoxABCD , ccpA , pyk and rot as being essential for full virulence as well as high-level NO· resistance[8 , 9 , 12 , 14] . Interestingly , we did not find mutants with moderate defects in NO· resistance ( e . g . , hmp and ldh1 ) as having significant defects in murine SSTIs . We have previously observed that many genes contributing to full NO· resistance in S . aureus have much larger virulence roles in sepsis models than in the SSTI model[34] . While difficult to quantify , it may be that the localized skin immune response generates less NO· than the systemic response to acute sepsis . However , the finding that ~30% of genes required for full NO· resistance in vitro are also required for persistence within murine skin infection suggests that immune radicals may still exert some selective pressure in this model . While many of these same gene products confer fitness in other respiration-limited environments as well and may be selected for in the skin for reasons other than NO· , we have shown that hypoxia sets in after day 7 when the wound closes[7] . We did not prolong our Tn-Seq selection beyond day 7 due to the significant loss of bacterial viability that would have imposed population bottlenecks , thereby complicating analyses . This was the first report to our knowledge of a Tn-Seq screen performed with a CA-MRSA isolate of S . aureus . We observed many differences in the fitness requirements in SSTIs using S . aureus LAC compared to a recent study that used the strain HG003 , a MSSA strain[20] . For example , the authors of this study found that a primary group of genes required for fitness at 48-hrs in SSTIs but not rich media were pyrimidine biosynthetic genes . In contrast , we did not observe any differences in fitness for the majority of pyrimidine biosynthetic genes between rich media and SSTIs . Furthermore , the former study found sarA Tn-insertion mutants to be supergrowers in SSTIs . This is in stark contrast to our study , in which sarA Tn-insertion mutants were highly compromised for fitness both in vitro and in vivo . Many other studies also support a role for sarA in S . aureus virulence and fitness , corroborating our result[35 , 36] . Finally , the previous report found a modest defect in fitness within murine abscesses for mutants in agrA , whereas we found insertional inactivation of the Agr system to improve fitness . This implies that in LAC , a strain known to robustly express Agr , production of numerous virulence factors imparts a fitness cost that can be overcome in mixed infections when the bulk bacterial population is Agr-positive . This could also explain the reduced fitness of insertions in codY . While ΔcodY mutants have been reported to be hypervirulent[37] , presumably due to heightened Agr-activity[38] , in a mixed infection enhanced Agr-activity could impose a fitness cost . These discrepancies are likely due to major differences in virulence between LAC and HG003 . LAC proliferates much more extensively in murine SSTIs than HG003 . Moreover , relative to other clinical isolates , CA-MRSA strains have been shown to produce elevated levels of alpha-toxin , PSMs , and secreted proteases[4] , all of which would increase inflammation and host cell lysis while decreasing the ability of the host to confine the lesion . As a result , the skin abscess environments encountered by LAC and HG003 would be very different . Differences in host cell lysis by S . aureus toxins would alter the pool of available nutrients and thus differentially affect the metabolic genes required for fitness . Furthermore , disparities in abscess morphology and inflammation would have major impacts on oxygen availability and levels of antimicrobial inflammatory mediators , also influencing which genes contribute to fitness for each strain . Additionally , CA-MRSA strains have altered virulence gene regulation compared to many HA-MRSA strains . Therefore , elevated expression of certain genes could greatly increase their relative importance during infection . The dissimilarities in results between the two Tn-Seq studies emphasize the importance of recognizing strain differences in S . aureus research and not generalizing results between strains . The most significant finding was the absolute necessity of the F1F0 ATPase for S . aureus growth in the absence of respiration and during infection . This was fully consistent with previous Tn-Seq approach reports [20] . Our initial hypothesis to explain this finding was that this enzyme complex was the only means of establishing PMF in non-respiring cells . PMF is an important cellular energy source used to perform work such as ATP synthesis and solute transport , and it is commonly generated during respiration at coupling sites in the electron transport chain where protons are pumped out of the cell . The only predicted proton pump in the S . aureus respiratory chain is the terminal oxidase QoxABCD ( Fig 9B ) . S . aureus possesses a type II NADH dehydrogenase that does not translocate protons[39] . Likewise , the high affinity terminal oxidase CydAB and nitrate reductase NarGH are also predicted to lack proton translocation activity . However , respiring cells can accomplish proton translocation via Q loops , when quinones are reduced on the cytoplasmic side of the cell membrane and acquire protons that are subsequently released extracellularly when quinol oxidation occurs on the opposite side of the membrane ( Fig 9A ) [40–42] . In the absence of respiration , proton extrusion via Qox or Q loops does not occur and PMF must be maintained in other ways . In many bacteria including E . coli , a major strategy for translocating protons in the absence of respiration is reversal of the F1F0 ATPase reaction , where ATP is hydrolyzed for energy to translocate protons out of the cell[32] . We therefore proposed that the F1F0 ATPase is necessary for functioning as a proton pump in the absence of respiration to contribute to PMF homeostasis ( Fig 9C ) . Consistent with this notion is the fact that ATP levels in the ΔatpG mutant were always higher than WT suggesting that the F1F0 ATPase is a major consumer of ATP rather than a source . However , the ΔatpG mutant was able to maintain PMF and even exhibited a hyperpolarized membrane suggesting that a compensatory ion ( e . g . K+ or Na+ ) was being used to maintain ΔΨ ( Fig 9D ) [43] . In this case , the compensatory ion would only contribute to ΔΨ and not ΔpH . Therefore , to attain constant PMF , a hyperpolarized membrane ( higher ΔΨ ) would be necessary . This would explain the enhanced susceptibility of the ΔatpG mutant to aminoglycoside antibiotics ( e . g . kanamycin , Fig 5C ) given that these drugs rely on ΔΨ to drive uptake[44] . Another consequence of exporting cations other than protons is that the cytosol would become acidic since protons are not efficiently extruded . Indeed , WT exhibited raised pH upon exposure to high NO· levels sufficient to inhibit respiration , whereas the cytosol of the ΔatpG mutant remained slightly acidic after NO· addition . This is problematic for enzymes that are critical for maintaining redox balance when respiration is inhibited , namely the three lactate dehydrogenases of S . aureus , which exhibit maximal activity above pH 8 . 0[33] . The acidic cytosol of the ΔatpG mutant would limit the activity of these enzymes , and potentially others that are critical to non-respiratory growth . The result is that the ΔatpG mutant simply cannot thrive under conditions in which respiration is limited . In other bacterial species , the importance of the F1F0 ATPase varies and has been attributed to a variety of mechanisms including meeting energy demands , balancing pH and serving as a proton “relief valve” [45–47] . Thus , the requirement for the F1F0 ATPase in non-respiring bacteria is not specific to S . aureus as it was also shown to be essential for fermenting Listeria monocytogenes[22] . Regardless of the role for the F1F0 ATPase in individual bacteria , targeting it with novel antimicrobials may represent a broad-spectrum treatment option . Importantly , there have been recent attempts to develop compounds that specifically target bacterial F1F0 ATPases , including diarylquinolone drugs specific for the Mycobaterium tuberculosis F1F0 ATPase that have been entered into phase IIb clinical trials[48 , 49] . While these compounds are fairly specific for mycobacteria , other derivatives developed to target Gram positive F1F0 ATPases have recently been reported as having bactericidal activity on both planktonic and biofilm-associated S . aureus[23] . Here , our results suggest that targeting the F1F0 ATPase may prove to be very effective in vivo given the environmental factors that limit bacterial respiration at sites of inflammation ( e . g . iron limitation , hypoxia and immune radicals ) . By applying a non-biased saturating Tn-Seq approach to identify S . aureus NO·-resistance genes that are critical for fitness within a mammalian SSTI , we have both validated previous reports describing enzymes and regulators essential for full NO· resistance in S . aureus and at the same time identified new genes . Future work will focus on illuminating the mechanisms by which these new genes contribute to NO· resistance in S . aureus as well as enhance fitness within the inflamed skin abscess environment .
Animal studies carried out in this work fall under an animal protocol approved by the University of Pittsburgh Institutional Animal Care and Use Committee ( protocol id 16027663 , PHS Assurance number: A3187-01 ) . The University of Pittsburgh is an AAALAC accredited institution and adheres to the standards set by the Animal Welfare Act and the NIH Guide for the Care and Use of Laboratory Animals . Bacterial strains , plasmids , and primers are listed in S4 Table . S . aureus LAC , a USA300 isolate , was used as background for all experiments . LAC Tn-library construction is described below . Deletion mutants to verify Tn-Seq results were created via allelic exchange using previously described methods[50 , 51] . TSB containing 5mg/ml glucose ( achieved by supplementation with an additional 2 . 5mg/ml glucose ) was used for in vitro growth experiments to ensure glucose was not limiting given the its requirement for NO· resistance[9] . Growth curves were performed using 200μl cultures within a 96-well plate . A Tecan Infinite M200 Pro microplate reader was used to detect change in absorbance ( 650nm ) at 15-min intervals . Growth curves were run for 24-hrs , or 96 cycles of 1mm orbital shaking for 830s followed by 1mm linear shaking for 30s . Overnight cultures were grown in TSB 5mg/ml glucose , washed with PBS , and diluted to an OD650 of 0 . 01 for each growth curve . The NO· donors used in this study were 2 , 2′- ( hydroxynitrosohydrazono ) bis-ethanimine ( DETA ) /NO , or a mixture of NOC-12 and diethylamine nitric oxide ( DEA/NO ) , each resuspended in 0 . 01 N NaOH . The transposon library was also constructed using S . aureus LAC . For generation of the transposon library , we used a modified version of the plasmids and protocol for bursa aurealis transposition in S . aureus described previously[25] . We modified the plasmid pFA545 by replacing the xylose-inducible promoter and transposase allele ( tnp ) ( contained within the NheI-digestible fragment ) with a new fragment containing the constitutive lgt promoter fused to tnp to form pMG020 . This plasmid was transformed into RN4220 and a Φ-11 phage lysate was generated immediately for future steps . A S . aureus LAC strain containing pBursa was transduced with pMG020 Φ-11 phage lysate , incubated at 30° C , and individual colonies from the transduction were resuspended in 100μl PBS . 10–15μl of this resuspension was plated on large petri dishes ( 150mm ) of TSB containing 10μg/ml erythromycin and incubated at 43° C for 48-hrs to allow for transposition to occur . 60 plates of approximately 2 , 500 colonies were each scraped by adding 2ml of TSB 10μg/ml erythromycin + 25% glycerol . Each aliquot was thoroughly vortexed and then 1ml from each was combined into a single pool representing approximately 150 , 000 transposon mutants ( ~2 colonies for each individual Tn insertion ) . 100μl aliquots of this pool were frozen at -80° C until use . Before using any aliquots for experiments , one aliquot was thawed , and DNA was extracted using the Epicentre MasterPure Gram Positive DNA Purification Kit and subjected to Tn-Seq Analysis . To verify an assay in which NO·-sensitive mutants could be selected against in vitro , we mixed known NO·-sensitive mutants ( ΔsrrAB , Δhmp , Δldh1 , and ΔsarA ) of varying sensitivities with WT LAC at a ratio of 1:100 ( more representative of a Tn-Library in which most of the culture will not be NO· sensitive ) . Cultures were diluted to a starting OD650 0 . 01 ( 107 cfu/ml ) containing a 1:100 ratio of mutant to WT in 5ml of TSB 5mg/ml glucose . For NO· selection , 10mM DETA/NO was added at inoculum , and every 12-hrs for a 48-hr period cultures were diluted 1:100 into fresh , warm TSB 5g/L glucose plus 10mM DETA/NO . For aerobic selection , cultures were grown in the absence of DETA/NO and were diluted 1:100 every 5-hrs for a 15-hr period . At each time point , serial dilutions of the cultures were plated both on plain TSB and TSB containing appropriate antibiotics to select for each mutant , allowing for the ratio of mutant:WT to be plotted over time . For all experiments using the Tn-library , a 100μl aliquot of library was thawed and then added to 100ml of TSB 5g/L glucose and grown for short overnight of 10-hrs ( to minimize selection during stationary phase ) . Composition of this inoculum culture was also analyzed by Illumina sequencing . For selection in the presence of NO· , an overnight culture was diluted to an OD650 0 . 02 ( approximately 107 cfu/ml ) in 5ml of TSB 5mg/ml glucose , and 5mM DETA/NO was added . The culture was grown shaking at 37° C . Every 12-hrs for a 48-hr period , the cultures were diluted 1:100 into fresh , warm TSB 5g/L glucose and fresh 10mM DETA/NO was added . Additionally , at each 12-hr dilution time point , the culture was plated for cfu to determine the number of generations that had occurred; on average there were 6–7 generations per 12-hrs for a total of 24–28 generations per experiment . For selection during aerobic growth , a starting OD650 0 . 02 was used and cultures were grown shaking at 37° C but in the absence of DETA/NO . To achieve approximately the same number of generations per serial passage as during growth in the presence of DETA/NO , cultures were diluted 1:100 into fresh TSB every 5-hrs for a 20-hr period and again plated for cfu at each 5-hr time point . For selection in murine SSTIs , six-week old C57/B6 mice were infected as previously described[34] . Briefly , mice were shaved and anesthetized with avertin ( 250 mg/kg ) , and inoculated subcutaneously with 107 cfu ( 20μl of overnight diluted in PBS to an OD650 1 ) . Four mice were sacrificed each at day 3 and day 7 post-infection . To collect bacteria from abscesses , we modified a recently published procedure[20] . Abscesses were removed and pushed through mesh 40 μm nylon cell strainers ( Falcon ) for homogenization , using 1 . 5ml TSB + erythromycin . Each 1 . 5ml tissue homogenate was diluted 1:10 into 15ml of TSB erythromycin and a 5-hr outgrowth was allowed by shaking at 37° C . After outgrowth , the cultures were centrifuged to pellet bacteria and samples were pooled into two groups ( each representing two mice ) to ensure enough bacteria for DNA extraction . Each pellet was resuspended in 1ml TSB Erythromycin + 25% glycerol and frozen at -80° until DNA extraction . DNA was extracted from each library using the Epicentre MasterPure Gram Positive DNA Purification Kit ( Madison , WI ) . DNA was then fragmented by nebulization to a range of ~300-700bp fragments using Rapid Nebulizers ( 454 Life Sciences , Branford , CT ) according to manufacturer instructions , followed by purification using a QIAquick PCR Purification Kit ( Qiagen ) . Libraries were prepped for sequencing according to a previously published protocol with minor modifications[52] . Poly-C tails were added to 1μg of fragmented DNA using terminal deoxynucleotidyl transferase ( Promega , Madison , WI ) in a 20μl reaction ( 47 . 5μM dCTP , 2 . 5μM ddCTP , 0 . 5μl TdT enzyme , 4μl 5x buffer ) run for 1-hr at 37° C . After the reaction , products were purified using Edge Biosystems spin columns ( Gaithersburg , MD ) . An initial PCR was performed to specifically amplify transposon-genome junctions with a poly-G primer ( olj376 ) and a transposon specific primer ( olj510 ) . The 50μl PCR reaction contained 5μl poly-C DNA , 1 . 8μM primer olj376 , 0 . 6μM primer olj510 , 0 . 4mM dNTPs , 1μM Easy-A Cloning Enzyme , and 5μl 10x buffer . The thermocycler was programmed as follows: cycle 1 ( 1x , 1-min at 95° C ) , cycle 2 ( 25x , 30-sec at 95° C , 30-sec at 58° C , 2-min at 72° C ) , and cycle 3 ( 1x , 2-min at 72° C ) . For additional specificity and to add Illumina adapter sequences and barcodes , a second PCR was performed with a nested transposon-specific primer ( olj511 ) and standard Illumina barcoding primers . The 50μl PCR reaction contained 0 . 5μl from first PCR reaction , 0 . 6μM primer olj511 , 0 . 6μM barcode primer , 0 . 4mM dNTPs , 1μM Easy-A Cloning Enzyme , and 5μl 10x buffer . The thermocycler was programmed as follows: cycle 1 ( 1x , 1-min at 95° C ) , cycle 2 ( 15x , 30-sec at 95° C , 30-sec at 52° C , 2-min at 72° C ) , and cycle 3 ( 1x , 2-min at 72° C ) . To remove primer-dimers and further size-select samples , we used Agencourt AMPure XP beads to purify PCR products . Prepped libraries were then multiplexed and sent for sequencing on an Illumina HiSeq2500 using a custom primer ( olj512 ) and 50x unpaired reads . Reads were mapped to the S . aureus USA300_FPR3757 genome with Geneious 8 . 0 using medium sensitivity over three iterations to generate . sam files for further analyses . The program Tn-Seq Explorer[53] was used for tabulating read counts per Tn-insertion , insertion number and insertion density ( # of actual insertions/# of possible insertions ) for each gene . Insertion density takes into account both gene length and AT content , both of which affect the number of insertions possible . Median read per insertion was determined for each gene using a python-based Tn-Seq code ( available upon request ) . There were a significant number of overrepresented reads in the output libraries from murine SSTIs . To exclude these “jackpots” from the analysis , we used median read counts per gene rather than total or average read counts . We generated “representation” or “R” values for each gene , where R = ( median read count ) x ( insertion density ) , thereby eliminating any influence from random jackpots . To examine the relative importance of every gene for fitness within each library , log transformed R values for each data set were used to define essential genes ≤ 3 SD below the mean R-value or genes that significantly contribute to fitness ( 3 SD ≤ R-value ≤ 2 SD below mean ) . Ratios between R-values of NO· cultured cells to aerobically cultured cells were calculated . First , all genes deemed essential for serial aerobic passage ( ≤ 3 S . D . below median log-transformed R-value ) were eliminated . The remaining were used to compare R-values between NO· and aerobic cultures . Similarly , ratios between day 3 and day 7 compared to overnight cultures were generated by initially removing genes essential for overnight growth . Since atpG is located roughly in the middle of the atpIBEFHAGDC operon and thus is likely polar on the distal genes , we complemented the ΔatpG operon with pEP01 harboring the entire operon under its native promoter on pLZ-Spec . A 7 . 2 kb fragment containing the atpIBEFHAGDC operon and its promoter were amplified with atp . 1a/1b ( S4 Table ) and cloned into the NdeI/XhoI sites in pLZ-Spec . The resulting pEP01 was electroporated into AR1524 ( LAC ΔatpG ) to yield the complemented strain , AR1591 . Aerobic , NO· , and iron chelation growth curves were performed as described above in 96-well plates on a Tecan M200 plate reader . For NO· exposure , strains were treated with 10mM DETA/NO at the time of inoculation or with the combination of 10mM NOC-12/ 1mM DEA/NO at mid-exponential phase ( OD650 = ~0 . 2 ) . For iron chelation , strains were treated with 2mM 2 , 2’-dipyridyl at the time of inoculation . Anaerobic growth curves were performed using a BioTek Synergy H1 plate reader with oxygen control . The percent oxygen was maintained at 1% throughout the experiment with the dissolved oxygen being even lower . Overnight cultures of WT and ΔatpG grown in TSB 5mg/ml glucose were diluted to an OD650 0 . 01 in 96-well plates . Two-fold serial dilutions of H2O2 , kanamycin , vancomycin , tetracycline , and chloramphenicol were added to wells of each strain at the time of inoculation . Cultures were shaken at 37° C in a Tecan M200 plate reader for 18-hrs . Minimum inhibitory concentration ( MIC ) was defined as the minimum amount of reagent required to inhibit growth above OD650 0 . 15 for 18-hrs . Intracellular ATP concentrations were measured for WT and ΔatpG using the BacTiter-Glo microbial cell viability assay ( Promega ) according to manufacturer instructions . ATP levels were normalized to the OD650 value for each time point . Strains were grown in TSB 5mg/ml glucose in 96-well plates , and aerobic ATP levels were measured at an OD650 0 . 25 , at which time NO· mix ( 10mM NOC-12 , 1mM DEA/NO ) was added to remaining wells ( t0 ) . ATP levels were again measured 1-hr following NO· addition . Membrane potential was measured for WT and ΔatpG using the BacLight Bacterial Membrane Potential Kit ( ThermoFisher ) . Strains were grown in TSB 5mg/ml glucose in 96-well plates and membrane potential was measured at OD650 0 . 25 ( t0 ) , just before NO· addition , and then 1-hr following NO· mix ( 10mM NOC12 , 1mM DEANO ) addition . At each time point , two 200μl wells of each strain were combined and concentrated in half the volume of PBS ( 200μl ) and transferred to a 96-well black , clear-bottom plate . 30μM DiOC2 ( 3 ) dye was added to the concentrated culture . Using a Tecan M200 plate reader , red and green fluorescence was detected ( emission 488 , excitation 525 and 613 ) every 5-min for a 30-min period . The maximum red:green fluorescence ratio was taken as the value for relative ΔΨ . Fluorescence ratios were converted to mV by interpolating data from a standard curve generated by addition of 1 μM valinomycin plus K+ at the following concentrations ( μM ) : 0 , 1 , 3 , 10 , 30 , 100 and 300 . These correspond to a ΔΨ of ( mV ) : -180 , -150 , -120 , -90 , -60 , -30 and 0 , respectively . Intracellular pH was determined for WT , ΔatpG and ΔatpG + pATPase using the pHrodo Red AM Intracellular pH Indicator Kit ( ThermoFisher ) . Strains were grown in TSB 5g/L glucose to an OD650 0 . 25 then exposed to NO· mix ( 10mM NOC12 , 1mM DEANO ) for 1-hr . At 1-hr post . 200μl of each sample was collected and washed with HEPES buffer pH 7 . 4 and then stained with 50 nM pHrodo Red AM staining solution and incubated at room temperature for 30 minutes . Samples were then washed with HEPES buffer pH 7 . 4 and fluorescence measurements were taken using a BioTek Synergy HI plate reader ( excitation 550 , emission 590 ) . Fluorescence measurements were converted to pH values using a standard curve of samples treated with 10μM valinomycin/nigericin at pH levels 4 . 5 , 5 . 5 , 6 . 5 , and 7 . 5 .
|
The human pathogen Staphylococcus aureus is remarkably resistant to many facets of the host immune response , including the antibacterial radical nitric oxide ( NO· ) . The mechanism underlying this resistance is complex and comprises many gene products . Here we employ an approach that involves transposon mutagenesis coupled to next-generation sequencing ( known as Tn-Seq ) to identify the complete set of genes required for S . aureus NO· resistance and virulence . While we identified many previously reported NO·-resistance determinants , new gene products were discovered from this untargeted approach . Specifically , we identified the F1F0 ATPase as being essential during NO· stress and virulence yet dispensable under normal culture conditions . The reason for this conditional fitness contribution stems from the fact that under fermentative conditions , the F1F0 ATPase functions in the ATP hydrolysis mode , effectively extruding protons and raising the intracellular pH above 8 . 0 . This happens to be the optimal pH for many fermentation enzymes . Without the F1F0 ATPase , proton extrusion is limited and the intracellular pH remains too low for efficient fermentation to continue . Thus , during infection when S . aureus must ferment due to the nature of inflamed tissue , the F1F0 ATPase becomes an essential enzyme complex and a valid target for the development of new antimicrobials .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"protons",
"medicine",
"and",
"health",
"sciences",
"neurochemistry",
"pathology",
"and",
"laboratory",
"medicine",
"chemical",
"compounds",
"respiratory",
"infections",
"enzymes",
"pathogens",
"microbiology",
"enzymology",
"carbohydrates",
"neuroscience",
"staphylococcus",
"aureus",
"organic",
"compounds",
"glucose",
"phosphatases",
"pulmonology",
"signs",
"and",
"symptoms",
"genetic",
"elements",
"bacteria",
"bacterial",
"pathogens",
"abscesses",
"staphylococcus",
"medical",
"microbiology",
"proteins",
"neurochemicals",
"nitric",
"oxide",
"microbial",
"pathogens",
"chemistry",
"nucleons",
"adenosine",
"triphosphatase",
"physics",
"biochemistry",
"diagnostic",
"medicine",
"organic",
"chemistry",
"nuclear",
"physics",
"genetics",
"transposable",
"elements",
"biology",
"and",
"life",
"sciences",
"monosaccharides",
"physical",
"sciences",
"genomics",
"mobile",
"genetic",
"elements",
"organisms"
] |
2018
|
Genetic requirements for Staphylococcus aureus nitric oxide resistance and virulence
|
Plasmacytoid dendritic cells ( pDC ) are innate immune cells that sense viral nucleic acids through endosomal Toll-like receptor ( TLR ) 7/9 to produce type I interferon ( IFN ) and to differentiate into potent antigen presenting cells ( APC ) . Engagement of TLR7/9 in early endosomes appears to trigger the IRF7 pathway for IFN production whereas engagement in lysosomes seems to trigger the NF-κB pathway for maturation into APC . We showed previously that HIV-1 ( HIV ) localizes predominantly to early endosomes , not lysosomes , and mainly stimulate IRF7 rather than NF-κB signaling pathways in pDC . This divergent signaling may contribute to disease progression through production of pro-apoptotic and pro-inflammatory IFN and inadequate maturation of pDCs . We now demonstrate that HIV virions may be re-directed to lysosomes for NF-κB signaling by either pseudotyping HIV with influenza hemagglutinin envelope or modification of CD4 mediated-intracellular trafficking . These data suggest that HIV envelope-CD4 receptor interactions drive pDC activation toward an immature IFN producing phenotype rather than differentiation into a mature dendritic cell phenotype .
Type I interferon ( IFN ) plays a dichotomous role in chronic viral infections such as Human Immunodeficiency Virus-1 ( HIV ) , contributing to the control of viral replication during the earliest stages of infection , yet fueling disease progression by activating target cells for infection , decreasing antiviral gene expression , enabling infection with increased reservoir size , and accelerating CD4 T-cell loss [1–8] . Plasmacytoid dendritic cells ( pDC ) are thought to play a significant role in IFN responses during HIV infection , arriving rapidly at sites of mucosal transmission [4] and relocating from blood to lymphoid tissues where they produce pro-apoptotic and pro-inflammatory IFN [9–11] . Cellular mechanisms underlying HIV-stimulated IFN production by pDC are only partially understood . We have previously shown that abundant IFN is produced by pDC upon HIV stimulation through endosomal recognition of genomic RNA by TLR7 . This response requires the presence of HIV envelope protein on viral particles , interactions between CD4 and the viral envelope protein , HIV endocytosis and endosomal acidification; however , co-receptor usage , viral fusion and viral replication are not required [12 , 13] . Cell-to-cell infection seems to amplify pDC responses to HIV , however precise mechanisms underlying differences between cell-free and cell-to-cell pDC activation are not clearly defined [14] . We and others have shown that pDC are highly resistant to HIV infection , and this block to replication is IFN-independent [15 , 16] . In addition to IFN production , pDC can act as antigen-presenting cells ( APC ) to activate T-cell–mediated adaptive immune responses [17–21] . Acquisition of an APC phenotype requires specific signals that are distinct from the signals that induce large amounts of IFN . We have previously shown that HIV stimulated pDC express low levels of the co-stimulatory molecule CD86 and express Indoleamine 2 , 3-dioxygenase ( IDO ) , a potent inducer of regulatory T cells , indicating that they do not differentiate into mature APC and fail to stimulate potent T cell responses [22 , 23] . However , pDC can differentiate into APC with influenza virus or the synthetic TLR7 agonist R837 and are able to cross-present antigens from HIV-1-infected apoptotic cells to HIV-specific CD8+ T lymphocytes , demonstrating that pDC do not have an intrinsic defect in presentation of HIV antigens , but rather that sensing of HIV does not provide the signals that are required for efficient differentiation of pDC into APC [17] . pDC sense single stranded RNA or unmethylated DNA containing Cytosine–Guanosine dinucleotides ( CpG ) through Toll-like receptors ( TLR ) 7 and 9 , respectively , located in endosomal compartments . Both TLR7 and TLR9 signal through the adapter protein myeloid differentiation primary response gene 88 ( MyD88 ) . Downstream IFN signaling occurs in response to activation of IFN genes through phosphorylation of interferon regulatory factor 7 ( IRF7 ) , whereas downstream signaling of nuclear factor kappa-light-chain-enhancer of activated B cells ( NF-κB ) leads to the transcriptional activation of proinflammatory kinases and upregulation of MHC and co-stimulatory molecules necessary for maturation into APC . [12 , 24] . The functional response of pDC to pathogens is flexible . As posited by the spatiotemporal model of pDC sensing [25] , differential pDC activation is likely related to the subcellular location where the TLR senses the pathogen . Thus , engagement of TLR 7/9 in the early endosomes of pDC preferentially triggers the IRF7 signal cascade , leading to type I IFN responses; whereas engagement of TLR7/9 in lysosomes preferentially triggers the NF-κB signal cascade , leading to the production of proinflammatory cytokines TNFα and IL6 , upregulation of co-stimulatory molecules , and an APC phenotype [25 , 26] . Differential trafficking and therefore sensing of synthetic TLR9-activating CpGs is attributed to sequence-related secondary and tertiary structural features of the CpGs . CpGs which contain phosphodiester backbones and palindromic motifs ( CpGA ) form multimeric complexes and traffic to early endosomes for IRF7 signaling whereas CpGs which contain phosphorothioate backbones and lack palindromic motifs ( CpGB ) traffic as monomers to lysosomes for NF-κB signaling . Intermediate CpGs ( CpGC ) combine structural elements of both CpGA and CpGB , traffic to both compartments , and stimulate both IRF7 and NF-κB signaling [27–29] . While the spatiotemporal model of pDC sensing has been most clearly evaluated using synthetic TLR9 agonists ( CpG ) , we have shown that the model also applies to HIV and TLR7 , whereby HIV traffics to early endosomes in pDC , activating IRF7 signaling rather than NF-κB signaling [22 , 23] . The upstream events that determine activation of each of these pathways , and in particular , HIV virion trafficking in pDC , are currently unknown , however , prior studies suggest that HIV envelope may play a major role [13 , 30 , 31] . Here we demonstrate that HIV trafficking and pDC phenotype is predominantly determined by envelope-CD4 interaction , such that manipulation of HIV envelope or CD4 intracellular trafficking enables modulation of divergent sensing of HIV .
We hypothesized that HIV envelope protein interactions with cell surface CD4 determine the intracellular trafficking of HIV and the resultant signaling in pDC , based on the spatiotemporal model of TLR signaling [25] . To test this , we replaced HIV envelope protein with envelope protein from a virus that activates pDC to differentiate into mature pDC , namely influenza virus . Influenza virus hemagglutinin envelope glycoprotein ( HA ) binds to sialic acids on the cell surface to trigger clathrin-dependent endocytosis [32] . We pseudotyped HIV virions with influenza hemagglutinin glycoprotein envelope ( HA-HIV ) and first compared the functional response of purified pDC to HIV , influenza , and HA-HIV , in terms of magnitude and kinetics of TNFα ( TNF ) and IFN production . TNF is produced downstream of NF-κB signaling and IFN is produced downstream of IRF7 signaling . TNF and IFN were measured by intracellular cytokine staining ( ICS ) at 30 minutes , 2 hours , 6 hours , and 12 hours and in the culture medium by cytokine bead array ( CBA ) and ELISA , respectively , at 2 hours , 6 hours , and 12–24 hours . As previously demonstrated [22] , the response of pDC to HIV was characterized by delayed IFN and TNF responses with IFN predominating at later time points . In comparison , both HA-HIV and influenza stimulated pDC to rapidly produce TNF within 30 minutes- 2 hours , an effect which plateaued by 12–24 hours . Influenza also stimulated early IFN secretion , within 2 hours . HA-HIV induced IFN secretion , albeit at lower levels than Flu itself ( Fig 1A and 1B ) , possibly due to faster trafficking kinetics and early global cytokine shutdown ( as evidenced by earlier IFN and TNF shutdown compared to Flu in ICS Fig 1A ) . Strikingly , TNF was always produced antecedent to IFN , as evidenced by ICS staining , and as has been observed in murine pDC [33] ( Fig 1A ) . After overnight incubation of pDC , HIV stimulated minimal upregulation of CD86 and HLA-DR while HA-HIV and influenza stimulated strong upregulation of CD86 and HLA-DR expression , providing further evidence that HA-HIV and influenza activate NF-κB signaling/maturation pathways in pDC while HIV does not ( Fig 1C ) . Similar maturation and IFN effects were seen whether pDC were stimulated with X4 lab strain , MN or HIV backbone pNL43-ΔEnv-vpr+-luc+ pseudotyped with X5 HIV envelopes ( JRFL , REJO , JOTO ) as compared to HIV backbone pNL43-ΔEnv-vpr+-luc+ pseudotyped with hemagglutinin envelopes H1 and H5 ( S1A Fig ) . A functional measure of pDC maturation is to test whether cells become refractory to re-stimulation by TLR agonists , known as TLR tolerance . We had previously shown that HIV-activated pDC maintain an immature phenotype and are not refractory to re-stimulation to produce IFN . This effect was not due to activation of pDC that had failed to become activated during the previous overnight incubation [22] . We therefore compared the effects of HIV , influenza , HA-HIV , and CpGB , a potent pDC TLR9 maturation stimulus , to inhibit re-stimulation , as a marker of complete pDC maturation . As compared to HIV , HA-HIV inhibited re-stimulation of pDC similarly to influenza and CpGB , thus signifying that HIV pseudotyped with HA matured pDC fully ( Fig 1D ) . Overall , swapping HIV envelope with influenza envelope induced a mature pDC phenotype similar to that induced by influenza activation . Notably , MN HIV , a CXCR4 lab strain of HIV was used in these experiments . However , HIV ( pNL43-ΔEnv-vpr+-luc+ ) pseudotyped with envelopes JRFL , REJO , or JOTO ( all R5-tropic ) all stimulate pDC to produce IFN , and not to mature , as expected since co-receptor usage is not essential for pDC sensing of HIV ( S1A and S1B Fig ) [12] . Because HA-HIV and influenza similarly stimulated pDC to mature , we sought to investigate whether HA-HIV traffics similarly to influenza in pDC . HIV virions ( pNL43-ΔEnv-vpr+-luc+ ) pseudotyped with HA and packaging green florescent protein ( GFP ) and HIV virions ( pNL43-ΔEnv-vpr+-luc+ ) pseudotyped with JRFL ( R5 envelope ) and packaging GFP were generated , using eGFP-Vpr plasmids , as previously described [22] . Influenza packaging GFP were also generated , as previously described [34 , 35] . We found that HA-HIV , similarly to influenza and unlike HIV itself , rapidly trafficked to lysosomes by 30 minutes as evidenced by co-localization with Lysotracker , a dye that traffics to these organelles ( Fig 2A and 2C ) . Both influenza and HA-HIV extensively co-localized with Lysotracker at 30 minutes and 2–4 hours , whereas HIV was barely visible inside the cell at these early time points . After overnight incubation HIV was well visualized inside the cell , but did not co-localize well with lysotracker . Influenza and HA-HIV still seemed to co-localize with lysotracker even though the florescent signal was faded , likely due to lysosomal degradation of the virions ( S2A–S2D Fig ) . We confirmed that HIV traffics to early endosomal ( EEA1 ) compartments by 18 hours as previously shown [22] , whereas influenza and HA-HIV traffic significantly less to these compartments and the green signal is faded at 18 hours ( Fig 2B and 2D ) . Thus , the nature of the viral envelope seems crucial to determining trafficking of virions in pDC , and for the downstream signaling pathways activated in different intracellular compartments . Our results indicate that viral envelope protein dictates early intracellular trafficking of virions in pDC , suggesting that trafficking is directed by interaction of HIV envelope protein with its cognate receptor . In pDC , sensing of HIV involves CD4-mediated endocytosis [12] . CD4 is a type I integral membrane glycoprotein that can be internalized through clathrin-mediated endocytosis . The intracellular tail of CD4 displays motifs important for its internalization: a dileucine motif that allows interaction with the clathrin adaptor 2 ( AP-2 ) [36] and an adjacent serine , whose phosphorylation augments the affinity of the dileucine motif for AP-2 [36] , thereby regulating CD4 endocytosis . In cells of macrophage-monocyte lineage , CD4 is constitutively endocytosed at low levels through clathrin-coated pits to early and recycling endosomes [37] , as CD4 is serine-phosphorylated to some extent even in unstimulated cells [36] . We first examined whether CD4 itself and CD4-associated targeting motifs are responsible for the predominant localization of HIV in early endosomes . We used HEK 293 T ( HEK ) cells and HEK reporter cells as a model because HEK cells do not express CD4 under native conditions and therefore manipulation of CD4 trafficking and viral-CD4 interactions can be studied more clearly . Moreover , due to technical limitations , it was not possible to transfect or transduce primary pDCs or the Gen2 . 2 pDC cell line to undertake these studies . We engineered hybrid CD4 molecules , mutating its intracytoplasmic tail or swapping it with the intracellular domain of CD205 ( DEC205 ) and Lamp-1 ( Fig 3A ) . CD205 and Lamp-1 contain distinct lysosome-targeting motifs in their intracytoplasmic tail that induce constitutive targeting to late endosomes/lysosomes [38–40] . DEC-205 expresses the coated-pit internalization sequence ( FSSVRY ) and lysosome-targeting motif ( EDE ) , whereas Lamp-1 expresses the lysosomal targeting motif ( YQTI ) . Several mutants were tested , as visualized in Fig 3A: ( 1 ) . CD4-WT ( wild type CD4 ) , ( 2 ) . CD4-STOP ( lacking the cytoplasmic domain ) , ( 3 ) . CD4-DEC ( replacing the CD4 cytoplasmic domain with the DEC-205 cytoplasmic domain to shuttle CD4 to the lysosomes ) , and ( 4 ) . CD4-LAMP ( replacing the CD4 cytoplasmic domain with the Lamp-1 cytoplasmic domain to shuttle CD4 to the lysosomes ) . The CD4 mutant sequences were introduced into lentiviral vectors for stable transduction of HEK cells for microscopy , and HEK-Blue hTLR7-expressing cells to measure NF-κB activation by HIV . HEK-Blue hTLR7 co-express human TLR7 and an NF-κB inducible secreted embryonic alkaline phosphatase ( SEAP ) reporter gene . CD4 expression was maintained in the presence of puromycin , and CD4 expression across cell lines was uniform at 65–75% . For these experiments we used JRFL HIV packaging GFP- HIV Gag-iGFP ( GFP HIV ) to track HIV intracellular trafficking . Following incubation with HIV for 2 to 4 hours , CD4-expressing cells bound and endocytosed HIV efficiently , as shown in Fig 3C . This time point was chosen because HIV was not visualized well before 2 hours , and the fluorescent signal was faded after overnight incubation . The main path of viral entry in CD4-expressing HEK cells is CD4-mediated endocytosis as CD4 blockade completely abrogated HIV uptake ( Fig 3B ) . HIV co-localized extensively with CD4 , whatever CD4 construct the cells expressed , in the range of 65% to 80% co-localization per cell , as measured by single cell Mander’s coefficient ( Fig 3D ) . HEK cells which are not expressing CD4 do not take up HIV as represented by the ( - ) condition in representative images ( Fig 3C ) . Although the intracellular distribution pattern appeared different between the different CD4 constructs , with CD4-WT and CD4-STOP appearing more cell-surface associated and CD4-LAMP and CD4-DEC appearing more internal compartment associated , the overall fluorescence intensities of cell-associated HIV were comparable . To better characterize the intracellular localization of GFP- HIV in HEK cells expressing CD4 hybrids , cells were exposed to GFP-HIV for 2–4 hours , then fixed and stained for the early endosomal marker EEA1 , the recycling endosomal marker transferrin receptor ( TfR ) , or the lysosomal marker LAMP1 . As an additional lysosomal marker , after 2–4 hours of GFP-HIV incubation , lysotracker was added to culture media , and live imaging was performed . Internalized HIV trafficked predominantly to early EEA1+ and TfR+ endosomes in CD4-WT , similar to trafficking of HIV in pDC ( Fig 4 ) . However , HIV trafficked mainly to LAMP-1+ and lysotracker+ lysosomes in CD4-DEC and CD4-LAMP ( Fig 5 ) . Surprisingly HIV trafficked to EEA1+ compartments in CD4-STOP instead of remaining on the cell surface , but did not traffic predominantly to TfR ( Fig 4A–4D ) . As endosomes containing EEA1 ultimately direct trafficking to the degradative machinery of the cell whereas endosomes containing TfR do not [41] , it is likely that some of the mutant CD4-STOP proteins are targeted for degradation to some extent . Overall , these data are consistent with the expected trafficking patterns of each CD4 construct , and indicate that replacing the intracellular tail of CD4 with those of CD205 or LAMP-1 targets CD4 and HIV to the late endosomes/ lysosomes , demonstrating that targeting motifs in CD4 drive the intracellular localization of HIV via CD4-mediated endocytosis . We next examined the functional consequences of CD4 altered trafficking . The lentiviral constructs ( Fig 3A ) were transduced into HEK-Blue cells to measure NF-κB activation . Notably this cell type is not optimized to produce IFN , therefore only the NF-κB signaling arm was tested in this system . R848 , a TLR7/8 agonist , was used as a positive control and strong NF-κB activator that rapidly accumulates in late endosomes [42] . The TLR9 agonist CpGB was used as a negative control in these TLR7-expressing cells and did not induce NF- κB activation in any condition ( Fig 6 ) . Whereas HIV did not induce NF- κB activation in CD4-WT and CD4-STOP expressing cells , it induced NF-κB activation in CD4-DEC and CD4-LAMP expressing cells . These results support the spatiotemporal model of cell signaling , whereby partition into early or late endosomes regulates IFN vs NF-κB signaling [25] . HIV traffics to early endosomes in pDC and in CD4-WT expressing cells , and induces weak NF-κB signaling , whereas retargeting CD4 to lysosomes allows for NF-κB activation by HIV . We also examined more precisely why native CD4 delivers HIV into early endosomes . Endocytosis and intracellular CD4 trafficking is dependent on a dileucine motif in its intracellular domain , and is regulated by phosphorylation of an adjacent Serine ( Ser408 ) . Phosphorylation of Ser408 , as occurs with phorbol ester ( PMA ) , dramatically enhances CD4 delivery from the cell surface to the lysosomes [43] . In contrast to PMA stimulation , HIV activation of human pDCs does not cause marked CD4 internalization at early timepoints . Whether pDC are unstimulated or stimulated with HIV , CD4 internalization is grossly unchanged , whereas PMA stimulation , in the absence or presence of HIV , causes internalization at early timepoints ( Fig 7 ) . To investigate whether CD4 phosphorylation on Ser408 targets CD4 and HIV to late endosomes/ lysosomes , we generated two mutations of the serine residue: one to alanine to abrogate phosphorylation ( CD4-SA ) , and one to the phospho-mimic Glutamic Acid ( CD4-SE ) , and produced HEK cells stably expressing these CD4 mutants . Intracellular localization of HIV and CD4 was monitored by microscopy . As shown in Fig 8 , HIV co-localized with CD4 and appeared less membrane associated in CD4-SE cells , whereas it remained predominantly on the surface of CD4-SA cells . In CD4-SA cells , intracellular HIV accumulated in EEA1+ and TfR+ compartments ( Fig 9 ) , similarly to WT-CD4 cells ( Fig 4 ) , whereas HIV accumulated in Lamp-1+ and lysotracker+ compartments in CD4-SE cells ( Fig 10 ) , similarly to CD4-DEC cells and CD4-Lamp cells ( Fig 5 ) . To test the functional consequences , we transduced HEK-Blue cells with CD4-SA and CD4-SE , as compared to CD4-WT , to measure NF-κB activation , as above . Whereas HIV did not induce NF-κB activation in CD4-WT and CD4-SA expressing cells , it induced NF-κB activation in CD4-SE expressing cells ( Fig 11 ) . Thus , CD4 phosphorylation on Ser408 appears to target CD4 and HIV to late endosomes/ lysosomes , whereas it is routed to early endosomes in its absence ( CD4-SA and CD4-WT ) . Poor specificity of available anti-phospho-Ser408 CD4 antibodies precluded the possibility of studying the phosphorylation state of Ser408 CD4 in primary pDCs after HIV activation . While our data strongly suggest that HIV trafficking and subsequent immune signaling in pDC is driven by HIV envelope/CD4 interactions , we tested an alternative hypothesis of pDC signaling , that the strength of TLR signaling in the early sorting endosomes determines the trafficking and subsequent signaling of TLR agonists . Previous work showed that TLR signaling in other cell types accelerates endosomal maturation through TLR-induced p38 mitogen-activated protein kinase ( p38 ) signaling , as the absence of a TLR/MyD88 signal diminishes phagosome maturation [44] . TLR activation also activates lysosomal function in myeloid DC , which could further influence TLR signaling pathways [45 , 46] . According to this model , strong TLR triggering in the early endosome would accelerate endosomal maturation , and endocytosed viruses or oligonucleotides would rapidly reach late endosomes for NF-κB signaling [25 , 33] . Using murine pDC differentiated from TLR7-/- and TLR9-/- bone marrow , we tested this hypothesis . As HIV cannot be used in murine immune systems , we tested TLR9 agonists , focusing on CpGB . CpGB has been shown to rapidly traffic to late endosomes/ lysosomes in mouse and human pDC [25 , 26] , and it induces strong NF-κB signaling and pDC maturation [22] . According to this model where TLR signaling induces endosomal maturation , CpGB would traffic rapidly to lysosomes in wild type ( WT ) or TLR7-/- pDC , but would fail to traffic to lysosomes in murine TLR9-/- pDC . Because MyD88 knockout DCs lacking TLR signaling have been described as having an altered constitutive rate of endosomal maturation [44 , 47] , we also constructed a control oligonucleotide lacking stimulatory CpG motifs ( due to inversion of the CpG motif into a GpC dinucleotide ) which lacks TLR triggering activity . Wild type , TLR7-/- , and TLR9-/- murine pDC were derived from bone marrow with Flt3 ligand , as previously described [48] and purified >85% ( S3A Fig ) . pDCs were incubated with FAM-labeled CpGB or nonactivating FAM-labeled GpC control and were imaged by flow cytometry and live microscopy . We first confirmed the specificity of the various pDC TLR knock out cells by evaluating the expression of the maturation molecule CD86 after overnight incubation of wild type , TLR9-/- , and TLR7-/- pDC with TLR7 agonist R848 and TLR9 agonist FAM-CpGB . As expected , R848 matured WT and TLR9-/- pDC but did not mature TLR7-/- pDC , whereas FAM-CpGB matured WT and TLR7-/- pDC but did not mature TLR9-/- pDC . pDC stimulated with FAM-GpC did not mature WT pDC due to the lack of CpG immunostimulatory motifs ( S3B Fig ) . We then monitored trafficking of FAM-CpGB to lysosomes in WT , TLR7-/- and TLR9-/- pDC by microscopy , by measuring co-localization with Lysotracker Red . Across a z stack spanning the midplane of the cell , we observed that both FAM-CpGB in WT , TLR7-/- and TLR9-/- pDC ( S4A and S4C Fig ) , and FAM-GpC in WT pDC ( S4B and S4D Fig ) , rapidly and extensively trafficked to lysosomes within 15-20min . Thus , TLR activation did not affect intracellular trafficking of TLR agonists in murine pDC , suggesting that an alternative model also exists for human pDC . Altogether , these data demonstrate that divergent HIV-1 sensing by pDC is mediated by CD4-HIV envelope interactions .
Although many receptors and signaling pathways have been shown to modulate TLR signaling pathways in pDC [49] , it is likely that the functional outcome of TLR signaling is determined upstream and early on . The spatiotemporal model of TLR signaling [25] was proposed to account for the functional flexibility of pDC in response to TLR signaling , and argues that the surface or intracellular localization of TLR signaling initiation determines which downstream signaling pathway is triggered , with differing functional outcomes [50] . This is because each compartment is associated with different adaptors and signaling platforms , specialized in inflammatory cytokine secretion and NF-κB activation , or signaling through IRF7 and IFN production [25 , 33 , 50] . As HIV steadily traffics to early endosomes in pDC , a compartment associated with IFN signaling [25] , our goal was to understand how early HIV trafficking is regulated in pDC and how it affects pDC functional response . We studied whether HIV trafficking in pDC involves envelope-receptor interaction and targeting signals in endocytic receptors . A hybrid virus was constructed , where HIV envelope was replaced by influenza hemagglutinin envelope , while maintaining all other HIV structural components unchanged ( HA-HIV ) . In contrast to HIV , influenza is rapidly endocytosed by pDC and triggers a strong NF- κB activation , secretion of inflammatory cytokines , and maturation of pDC [22 , 51] . Strikingly , HA-HIV was rapidly routed to late endosomes/ lysosomes in pDC , contrary to HIV with its native envelope . Furthermore , it induced early secretion of inflammatory cytokines and strong pDC maturation , in a manner and kinetics similar to influenza virus . This demonstrates that virus envelope directly determines HIV trafficking and pDC phenotype . Despite different structural components and nucleic acids , HA-HIV and influenza induced a similar functional response in pDC , which strengthens the importance of viral envelope in determining pDC phenotype . This latter extended to the unresponsiveness of pDC to further stimulation , whereas HIV stimulated pDC could be re-stimulated to produce IFN . HA-HIV triggers IFN secretion , although at lower levels than Flu itself . This may be due to kinetics differences in trafficking to late endosomes and activation of negative signaling pathways or exhaustion . Trafficking to late endosomes , NF-κB activation and pDC maturation correlates with a state of refractoriness , likely established early during stimulation , and already evidenced by a global cytokine shutdown after a few hours . As shown in Fig 1A , this shutdown occurs earlier for HA-HIV than for Flu , at the time when IFN secretion is starting to be amplified . It is likely HA-HIV traffics significantly faster to late endosomes and triggers early cytokine shutdown . Another example of HIV pseudotyping is VSV-G pseudotyped HIV , where VSV-G from vesicular stomatitis virus is used to allow HIV uptake and infection of many cellular subtypes . The putative receptor for VSV-G has been recently suggested to be LDL receptor [52] , which traffics mainly to recycling endosomes but rarely to lysosomes . In accordance with this trafficking pattern and localization in early endosomes , VSV-G-pseudotyped HIV behaves mostly like HIV with a native envelope to trigger high levels of IFN but little pDC maturation [12] . In addition to demonstrating that viral envelope determines HIV localization and pDC phenotype , HA-pseudotyped HIV may also provide a tool to study HIV antigen presentation and vaccine design [53] , as it enhances expression of MHC and co-stimulatory molecules on pDC , and influenza itself triggers a developmental program suited for antigen presentation in pDC [21 , 54] . If the nature of the viral envelope dictates HIV trafficking in pDC , it may be due to intracellular targeting motifs present in the viral receptor ( s ) [38] . HIV is mainly taken up through CD4 in pDC , and we tested whether altering the intracytoplasmic domain of CD4 would affect HIV trafficking and TLR signaling . Indeed , we observed that swapping CD4 intracytoplasmic domain for DEC205 or Lamp1 intracytoplasmic domain dramatically re-routed HIV into late endosomes/ lysosomes in CD4 expressing cells , whereas intracellular HIV was localized predominantly in early endosomes in cells expressing native CD4 . DEC205 and Lamp1 contain lysosomal targeting motifs which are likely responsible for constitutive CD4 and HIV targeting to late compartments . In addition , redistribution of HIV to late endosomes was accompanied by activation of NF-κB , not observed when HIV accumulates in early endosomes , again consistent with the spatiotemporal model of TLR signaling . CD4 contains a dileucine motif in its intracytoplasmic domain , which is essential for CD4 endocytosis [36] . In addition , two adjacent Serines , Ser408 and Ser415 , can be phosphorylated and impact CD4 endocytosis and distribution . Completely deleting these motifs by removing the whole CD4 intracytoplasmic domain indeed almost completely abrogated CD4-mediated HIV endocytosis . However , phosphorylation of Ser408 not only enhances CD4 endocytosis [36] , but also redirects CD4 to lysosomal compartments [43] . Ser408 phosphorylation enhances its association with clathrin Adaptor protein AP-1 and AP-2 [36] . In our experiments , mutating Ser408 to Glutamic acid ( CD4-SE ) to mimic Ser408 phosphorylation , induced a complete redistribution of CD4 and HIV into lysosomes ( Fig 10 ) suggesting that in pDC , HIV traffics by default to recycling endosomes due to the CD4 dileucine motif , in the absence of Ser408 phosphorylation . These results are supported by our CD4 internalization studies where we found that HIV-activation of pDC does not seem to alter CD4 internalization , as compared to PMA-activation . Similarly , in HIV infected cells , Nef triggers endocytosis and degradation of CD4 through a dileucine based motif [55 , 56] and CD4 endocytosis and targeting to lysosomes are encoded in different regions of the Nef protein [56 , 57] . The relatively stable localization of HIV in early endosomes , observed for as long as 18 to 24h , remains unexplained . Furthermore , we observed strong co-localization of CD4 with HIV throughout the course of the study . Although the interaction between CD4 and HIV envelope may be stable enough to maintain CD4-HIV co-localization for a prolonged period of time , other explanations are possible . A recent study described in detail how endocytosis of HIV is coupled to dynamin-dependent endocytosis and partial fusion with plasma and endosomal membrane [58] , which may tether HIV envelope and CD4 to the endosomal membrane in the absence of fusion . Furthermore , pDC possess specialized large perinuclear intracellular stores of MHC-I molecules , with characteristics of recycling endosomes in immature pDC , that can be used as sites for rapid MHC-I loading and peptide presentation [21] . These intracellular stores may represent the stable compartment in which HIV accumulates in pDC , and prolonged localization in these early recycling endosomes may ultimately have important consequences for HIV antigen presentation . pDC may harbor HIV in these structures until activated by a maturation stimulus . Indeed , pDC are capable of HIV antigen cross-presentation [17] , and cross-presentation is strongly enhanced by maturation-inducing stimuli [59 , 60] . Upon influenza activation , stored MHC-I molecules are translocated to the cell surface for efficient cross-presentation by pDC [21] , indicating that the process of maturation drives antigen presentation . On the other hand , the non-acidic environment and limited access to MHC-II compartments may prevent efficient MHC-II peptide generation and association with MHC-II molecules . In addition , MHC-II clustering and antigen presentation by pDC is dependent on NF-κB [51 , 61] , which HIV weakly induces due to localization in early endosomes . The lack of pDC maturation induced by HIV might prevent effective cross-presentation and MHC-II restricted presentation , due to localization in early endosomes and weak NF-κB activation . Thus , the particular compartmentalization of HIV can affect HIV antigen presentation . As shown here , HA-pseudotyped HIV , which traffics to late endosomes and activates NF- κB , may serve as a tool to enhance cross-presentation of HIV antigens by pDC . We also tested whether TLR signaling itself alters maturation of endocytic compartments in pDC as was previously demonstrated in the case of murine macrophages , where TLR/ MyD88 signaling induced marked phagosome maturation , possibly through p38 MAP kinase activation [44] . We tested this model in murine pDC , and compared trafficking of CpGB in WT , TLR7-/- or TLR9-/- pDC . We did not observe any difference in trafficking of CpGB in these conditions , and internalized CpGB was rapidly found within late endosomes/ lysosomes whether TLR signaling had occurred or not . As MyD88-/- cells have altered endosomal maturation [47] , we also used a non-stimulatory type B oligonucleotide ( lacking immunostimulatory CpG motifs ) in WT pDC . However , its trafficking was identical to the stimulatory CpGB oligonucleotide . Furthermore , as described previously [44 , 62] , the lack of TLR signaling , as with nonstimulatory GpC , decreased the speed of CpG endocytosis ( S4E Fig ) , however all oligonucleotides localized identically . These data argue against the above model , whereby rate of endocytosis and/ or agonist potency drive its intracellular localization , and establish trafficking as independent of TLR signaling in pDC [47 , 63] . Using purified human pDCs and viruses , we demonstrate that HIV trafficking in pDC at early timepoints is determined by the initial envelope-receptor ( CD4 ) interaction , and is regulated by receptor targeting motifs . Whether this is still valid in the case of cell-associated virus , this remains to be determined . Engineering of viral envelope for increased pDC maturation and antigen presentation ( e . g . HA-HIV ) or for increased IFN secretion [64] may prove useful for vaccine design and modulation of chronic immune activation in HIV disease .
PBMCs were separated on Ficoll-Hypaque ( Amersham Biosciences ) from buffy coats ( New York Blood Center ) . pDC were purified by BDCA-4 magnetic bead separation ( Miltenyi Biotec ) as described previously [12] , with a purity ranging from 80 to 95% . Cells were cultured in RPMI 1640 Glutamax ( Invitrogen ) with 5% PHS ( Innovative Research , MI ) , gentamycin , and HEPES . HIV-1MN ( X4-tropic ) were produced at the AIDS Vaccine Program , National Cancer Institute as previously described [12 , 23 , 65] . Plasmids encoding JRFL HIV-1 envelope , JOTO HIV-1 envelope , REJO HIV-1 envelope , pNL43-ΔEnv-vpr+-luc+ and pCAGGS ( human airway trypsin-like protease to cleave HA0 to HA1 and HA2 ) were provided by Carol Weiss ( FDA , Silver Spring , MD ) , plasmids encoding CMV/R Influenza H1 ( A/PR8/8/34 ) ( VRC 7702 ) , CMV/R influenza A/PR/8/1934 NA ( VRC 9776 ) , CMV/8R A/Thailand/1Kan-1/2004 H5 ( VRC 7705 ) , and CMV/8R A/Thailand/1Kan-1/2004 NA ( VRC 7708 ) , and were provided by Gary Nabel and Chih-Jen Wei ( Vaccine Research Center , NIH , Bethesda , MD ) . Plasmids encoding vpr-gfp and MN HIV pNL4-3 Δvpr were obtained from David Ott and Jeffrey Lifson ( AIDS Vaccine Program , Frederick , MD ) . Plasmids encoding HIV Gag-iGFP JRFL were obtained from Benjamin Chen ( Mount Sinai School of Medicine , NY , NY ) . The influenza hemagglutinin viral pseudotypes were generated by calcium phosphate co-transfection of 3 . 0x106 HEK cells in a 10cm2 dish with 10ug HIV core ( pNL43-ΔEnv-vpr+-luc+ ) , 400ng hemagglutinin envelope ( VRC 7702 , VRC 9776 , VRC 7705 , or VRC 7708 ) , 100ng PR8 NA ( VRC9776 ) , and 100ng HAT pCAGGS , with media change after 6 hours , and viral harvest at 48 hours . HIV pseudotypes were generated by calcium phosphate co-transfection of 3 . 0x106 HEK cells in a 10cm2 dish with 10ug HIV core ( pNL43-ΔEnv-vpr+-luc+ ) and 6ug Env in pCDNA3 . 1 ( e . g . REJO , JOTO , JRFL ) . HIV Gag-iGFP is a full-length molecular clone of HIV derived from pNL4-3 that packages GFP inserted into the Gag protein between the MA and CA domains of Gag , with JRFL Env cloned into the place of the NL4-3 Env . To generate Gag-iGFP virions for CD4-expressing HEK experiments , 20ug of plasmid was transfected using the calcium phosphate method , with media change at 6 hours and transfection for 48 hours . For all viruses , transfection supernatants were filtered through a 0 . 45uM membrane , pelleted through a 20% sucrose cushion at 25 , 000g for 2hrs at 4°C . Pelleted viruses were resuspended in PBS , aliquoted , and stored at -80°C until use . HIV virions were quantified using p24 ELISA ( AIDS Vaccine Program ) and HA-HIV viruses were quantified using turkey hemagglutination inhibition assay for hemagglutination unit ( HAU ) as well as p24 ELISA . PR8 influenza was provided by David Levy ( NYU ) and influenza packaging GFP was provided by Jesse Bloom ( Fred Hutchinson Cancer Research Center , Seattle , WA ) as previously described [34 , 35] and were quantified by using turkey hemagglutination inhibition assay for hemagglutination unit . Purified pDC were stimulated at 50 , 000 cells/100uL media at 37° C with 5% CO2 with: MN HIV 300ng ( AIDS Vaccine Program , National Cancer Institute ) , HIV pseudotyped HIV envelopes 300ng , type B CPG oligodeoxyribonucleotide ( CpGB ) 2 ug 5’ T*C*G*T*C*G*T*T*T*T*G*T*C*G*T*T*T*T*G *T*C*G*T*T*-3’ where asterisks indicate a phosphorothioate bond ( IDT ) , Resiquimod ( R848 ) 10μM ( 3M Corporation , St . Paul , MN ) , influenza virus PR8 HAU 10 heat inactivated at 56°C for 30 minutes in a water bath , or HA-HIV HAU 10 . For intracellular staining , brefeldin A was added after 30 minutes , 2 hours , 6 hours , or 12 hours; at 24 hours cells were washed , fixed , and stained with PE-conjugated CD123 PE ( BD Biosciences ) , APC-conjugated TNFα ( eBioscience ) , and FITC-conjugated IFN-α ( BD Biosciences ) in 0 . 05% saponin , and analyzed by FACS . Following incubation of pDCs ( 50 , 000 cells/100uL ) with HIV MN and HIV JRFL , influenza , and HA-HIV , culture supernatants were tested for IFN and TNF by IFNα by ELISA ( PBL Interferon Source ) and human inflammatory kit CBA ( Abcam ) , respectively , following manufacturer instructions . For surface staining after overnight incubation , cells were stained with CD123 PE , CD86 APC , and HLA-DR PerCP ( BD Pharmigen ) , washed , fixed with 4%PFA , and analyzed by FACS . For IFNα restimulation experiment , pDCs were incubated overnight with influenza , HIV , HA-HIV , or CpGB , washed , and then incubated again with the same agonists . Culture supernatants were tested after the first and second overnight incubation for IFNα by ELISA ( PBL Interferon Source ) . CD4 expression plasmid ( pcDNAI ) was provided by Nathaniel Landau ( NYU , NY , NY ) . It was used as a template to mutate CD4 by overlapping PCR . Mutated CD4 were then inserted into pLenti vectors , and lentiviruses were produced using Mirus TransIT co-transfection of 3 . 0x106 HEK cells in a 10cm2 dish for 72 hours with 1 . 45 ug VSVg , 2 . 05 ug RSR-Rev , 2 . 9 ug pMDL plasmids ( provided by Dr Landau ) and 8 . 7 ug pLenti-CD4 constructs ( WT , STOP , DEC , LAMP , SE , SA ) . CD4-STOP was encoded of the first 425 amino-acids , including the transmembrane domain of CD4 . CD4-DEC contained the full extracellular domain of CD4 ( 397 amino acid in the immature form ) and the transmembrane and intracytoplasmic domain of DEC205 ( 56 amino acids ) . CD4-Lamp encoded the full extracellular and transmembrane domain of CD4 , and 12 amino acids of Lamp intracytoplasmic domain . CD4-SE and CD4-SA was identical to CD4 WT , except for a mutation of Serine 408 ( mature form ) for glutamate and alanine , respectively . CD4-expressing HEK cell lines were generated by infecting 3 . 0x106 HEK cells in a 10cm2 dish with lentivirus transfection supernatants and CD4 expression was maintained in the presence of puromycin . C57BL/6 wild type , TLR7-/- and TLR9-/- double-knockout mouse femurs were provided by B . Pulendran , Emory University . Mice were maintained in specific-pathogen-free conditions at the Emory Vaccine Center vivarium in accordance with all animal protocols reviewed and approved by the Institute Animal Care and Use Committee of Emory University . Bone marrow cells were isolated by flushing femurs with PBS supplemented with 2% heat inactivated FBS . BM cells were resuspended in Tris-ammonium chloride at room temperature for 1 minute to lyse RBC , washed , then cultured in RPMI 1640 Glutamax ( Invitrogen ) with 10%FBS , 1nM sodium pyruvate , 10mM HEPES buffer , 100 units/mL penicillin , 100ug/mL streptomycin , 2mM L-glutamine , 1% MEM nonessential amino acids . BM cells were cultured for 8 days at 1 × 106 cells/ml in 24-well plates in culture medium supplemented with 200 ng/ml recombinant murine Flt-3 ligand ( Peprotech ) as previously described [48] . After 8 days pDCs were purified from BM cells using the mouse plasmacytoid dendritic cell isolation kit II ( Miltenyi ) with a purity ranging from 85 to 95% and cells were phenotyped by CD11c PerCP and PDCA1 APC ( BD Pharmigen ) . Purified murine pDC ( WT , TLR7-/- , and TLR9-/- ) were stimulated overnight at 50 , 000 cells/100uL media at 37°C with 5% CO2 with R848 10μM ( 3M ) , 5’FAM-CpGB 5’ TCGTCGTTTTGTCGTTTTGTCGTT-3’ 2 ug ( IDT ) , or 5’FAM-GpC 5’-TGCTGCTTTTGTGCTTTTGTGCTT-3' 2ug ( IDT ) , both with phosphodiester backbones , and were stained with CD11c PerCP , PDCA1 APC , and CD86 PE ( BD Pharmigen ) . Purified murine pDC ( 200 , 000 cells/200uL ) were stimulated in 0 . 01% Poly L-lysine ( Sigma ) coated 8 chamber polystyrene vessel tissue culture treated glass slides ( CultureSlides BD Falcon ) in 10% FBS culture media ( 200uL ) with FAM-CpGB 2ug or FAM-GpC 2ug and stained with 1uM lysotracker ( Invitrogen ) . Cells were imaged by live microscopy for one hour using the Advanced Precision PersonalDV Imaging system , with temperature ( 37°C ) and C02 ( 5% ) humidity control , at a size of 512 X 512 pixels and a bit depth of 16 using a 60X , 1 . 4 N . A . oil objective lens . Images were deconvoluted using the DeltaVision deconvolution system and analyzed using ImageJ . Purified human pDC ( 50 , 000 cells/50uL ) were stimulated in Fibronectin ( Corning ) coated Ibidi imaging chambers μ-Slide VI0 . 4 ( Ibidi , Madison , WI ) in 5% PHS culture media ( 200uL ) with GFP PR8 influenza 10 HAU , GFP HA-HIV ( 10HAU/500ng p24 ) or GFP JRFL pseudotyped HIV 500ng p24 and stained with 1uM lysotracker ( Invitrogen ) . Live imaging was carried out after 30 minutes , 2–4 hours , or after overnight stimulation of cells , using a Yokogawa CSU-X1 spinning disk mounted on a Zeiss AxioObserver Z1 and controlled by MetaMorph under conditions that were reproduced across all experiments . 488nm and 561nm laser lines were generated by a Prairie Technologies Aurora solid state laser and fluorescence and brightfield ( phase ) images were captured with a Hamamatsu EM-CCD C9100 digital camera set at an EM gain of 11 MHz , a size of 512 X 512 pixels and a bit depth of 16 using a 63X , 1 . 4 N . A . oil objective lens . Temperature ( 37°C ) , C02 ( 5% ) , and humidity were controlled using a Tokai Hit incubator . CD4 expressing HEK cells were incubated ( 200 , 000 cells/200 μl ) with purified NA/LE mouse anti-human CD4 ( 10 μg/ml; BD Pharmigen ) or isotype control purified NA/LE mouse IgG1κ for 30 minutes . HIV Gag-iGFP was then added , and cells were placed back in culture for 18 hours . Cells were washed , fixed with 4% PFA , and analyzed by FACS . Purified human pDC ( 100 , 000 ) were incubated for 18 hours with GFP PR8 influenza 10 HAU , GFP HA-HIV ( 10HAU/500ng p24 ) or GFP JRFL pseudotyped HIV ( HIV Gag-iGFP JRFL ) 500ng p24 in 8 chamber polystyrene vessel tissue culture treated glass slides ( CultureSlides BD Falcon ) in 10% FBS ( Gibco ) culture media ( 200uL ) then cells were washed x2 , fixed with 4% paraformaldehyde , permeabilized with 0 . 1% Triton X-100 , blocked with 0 . 5% BSA in PBS , stained with mouse anti-EEA1 ( 0 . 5 ug; BD biosciences ) at 4°C overnight , washed x 2 then stained with donkey anti-mouse TRITC ( Jackson Immunoresearch ) for one hour at RT , then washed , dried , and mounted in DAPI Anti-fade GOLD ( Vector labs ) . Alternatively pDCs were incubated with media , phorbol 12-myristate 13-acetate 100ng/mL ( PMA ) , HIV , or PMA and HIV for 4 hours then cells were washed x2 , fixed with 4% paraformaldehyde , permeabilized with 0 . 1% Triton X-100 , blocked with 0 . 5% BSA in PBS , and then stained with CD4 PE ( BD Pharmigen ) for one hour , washed x2 , dried , and mounted in DAPI Anti-fade GOLD ( Vector labs ) . HEK cells transduced with CD4 mutants ( 25 , 000 ) were incubated overnight in 8 chamber polystyrene vessel tissue culture treated glass slides ( CultureSlides BD Falcon ) in 10% FBS ( Gibco ) culture media ( 200uL ) . iGFP HIV 500ng was added to culture media for 2–4 hours then cells were washed x2 , fixed with 4% paraformaldehyde , permeabilized with 0 . 1% Triton X-100 , blocked with 0 . 5% BSA in PBS , and then stained with CD4 PE ( BD Pharmigen ) for one hour . Alternatively , cells were stained with mouse anti-EEA1 ( 0 . 5 ug; BD biosciences ) , mouse anti-transferrin receptor ( 0 . 5ug; Invitrogen ) , mouse anti LAMP-1 H4A3 ( 0 . 5ug; Developmental Studies Hybridoma Bank University of Iowa ) at 4°C overnight , washed x 2 then stained with donkey anti-mouse TRITC ( Jackson Immunoresearch ) for one hour at RT , then washed , dried , and mounted in DAPI Anti-fade GOLD ( Vector labs ) . Fixed cells were imaged using a Zeiss LSM 880 confocal microscope configured to generate laser lines at 488nm and 561nm , as well as transmitted light . Images were scanned using a 100X , 1 . 46 N . A . oil objective lens at a size of 1024 X 1024 pixels and a bit depth of 12 . Images were captured using sequential ( multitrack ) acquisition to avoid bleedthrough of signal between channels . Colocalization of fluorescently labeled ligands ( green ) with endosomal markers or CD4 receptor ( red ) were analyzed quantitatively using JACoP plugin on ImageJ software or MetaMorph colocalization analysis . Image pair channel files of single image mid-planes were opened in as separate 16-bit ( grey scale ) image files . Individual cell regions were identified using the corresponding brightfield image , images were thresholded , and colocalization analysis was performed using the MetaMorph “measure colocalization” application , which measures the localization correlation between corresponding pixels in two paired images and provides the Manders correlation [66] . 50–100 cells were analyzed to generate the individual coefficients and data was plotted using GraphPad Prism software . CD4-expressing HEK-Blue hTLR cell lines were generated by infecting 3 . 0x106 HEK-Blue hTLR cells in a 10cm2 dish with CD4 lentivirus transfection supernatants . Stable CD4 expression of cell lines ( WT , STOP , DEC , LAMP , SA , and SE ) were cultured in the presence of puromycin and CD4 expression was verified by FACS and microscopy . 200 , 000 cells were placed in HEK-blue culture media overnight ( 200uL ) with CpGB , HIV , or R848 and NF-KB activation was measured using the HEK-blue detection system , with blue color depicting NF-κB activation , as measured by ELISA according to manufacturer specification ( Invivogen ) . Statistical significance was determined by the unpaired Student’s t test and analysis of variance .
|
Plasmacytoid dendritic cells ( pDC ) are innate immune cells that are specialized to produce type I interferon ( IFN ) and to activate adaptive immune responses . Although IFN is an anti-viral cytokine , it may contribute more to pathogenesis than to protection during chronic viral infections , including chronic HIV infection . pDC sense HIV to produce abundant IFN but minimal NF- κB–dependent production of TNFα and minimal up-regulation of co-stimulatory molecules , suggesting that HIV promotes pDC to become interferon producing cells ( IPC ) rather than antigen presenting cells ( APC ) . Here , we use florescent HIV virions pseudotyped with influenza hemagglutinin ( HA ) envelope and a cell system expressing CD4 molecules with modified intracellular trafficking . We found that HIV virions pseudotyped with HA stimulate pDC to mature , similar to influenza-stimulated pDC , and traffic intracellularly similarly to influenza . We also find that CD4-mediated intracellular trafficking guides HIV trafficking and downstream signaling . Our study presents new and important findings which demonstrate that divergent HIV sensing by pDC to produce IFN , rather than to become mature antigen presenting cells , is mediated specifically by CD4-HIV envelope interactions .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"lysosomes",
"vesicles",
"pathology",
"and",
"laboratory",
"medicine",
"hek",
"293",
"cells",
"influenza",
"pathogens",
"biological",
"cultures",
"microbiology",
"retroviruses",
"viruses",
"immunodeficiency",
"viruses",
"immune",
"receptor",
"signaling",
"rna",
"viruses",
"membrane",
"receptor",
"signaling",
"cellular",
"structures",
"and",
"organelles",
"endosomes",
"research",
"and",
"analysis",
"methods",
"specimen",
"preparation",
"and",
"treatment",
"infectious",
"diseases",
"staining",
"proteins",
"medical",
"microbiology",
"hiv",
"microbial",
"pathogens",
"cell",
"lines",
"biochemistry",
"signal",
"transduction",
"cell",
"staining",
"cell",
"biology",
"viral",
"pathogens",
"interferons",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"lentivirus",
"cell",
"signaling",
"organisms"
] |
2016
|
CD4 Receptor is a Key Determinant of Divergent HIV-1 Sensing by Plasmacytoid Dendritic Cells
|
Rabies in China remains a public health problem . In 2014 , nearly one thousand rabies-related deaths were reported while rabies geographic distribution has expanded for the recent years . This report used surveillance data to describe the epidemiological characteristics of human rabies in China including determining high-risk areas and seasonality to support national rabies prevention and control activities . We analyzed the incidence and distribution of human rabies cases in mainland China using notifiable surveillance data from 1960–2014 , which includes a detailed analysis of the recent years from 2004 to 2014 . From 1960 to 2014 , 120 , 913 human rabies cases were reported in mainland China . The highest number was recorded in 1981 ( 0 . 7/100 , 000; 7037 cases ) , and in 2007 ( 0 . 3/100 , 000; 3300 cases ) . A clear seasonal pattern has been observed with a peak in August ( 11 . 0% of total cases ) , Human rabies cases were reported in all provinces with a yearly average of 2198 from 1960 to 2014 in China , while the east and south regions were more seriously affected compared with other regions . From2004 to 2014 , although the number of cases decreased by 65 . 2% since 2004 from 2651 to 924 cases , reported areas has paradoxically expanded from 162 prefectures to 200 prefectures and from southern to the central and northern provinces of China . Farmers accounted most of the cases ( 65 . 0% ) ; 50–59 age group accounted for the highest proportion ( 20 . 5% ) , and cases are predominantly males with a male-to-female ratio of 2 . 4:1 on average . Despite the overall steady decline of cases since the peak in 2007 , the occurrence of cases in new areas and the spread trend were obvious in China in recent years . Further investigations and efforts are warranted in the areas have high rabies incidence to control rabies by interrupting transmission from dogs to humans and in the dog population . Furthermore , elimination of rabies should be eventually the ultimate goal for China .
Rabies is a zoonotic disease caused by viruses of genus Lyssavirus , it’s also a vaccine-preventable viral disease which occurs in more than 150 countries and territories . Although a number of carnivore and bat species serve as natural reservoirs , worldwide rabies in dogs is the source more than 95% of human infections [1–3] . A recent study estimates that globally canine rabies causes approximately 59 , 000 ( 95% Confidence Intervals: 25–159 , 000 ) human deaths [4] , As WHO reported , most cases occur in Asia and Africa [5 , 6] . In Asia , rabies has been controlled or eliminated for decades in Malaysia , Japan and many island countries or regions [7] . However , India is reported to have the highest incidence of rabies globally [6 , 8] , and China is a high-risk environment for rabies , with human rabies cases second only to India [7 , 9] . Here we describe the rabies surveillance system and characterize the epidemiology of the disease in China . We focused on the recent years , to identify high-risk areas and their specificities to help plan resource allocation for rabies interventions .
The Chinese Notifiable Disease Reporting System ( NDRS ) which was initiated in 1950s is the fundamental communicable disease surveillance system in China [10] . In the 1950s to the middle of 1980s , case numbers of communicable diseases in counties were monthly reported by mail through municipal and provincial health agencies until they finally reach the Ministry of Health taking a rather long time of 30 to 40 days . From the middle of 1980s to the end of 2003 , although computers started to be used in some areas to collect and aggregate data of infectious disease , the whole country’s data still had to go through all the levels to reach the MOH which caused a time lag in reporting , making it hard for the early detection of outbreaks . After SARS outbreak in 2003 , the Chinese government strengthened the construction of public health information system . On January 1st , 2004 , the Real-time Notifiable Infectious Disease Reporting System was put into use nationwide , realizing the timely online monitoring of individual cases which marks a leap in the surveillance of communicable diseases in China . WHO classified human rabies cases to suspected , probable and confirmed cases . A suspected rabies case-patient was defined as a case that is compatible with a clinical case definition; a probable rabies case-patient was defined as a suspected case plus a reliable history of contact with a suspected rabid animal; a confirmed case-patient was defined as a suspected or probable case that is laboratory-confirmed [6] . In China , from beginning , Rabies cases were diagnosed according to the unified diagnostic criteria issued by Chinese Ministry of Health , this diagnostic criteria was modified in 2008 . Before modification , a probable rabies case-patient was defined as a patient licked , bit or scratched by dog , cat or other mammals with clinical symptoms of prickling or itching sensation at the site of bite , progressing within days to agitation , anxiety , confusion , hydrophobia , aerophobia , and paralysis of muscles or cranial nerves . A confirmed case-patient was defined as a probable rabies patient with laboratory evidence of rabies virus infection detected by direct fluorescent antibody test ( DFA ) , or by virus isolation testing of clinical specimens . After modification , clinical symptoms of case-patient were classified into two forms according to the case definition of World Health Organization ( WHO ) , furious rabies and paralytic rabies . Symptoms of furious rabies were similar to the definition before modified , while paralytic rabies without exhibit signs of hyperactivity or hydrophobia , starting at the site of the bite or scratch , muscles gradually become paralyzed , progressing with the systemic flaccid paralysis . With epidemiological history , patient either of furious rabies or paralytic rabies was defined as a probable case-patient . A confirmed case-patient was defined as a probable rabies patient with laboratory evidence of rabies virus infection detected by direct fluorescent antibody test ( DFA ) , reverse-transcriptase polymerase chain reaction ( RT-PCR ) , or by virus isolation testing of clinical specimens . Rabies was on the initial list of notifiable diseases in mainland China from 1950s . All probable rabies cases and those laboratory-confirmed were required to be mandatory reported when those patients sought medical consultations . Appropriate clinical specimens , including saliva , cerebrospinal fluid ( CSF ) , urine , nuchal skin biopsies , or brain tissues post mortem , were collected upon the patient’s syndromes by nurses and shipped immediately for virus testing following standardized procedures . Specimens were sent to the local or provincial level CDCs and there they were tested for rabies virus by DFA , RT-PCR , or virus isolation using the protocols and kits released by the China CDC . Only very small portion of specimens were sent to the China CDC tested by virus isolation for difficulty in testing in province CDC or further characterization of the results . From 1950 to 2003 , number of cases and deaths by province were reported monthly to the China CDC , age and sex aggregate data was added from 1988 . After 2004 , each probable and confirmed case-patient with individual data was required to be reported from all health care facilities nationwide to the China CDC by clinicians within 24 hours after diagnosis using a standardized form to collect data about demographic information ( gender , date of birth , and location ) , residence type ( rural or urban ) , case classification ( probable or confirmed case ) , date of onset and , if applicable , date of death; all data were reported electronically online to the China CDC . Descriptive statistics included frequency analysis for categorical variables , means and standard deviations for normal distributions , or medians and inter quartile ranges ( IQR ) for continuous variables . As the quality of data from 1950 to 1959 was not very stable , we calculated temporal trends of cases and deaths from 1960 to 2014 , geographic distribution of total cases was shown and patterns of seasonal distribution , sex , age and occupation was counted . Annual population denominators during the study period were obtained from National Bureau of Statistics of China . Statistical analysis was performed using SPSS ( v17 . 0 , SPSS Inc , Chicago , IL , USA ) . ArcGIS 10 . 0 ( ESRI , Redlands , CA , USA ) was used to assess the geographical distribution of cases across all the provinces in mainland China . For all analyses , probabilities were 2-tailed and a p-value of <0 . 05 was considered statistically significant . It was determined by the National Health and Family Planning Commission , China , that the collection of data from human rabies cases was part of a continuing public health surveillance of a notifiable infectious and was exempt from institutional review board assessment .
From 1960 to 2014 , China reported 120 , 913 human rabies cases with a yearly average of 2198 , including 23 , 932 ( 19 . 8% ) cases were reported during 2004–2014 ( Fig 1 ) . The highest incidence was observed in 1981 ( 0 . 71/100 , 000 , 7037 cases ) . When comprehensive control measures were conducted , such as management of stray dogs , vaccination of dogs , PEP of the exposures , the incidence rate decreased from the 1980s to the 90s and reached the lowest point in 1996 ( 0 . 01/100 , 000 , 159 cases ) . However , the incidence rate increased again in 1997 to reach a second peak in 2007 ( 0 . 25/100 , 000 , 3300 cases ) . Since then the incidence rate has gradually decreased to reach 0 . 07/100 , 000 in 2014 ( n = 924 cases ) . Historically—during 1960–2014 –all provinces reported human rabies cases were reported in all provinces with a predominance in the eastern and southern regions . Before 1996 ( the lowest reporting year; n = 159 cases ) , human rabies were mainly concentrated in southern , central and northeast of China , while the northeast area maintained at a relatively low reporting level . From 2004 to 2014 , cases were mainly prevalent in southern regions , namely in Guangxi , Guizhou , Guangdong and Hunan provinces which accounted for 52% of the total cases . However , cases in these provinces decreased after 2007 , while in the north , rabies cases still increased in recent years , such as Shanxi , Inner Mongolia and Beijing by 4800% , 600% and 600% . For Yunnan and Shaanxi , increased in 2011 and decreased in 2014 . In 2004 , 2 , 651 cases were reported in 162 prefectures mainly in the eastern and southern regions while 924 ( -65 . 2% ) cases were reported in 200 ( +23 . 5% ) prefectures in 2014 ( Prefecture is administrative subdivision of provincial-level division , which is a level between the provincial and the county level in China ) . Although the number of cases decreased in the past recent years , the number of reported areas expanded paradoxically with a spread into the northern and western regions of China . For example , Shanxi province had only had one prefecture that reported only one case in 2004 while—in 2014 –there were 8 prefectures that reported cases ( n = 43 ) . Hebei has 4 prefectures reported 7 cases in 2004 while 10 prefectures reported 56 cases . In the northeast region and northwest region , the number of reported prefectures were both very low , but Shaanxi had a large rise after 2008 , no case were reported in 2008 and 3 prefectures reported 20 cases in 2014 and reported 26 cases in Hanzhong and Shangluo prefectures peaking in 2009 . However , south and central regions still remained high rabies incidence for a number of years ( Fig 2 ) . When focusing on prefectures with relatively high number of reported cases—during 2010–2014 –we identified 15 prefectures that accumulatively reported more than 100 cases ( Fig 3 ) . 15 prefectures were located in 5 provinces , Guangxi ( 6 prefectures ) , Guizhou and Guangdong ( 3 prefectures each ) , Yunnan ( 2 prefectures ) and Shanxi ( 1 prefecture ) . These provinces were all concentrated in southern and central areas of China . The top 3 prefectures were Qingyuan , Maoming and Guigang . All these provinces and prefectures were in southern China . In addition , of these 15 prefectures , the number of onset decreased in 12 by 53 . 85% on average when comparing 2004–2009 with that of 2010–2014 ( highest 70 . 1% in Qianxinan , in Guizhou province ) . In contrast cases increased in Honghe ( average of 13 cases , 144 . 4% ) and Wenshan ( 21 cases , 122% ) , Yunnan province , and in Linfen ( 21 cases , 2100% ) from Shanxi province . Since 1960 , Cases occurred throughout the year; however , most cases ( 59% of total cases ) were reported from June to November and the numbers peaked usually in August ( 11 . 0% of total cases ) ( Fig 4A ) . During these 55 years , August maintained the peak month in 29 years , followed by September ( 9 years ) and October ( 8 years ) . Following the peak year of 2007 , as the reported number of cases steadily decreased year by year the seasonal pattern was less apparent especially from 2012 to 2014 ( Fig 4B ) . The range of cases in the highest and lowest month in 2007 and 2014 was 208 and 61 , respectively . Demographic data has only been available since 1988 . During 1988–2014 , the overall male-to-female ratio of 2 . 4:1; this ratio was similar in different years ( P>0 . 05 ) . Overall , the median age was 46 years ( IQR: 23–59 ) ; females were on average older than males ( 49 years vs 46 years of age ) . From 1988 to 2003 , 0–9 age group reported the highest proportion of the total cases ( 24 . 4% ) , then followed by 10–19 age group ( 17 . 2% ) , and after 70yrs group ( 3 . 3% ) accounted smallest proportion . However , from 2004 to 2014 , 50–59 age group accounted for the highest proportion ( 20 . 5% ) , then followed by 40–49 age group ( 16 . 1% ) , while 20–29 age group ( 4 . 9% ) had the lowest proportion as shown in Fig 5A and 5B . Farmers accounted for 65 . 0% of the total cases , and then followed by students and children ( 24 . 1% ) . This pattern has been consistent year by year since 1992 in China . ( Fig 5C ) . Rural cases accounted for 78 . 7% of all cases since 2004 , this proportion was lowest of 75 . 2% in 2004 and highest of 80 . 9% in 2005 , no statistical difference between different years ( P>0 . 05 ) . The proportion of lab-confirmed cases was very low in China , and average of 1 . 2% cases were tested from 2004 to 2014 , however , this proportion increased in recent years ( P<0 . 01 ) ( Fig 5D ) . The median time of onset-to-death interval of rabies cases was 3 days , the median time of onset-to diagnosis interval was 2 days , the median of diagnosis-to-death interval was 1 day , we analyzed the three median by year , and the data maintained constant in every year from 2004 .
Rabies is notifiable in animals and humans in China , despite the overall steady decline of cases since the peak in 2007 , the occurrence of cases in new areas and the spread trend were obvious in China in recent years . Further investigations and efforts are warranted in the high spot area to control rabies by interrupting transmission from dogs to humans and in the dog population . Elimination of rabies should be eventually the ultimate goal for China .
|
China is a high-risk environment for rabies , with human rabies cases second only to India globally . This paper reviews 55 years of rabies epidemiology in mainland China , and detailed analysis of data in recent years . In this study , notifiable surveillance data were analyzed and found that rabies still remains a serious public health problem in China , the east and south regions were more seriously affected compared with other regions , however , the occurrence of cases in new areas and the spread trend were obvious in China in recent years . Moreover , males in rural areas had higher risk of infection than residents in urban areas , a clear seasonal pattern has been observed with a peak in August . These findings indicated a clear need to increase government and public consciousness with regard to the potential risk of rabies and the means of avoiding the disease . Further efforts should be strengthen specially in the high spot areas .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"china",
"pathogens",
"immunology",
"tropical",
"diseases",
"geographical",
"locations",
"microbiology",
"vertebrates",
"animals",
"mammals",
"dogs",
"viruses",
"preventive",
"medicine",
"age",
"groups",
"rabies",
"rna",
"viruses",
"neglected",
"tropical",
"diseases",
"infectious",
"disease",
"control",
"vaccination",
"and",
"immunization",
"rabies",
"virus",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"zoonoses",
"medical",
"microbiology",
"epidemiology",
"microbial",
"pathogens",
"people",
"and",
"places",
"infectious",
"disease",
"surveillance",
"lyssavirus",
"asia",
"viral",
"pathogens",
"disease",
"surveillance",
"biology",
"and",
"life",
"sciences",
"population",
"groupings",
"viral",
"diseases",
"amniotes",
"organisms"
] |
2016
|
Human Rabies in China, 1960-2014: A Descriptive Epidemiological Study
|
Chikungunya virus causes mosquito-transmitted infection that leads to extensive morbidity affecting substantial quality of life . Disease associated morbidity , quality of life , and financial loss are seldom reported in resources limited countries , such as Bangladesh . We reported the acute clinical profile , quality of life and consequent economic burden of the affected individuals in the recent chikungunya outbreak ( May to September 2017 ) in Dhaka city , Bangladesh . We conducted a cross-sectional study during the peak of chikungunya outbreak ( July 24 to August 5 , 2017 ) to document the clinical profiles of confirmed cases ( laboratory test positive ) and probable cases diagnosed by medical practitioners . Data related to clinical symptoms , treatment cost , loss of productivity due to missing work days , and quality of life during their first two-weeks of symptom onset were collected via face to face interview using a structured questionnaire . World Health Organization endorsed questionnaire was used to assess the quality of life . A total of 1 , 326 chikungunya cases were investigated . Multivariate analysis of major clinical variables showed no statistically significant differences between confirmed and probable cases . All the patients reported joint pain and fever . Other more frequently reported symptoms include headache , loss of appetite , rash , myalgia , and itching . Arthralgia was polyarticular in 56 . 3% of the patients . Notably , more than 70% patients reported joint pain as the first presenting symptom . About 83% of the patients reported low to very low overall quality of life . Nearly 30% of the patients lost more than 10 days of productivity due to severe arthropathy . This study represents one of the largest samples studied so far around the world describing the clinical profile of chikungunya infection . Our findings would contribute to establish an effective syndromic surveillance system for early detection and timely public health intervention of future chikungunya outbreaks in resource-limited settings like Bangladesh .
Chikungunya is a mosquito-borne ( Aedes species ) , self-limiting , febrile illness with severe debilitating arthropathy caused by chikungunya virus ( CHIKV ) . This virus was first identified during an epidemic of febrile polyarthralgia in Tanzania in 1953 [1] . Since then , CHIKV has been reported to cause several large-scale outbreaks in Africa , India , Southeast Asia , Western Pacific and Americas [2–4] . Before 2000 , chikungunya outbreaks were mostly sporadic and limited . But thereafter , the virus has been frequently causing severe forms of epidemics imposing heavy economic burden and productivity loss [5 , 6] . The CHIKV outbreaks are characterized by a sudden disappearance for a considerably long period of time from a particular geographic area before re-emergence . As such , chikungunya has re-emerged in devastating form of epidemics in and around the Indian Ocean in 2005 , nearly after 30 years of quiescence [7] . In Bangladesh , the first recognized outbreak of chikungunya was reported in 2008 in two villages in the northwest part of the country adjacent to Indian border [8] . Two small-scale outbreaks were documented in rural communities in 2011 [9] and 2012 [10] . Dhaka , the capital of Bangladesh , is one of the most densely populated cities in the world with approximately 18 million inhabitants [11] , have experienced a large-scale chikungunya outbreak in 2017 . Mainstream local media outlets have extensively covered the epidemic [12 , 13] . Despite a lower middle-income country , Bangladesh has achieved remarkable progress in reducing maternal and child mortality by improving primary healthcare access [14] . However , country’s overall public healthcare facilities are not adequate . Approximately 70% patients prefer private clinics/hospitals for seeking medical advice [15 , 16] . It allocates only US$31 per capita on health expenditure [17] . Thus , continuous surveillance of any unprecedented disease like chikungunya is a huge challenge in Bangladesh since most public hospitals lack modern diagnostic facilities and medical documentation system [18] . Therefore , data from public medical facilities alone cannot provide an overall picture of an outbreak resulting in underreporting of cases . A state-run institute named Institute of Epidemiology , Disease Control and Research ( IEDCR ) has monitored the recent chikungunya outbreak based on RT-PCR data sourced from three diagnostic laboratories including the institute itself and two other private facilities . According to IEDCR’s own data on 1003 RT-PCR confirmed cases , the peak of the outbreak was between early May and end of July 2017 [19] . Also , according to their newsletter , about 87 . 3% ( 12060 out of 13814 cases ) of suspected chikungunya patients visited three public hospitals ( DMCH , ShMCH , SSMCMH ) in addition to IEDCR for treatment . Taking other public hospitals and more than 2000 registered private clinics/hospitals in the Dhaka metropolitan are into account [20 , 21] , arguably , there was a widespread chikungunya outbreak in the city . CHIKV infection has emerged as a major public health concern since it often affects a large proportion of the population within an outbreak area and causes considerable pain , distress , and anxiety as well as significant economic burden due to severe clinical manifestations [22–25] . CHIKV infection is usually non-fatal and the sign of clinical symptoms resolves over time . The clinical severity of this disease is associated with reduced quality of life ( QoL ) . It is found that QoL drops dramatically when patients are severely affected , especially during the early phase of chikungunya infection lasting for one or two weeks , while prolonged musculoskeletal manifestations ( chronic arthralgia ) may last for months to years [23] . A study , for instance , has observed a 20-fold reduction of QoL in the non-recovered CHIKV positive group while 5-folds reduction was in recovered group as compared to healthy control group [26] . Poor QoL resulting from CHIKV is linked to immediate and long-term economic burden [24 , 27 , 28] . Due to lack of health insurance system and other socio-economic factors , it is challenging to determine the actual economic burden of chikungunya outbreaks in developing countries . It appears that the actual economic loss is often overlooked because of non-fatality of the disease . CHIKV outbreaks have caused significant financial burdens on society level in India and across Latin America [24 , 29–31] . The impact on household finance was substantial , particularly during acute phase of infection when QoL was found to be at the lowest level . The poor segment of the population generally bears the economic burden most . Consequently , it may aggravate poverty . Long-term arthralgia secondary to CHIKV has also reported to cause significant economic impact [24] . A higher health care utilization has been reported for patients up to six years after acute infection of CHIKV [27] . In this study , we are interested to find out the answers for two primary research questions: a ) the impact of clinical severity on QoL during the acute phase , and b ) the impact of treatment cost on the economic conditions of CHIKV patients during acute phase .
In our study , we investigated patients who experienced typical clinical symptoms of CHIKV infection [32] ( febrile illness with arthralgia/arthritis ) during the peak of recent Dhaka-outbreak ( May-July , 2017 ) . Patients with a positive RT-PCR or serological test ( CHIKV positive but not dengue ) were categorized as confirmed cases . For probable case , we followed the recommendation of the World Health Organization ( WHO ) [33] and standardized European case definition [34] where during an established outbreak , a patient meeting both clinical and epidemiological criteria are considered as probable case . Thus , patients diagnosed by medical practitioners based on characteristic clinical features of CHIKV infections without a RT-PCR or serological tests were documented as probable cases . As discussed earlier , because of the magnitude of the outbreak , the local health authority ( IEDCR ) monitored the outbreak and released health bulletin on a daily basis . There was no sign or warning of any other arboviruses related outbreaks ( such as dengue ) by IEDCR . Unlike developed countries , our healthcare system lacks a functional referral system [18] and organized medical record-keeping system . Therefore , we reached out to every enrolled patient to verify the documents of CHIKV diagnosis . Scanned copies of the laboratory test results were collected from the confirmed cases as a proof . Patients claimed to have CHIKV infection without physicians’ confirmation ( clinical and or laboratory ) were excluded from the study . It is important to note that molecular and serological diagnostic tests for CHIKV infection were available only in a handful of diagnostic facilities ( mostly private ) . The cost of diagnostic test was relatively high considering the average income of city dwellers . In this context , given the insignificant mortality related to CHIKV , the government health regulatory authority ( DGHS , Bangladesh ) discouraged ( non-obligatory instruction ) suspected patients for laboratory tests in order to turn down the panic among city dwellers [35] . As a result , we had more probable cases than confirmed cases . A cross-sectional study was conducted between July 24 and August 5 , 2017 , to investigate the clinical profiles , economic burden , and quality of life of chikungunya affected individuals . Biomedical Research Foundation ( BRF ) , Bangladesh took the initiative to study the impact of this widespread outbreak . The study was conducted by a team of 111 volunteer researchers comprising of clinicians , public health professionals , statisticians as well as undergraduate and postgraduate students . Considering the urgency of the issue and time-sensitive nature of the outbreak , we relied entirely on voluntary services rather than financial support from external sources . Participants were conveniently selected by the members of the research team from within their known circles ( e . g . , friends , friends of friends ) living in Dhaka city , where the outbreak has occurred . Data were collected via face-to-face interview using a structured questionnaire . Enumerators were given one-day training on the use of the questionnaire for face-to-face interviews . Respondents were asked about clinical symptoms , loss of productivity due to missing work hours , and quality of life during their first two-weeks of symptom onset . Data on economic variables were collected from only the earning members of the chikungunya affected families . Descriptive and inferential statistical procedures were used to ascertain the clinical profile of CHIKV infection and its impact on quality of life and economic well-being . Descriptive statistics were reported as percentage and means when applicable along with standard deviation . The intensity of joint pain was evaluated by a 10-point numerical rating scale ( NRS ) , which was categorized as mild ( score between 1 and 3 ) , moderate ( score between 4 and 6 ) and severe ( score between 7 and 10 ) in the analytic stage [36] . To determine the impact of chikungunya infection on quality of life ( QoL ) , World Health Organization ( WHO ) endorsed quality of life questionnaire ( brief version ) , known as WHOQOL-BREF , was used . We used the validated Bengali version of the questionnaire [37] . The responses from WHOQOL-BREF questionnaire were analyzed as per the recommendation and scoring guidelines [38] . Cronbach's alpha coefficient was calculated to check the internal consistency of scores . For assessing the impact of CHIKV on the economic well-being of a family , a 10-point rating scale questions were used to capture granular responses . Missing cases were excluded from bivariate analysis . Completed data collection forms were scrutinized by data collection supervisors . Follow up calls were made within two days of data collection to 10% randomly selected participants to ensure the authenticity of the responses . Verified data were entered and subsequently managed using REDCap electronic data capture tool hosted at BRF [39] . Data were analyzed using R statistical software . For testing association between categorical data , Pearson’s chi-square test was used , and Yate’s correction for continuity was applied where appropriate . We performed independent sample t-test when comparing means of continuous variables . A two-tailed p-value smaller than 0 . 05 was considered statistically significant . The study was approved by the Ethical Review Committee ( ERC ) of Bangladesh University of Health Sciences ( Memo no: BUHS/BIO/EA/17/077 ) . As approved by ERC , verbal and/or written informed consent was obtained from every participant as per their convenience . Trained enumerators first approached and explained consent form to the prospective participants and study questionnaire was shared or discussed with them . After obtaining consent ( oral and/or written ) , participants were registered for face-to-face interview . Some respondents could not sign their names , in which case the questionnaire was marked indicating a case with verbal consent . All adult participants 18 years or above provided informed consent . Although children ( 18 years and below ) were included in the survey , no child was interviewed . Parents or guardians provided information about children’s clinical symptoms . Children were also not included in the economic wellbeing and quality of life section of the study .
A total of 1 , 474 patients were enrolled in our study . After rigorous verification and cross-checking , 148 cases ( 10% ) were discarded due to the incompleteness of data . Finally , of all 1326 verified cases , 18% ( 239/1 , 326 ) constituted confirmed cases ( 214 cases were serologically confirmed , while 25 cases by RT-PCR ) and 82% ( 1087/1326 ) were probable cases . The geospatial distribution of 855 patients ( who provided address ) out of 1326 showed that our study represented most of the administrative zones ( 20 out of 25 ) of Dhaka City ( Fig 1 ) . The mean age of the participants was 33 . 74 years ( SD = 14 . 83 ) and male to female ratio was 1 . 33:1 . Children ( <15 years of age ) represented 6 . 4% of the study subjects , 42 . 5% were adolescents and young adults ( AYA , 15–29 years ) , 44 . 3% were adults ( 30–59 years ) and 6 . 8% of the cases aged over 60 years ( Table 1 ) . About 43 . 7% of the participants were graduates , 33 . 3% completed high school and 4 . 9% had no education . A total of 339 ( 25 . 6% ) participants had comorbidities ( Table 1 ) . Seventy-six patients ( 5 . 7% ) were hospitalized with chikungunya infection , of which 35 were confirmed cases . Over 90% of the respondents reported high-grade fever with an average maximum temperature of 103 . 6°F ( SD = 0 . 89 ) . The mean duration of fever was 4 . 88 days ( SD = 2 . 7 ) . As the first clinical symptom , 74 . 6% ( n = 1326 ) of the respondents experienced pain ( joint and/or muscle pain ) prior to fever ( Table 2 ) . This unique clinical feature was consistent irrespective of age and sex of the patients . About 85% of the patients ( both confirmed and probable cases ) complained severe pain with a mean pain score of 8 . 3 ( out of 10 ) ( SD = 1 . 63 ) , and about 65% patients suffered more than 10 days during the acute phase ( Table 3 ) . Arthralgia was oligoarticular ( 2–4 joints ) in 40 . 1% , and polyarticular ( >5 joints ) in 56 . 3% of the patients . Peripheral small joints were the most common site of involvement . The joint pain was symmetrical in 64 . 8% of the patients ( Table 3 ) . Joint swelling and skin rash were significantly higher among confirmed cases ( 62 . 6% and 78 . 2% respectively ) . Other common symptoms reported were redness of eyes ( over 56 . 5% ) , nausea ( 60% ) , oral ulcer ( 31 . 2% ) , diarrhea ( 25% ) , and edema ( 18% ) ( S1 Table ) . Bivariate analysis showed that the severity of some clinical symptoms was gender ( S1A Fig ) and age group specific ( S1B–S1E Fig ) . For other clinical symptoms analyzed for Dhaka outbreak , see S1 Table and S2 Table . Multivariate analysis of 20 major clinical variables between confirmed and probable cases showed no statistically significant differences between these two patient categories except for rash and swollen joint ( S3 Table ) . This finding suggests that physicians were able to diagnose chikungunya cases effectively based only on typical clinical features during the outbreak . We estimated the overall treatment cost ( including consultation fee , cost of laboratory tests , medicine , transport and special food ) during the acute phase of chikungunya . We found that confirmed cases had to spend around BDT 8 , 192 ( SD = 12 , 127 . 63 ) on an average compared to BDT 2 , 122 ( SD = 3 , 422 ) for probable cases which are equivalent to $99 . 3 ( SD = 147 ) compared to $26 ( SD = 41 . 5 ) , respectively . Approximately 70% ( n = 1 , 302 ) of the patients lost more than 7 productive days while 29 . 6% of them lost more than 10 days in the acute phase of the disease . Our analysis relying on 424 family heads suffering from chikungunya showed that economic impact was most prominent in low-income ( < 10 , 000 BDT ) categories ( Fig 2 , S4 Table ) . We assessed the quality of life ( QoL ) of 1 , 216 respondents who were 18 years and older . The Cronbach's alpha of WHOQOL-BREF was adequate ( 0 . 89 ) for all 26 questions . The average score was highest in the environmental health domain ( mean = 11 . 43 , SD = 2 . 52 ) followed by psychological domain ( mean = 10 . 03 , SD = 2 . 75 ) , social relationship domain ( mean = 10 . 02 , SD = 2 . 94 ) , and physical domain ( mean = 8 . 32 , SD = 2 . 33 ) ( S5 Table ) . Pearson’s correlation coefficient showed a positive linear relationship between Q1 ( represents overall QoL of WHOQOL-BREF ) and all four domains separately ( S6 Table ) . The strongest correlation was found between Q1 and the physical domain ( r = 0 . 46 ) , followed by the psychological domain ( r = 0 . 36 ) . Overall 83 . 2% patients responded ‘very low’ or ‘low’ on Q1 indicating a catastrophic impact on the quality of life during acute-phase CHIKV infection . Average Q1 scores were significantly affected by monthly income ( p = 0 . 0032 ) , marital status ( p<0 . 0001 ) , age ( p<0 . 0001 ) , employment type ( p<0 . 0001 ) . Patients with severe arthralgia had significantly lower QoL compared to mild to moderate arthralgia ( p<0 . 0001 ) during the acute phase ( Fig 3 ) .
The first ever large-scale outbreak of chikungunya in Dhaka led to a rapid spread of infection across the city . To our knowledge , we have analyzed one of the largest samples of 1 , 326 cases to demonstrate the detail clinical profile of acute chikungunya infection . Our study also sheds light on the extent of economic impact on victim families and quality of life of the patients . Both dengue and chikungunya viruses are transmitted by the same mosquito vector and often difficult to differentiate clinically . As discussed earlier , there was no warning of dengue outbreak by health authorities during our study period . Moreover , dengue incidences have been declining since the first large-scale dengue outbreak occurred in Dhaka in 2000 [40] . A seroprevalence study showed that around 80% individuals of Dhaka city had a past history of dengue infection [41] . Furthermore , a serology-based study revealed that approximately 6 . 1% ( n = 271 ) of acute clinical cases ( <7 days ) of chikungunya was positive for dengue infection during recent chikungunya outbreak . Nearly 3 . 1% of these dengue positive patients were co-infected with CHIKV ( obtained from personal communication; Professor Md . Akram Hossain , National Institute of Preventive & Social Medicine , Dhaka , Bangladesh ) . On multivariate analysis , all clinical symptoms ( except skin rash and swollen joints ) were similar among confirmed and probable cases ( Table 2 , Table 3 and S3 Table ) . Compared to other studies conducted on self-reported [22 , 42–44] or hospitalized patients [45–47] , a higher frequency of rash ( 69 . 6% ) and swollen joints ( 52 . 1% ) were documented in the present study . In most reported outbreaks , sudden onset of fever was found to be the initial clinical symptom of chikungunya . However , joint ( and/or muscle ) pain preceded fever in more than 70% of patients in our study irrespective of age and sex ( Table 2 ) . In Kerala outbreak , joint pain was reported as the initial symptom in 17% patients , which varied among different age group reaching up to 77% in patients over 60 years [48] . This initial symptom of pain could be considered as the hallmark of chikungunya infection in Dhaka outbreak . Severe arthropathy is the most consistent clinical feature of chikungunya infection . In our study , all patients experienced joint pain . Polyarthralgia was documented in about 56 . 3% of the patients while oligoarthralgia was present in 40 . 1% cases . In the present study , ankle ( 82 . 6% ) and wrist ( 74 . 8% ) were the most affected joints . In the acute phase , the frequency of incapacitating pain involving certain peripheral joints ( ankle 82 . 6% , feet 63 . 8% , wrist 74 . 8% , fingers 73 . 2% and knee 74 . 5% ) was found to be similar with that of French soldiers cohort and another study conducted in La Reunion [22 , 44] ( S2 Table ) . In contrast , lower frequencies were reported in India and Suriname [42 , 45 , 48] . Over 85% of the patients ( n = 1 , 326 ) experienced severe pain with a median NRS score of 8 . 3 throughout the acute phase . This finding is consistent with a study carried out in La Reunion [44] . About 70% of our patients faced problem in doing routine daily activities and about 65 . 7% reported sleep disturbance due to severe arthropathy ( Table 3 ) . The actual impact of chikungunya fever on daily life activities might be much higher than reported here because of poor health literacy in developing countries . Moreover , socio-culturally , patients often fail to explain their health conditions clearly; rather express them in terms of overall satisfaction from a spiritual standpoint . What it means is that even though they are clinically suffering , people would report that they are alright . The overall severity and the extent of arthralgia related manifestations suggest an aggressive strain of chikungunya virus probably circulated in Dhaka . The sequencing of the viral strain is warranted to find out the lineage of chikungunya virus . The severity of certain clinical manifestations of chikungunya might depend on several factors including age , gender , immune status , genetic predisposition and co-morbid conditions [49] . Bivariate analysis showed that children ( <15 years ) tended to have a higher proportion of oligo-arthralgia and skin rash; while morning stiffness , severity , and duration of pain were proportionally lower among children as compared to other age groups . Joint swelling was most commonly noted in elderly patients ( 60+ years ) , while the severity of pain was highest among adults ( 30–59 years ) . In our study , the number of male participants was higher than female participants . Female patients experienced a higher incidence of skin rash , itching , joint-swelling compared to their male counterparts ( S1 Fig ) . We found that chikungunya infection caused significant loss of productivity due to absenteeism from job , household work and school . Over 95% of the respondents ( including confirmed and probable cases , n = 1 , 302 ) were mostly confined to sickbed . As a consequence , 29 . 6% of the patients lost more than 10 days of productivity during the acute phase ( Fig 2 ) . Notably , no national health insurance system exists in Bangladesh and therefore , all treatment costs were considered as out-of-pocket expenditures of the patients . Considering socio-economic conditions , a significant amount of money had to be spent for treatment purposes . We have not estimated the overall economic burden of the loss of productivity due to this chikungunya outbreak . However , we have attempted to paint a picture of the economic impact on victim families based on responses to a rating scale . Our analysis suggests that low income ( <$303 per month ) families are more likely to face significant economic pressure . Particularly , families of daily workers were the worst hit while there was no significant impact on the families in the higher income category ( >$606 per month ) . Anecdotal evidence suggests that many daily or menial workers lost their jobs due to long absenteeism . It is indispensable to estimate the overall disease burden through systematic epidemic studies to determine the real economic burden of an outbreak like this . Majority of the patients in our study reported low to very low overall quality of life . Although males had slightly higher average scores than females , the difference was not statistically significant ( p = 0 . 138 ) . QoL was significantly lower in patients with severe pain compared to those with moderate pain . Elderly patients reported lower average QoL scores compared to <60 years . In particular , patients in the highest income bracket ( BDT 50 , 000 per month; >$606 per month ) reported the lowest average overall score ( 1 . 66 , p = 0 . 003 ) . This could be due to the fact that most of the respondents in the higher income group are older . Overall QoL scores differed significantly between different job categories ( Fig 3 ) . Not surprisingly , students reported the highest average score ( 2 . 0 ) which is consistent with our findings that younger patients who are mostly students , reported higher scores compared to their older counterparts . The jobless and dependents showed an overall score of 1 . 93 . Interestingly , housewives reported higher QoL score ( 1 . 86 ) compared to those of businessmen ( 1 . 78 ) and service holders ( 1 . 70 ) . It would be of interest to explore the interplay between gender and various socio-cultural factors in terms of quality of life . Our study has some limitations . First , the participants were purposively selected using social connections . Even though it is not uncommon to adopt a non-random sampling design in such circumstances ( considering Bangladesh being one of the healthcare-resource-limited countries ) , the results should be interpreted after taking this into consideration . Second , the representation of confirmed cases was relatively low ( 18% ) . This could be explained by the limited availability of diagnostic facility along with high cost of the tests during CHIKV outbreak . Third , because of the retrospective data collection scheme , there could be recall bias . However , since our study was conducted during the very end of the peak of outbreak , it is likely that the recall bias was minimum . Lastly , it is possible that some participants may have overvalued some clinical symptoms due to massive media coverage of the outbreak . Our work represents one of the largest samples ( n = 1 , 326 ) studied so far around the world describing the clinical profile of chikungunya infection . This study demonstrates the severity of clinical symptoms during the acute phase and how it has impacted on productivity and quality of life of the affected individuals . We found joint pain prior to fever as a unique symptom in the Dhaka ( 2017 ) outbreak . A possible reason could be a novel viral strain that warrants future molecular investigations . Our findings support that during an established outbreak , CHIKV patients can effectively be identified using a set of easily recognisable clinical criteria ( i . e . syndromic approach ) without lab confirmation; an approach also suggested by others [50 , 43]for resource-constrained developing countries . We believe our study would play a pivotal role in devising an effective syndromic surveillance system for CHIKV in Bangladesh that would allow early detection of future outbreaks and timely public health interventions .
|
A major outbreak of chikungunya virus occurred for the first time in Dhaka , Bangladesh between May and September 2017 . In this study , a face-to-face interview with a structured questionnaire was conducted to collect data to investigate the clinical symptoms , quality of life , and economic aspects of 1 , 326 chikungunya patients during the first two weeks of infection . The severity of the disease was similar to previously reported severe outbreaks elsewhere but joint pain prior to fever emerged as a unique symptom in the Dhaka outbreak . This unique clinical feature was consistent across age and sex of the patients . Some clinical symptoms varied with age . For instance , a higher proportion of skin rash were found among children ( under 15 ) while morning stiffness , severity , and duration of pain were proportionally higher among other age groups . Joint swelling was most commonly noted in elderly patients ( 60+ years ) . About 83% of the patients reported low to very low overall quality of life ( QoL ) during first two weeks of chikungunya infection . Elderly patients reported lower average QoL scores compared to <60 years . Interestingly , housewives reported higher QoL score compared to those of businessmen and service holders . In particular , patients in the highest monthly income category bracket ( BDT 50 , 000 per month; >$606 per month ) reported the lowest average overall score . Nearly 95% of the patients have mostly confined to sickbed and approximately 30% of them lost more than 10 days of productivity due to severe arthropathy . Our study would contribute to establishing an effective syndromic surveillance system for early detection and timely public health intervention of future chikungunya outbreaks in resource-limited countries like Bangladesh .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"togaviruses",
"chikungunya",
"infection",
"economic",
"analysis",
"pathogens",
"tropical",
"diseases",
"microbiology",
"geographical",
"locations",
"social",
"sciences",
"alphaviruses",
"health",
"care",
"viruses",
"chikungunya",
"virus",
"rna",
"viruses",
"neglected",
"tropical",
"diseases",
"bangladesh",
"infectious",
"diseases",
"economic",
"impact",
"analysis",
"health",
"economics",
"medical",
"microbiology",
"epidemiology",
"microbial",
"pathogens",
"economics",
"people",
"and",
"places",
"quality",
"of",
"life",
"asia",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"organisms"
] |
2018
|
Chikungunya outbreak (2017) in Bangladesh: Clinical profile, economic impact and quality of life during the acute phase of the disease
|
Mycetoma is a chronic granulomatous disease . It is classified into eumycetoma caused by fungi and actinomycetoma due to filamentous actinomycetes . Mycetoma can be found in geographic areas in close proximity to the Tropic of Cancer . Mexico is one of the countries in which this disease is highly endemic . In this retrospective study we report epidemiologic , clinical and microbiologic data of mycetoma observed in the General Hospital of Mexico in a 33 year-period ( 1980 to 2013 ) . A total of 482 cases were included which were clinical and microbiology confirmed . Four hundred and forty four cases ( 92 . 11% ) were actinomycetomas and 38 cases ( 7 . 88% ) were eumycetomas . Most patients were agricultural workers; there was a male predominance with a sex ratio of 3∶1 . The mean age was 34 . 5 years old ( most ranged from 21 to 40 years ) . The main affected localization was lower and upper limbs ( 70 . 74% and 14 . 52% respectively ) . Most of the patients came from humid tropical areas ( Morelos , Guerrero and Hidalgo were the regions commonly reported ) . The main clinical presentation was as tumor-like soft tissue swelling with draining sinuses ( 97 . 1% ) . Grains were observed in all the cases . The principal causative agents for actinomycetoma were: Nocardia brasiliensis ( 78 . 21% ) and Actinomadura madurae ( 8 . 7% ) ; meanwhile , for eumycetomas: Madurella mycetomatis and Scedosporium boydii ( synonym: Pseudallescheria boydii ) were identified . This is a single-center , with long-follow up , cross-sectional study that allows determining the prevalence and characteristics of mycetoma in different regions of Mexico .
Mycetoma is a chronic granulomatous disease , associated with a progressive , inflammatory reaction that clinically presents as tumor-like soft tissue swelling with sinus tract formation that drains purulent material containing grains . Mycetoma usually results of traumatic implantation of soil organisms on subcutaneous tissue; can be classified as eumycetoma or actinomycetoma depending on whether the infection is caused by filamentous fungi or aerobic filamentous actinomycetes , respectively [1] , [2] , [3] . Mycetoma represents a classical neglected disease that primarily affects the poorer populations and rural regions of Africa , Latin America , and Asia at latitudes defined as the “mycetoma belt” where higher mycetoma frequencies are observed . This region is located around the Tropic of Cancer , between latitudes 15° South and 30° North , encompassing the countries with the highest rates of infection including Sudan , Somalia , Senegal , India , Yemen , Mexico , and Venezuela [1] , [4] , [5] . The predominant climate of the “mycetoma belt” is subtropical and dry tropical with an annual average rainfall of about 500–1000 mm and temperatures ranged from 10–20°C to 20–40°C , respectively . This region is characterized by low humidity and low annual rainfall with well-defined alternating rainy and dry seasons . Actinomycetomas caused by Nocardia spp . occur mostly in regions with higher humidity , while actinomycetoma caused by Actinomadura spp . and Streptomyces spp . or eumycetoma occur in drier areas with low relative humidity [1] , [3] , [4] , [5] . Most causative agents of mycetoma , including fungi and actinomycetes , have been isolated from soil , decaying organic matter , plants and thorns; and , the disease is usually associated with traumatic injury followed by inoculation of the microorganism propagule . There are three main factors associated with the establishment of disease: inoculum size , immune status of the host , and hormonal adaptation ( based on the observation that men typically develop the disease ) [1] , [3] , [6] , [7] , [8] , [9] . Epidemiological data from different areas demonstrate that males are more affected ( sex ratio 3–4∶1 ) , ranging in age between the third and fourth decades of life ( 20–40 years ) . Some studies have reported that 3–5% of cases affect children committed to field-work [6] , [10] , [11] . Mycetoma is common in persons that work in rudimentary conditions without protective garments or shoes leading to the presentation of the illness primarily in poor rural workers or homemakers that participate in outdoor activities . Nearly all cases affect the lower limbs ( 75% ) , especially the foot and lower limbs . The nature of the patient's occupation also influences disease presentation , for example lumberjacks and sugarcane carriers generally present with mycetoma on the back [6] , [10] , [11] . The incubation period is unknown , disease symptoms present months to years after traumatic inoculation , depending on the inoculum size , strain virulence , and the host's immune response [7] . Because reporting mycetoma cases is not mandatory , the worldwide incidence is unknown; however , a recent published meta-analysis by van de Sande [11] reported incidence rates of 3 . 49 and 1 . 81 per 100 , 000 habitants in Mauritania and Sudan , respectively; while Mexico showed the highest rate in Latin American . Although a separate study reported that the number of cases per year in Sudan , Mauritania , and Mexico were 106 , 80 . 7 , and 73 , respectively [6] , several cases are probably not reported . The objective in this study was to provide epidemiological , clinical and microbiological data of mycetoma in different regions on Mexico , presenting at a public single center , the General Hospital of Mexico ( specialty hospital ) .
The Institutional Review Board approved the retrospective ( cross-sectional ) analysis of the database and clinical records of the Mycology Department of the Dermatology Service at the General Hospital of Mexico , patients were enrolled between January 1980 and December 2013 ( 34 years ) . We included all cases of mycetoma confirmed by microscopic observation of grains by direct examination with 10% potassium hydroxide ( KOH ) , saline solution , and lugol solution . The culture media used were Sabouraud dextrose agar and Yeast extract agar , however , when infection by Actinomadura madurae was suspected , Lowenstein-Jensen agar , and BHI agar ( Brain Heart Infusion ) were used . Histological examination was performed in some cases using hematoxylin and eosin ( H&E ) , Grocott's methenamine silver ( GMS ) , and Periodic acid–Schiff ( PAS ) Actinomycetes identification was carried out using micro morphological criteria ( Gram and Kinyoun stains ) as well as biochemical and major phenotypic tests such as urease production , hydrolysis of casein , gelatin , tyrosine , xanthine , hypoxanthine substrates and , growth at 45°C [12] , [13] . Fungal agent identification was based on morphological and reproductive form criteria and on biochemical tests . Some strains were identified using molecular techniques ( by amplification and sequence analysis of ribosomal DNA . the internal transcribed spacer region ( ITS ) , the translation elongation factor 1 alpha ( EF1-α ) , the partial beta tubulin gene ( TUB ) , and the small subunit of the nuclear ribosomal RNA gene [nucSSU] ) . General epidemiologic and clinical data were extracted from clinical records . All patients remained anonymous and descriptive statistics were used to analyze the data .
A total of 482 mycetoma cases were included in the present study , one patient presented simultaneously two different mycetomas . Demographic data are described in Table 1 . Figure 1 illustrates the age distribution of mycetoma , and Figure 2 the geographic mycetoma distribution in Mexico . The Pacific Ocean zone including Morelos , Guerrero , and Oaxaca accounted for 284 cases ( 58 . 92% ) and the Gulf of Mexico zone ( including Hidalgo and Veracruz ) accounted for 162 cases ( 33 . 60% ) . The remaining 36 cases presented in other states including Chiapas , Tabasco , Puebla , Michoacan , Colima , Jalisco , San Luis Potosi , Durango , Chihuahua , and Baja California . Most cases were actinomycetomas ( 92 . 11% ) with a male to female ratio of 2 . 8∶1 . However , this ratio changes in cases due to Actinomadura madurae , of the 42 cases 13 ( 30 . 06% ) were males and 29 ( 69 . 04% ) were females with a male∶female ratio of 1∶2 . 2 . Age ranges were classified in decades , the mean age was 34 . 5 years old ( range 7–92 years ) . Pediatric cases ( <18 years old ) were 20/482 ( 4 . 14% ) and 5/482 cases ( 1 . 083% ) were younger than 15 years old . Lower limbs were affected in 341 cases ( 70 . 74% ) . Three hundred one cases ( 62 . 44% ) occurred in the foot , 70 cases ( 14 . 52% ) affected the upper limbs ( 36 cases on the hands [7 . 46%] and 34 on the arms [7 . 05%] ) . The trunk was involved in 49 cases ( 10 . 16% ) , 38 ( 7 . 88% ) of those included back and shoulders . All cases involving multiple sites were associated with multiple traumatic inoculations . In regards to clinical presentation , mycetomas were classified as follows: 468 cases as tumor-like with draining sinuses ( 97 . 1% ) ; eight cases as tumor–like without sinuses ( 1 . 65% ) ; four cases as verrucous plaque ( 0 . 82% ) and two as cystic form ( 0 . 41% ) . Eight patients ( 1 . 6% ) presented lymphatic spread from the original mycetoma lesion: six from the foot to the inguinal area and two from the back to the axillary region . ( Figure 3 ) One patient presented with two mycetomas , each with a distinct causative agent: the one affecting the right foot was caused by Madurella mycetomatis , and the second one affecting the left foot was caused by Fusarium solani complex . Etiological agents were identified in 472/482 cases ( 98 . 34% ) . The agent was found in all of the actinomycetoma cases ( n = 444 cases ) . In 430 cases ( 89 . 2% ) the microorganisms were isolated and identified , and the remaining 14 cases were classified according to the grains observed during direct examination and/or histopathologic analysis . Two cases had double concurrently causative agents: N . asteroides s . l . +N . brasiliensis [14] and N . brasiliensis+A . madurae . Thirty-eight cases were eumycetomas and in 30/38 cases the etiological agents were isolated . Of the remaining seven cases ( four hyaline and three melanized-type ) only grain observation at direct microscopy was detected without identifying the causative fungi ( Table 2 ) [15] , [16] .
Mycetoma is a chronic granulomatous disease generally affecting low-income people including agricultural workers , peasants , or rural workers laboring with limited or no protective garments and soiled tools . The majority of cases ( 62% ) described in this report affected the foot , supporting previous reports [3] , [5] , [6] , [10] and one meta-analysis [11] that described foot as the most common site of infection ( 68 . 7% of cases ) . Since mycetoma presented most often on the feet of individuals living in the Indian endemic region , it explains why initial reports mentioned it as “Madura foot” [11] , [17] , [18] , [19] . Due to the predilection for feet , mycetoma control could be achieved by using appropriate footwear and clothing that protects the limbs . However , it should be emphasized that people living in endemic regions sometimes wear open-toed shoes , mainly due to the warm climate and therefore are less protected against potential trauma [1] , [5] , [20] , [21] . Mycetoma is associated with high morbidity and low mortality; however , the socioeconomic impact is significant; therefore patients are unable to work , resulting in decreased family income . In addition , treatments are expensive and difficult to maintain due to prolonged course of the disease . Almost all countries located in the “mycetoma belt” do not provide free quality health services or medical insurance [5] , [20] , [21] . The classical clinical presentation of mycetoma should lead to simple diagnosis based on the identification of a swelling zone with multiple sinus tracts; however , there is a significant lack of information between patients and clinicians , leading to delayed diagnosis and late referral to hospital , and consequently inadequate therapeutic response [5] , [7] , [8] . This study examined nearly 500 mycetoma cases from a single public hospital , helping to control for variables and facilitated the isolation and identification of the causative agents ( the causative agent was classified in most of the cases ) . Ninety two percent were actinomycetoma and 8% were eumycetoma , in accordance with previous studies [10] . This result differed from one report [6] that identified 3 . 5% eumycetoma cases this is probably due to the difficult diagnosing these cases until they are finally referred to specialty hospitals . Reports from Latin America [22] , [23] , [24] , [25] and , particularly , studies conducted in Mexico show a predominance of actinomycetoma , in contrast to those made in Africa , India , and Asia where cases of eumycetoma predominate [17] , [20] , [26] , [27] , [28] , [29] . This epidemiological difference can be explained by differences in climate and other environmental factors . The effect of climate is observed in mycetoma cases reported in India [17] , [18] , [19] where the majority of actinomycetoma occur in the northern region , where the climate is subtropical and has a higher annual rainfall; while , eumycetomas occur more often in the southern where the climate is dry tropical , has a low relative humidity , and more constant temperatures . In Mexico , eumycetomas occur in drier areas . This study provides a more accurate number of cases of mycetoma in Mexico ( 73 new cases per-year ) [6] , we believe that mycetoma remains difficult for clinicians to diagnose , for that several cases may be under diagnosed and therefore underreported [1] , [3] , [11] . In our study , we observed that mycetoma primarily affects men , with a male∶female ratio of almost 3∶1 in our study , in concordance with a previous report about mycetoma incidence in Mexico [6] . This male predominance of micetoma can be attributed to occupational and hormonal aspect [8] , [9] . The role of hormones in disease susceptibility may be explained by the few cases of mycetoma in children and the rapid growth of the lesions and increased in severity during pregnancy . Interestingly , the observed change of male∶female ratio in cases of mycetoma caused by A . madurae ( male∶female ratio of 1∶2 . 2 ) is partly due to this microorganism is not affected by progesterone and testosterone as with Nocardia brasiliensis [8] , [9] . Figure 1 shows mycetoma is most prevalent in the third decade of life ( 63 . 26% ) , which represents the most productive ages . The mean age was 34 . 5 years , similar to observations were made by van de Sande [7] . Some cases were reported in elderly , however we must consider that the infection may have started many years ago , suggesting that these individuals may have acquired the disease in youth . Moreover , only 4% of the cases were reported in patients <18 years old similar to previously reported studies [6] , [30] . The percentage of children infected in our study ( and other reports [6] , [30] ) differed from a report by Fahal et al . [31] that described a 15% infection rate ( n = 722 ) in children in Sudan , this was probably due to their outdoor work activities . However , the same study reported trauma in only 22 . 5% patients suggesting that different mechanisms of infection that deserve to be clarified [1] , [31] . The main clinical presentation was tumor-like with draining sinuses; cases presented as tumor-like without sinuses and cystic form were all eumycetomas; verrucous-plaque presentation was rare , the last one is very important since its differential diagnosis include verrucous-tuberculosis , chromoblastomycosis and nontuberculous mycobacterial diseases . Although lower limbs ( predominantly feet ) were most commonly affected by mycetoma in our population ( similar to the majority of previous reports [6] , [10] , [25] ) it is interesting that the trunk was affected in about 10% of cases ( predominantly back and shoulders ) . A previous Mexican report [6] described and incidence rate of the trunk in 19% of the cases , which was significantly different from the 1 . 4% rate described for these cases in Sudan [11] . These differences in anatomic regions affected may be explained due to occupation differences; patients in Mexico usually carry wood , sugarcane , or diverse materials on their backs . Mycetoma affecting the trunk should be considered of poor prognosis because of the proximity of lungs , spinal cord , and viscera [32] . It is also important to emphasize that cases presenting with multiple infections were described in immunocompetent patients that suffered multiple traumatic inoculations . Cases associated with lymphatic spread ( 1 . 65% ) are typically seen in immunocompromised patients ( malnutrition , immunosuppressants , malignant tumors , and chronic alcoholism ) . Nocardia spp . was the main etiological agent ( 82 . 32% of cases ) , being N . brasiliensis the predominant species ( 78 . 21% ) . The strains were identified using phenotypic tests that lead to the identification of two more species: N . asteroides complex and N . otitidiscaviarum ( formerly N . caviae ) [14] , [33] . Molecular biology techniques such as PCR and sequencing of 16S rRNA and the hsp65 gene allow the correct identification and classification of Nocardia species [13] . For example , N . mexicana , N . harenae , and N . takedensis were isolated from Mexican patients [34] , [35] , [36] . The second most common causative agent was A . madurae; distributed worldwide [11] , which is easily identified due to the larger ( 1–3 mm ) white-yellowish , soft , wide-fringed border grains . Regardless , A . madurae is more difficult to isolate than Nocardia species , which regularly grow in rich culture media such as Lowenstein-Jensen , BHI-agar . Morphologic analyses leads to the identification of the causative agent , however , confirmation is carried out using phenotypic and temperature tests [12] . Other etiologic agents include A . pelletieri and S . somaliensis commonly found in Africa and Asia . Remarkably , the two cases presented with mixed infections following traumatic inoculation were due to N . brasiliensis+N . asteroides s . l . and N . brasiliensis+A . madurae , respectively , both presented in immunocompetent patients [14] . The causal agents of eumycetoma represented only 8% of our series , melanized fungi was the most commonly observed ( 26 cases , 5 . 39% ) . Of these , Madurella mycetomatis was the foremost isolated fungi , found in 15 cases , and also considered responsible for 25% of cases worldwide , mainly in Africa and Asia [11] . Some cases have a well-defined history of trauma ( e . g . , thorn pricks ) prior to mycetoma development; however , in some cases a well-defined traumatic event was not identified , de Hoog [37] recently noted that M . mycetomatis is a close relative to dung-inhabiting fungi and suggested that the natural habitat of this fungus could therefore also be dung; trauma or repeated contact with cattle dung could act as an adjuvant for inoculation of causative agents of mycetoma . Identification of Madurella species has been hampered by the absence of sporulation leading to confusion during the identification process . The use of molecular techniques for the identification of specific regions and genes ( e . g . , rRNA , ITS , parcial β-tubulin gene , RNA polymerase II subunit 2 gene ) has defined Madurella species as a cryptic complex belonging to the order Sordariales , consists primarily of M . mycetomatis , M . fahalii , M . tropicana , and M . pseudomycetomatis [38] . Identification of the infecting agent is critical since differences in thermal adaptation and susceptibility to antifungal agents exist between strains . M . grisea has been reclassified and now belongs to the order Pleosporales and named Trematospheria grisea . It should be noted that the latter was more frequently reported as a cause of mycetoma than M . mycetomatis in a recent report from Mexico [6]; however , difficulties in morphological and phenotypic identification could have led to confusion . Other melanized fungi further characterized by molecular biology techniques was Exophiala jeanselmei and Cladophialophora bantiana , which are common agents of phaeohyphomycosis [15]; and Cladophialophora mycetomatis , considered new specie [16] . Regarding hyaline fungi , Scedosporium boydii ( Syn: Scedosporium apiospermum , Pseudallescheria boydii ) , was the foremost isolated strain ( similar to other studies ) [6] , [11] , [24] , [26] . Mycetomas due to Fusarium have been previously described [39] , [40] , we found F . solani in two cases; interestingly , one of them was microscopically classified as F . chlamydosporum , but was reclassified using molecular biology as part of the Fusarium solani complex ( CBS 135554 ) . One case resulted from infection with Aspergillus nidulans , an agent rarely reported [41] and identified morphologically for the presence of Hülle cells . A case resulting from Microsporum canis infection was classified as pseudomycetoma [42] because of the rarity of the agent as a mycetoma pathogen and usually development as consequence of a chronic tinea capitis , typically seen in immunocompromised patients in the absence of traumatic inoculation [42] , [43] . The study has limitations inherent to its design , however , provides important information about the status of mycetoma in Mexico . The study results can be generalized only to our population ( Mexico ) ; although the geographical areas studied has similarities with other world regions in terms of climate , distribution of etiologic agents and sociocultural conditions . Mycetoma fulfills all the criteria of a neglected tropical disease [20] , [21] , [44] . It is extremely important to monitor cases and their causative agents , as a mean to understand the epidemiology of the disease , and to establish interventions for prevention , treatment and rehabilitation .
|
Mycetoma is a chronic , subcutaneous granulomatous disease that usually begins after traumatic inoculation with causative microorganisms . Based on its etiology , mycetoma is referred to eumycetoma when the infection is caused by filamentous fungi , and actinomycetoma when the infection is due to aerobic actinomycetes ( in Mexico predominantly Nocardia brasiliensis ) . Establishing the etiology is extremely important since it impacts treatment regimens . Mycetoma typically presents around the Tropic of Cancer between latitude 15° South and 30° North ( also known as “mycetoma belt” ) affecting poor populations in Africa , Asia , and Latin America , including Mexico , which represents a highly endemic area with higher frequencies of actinomycetomas . Mycetoma usually affects males ( male∶female ratio of 3∶1 ) , agricultural or rural workers ( age range 20–40 years ) that typically do not have access to protective equipment . The main clinical presentation is as soft tissue swelling with sinus tract formation draining grains , which leads to diagnosis . The foot is the most commonly affected localization; however , when disease presents in high risk areas , such as the trunk , it can disseminate to the lungs and spinal cord . This report represents a single center study which provides epidemiologic , clinical , and microbiological data of mycetoma cases in different regions of Mexico .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases",
"subcutaneous",
"mycoses",
"medicine",
"and",
"health",
"sciences",
"mycetoma",
"neglected",
"tropical",
"diseases",
"tropical",
"diseases",
"fungal",
"diseases"
] |
2014
|
Mycetoma: Experience of 482 Cases in a Single Center in Mexico
|
Epstein-Barr virus ( EBV ) , a ubiquitous B-lymphotropic herpesvirus , ectopically infects T or NK cells to cause severe diseases of unknown pathogenesis , including chronic active EBV infection ( CAEBV ) and EBV-associated hemophagocytic lymphohistiocytosis ( EBV-HLH ) . We developed xenograft models of CAEBV and EBV-HLH by transplanting patients' PBMC to immunodeficient mice of the NOD/Shi-scid/IL-2Rγnull strain . In these models , EBV-infected T , NK , or B cells proliferated systemically and reproduced histological characteristics of the two diseases . Analysis of the TCR repertoire expression revealed that identical predominant EBV-infected T-cell clones proliferated in patients and corresponding mice transplanted with their PBMC . Expression of the EBV nuclear antigen 1 ( EBNA1 ) , the latent membrane protein 1 ( LMP1 ) , and LMP2 , but not EBNA2 , in the engrafted cells is consistent with the latency II program of EBV gene expression known in CAEBV . High levels of human cytokines , including IL-8 , IFN-γ , and RANTES , were detected in the peripheral blood of the model mice , mirroring hypercytokinemia characteristic to both CAEBV and EBV-HLH . Transplantation of individual immunophenotypic subsets isolated from patients' PBMC as well as that of various combinations of these subsets revealed a critical role of CD4+ T cells in the engraftment of EBV-infected T and NK cells . In accordance with this finding , in vivo depletion of CD4+ T cells by the administration of the OKT4 antibody following transplantation of PBMC prevented the engraftment of EBV-infected T and NK cells . This is the first report of animal models of CAEBV and EBV-HLH that are expected to be useful tools in the development of novel therapeutic strategies for the treatment of the diseases .
Epstein-Barr virus ( EBV ) is a ubiquitous γ-herpesvirus that infects more than 90% of the adult population in the world . EBV is occasionally involved in the pathogenesis of malignant tumors , such as Burkitt lymphoma , Hodgkin lymphoma , and nasopharyngeal carcinoma , along with the post-transplantation lymphoproliferative disorders in immunocompromised hosts . Although EBV infection is asymptomatic in most immunologically competent hosts , it sometimes causes infectious mononucleosis ( IM ) , when primarily infecting adolescents and young adults [1] . EBV infects human B cells efficiently in vitro and transform them into lymphoblastoid cell lines ( LCLs ) [2] . Experimental infection of T and NK cells , in contrast , is practically impossible except in limited conditions [3] , [4] . Nevertheless , EBV has been consistently demonstrated in T or NK cells proliferating monoclonally or oligoclonally in a group of diseases including chronic active EBV infection ( CAEBV ) and EBV-associated hemophagocytic lymphohistiocytosis ( EBV-HLH ) [5] , [6] , [7] , [8] , [9] , [10] . CAEBV , largely overlapping the systemic EBV+ T-cell lymphoproliferative diseases of childhood defined in the WHO classification of lymphomas [11] , is characterized by prolonged or relapsing IM-like symptoms , unusual patterns of antibody responses to EBV , and elevated EBV DNA load in the peripheral blood [12] , [13] , [14] . CAEBV has a chronic time course with generally poor prognosis; without a proper treatment by hematopoietic stem cell transplantation , the majority of cases eventually develop malignant lymphoma of T or NK lineages , multi-organ failure , or other life-threatening conditions . Monoclonal or oligoclonal proliferation of EBV-infected T and NK cells , an essential feature of CAEBV , implies its malignant nature , but other characteristics of CAEBV do not necessarily support this notion . For example , EBV-infected T or NK cells freshly isolated from CAEBV patients , as well as established cell lines derived from them , do not have morphological atypia and do not engraft either in nude mice or scid mice ( Shimizu , N . , unpublished results ) . Clinically , CAEBV has a chronic time course and patients may live for many years without progression of the disease [15] . Although patients with CAEBV do not show signs of explicit immunodeficiency , some of them present a deficiency in NK-cell activity or in EBV-specific T-cell responses , implying a role for subtle immunodeficiency in its pathogenesis [16] , [17] , [18] . EBV-HLH is the most common and the severest type of virus-associated HLH and , similar to CAEBV , characterized by monoclonal or oligoclonal proliferation of EBV-infected T ( most often CD8+ T ) cells [5] , [6] . Clinical features of EBV-HLH include high fever , pancytopenia , coagulation abnormalities , hepatosplenomegaly , liver dysfunction , and hemophagocytosis [19] . Overproduction of cytokines by EBV-infected T cells as well as by activated macrophages and T cells reacting to EBV is thought to play a central role in the pathogenesis [20] . Although EBV-HLH is an aggressive disease requiring intensive clinical interventions , it may be cured , in contrast to CAEBV , by proper treatment with immunomodulating drugs [21] . No appropriate animal models have been so far developed for either CAEBV or EBV-HLH . NOD/Shi-scid/IL-2Rγnull ( referred here as NOG ) is a highly immunodeficient mouse strain totally lacking T , B , and NK lymphocytes , and transplantation of human hematopoietic stem cells to NOG mice results in reconstitution of human immune system components , including T , B , NK cells , dendritic cells , and macrophages [22] , [23] . These so called humanized mice have been utilized as animal models for the infection of certain human viruses targeting the hemato-immune system , including human immunodeficiency virus 1 ( HIV-1 ) and EBV [24] , [25] , [26] , [27] , [28] , [29] , [30] . Xenotransplantation of human tumor cells to NOG mice also provided model systems for several hematologic malignancies [31] , [32] , [33] . To facilitate investigations on the pathogenesis of CAEBV and EBV-HLH and assist the development of novel therapeutic strategies , we generated mouse models of these two EBV-associated diseases by transplanting NOG mice with PBMC isolated from patients with the diseases . In these models , EBV-infected T , NK , or B cells engrafted in NOG mice and reproduced lymphoproliferative disorder similar to either CAEBV or EBV-HLH . Further experiments with the models revealed a critical role of CD4+ T cells in the in vivo proliferation of EBV-infected T and NK cells .
Depending on the immunophenotypic subset in which EBV causes lymphoproliferation , CAEBV is classified into the T-cell and NK-cell types , with the former being further divided into the CD4 , CD8 , and γδT types . The nine patients with CAEBV examined in this study are characterized in Table 1 and include all these four types . Intravenous injection of 1−4×106 PBMC isolated from these nine patients resulted in successful engraftment of EBV-infected T or NK cells in NOG mice in a reproducible manner ( Table 1 ) . The results with the patient 1 ( CD4 type ) , patient 3 ( CD8 type ) , patient 5 ( γδT type ) , and patient 9 ( NK type ) are shown in Figure 1 . Seven to nine weeks post-transplantation , EBV DNA was detected in the peripheral blood of recipient mice and reached the levels of 105–108 copies/µg DNA ( Figure 1A ) . By contrast , no engraftment of EBV-infected cells was observed when immunophenotypic fractions containing EBV DNA were isolated from PBMC and injected to NOG mice ( Figure 1A and Table 2 ) . An exception was the CD4+ T-cell fraction isolated from patients with the CD4 type CAEBV , that reproducibly engrafted when transplanted without other components of PBMC ( Figure 1A , Table 2 ) . Flow cytometry revealed that the major population of engrafted cells was either CD4+ , CD8+ , TCRγδor CD16+CD56+ , depending on the type of the donor CAEBV patient ( Figure 1B ) . EBV-infected cells of identical immunophenotypes were found in the patients and the corresponding mice that received their respective PBMC ( Figure 1B ) . Although human cells of multiple immunophenotypes were present in most recipient mice , fractionation by magnetic beads-conjugated antibodies and subsequent real-time PCR analysis detected EBV DNA only in the predominant immunophenotypes that contained EBV DNA in the original patients ( Figure 1B , Table 1 ) . The EBV DNA load observed in individual lymphocyte subsets in the patient 3 and a mouse that received her PBMC is shown as supporting data ( Table S1 ) . When PBMC from three healthy EBV-carriers were injected intravenously to NOG mice , as controls , no EBV DNA was detected from either the peripheral blood , spleen , or liver ( data not shown ) . Histological analyses of the spleen and the liver of these control mice identified no EBV-encoded small RNA ( EBER ) -positive cells , although some CD3-positive human T cells were observed ( Figure S2 ) . Analysis of TCR Vβ repertoire demonstrated an identical predominant T-cell clone in patients ( patients 1 and 3 ) and the corresponding mice that received their PBMC ( Figure 1C ) . The general condition of most recipient mice deteriorated gradually in the observation period of eight to twelve weeks , with loss of body weight ( Figure S1 ) , ruffled hair , and inactivity . NOG mice engrafted with EBV-infected T or NK cells were sacrificed for pathological and virological analyses between eight and twelve weeks post-transplantation . On autopsy , the majority of mice presented with splenomegaly , with slight hepatomegaly in occasional cases ( Figure 2A ) . Histopathological findings obtained from a representative mouse ( recipient of PBMC from the patient 3 ( CD8 type ) ) are shown in Figure 2B and reveal infiltration of human CD3+CD20− cells to major organs , including the spleen , liver , lungs , kidneys , and small intestine . These cells were positive for both EBER and human CD45RO , indicating that they are EBV-infected human T cells ( Figure 2B ) . In contrast , no EBV-infected T cells were found in mice transplanted with PBMC isolated from a normal EBV carrier ( Figure S2 ) . Histopathology of a control NOG mouse is shown in Figure S2 . Morphologically , EBV-infected cells are relatively small and do not have marked atypia . The infiltration pattern was leukemic and identical with chronic active EBV infection in children [34] . The architecture of the organs was well preserved in spite of marked lymphoid infiltration . The spleen showed marked expansion of periarterial lymphatic sheath owing to lymphocytic infiltration . In the liver , a dense lymphocytic infiltration was observed in the portal area and in the sinusoid . The lung showed a picture of interstitial pneumonitis and the lymphocytes often formed nodular aggregations around bronchioles and arteries . In the kidney , dense lymphocytic infiltration caused interstitial nephritis . In the small intestine , mild lymphoid infiltration was seen in mucosa . Quantification of EBV DNA in the spleen , liver , lymph nodes , lungs , kidneys , adrenals , and small intestine of this mouse revealed EBV DNA at the levels of 1 . 5–5 . 1×107 copies/µg DNA . Mice transplanted with PBMC derived from CAEBV of other types exhibited similar infiltration of EBV-infected T or NK cells to the spleen , liver , and other organs ( Figure 2C and data not shown ) . We established EBV-positive cell lines of CD4+ T , CD8+ T , γδT , and CD56+ NK lineages from PBMC of the patients listed in Table 1 by the method described previously [35] , and confirmed by flow cytometry that the surface phenotypes of EBV-infected cells in the original patients were retained in these cell lines ( data not shown ) . To test whether these cell lines engraft in NOG mice , 1–4×106 cells were injected intravenously to NOG mice . The results are shown in Figure 3A and indicate that CAEBV-derived cell lines of the CD8+ T , γδT , and CD56+ NK phenotypes do not engraft in NOG mice . Neither human CD45-positive cells nor EBV DNA were detected in the peripheral blood of the mice up to twelve weeks post-transplantation . When the recipient mice were sacrificed at twelve weeks post-injection , no EBV DNA could be detected in the spleen , liver , bone marrow , mesenteric lymph nodes , and kidneys . In contrast , the CD4+ T cell lines derived from the CD4-type patients 1 and 2 engrafted in NOG mice and induced T lymphoproliferation similar to that induced by PBMC isolated freshly from these patients ( Figure 3A and data not shown ) . These results , together with the results of transplantation with EBV-containing subsets of PBMC , indicate that EBV-infected T and NK cells , with the exception of those of the CD4+ subset , are not able to engraft in NOG mice , when they are separated from other components of PBMC , suggesting that some components of PBMC are essential for the outgrowth EBV-infected T and NK cells in NOG mice . To identify the cellular component required for the engraftment of EBV-infected T and NK cells in NOG mice , we transplanted PBMC of CAEBV patients after removing individual immunophenotypic subsets by magnetic beads-conjugated antibodies . The results are shown in Figure 3B and summarized in Table 2 . With respect to the patients 3 and 4 , in whom CD8+ T cells are infected with EBV , removal of CD8+ cells from PBMC , as expected , resulted in the failure of engraftment , whereas elimination of CD19+ , CD56+ , or CD14+ cells did not affect engraftment . Importantly , elimination of CD4+ cell fraction , that did not contain EBV DNA , resulted in the failure of engraftment of EBV-infected T cells ( Figure 3B and data not shown ) . In the experiments with the patients 5 and 6 , in whom γδT cells were infected , removal CD4+ cells that did not contain EBV DNA , as well as that of γδT cells , resulted in the failure of engraftment . Removal of CD8+ , CD14+ , CD19+ , or CD56+ cells did not have an influence on the engraftment ( Figure 3B and data not shown ) . Regarding the patients 8 an 9 in whom EBV resided in CD56+ NK cells , removal of CD4+ as well as CD56+ cells resulted in the failure of engraftment , whereas that of CD8+ , CD19+ , or CD14+ cells did not affect engraftment ( Figure 3B and data not shown ) . In the patients 1 and 2 , in whom CD4+ T cells were infected , only the removal of CD4+ cells blocked the engraftment of EBV-infected cells and depletion of either CD8+ , CD19+ , or CD14+ cells had no effect ( Figure 3B and data not shown ) . These results suggested that EBV-infected cells of the CD8+ , γδT , and CD56+ lineages require CD4+ cells for their engraftment in NOG mice . To confirm this interpretation , we performed complementation experiments , in which EBV-containing fractions of the CD8+ ( patient 4 ) , γδT ( patient 5 ) , or CD56+ ( patient 7 ) phenotypes were transplanted together with autologous CD4+ cells . The results are shown in Figure 3A and indicate that EBV-infected CD8+ , γδT , or CD56+ cells engraft in NOG mice when transplanted together with CD4+ cells . Similarly , when EBV-infected cell lines of the CD8+ , γδT , and CD16+ lineages were injected intravenously to NOG mice together with autologous CD4+ cells , these cell lines engrafted to the mice ( Figure 3A ) . Finally , to further confirm the essential role of CD4+ cells , we examined the effect of the OKT-4 antibody that depletes CD4+ cells in vivo [24] . PBMC isolated from the CAEBV patient 3 ( CD8 type ) and the patient 8 ( NK type ) were injected intravenously to NOG mice and OKT-4 was administered intravenously for four consecutive days starting from the day of transplantation . The results are shown in Figure 4 and indicate that OKT-4 can strongly suppress the engraftment of EBV-infected T and NK cells . In the mice treated with OKT-4 , no splenomegaly was observed and EBV DNA was not detected either in the peripheral blood , spleen , liver , or lungs at eight weeks post-transplantation . Previous analysis of EBV gene expression in patients with CAEBV revealed the expression of EBNA1 , LMP1 , and LMP2A with the involvement of the Q promoter in the EBNA genes transcription and no expression of EBNA2 , being consistent with the latency II type of EBV gene expression [36] , [37] , [38] . To test whether EBV-infected T and NK cells that proliferate in NOG mice retain this type of viral gene expression , we performed RT-PCR analysis in the spleen and the liver of mice that received PBMC from the CAEBV patient 3 ( CD8 type ) . The results are shown in Figure 5A and demonstrate the expression of mRNAs coding for EBNA1 , LMP1 , LMP2A , and LMP2B , but not for EBNA2 . Expression of the EBV-encoded small RNA 1 ( EBER1 ) was also demonstrated . EBNA1 mRNAs transcribed from either the Cp promoter or the Wp promoter were not detected , whereas those transcribed from the Q promoter was abundantly detected . These results indicate that EBV-infected T cells retain the latency II pattern of latent EBV gene expression after engraftment in NOG mice . Similar analyses with NOG mice engrafted with EBV-infected NK cells also showed the latency II type of EBV gene expression ( data not shown ) . In patients with CAEBV , high levels of cytokines have been detected in the peripheral blood and are thought to play important roles in the pathogenesis [20] , [39] , [40] . To test whether this hypercytokinemia is reproduced in NOG mice , we examined the levels of various human cytokines in the sera of transplanted mice using ELISA kits that can quantify human cytokines specifically . The results are shown in Figure 5B and indicate that the mice transplanted with PBMC of the patient 3 ( CD8 type ) or the patient 8 ( NK type ) contained high levels of RANTES , IFN-γ , and IL-8 in their sera . To extend the findings obtained from the CAEBV xenograft model to another disease with EBV+ T/NK lymphoproliferation , we transplanted NOG mice with PBMC isolated from patients with EBV-HLH . Characteristics of the four EBV-HLH patients examined in this study and the results of transplantation with their PBMC are summarized in Table 1 . EBV DNA was detected in the peripheral blood three to four weeks post-transplantation and rapidly reached the levels of 1×104 to 1×106 copies/µg DNA ( results of typical experiments are shown in Figure 6A ) . Similar to the findings in CAEBV , EBV DNA was not detected in the recipient mice , when CD4+ cell fraction was removed from PBMC ( Figure 6A ) . Immunophenotypic analyses on the peripheral blood lymphocytes isolated from EBV-HLH patients and corresponding recipient mice revealed that cells of an identical immunophenotype ( CD3+CD8+CD45RO+CD19−CD4−CD45RA−CD16−CD56− ) were present and contained EBV DNA in both the patients and corresponding mice ( Figure 6C and data not shown ) . The EBV DNA load observed in individual lymphocyte subsets in the patient 10 and a mouse that received his PBMC is shown as supporting data ( Table S2 ) . General condition of the recipient mice deteriorated consistently more quickly , with the loss of body weight ( Figure S1 ) , ruffling of hair , and general inactivity , than those mice engrafted with EBV-infected T or NK cells derived from CAEBV . The mice were sacrificed around four weeks post-transplantation for pathological analyses . Macroscopical observation revealed moderate to severe splenomegaly ( Figure 6D ) in the majority of recipient mice , and slight hepatomegaly in a limited fraction of them . A finding characteristic to these mice were massive hemorrhages in the abdominal and/or thoracic cavities , that were not seen in the mice transplanted with CAEBV-derived PBMC ( Figure 6D and data not shown ) . These hemorrhagic lesions may reflect coagulation abnormalities characteristic to HLH . Histopathological analyses revealed a number of EBER+ cells in the spleen and the liver ( Figure 6E ) and quantification of EBV DNA in these tissues revealed 1 . 4×101 to 2 . 4×102 copies/µg of EBV DNA . When the tissues were examined by immunostaining and EBER ISH , the EBER+ cells were shown unexpectedly to be mostly CD45RO- and CD20+ in all five transplantation experiments with four different patients , indicating that the majority of EBV-infected cells in these tissues are of the B-cell lineage ( Figure 6E and data not shown ) . EBER+ large B cells were seen scattered among numerous reactive small T cells , most of which are CD8+ , in the tissues of the spleen , liver , lungs and kidneys . A number of macrophages were also seen in these tissues . Fractionation of mononuclear cells obtained from the liver of a mouse transplanted with PBMC of the EBV-HLH patient 10 , followed by real-time PCR , detected EBV DNA ( 1 . 4×101 copies/µg DNA ) only in the CD19+ B-cell fraction . In addition , an EBV-infected B lymphoblastoid cell line , but not an EBV-positive T cell line , could be established from this liver . Thus the presence of EBV in B cells were demonstrated by three independent methods in the tissues of EBV-HLH mice . Enzyme-linked immunosorbent assay revealed extremely high levels of human cytokines , including IL-8 , IFN-γ , and RANTES , in the sera of both the original patients and the recipient mice ( Figure 6B ) . The levels of IL-8 and IFN-γ were much higher than those observed in the peripheral blood of patients with CAEBV and mice that received their PBMC . Thus , NOG mice transplanted with EBV-HLH-derived PBMC are distinct from those transplanted with CAEBV-derived PBMC in the aggressive time course of the disease , internal hemorrhagic lesions , extremely high levels of IL-8 and IFN-γ in the peripheral blood , and the presence of EBV-infected B cells in lymphoid tissues .
The mouse xenograft models of CAEBV and EBV-HLH developed here represent the first recapitulation of EBV-associated T/NK lymphoproliferation in experimental animals . Previously , Hayashi and others inoculated rabbits with Herpesvirus papio and succeeded in the generation of T-cell lymphoproliferative disorder with pathological findings suggestive of EBV-HLH [41] . This model , however , is based on an EBV-related virus and not EBV itself , and therefore may contain features irrelevant to the original human disease . Although the CAEBV and EBV-HLH models described here exhibited some common features , including the abundant presence of EBV-infected T or NK cells in the peripheral blood , there were some critical differences between the two models , probably reflecting the divergence of the pathophysiology of the original diseases . First of all , in the EBV-HLH model mouse , EBV was detected mainly in B cells in the spleen and the liver , while it was found mainly in T cells in the peripheral blood . This makes an obvious contrast with the CAEBV model mouse , where EBV was detected in T or NK cells in both the peripheral blood and lymphoid tissues . We do not have an explanation for the apparent discrepancy in the host cell type of EBV infection between the peripheral blood and lymphoid tissues of the EBV-HLH model . It should be , however , noted that histopathology of EBV-HLH tissues has not been fully investigated and therefore it is still possible that significant number of EBV-infected B cells are present in the lymphoid tissues of EBV-HLH patients . Other differences between the two models include much higher plasma levels of IL-8 and IFN-γ more aggressive and fatal outcome , and internal hemorrhagic lesions in EBV-HLH model mice , probably reflecting the differences in the pathophysiology of the original diseases . EBV-positive B-cell proliferation was not seen in CAEBV model mice even in long-term observation beyond twelve weeks . This seems puzzling since low but significant amount of EBV DNA was found also in B19+ B-cell fraction in most patients with CAEBV . It should be noted that EBV-infected T or NK cell lines could be established relatively easily from patients with CAEBV by adding recombinant IL-2 in the medium . In contrast , establishment of EBV-infected B LCLs from these patients has been extremely difficult . In fact , we could establish B-LCLs from a few patients with CAEBV only when their PBMC were cultured on feeder cells expressing CD40 ligand . Therefore , we speculate that in the particular context of CAEBV , both in the patient and the model mouse , proliferation of EBV-infected B cells are somehow inhibited by an unknown mechanism . Analysis on the conditions of engraftment of EBV-infected T/NK cells using these new xenograft models revealed that EBV-infected T and NK cells of the CD8+ T , TCRγδT and CD56+ NK lineages and cell lines derived from them require CD4+ T cells for their engraftment in NOG mice . Only those EBV-infected cells and cell lines of the CD4+ T lineage could engraft in NOG mice on their own . These findings suggest that some factor ( s ) provided by CD4+ cells are essential for engraftment . Soluble factors produced by CD4+ T cells may be responsible for this function and we are currently examining cytokines , including IL-2 , for their ability to support the engraftment of EBV-infected T and NK cells . It is also possible that cell to cell contact involving CD4+ cells is critical for engraftment . This dependence on CD4+ cells represents an interesting consistency with the previous finding that engraftment of EBV-transformed B lymphoblastoid cells in scid mice required the presence of CD4+ cells [42] , [43] . It has been speculated that T cells activated by an EBV-induced superantigen may be involved in the engraftment of EBV-infected B lymphoblastoid cells in scid mice [44] . Although a similar superantigen-mediated mechanism might also be assumed in T- and NK-cell lymphoproliferation in NOG mice , the data of TCR repertoire analyses ( Figure 1C and data not shown ) show no indication for clonal expansion of Vβ13 T cells that are known to be specifically activated by the EBV-induced superantigen HERV-K18 . It seems therefore unlikely that this superantigen is involved in the CD4+ T cell-dependent engraftment of EBV-infected T and NK cells . We expect CD4+ T cells and/or molecules produced by them may be an excellent target in novel therapeutic strategies for the treatment of CAEBV and EBV-HLH . In fact , administration of the OKT-4 antibody that depletes CD4+ cells in vivo efficiently prevented the engraftment of EBV-infected T cells . As a next step , we plan to test the effect of post-engraftment administration of OKT-4 . The dependence of EBV-infected T and NK cells on CD4+ T cells for their engraftment in NOG mice suggests the possibility that these cells are not capable of autonomous proliferation . Consistent with this notion , EBV-infected T and NK cell lines , including that of the CD4+ lineage , are dependent on IL-2 for their in vitro growth and do not engraft in either nude mice or scid mice when transplanted either s . c . or i . v ( Shimizu , N . , unpublished results ) . Clinically , CAEBV is a disease of chronic time course and patients carrying monoclonal EBV-infected T or NK cell population may live for many years without progression of the disease [15] . Overt malignant T or NK lymphoma usually develops only after a long course of the disease . Taking all these findings in consideration , we suppose that EBV-infected cells are not truly malignant at least in the early phase of the disease , even when they appear monoclonal . Because infection of EBV in T or NK cells is not unique to CAEBV and has been recognized also in infectious mononucleosis [45] , [46] , the critical deficiency in CAEBV may be its inability to immunologically remove EBV-infected T and NK cells . In this context , it should be emphasized that EBV-infected T or NK cells usually exhibit the latency II pattern of EBV gene expression and do not express EBNA3s , that possess immuno-dominant epitopes recognized by EBV-specific T cells [47] . EBV-infected T and NK cells are thus not likely to be removed by cytotoxic T cells as efficiently as EBV-infected B cells that express EBNA3s . The reported lack of cytotoxic T cells specific to LMP2A [17] , one of the few immuno-dominant EBV proteins expressed in the virus-infected T and NK cells , may therefore seriously affect the host's capacity to control their proliferation . A genetic defect in the perforin gene was recently identified in a patient with clinical and pathological features resembling CAEBV , suggesting that defects in genes involved in immune responses can result in clinical conditions similar to CAEBV [48] . Engraftment of EBV-infected T and NK cells in NOG mice was in most cases accompanied by co-engraftment of un-infected cell populations . These un-infected cells might have been maintained and induced to proliferate by certain factors produced by EBV-infected T or NK cells . Abundant cytokines produced by these cells may be responsible for this activity . It is also possible that the proliferation of these un-infected cells represents immune responses . Experiments are underway to test whether these un-infected T cells contain EBV-specific cells . These un-infected T cells might also be reacting to host murine tissues . Intravenous injection of PBMC obtained from normal humans to immunodeficient mice including NOG mice has been shown to induce acute or chronic graft versus host disease ( GVHD ) [49] , [50] . However , because much less PBMC were injected to mice in the present study as compared to those previous studies , it is not likely that major GVHD was induced in NOG mice transplanted with PBMC of patients with CAEBV or EBV-HLH . CAEBV has been treated by a variety of regimens , including antiviral , cytocidal , and immunomodulating agents with more or less unsatisfactory results . Although hematopoietic stem cell transplantation , especially that with reduced intensity conditioning can give complete remission in a substantial number of patients [51] , [52] , it is still desirable to develop safer and more effective treatment , possibly with pharmaceutical agents . The xenograft model of CAEBV generated in this study may be an excellent animal model to test novel experimental therapies for the disease . In fact , the OKT-4 antibody that depletes CD4+ T cells in vivo gave a promising result implying its effectiveness as a therapeutic to CAEBV .
Protocols of the experiments with materials obtained from patients with CAEBV and EBV-HLH and from control persons have been reviewed and approved by the Institutional Review Boards of the National Center for Child Health and Development and of the National Institute of Infectious diseases ( NIID ) . Blood samples of the patients and control persons were collected after obtaining written informed consent . Protocols of the experiments with NOG mice are in accordance with the Guidelines for Animal Experimentation of the Japanese Association for Laboratory Animal Science and were approved by the Institutional Animal Care and Use Committee of NIID . Characteristics of the nine patients with CAEBV and the four patients with EBV-HLH examined in this study are summarized in Table 1 . Diagnosis of CAEBV and EBV-HLH was made on the basis of the published guidelines [19] , [53] and confirmed by identification of EBV-infected T or NK cells in their peripheral blood by flow cytometry and real-time PCR . Mice of the NOD/Shi-scid/IL-2Rγnull ( NOG ) strain [22] were obtained from the Central Institute for Experimental Animals ( Kawasaki , Japan ) and maintained under specific pathogen free ( SPF ) conditions in the animal facility of NIID , as described [22] . PBMC were isolated by centrifugation on Lymphosepar I ( Immuno-Biological Laboratories ( IBL ) ) and injected intravenously to the tail vein of NOG mice at the age of 6–8 weeks . Depending on the recovery of PBMC , 1–4×106 cells were injected to 2 to 4 mice in a typical experiment with a blood sample . For transplantation with individual cellular fractions containing EBV DNA , CD4+ T cells , CD8+ T cells , and CD56+ NK cells were separated with the IMag Cell Separation Systems ( BD Pharmingen ) following the protocol supplied by the manufacturer . To isolate γδT cells , CD19+ , CD4+ , CD8+ , CD56+ , and CD14+ cells were serially removed from PBMC by the IMag Cell Separation Systems . From the remaining CD19−CD4−CD8−CD56−CD14− population , CD3+ cells were positively selected by the same kit and defined as the γδT cell fraction . To transplant PBMC lacking individual immunophenotypic subsets , CD19+ , CD4+ , CD8+ , CD56+ or CD14+ cells were removed from PBMC by the IMag Cell Separation Systems and the remaining cells were injected to mice . To prepare PBMC lacking γδT cells , CD19+ , CD4+ , CD8+ , CD56+ , and CD14+ cells isolated from PBMC in the process of obtaining γδT cell fraction ( see above ) were pooled and mixed with the CD19−CD4−CD8−CD56−CD14− cells that did not react with anti-CD3 antibody . For complementation experiments , an EBV-containing cell fraction and the CD4+ cell fraction were isolated from a sample of PBMC as described above and the mixture of these two fractions were injected to NOG mice . The approximate numbers of injected cells are shown in Table 2 . PBMC isolated from the patients and the recipient NOG mice as described above were incubated for 30 min on ice with a mixture of appropriate combinations of fluorescently labeled monoclonal antibodies . After washing , five-color flow-cytometric analysis was carried out with the Cytomics FC500 analyzer ( Beckman Coulter ) . The following directly labeled antibodies were used: phycoerythrin ( PE ) -conjugated antibodies to CD3 , CD8 , and TCRα/β , fluorescein isothiocyanate ( FITC ) -conjugated antibodies to CD3 , CD4 , CD8 , CD19 , TCRVγ9 , TCRVδ2 , and TCRγ/δ , and Phycoerythrin Texas Red ( ECD ) -conjugated antibody to CD45RO from Beckman Coulter; PE-conjugated antibodies to CD16 , CD40 , and CD40L , and FITC-conjugated antibody to CD56 from BD Pharmingen . TCR Vβ repertoire analysis was performed with the Multi-analysis TCR Vβ antibodies Kit ( Beckman Coulter ) according to the procedure recommended by the manufacturer . NOG mice were injected intravenously with 5×106 PBMC isolated from the CAEBV patient 3 ( CD8 type ) or 8 ( NK type ) and were subsequently injected intravenously with 100 µg of the OKT-4 antibody on the same day of transplantation . Additional administration of the antibody was carried out by the same dose and route for the following three consecutive days . Peripheral blood EBV DNA load was then monitored every week . Mice were finally sacrificed four weeks post-transplantation and applied for pathological and virological analyses . Quantification of EBV DNA was carried out by real-time quantitative PCR assay based on the TaqMan system ( Applied Biosystems ) , as described [54] . Analysis of EBV gene expression by RT-PCR was carried out as previously described with the following primers [55] . EBNA1: sense , gatgagcgtttgggagagctgattctgca; antisense , tcctcgtccatggttatcac . EBNA2: sense , agaggaggtggtaagcggttc; antisense , tgacgggtttccaagactatcc . LMP1: sense , ctctccttctcctcctcttg; antisense , caggagggtgatcatcagta . LMP2A: sense , atgactcatctcaacacata; antisense , catgttaggcaaattgcaaa . LMP2B: sense , cagtgtaatctgcacaaaga; antisense , catgttaggcaaattgcaaa . EBER1: sense , agcacctacgctgccctaga; antisense , aaaacatgcggaccaccagc . Cp-EBNA1: sense , cactacaagacctacgcctctccattcatc; anti sense , ttcggtctcccctaggccctg . Wp/Cp-EBNA1: sense , tcagagcgccaggagtccacacaaat; antisense , ttcggtctcccctaggccctg . Qp-EBNA1: sense , aggcgcgggatagcgtgcgctaccgga; antisense , tcctcgtccatggttatcac . RT-PCR primers for β-actin were purchased from Takara ( Osaka , Japan ) . Tissue samples were fixed in 10% buffered formalin , embedded in paraffin , and stained with hematoxylin and eosin . For phenotypic analysis of engrafted lymphocytes , immunostaining for CD3 , CD8 ( Nichirei ) , CD45RO , and CD20 ( DAKO ) was performed on paraffin sections . EBV was detected by in situ hybridization ( ISH ) with EBV small RNA ( EBER ) probe . Immunohistochemistry and ISH were performed on an automated stainer ( BENCHMARK XT , Ventana Medical Systems ) according to the manufacturer's recommendations . To determine the cell lineage of EBV infected cells , paraffin sections were applied to double staining with EBER ISH and immunohistochemistry . Immediately following EBER ISH , immunostaining for CD45RO or CD20 was performed . Photomicrographs was acquired with a OLYMPUS BX51 microscope equipped with 40x/0 . 75 and 20x/0 . 50 Uplan Fl objective lens , a Pixera Penguin 600CL digital camera ( Pixera ) , and Viewfinder 3 . 01 ( Pixera ) for white balance , contrast , and brightness correction . The levels of human IL-8 , IFN-γ , and RANTES in plasma samples were measured with the Enzyme-linked immunosorbent assay ( ELISA ) kit provided by R&D Systems following instructions provided by the manufacturer . The Swiss-Prot accession numbers for the proteins described in this article are as follows: P13501 for RANTES; P10145 for IL-8; P01579 for IFN-γ; P03211 for EBNA1; P12978 for EBNA2; P12977 for EBNA3; P03230 for LMP1; and Q66562 for LMP2 . The DDBJ accession number for EBER is AJ315772 .
|
Epstein-Barr virus ( EBV ) is a ubiquitous human herpesvirus that infects more than 90% of the adult human population in the world . EBV usually infects B lymphocytes and does not produce symptoms in infected individuals , but in rare occasions it infects T or NK lymphocytes and causes severe diseases such as chronic active EBV infection ( CAEBV ) and EBV-associated hemophagocytic lymphohistiocytosis ( EBV-HLH ) . We developed mouse models of these two human diseases in which EBV-infected T or NK lymphocytes proliferate in mouse tissues and reproduce human pathologic conditions such as overproduction of small proteins called “cytokines” that produce inflammatory responses in the body . These mouse models are thought to be very useful for the elucidation of the pathogenesis of CAEBV and EBV-HLH as well as for the development of therapeutic strategies for the treatment of these diseases . Experiments with the models demonstrated that a subset of lymphocytes called CD4-positive lymphocytes are essential for the proliferation of EBV-infected T and NK cells . This result implies that removal of CD4-positive lymphocytes or suppression of their functions may be an effective strategy for the treatment of CAEBV and EBV-HLH .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"hematologic",
"cancers",
"and",
"related",
"disorders",
"medicine",
"infectious",
"diseases",
"myelomas",
"and",
"lymphoproliferative",
"diseases",
"lymphomas",
"viral",
"diseases",
"hematology",
"epstein-barr",
"virus",
"infectious",
"mononucleosis"
] |
2011
|
Novel Mouse Xenograft Models Reveal a Critical Role of CD4+ T Cells in the Proliferation of EBV-Infected T and NK Cells
|
Triggering receptor expressed on myeloid cells-1 ( TREM-1 ) is a potent amplifier of pro-inflammatory innate immune reactions . While TREM-1-amplified responses likely aid an improved detection and elimination of pathogens , excessive production of cytokines and oxygen radicals can also severely harm the host . Studies addressing the pathogenic role of TREM-1 during endotoxin-induced shock or microbial sepsis have so far mostly relied on the administration of TREM-1 fusion proteins or peptides representing part of the extracellular domain of TREM-1 . However , binding of these agents to the yet unidentified TREM-1 ligand could also impact signaling through alternative receptors . More importantly , controversial results have been obtained regarding the requirement of TREM-1 for microbial control . To unambiguously investigate the role of TREM-1 in homeostasis and disease , we have generated mice deficient in Trem1 . Trem1−/− mice are viable , fertile and show no altered hematopoietic compartment . In CD4+ T cell- and dextran sodium sulfate-induced models of colitis , Trem1−/− mice displayed significantly attenuated disease that was associated with reduced inflammatory infiltrates and diminished expression of pro-inflammatory cytokines . Trem1−/− mice also exhibited reduced neutrophilic infiltration and decreased lesion size upon infection with Leishmania major . Furthermore , reduced morbidity was observed for influenza virus-infected Trem1−/− mice . Importantly , while immune-associated pathologies were significantly reduced , Trem1−/− mice were equally capable of controlling infections with L . major , influenza virus , but also Legionella pneumophila as Trem1+/+ controls . Our results not only demonstrate an unanticipated pathogenic impact of TREM-1 during a viral and parasitic infection , but also indicate that therapeutic blocking of TREM-1 in distinct inflammatory disorders holds considerable promise by blunting excessive inflammation while preserving the capacity for microbial control .
Innate immune cells express several cell surface receptors and intracellular sensing molecules that allow for autonomous recognition of pathogen- and danger-associated molecular patterns ( PAMPs and DAMPs ) and initiation of pro-inflammatory anti-microbial reponses . Toll-like receptors ( TLR ) and nucleotide-binding oligomerization domain ( NOD ) -like receptors , which recognize a diverse group of highly conserved microbial structures , represent only two examples of large innate immune receptor families with activating functions . Over the last decade , an additional family of evolutionary conserved innate immune receptors has been identified and characterized , the so-called triggering receptors expressed on myeloid cells ( TREMs ) . TREMs belong to the immunoglobulin ( Ig ) superfamily of receptors and contain both inhibitory and activating receptors [1] , [2] , [3] . In contrast to the fairly ubiquitously expressed TLRs and NOD-like receptors , expression of TREMs is restricted to cells of the myeloid lineage [4] . Moreover , based on their capacity to integrate and potently modulate TLR- and NOD-induced signals , TREMs appear to mainly act as fine-tuners rather than initiators of inflammatory responses [3] , [5] . While TREM-1 , TREM-2 , TREM-3 ( in the mouse ) receptors [4] , [6] , and the TREM-1 like transcripts TLT-1 and TLT-2 have been described [7] , [8] , TREM-1 is the first identified and best characterized receptor of the TREM family with activating functions . TREM-1 consists of an ectodomain , composed of a single Ig V-type domain , a transmembrane region and a short cytoplasmic tail that recruits DAP12 for signaling [4] . TREM-1 is constitutively expressed on neutrophils and on subsets of monocytes and macrophages , and TREM-1 expression is further upregulated upon exposure of cells to microbial products [4] . Whereas crosslinking of TREM-1 with agonistic antibodies alone induces only modest cellular activation , TREM-1 potently synergizes with distinct TLR ligands for a substantial amplification of oxidative burst and production of pro-inflammatory mediators such as TNF , IL-1β , IL-6 , IL-8 , MCP-1 and Mip-1α [4] , [9] , [10] . In vivo , the role of TREM-1 has been mostly characterized in experimental models of endotoxin-induced shock or microbial sepsis where blockade of TREM-1 signaling conferred significant protection [9] , [11] , [12] . The detection of TREM-1 in inflammatory lesions caused by bacterial or fungal agents , but not in psoriasis or immune-mediated vasculitis [9] , has further led to the general concept that TREM-1 is primarily involved in microbial diseases , particularly , since elevated levels of the serum soluble form of the shed TREM-1 surface receptor ( sTREM-1 ) also appear to associate with bacterial infections in patients with pneumonia or suspected sepsis [13] , [14] . However , increasing evidence is now emerging that TREM-1 may additionally play a role in non-infectious inflammatory conditions . Thus , expression of TREM-1 can also be induced by the non-microbial agent monosodium urate monohydrate crystals ( MSU ) or by hypoxic cell culture conditions in vitro [15] , [16] . Augmented sTREM-1 levels have been reported for patients with rheumatoid arthritis , acute pancreatitis , chronic obstructive pulmonary disease and cardiac arrest [17] , [18] , [19] , [20] . Furthermore , we have previously described an involvement of TREM-1 in human inflammatory bowel diseases ( IBD ) and in models of experimental colitis [21] , [22] , [23] . Investigations on the precise function of TREM-1 in distinct diseases have so far been complicated by the still unidentified ligand ( s ) for TREM-1 . Putative ligands for TREM-1 have been described on the surface of human platelets and on murine granulocytes during experimental peritonitis and endotoxaemia [12] , [24] , [25] . In addition , necrotic cell lysates also appear to stimulate pro-inflammatory responses in a TREM-1-dependent manner , which may relate to association of TREM-1 with the High Mobility Group Box 1 ( HMGB1 ) protein [26] , [27] . Hence , it can be speculated that not only PAMPs but also DAMPs induce signaling via TREM-1 and that several ligands for TREM-1 may exist . In the absence of clearly defined ligands for TREM-1 , studies addressing the impact of TREM-1 in disease have so far mostly relied on the use of TREM-1/Ig fusion proteins or synthetic peptides mimicking part of the extracellular domain of TREM-1 . Although by the use of these agents substantial protection from endotoxin-induced shock , microbial sepsis or experimental colitis could be conferred [9] , [11] , [12] , [22] , several aspects regarding the true biological role of TREM-1 remain unclear . First , considering the redundancy of innate immune receptor-ligand interactions , the possibility exists that in these previous studies not only signaling through TREM-1 but through additional , potentially more relevant receptors was prevented . Second , controversial findings have been obtained with respect to the impact of impaired TREM-1 signaling on microbial control [9] , [28] , [29] , [30] . In order to investigate the role of TREM-1 in homeostasis and disease , we have generated a TREM-1-deficient ( Trem1−/− ) mouse by targeted deletion of exon 2 . Here we show , employing distinct inflammation and infection models ranging from experimental colitis to infections with Leishmania major , influenza virus and Legionella pneumophila , that complete absence of TREM-1 significantly attenuates morbidity and immune-mediated pathologies while microbial control remains unimpaired . These findings not only demonstrate an unanticipated clear role for TREM-1 in chronic inflammatory disorders , parasitic and viral infections , but also illustrate the potential for a novel therapeutic intervention in various disease settings .
To account for potential embryonically lethal effects of a total deletion of the Trem1 gene , a targeting vector was designed for conditional deletion of exon 2 ( Fig . S1 ) . Exon 2 encodes the extracellular domain of TREM-1 and also contains the putative ligand binding site [31] . Breeding of Trem1+/flox chimeric offspring mice with deleter mice that expressed Cre ubiquitously yielded viable Trem1+/− x Cre+/− offspring . Moreover , interbreeding of Trem1+/− mice gave rise to Trem1−/− mice at the expected Mendelian frequencies , and Trem1−/− mice were equal in size , weight and fertility to littermate Trem1+/+ controls . We thus continued to characterize Trem1−/− mice with a ubiquitously deleted Trem1 gene by elementary flow cytometry analyses . Deletion of Trem1 resulted in a gene-dose-dependent loss of TREM-1 surface expression by peripheral blood neutrophils and Ly6Clo monocytes ( Fig . 1 ) . Accordingly , TREM-1 was still expressed at ∼2-fold reduced levels in Trem1+/− mice , while surface TREM-1 expression was absent on myeloid cells in Trem1−/− mice ( Fig . 1 ) . Absence of Trem1 did not appear to affect the composition of various immune compartments , since numbers of distinct myeloid and lymphoid cell subsets isolated from the peripheral blood , bone marrow ( BM ) and spleen of Trem1−/− and Trem1+/+ mice were identical ( Fig . S2 ) . However , to formally exclude a potential effect of TREM-1 on hematopoiesis , the BM of Trem1−/− and Trem1+/+ mice was analysed in more depth with respect to hematopoietic stem cell and myeloid progenitor numbers following lineage depletion and depletion of lymphoid progenitors ( Fig . 2 ) . Stem cell-enriched cells were identified by their lineage− ( lin− ) Sca-1+ c-kithi phenotype ( LSK cells ) while common myeloid progenitors ( CMP ) , granulocyte/macrophage progenitors ( GMP ) and megakaryocyte/erythrocyte precursors ( MEP ) were discriminated within the Sca-1− c-kithi population according to their differential expression of FcγR and CD34 , respectively ( Fig . 2a ) . Compared to Trem1+/+ mice , Trem1−/− mice exhibited equal numbers of LSK cells , CMP , GMP and MEP ( Fig . 2B ) . Moreover , similar numbers of colony forming units could be observed in lineage-depleted ( lin− ) BM cells isolated from Trem1−/− mice ( Fig . 2B ) . Although these analyses indicated again that TREM-1 is unlikely to play a substantial role in hematopoietic processes , we were intrigued by the selective expression of surface TREM-1 by GMP , but not by CMP ( Fig . 2A ) . As a final measure , we therefore established mixed bone marrow chimeras with either Trem1−/− ( x GFP−/− ) and Trem1+/+ x GFP+/+ BM cells or , as a control , Trem1+/+ ( x GFP−/− ) and Trem1+/+ x GFP+/+ BM cells . Analysis of chimeric mice at 10 and 31 weeks post reconstitution and calculation of the respective ratios of GFP− to GFP+ peripheral blood neutrophil , Ly6Chi or Ly6Clo monocyte numbers demonstrated an equal capacity of Trem1−/− BM to give rise to distinct myeloid subsets as Trem1+/+ BM . Thus , while the potential role of TREM-1 expression by GMP still remains to be explored , deficiency in Trem1 does not appear to affect hematopoietic processes under homeostatic conditions . We next addressed whether absence of Trem1 could affect other receptors that use DAP12 for signaling , either by the potential presence of increased levels of intracellularly available DAP12 or by the lack of counterregulatory signals conferred by TREM-1 . Indeed , the hyperresponsive phenotype of DAP12-deficient macrophages is largely ascribed to a lack of inhibitory signals by TREM-2 which also employs DAP12 [32] . Due to the important role of TREM-2 in osteoclast formation and function [33] , [34] , we reasoned that lack of TREM-1 expression in Trem1−/− mice could possibly manifest in altered osteoclastogenesis . However , as determined by Xray and MicroCT analyses , no differences in bone density could be detected between Trem1−/− mice and their age- and sex-matched Trem1+/+ controls ( data not shown ) . Taken together , these analyses revealed no apparent phenotype of Trem1−/− mice under homeostatic conditions . We have previously demonstrated a substantial accumulation of TREM-1 expressing macrophages in the inflamed , but not healthy intestinal mucosa of patients with IBD and of mice with experimental colitis [22] , [23] . Hence , one of our major interests in the characterization of the Trem1−/− mouse was to unambigously investigate the role of Trem1 in the pathogenesis of IBD . To this end , CD4+ CD25− CD45RBhi T cells were adoptively transferred into Helicobacter-positive Trem1+/+ x Rag2−/− and Trem1−/− x Rag2−/− recipient mice and animals were monitored regularly for clinical signs of colitis . Importantly , whereas Trem1+/+ x Rag2−/− mice lost ∼20% of their initial body weight at the end of the observation period , weight loss in Trem1−/− x Rag2−/− mice was minimal and only transient ( Fig . 3A ) . Furthermore , shortening of the colon was substantially attenuated in Trem1−/− x Rag2−/− mice compared to controls ( Fig . 3B ) . While some of the Trem1−/− x Rag2−/− mice still exhibited moderate histopathological signs of intestinal inflammation , individual parameters of the histopathological scoring as well as the overall histopathological score were significantly reduced ( Fig . 3C–E ) . In order to gain insight in the potential underlying mechanism of the highly attenuated colitis in Trem1−/− x Rag2−/− mice , colonic lamina propria cells that were isolated from both groups of mice in the absence of an adoptive CD4+ T cell transfer ( healthy colon ) or 12–13 days post colitis induction were analysed by FACS . As depicted in Fig . 4A , the colonic lamina propria of healthy Trem1+/+ x Rag2+/+ mice and Trem1−/− x Rag2−/− mice contained a similar proportion of CD11b+ MHCIIhi cells and Gr1+ cells were virtually absent . In contrast , Gr1+ cells , representing infiltrating Ly6Chi Gr1int monocytes and Ly6Cint Gr1hi neutrophils , were readily detected in Trem1+/+ x Rag2+/+ mice and Trem1−/− x Rag2−/− mice at 12–13 days post colitis induction ( Fig . 4A ) . Notably , the relative frequency of Gr1+ cells among CD45+ CD11b+ colonic LP cells was ∼5-fold lower in Trem1−/− x Rag2−/− mice ( Fig . 4A ) . In further contrast to the colonic LP of healthy mice , among LP MHCII+ Gr1− cells of colitic mice two populations of MHCIIhi Ly6Clo and MHCIIint Ly6Chi cells could be discriminated , likely representing intestinal macrophages and monocytes in the process of differentiation ( Fig . 4A ) [35] , [36] . Also within this gate of MHCII+ Gr1− cells , substantial differences could be detected between the two groups of mice . Accordingly , the relative frequency of MHCIIint Ly6Chi cells was increased in colitic Trem1+/+ x Rag2−/− mice whereas Trem1−/− x Rag2−/− mice exhibited a larger proportion of MHCIIhi Ly6Clo macrophages ( Fig . 4A ) . When colonic LP cells of n = 9 mice of both groups were systematically analysed at 12–13 days post colitis induction , substantially reduced numbers of various cell subsets could be seen in Trem1−/− x Rag2−/− mice ( Fig . 4B ) . Thus , Trem1−/− x Rag2−/− mice not only exhibited reduced infiltrating CD4+ T cells but also significantly decreased numbers of neutrophils , Ly6Chi monocytes and MHCIIint Ly6Chi cells ( Fig . 4B ) . These differences were not apparent for the colonic LP of healthy Trem1+/+ x Rag2−/− and Trem1−/− x Rag2−/− mice which mainly contained MHCIIhi Ly6Clo cells anyway ( Fig . 4A and 4C ) . To gain more insight which myeloid TREM-1-expressing cell subset could potentially be involved in driving intestinal inflammation in Trem1+/+ x Rag2−/− mice , TREM-1 surface expression was analysed on colonic LP neutrophils , Ly6Chi monocytes as well as CD11b+ Ly6C+ Gr1− and CD11b+ Ly6C− Gr1− cells . As reported previously [23] , [37] , in the healthy colonic LP TREM-1 expression was hardly detectable owing to the absence of infiltrating neutrophils and Ly6C+ cells ( Fig . 4A and 4D ) . In colitic Trem1+/+ x Rag2−/− mice , TREM-1 expression was observed on neutrophils , Ly6Chi monocytes and CD11b+ Ly6C+ Gr1− cells ( Fig . 4D ) . Moreover , CD11b+ Ly6C− Gr1− cells , likely representing intestinal macrophages , that were isolated from colitic Trem1+/+ x Rag2−/− mice exhibited a ∼3-fold upregulated expression of surface TREM-1 ( Fig . 4D ) . In line with the reduced infiltrating cell numbers , mRNA expression for various innate and adaptive pro-inflammatory chemokines and cytokines was significantly decreased in the lamina propria of Trem1−/− x Rag2−/− compared to Trem1−/− x Rag2−/− mice ( Fig . 4E ) . While protection from colitis in Trem1−/− x Rag2−/− mice was associated with reduced expression of several pro-inflammatory mediators , previous data generated in our laboratory have demonstrated that TNF produced by nonlymphoid cells plays a non-redundant pathogenic role in the CD4+ T cell transfer model of colitis since Tnf−/− x Rag2−/− mice are completely protected from colitis induction [38] . However , in acute models of intestinal inflammation such as the DSS-induced colitis , Tnf−/− mice exhibit aggrevated disease [39] , [40] , presumably , because early anti-microbial and repair responses following DSS-induced breaching of the epithelial barrier are fundamentally impaired . Due to the central function of TREM-1 in amplifying pro-inflammatory cytokine production and oxidative burst , we hypothesized that during acute intestinal inflammation complete absence of Trem1 could also prove detrimental . Intriguingly , although upon administration of 3% DSS Trem1−/− mice initially lost weight to a similar extent as Trem1+/+ mice , weight loss was considerably attenuated at 7 days post colitis induction and at 9 days Trem1−/− mice had already improved again ( Fig . 5A ) . In Trem1−/− mice , shortening of the colon was markedly attenuated and total histopathological colitis scores were significantly decreased ( Fig . 5B and 5C–E ) . Furthermore , analogous to the CD4 T cell-induced colitis ( Fig . 4E ) , TREM-1 deficiency resulted in reduced colonic mRNA expression of several pro-inflammatory mediators ( Fig . 5F ) . Hence , in contrast to Tnf−/− mice [39] , [40] , Trem1−/− mice still showed an adequate host response to DSS-induced epithelial injury while exhibiting reduced immune-mediated pathologies . The observations made in the acute DSS model of colitis raised our interest whether Trem1−/− mice would also be able to control bona fide microbial infections , in particular , since maximal silencing of TREM-1 by a siRNA approach had proven deleterious in a fecal peritonitis model [29] . Since the rapid kinetics of this model hardly allowed to simultaneously look at beneficial effects of the Trem1 deficiency on immune-mediated tissue damage or to assess potential adverse consequences for the priming of adaptive immune responses , we chose the Leishmania major infection model . Following infection with L . major , C57BL/6 mice develop local cutaneous lesions that spontaneously resolve within 4–8 weeks . Central to the resolution is the TNF-mediated control of the early inflammatory response or the clearance of neutrophils and the later IFNγ-mediated and Th1-driven elimination of the parasite by infected macrophages [41] , [42] . Upon injection of 3×106 L . major promastigotes s . c . in the footpad of Trem1+/+ and Trem1−/− mice , an attenuation in lesion development was apparent in Trem1−/− mice already at 14 days post infection . From thereof , Trem1−/− mice showed a significantly decreased lesion size ( Fig . 6A ) . Notably , however , parasite counts did not differ between Trem1−/− and Trem1+/+ mice ( Fig . 6B ) . We further analysed the cellular composition of infected footpads from Trem1−/− and Trem1+/+ mice at 21 days post infection . While the overall cell counts were comparable , the cellular infiltrate in Trem1−/− mice exhibited ∼3-fold reduced neutrophil numbers ( Fig . 6C ) . To look at the potential impact of the Trem1-deficiency on the priming of Th1 responses , expression of IFNγ was analysed in cells isolated from the draining lymph node . The frequency of IFNγ-secreting CD4+ T cells was similar in Trem1−/− and Trem1+/+ mice ( Fig . 6D ) . In addition , comparable levels of IFNγ were detected in cells of both groups of mice upon re-stimulation in vitro with the parasitic antigen ( Fig . 6D ) . These data are in line with comparable parasite killing observed in Trem1−/− and Trem1+/+ mice . Thus , in the L . major infection model , absence of Trem1 does not appear to have adverse consequences on the priming of adaptive immune responses and on parasite control while neutrophil-mediated inflammatory lesion development is substantially reduced . The reduced neutrophil numbers at L . major-infected sites and the decreased lesion size in Trem1−/− mice agree with the notion that neutrophils play a central role in inflammatory lesion development . Indeed , the presence of non-healing lesions in L . major susceptible BALB/c strains is associated with elevated numbers of neutrophils [43] . Since neutrophils constitutively express high levels of TREM-1 ( Fig . 1B , 4D ) , we aimed to investigate the consequences of TREM-1-mediated stimulation on their functional responses in more detail . Analysis of mouse neutrophils has so far been complicated by the limited numbers of cells that can be retrieved from the peripheral blood , their short life span or the distinct differentiation stages of BM-derived neutrophils . Hence , we made use of a recently described system by which neutrophils can be differentiated ex vivo in large numbers using conditional Hoxb8 [44] . Using a slightly modified protocol , BM-derived progenitors were lentivirally transduced with conditional Hoxb8 in the presence of SCF , resulting in immortalized myeloid progenitor lines , termed SCF-condHoxb8 . Upon shutdown of Hoxb8 expression by withdrawal of 4-OHT , cells differentiate into mature neutrophils in the presence of SCF . As shown in Figure 7A and 7B , withdrawal of 4-OHT in fact induced the appearance of cells bearing the characteristic phenotype of mouse neutrophils after in vitro differentiation for 5 days . Moreover , Trem1+/+ SCF-condHoxb8-derived mature neutrophils also expressed distinct levels of surface TREM-1 ( Fig . 7B ) . Stimulation of Trem1+/+ , but not of Trem1−/− , SCF-condHoxb8-derived mature neutrophils with a plate-bound agonistic anti-TREM-1 antibody resulted in pronounced mRNA expression of iNOS and TNF ( Fig . 7C ) and induced rapid secretion of TNF at levels comparable to those triggered by LPS ( Fig . 7D ) . Since neutrophil survival versus apoptosis could represent a deciding factor in the control of inflammation not only in the L . major infection model but also in the pathogenesis of experimental colitis [45] , we compared the susceptibility of Trem1+/+ and Trem1−/− neutrophils to spontaneous apoptosis . Following prolonged culture of fully differentiated SCF-condHoxb8-derived neutrophils in vitro , an increasing and comparable frequency of AnnexinV+DAPI+ cells was detected for Trem1+/+ and Trem1−/− neutrophils ( Fig . 7E ) . However , in the presence of TREM-1-mediated stimulation a reduced apoptosis rate based on diminished Caspase 3/7 activity could be observed for Trem1+/+ neutrophils ( Fig . 7F ) . Furthermore , agonistic anti-TREM-1 stimulation of Trem1+/+ but not Trem1−/− neutrophils resulted in ∼2-fold upregulated mRNA expression of Myeloid Cell Leukemia-1 ( Mcl-1 ) ( Fig . 7G ) , analogously to what has recently been described for human TREM-1-stimulated monocytes [46] . Thus , TREM-1-mediated activation of mouse neutrophil appears to contribute to their survival . Intrigued by the substantially diminished inflammatory lesions , yet intact parasite clearance in L . major-infected Trem1−/− mice , we aimed to substantiate these findings in an altogether different infection model . Since high expression of TREM-1 by alveolar macrophages and previously published data [28] , [30] , [47] , [48] suggest a potential role for TREM-1 in lung inflammatory responses , we infected Trem1+/+ and Trem1−/− mice intratracheally with 50 PFU influenza A virus PR8 . Hypothermia and body weight loss , which are characteristically associated with experimental influenza virus infection , were observed in Trem1+/+ mice at 7 days post infection ( Fig . 8A and 8B ) . While the body temperature also dropped in Trem1−/− mice , a quicker recovery from hypothermia was seen in the Trem1−/− group ( Fig . 8A ) . Moreover , weight loss in Trem1−/− mice was significantly attenuated and Trem1−/− mice further exhibited reduced levels of IL-6 in bronchoalveolar lavage fluid ( BALF ) at 10 days post infection ( Fig . 8B and 8C ) . Notably , in spite of the reduced morbidity observed , Trem1−/− mice were equally competent in clearing the influenza virus infection as Trem1+/+ controls ( Fig . 8D ) . After having established that deficiency in TREM-1 attenuates disease but does not impair pathogen control during a parasitic and viral infection , we last sought to address the significance of TREM-1 in a bacterial infection model . Indeed , controversial results have been obtained with respect to the importance of TREM-1 in microbial control following infection of experimental animals with Pseudomonas aeruginosa [28] , [30] . Here , we employed a Legionella pneumophila infection model which also causes severe upper airway inflammation in permissive mice and critically depends on neutrophil-mediated control [49] , [50] , [51] . As shown in Figure 9 , at 3 days post infection with 5×106 CFU of L . pneumophila , CFU and neutrophil numbers in the BALF did not significantly differ between Trem1+/+ and Trem1−/− mice . Furthermore , at day 5 post infection CFU were reduced to < 500/BALF in both groups of mice with no significant differences observed between Trem1+/+ and Trem1−/− mice ( data not shown ) .
The significance of TREM-1 as a central amplifier of acute pro-inflammatory responses during endotoxin-induced shock and microbial sepsis is well established . However , increasing evidence , including the recently reported association of TREM-1 with the DAMP protein HMGB1 [26] , [27] , also suggests a potential role for TREM-1 during non-infectious and chronic inflammatory conditions . In line with this notion , we have previously described a crucial involvement of TREM-1 in IBD as based on the significant amelioration of experimental colitis upon blockade of TREM-1 with the antagonistic LP17 peptide [22] . Blocking TREM-1 signaling by daily administration of TREM-1-Ig fusion proteins or synthetic analogues in chronic disease models is not only costly and straining but also fails to cover for the possibility that the yet unidentified TREM-1 ligand may signal through alternative receptors . Here , we have generated a Trem1−/− mouse to unambiguously investigate the impact of a complete TREM-1-deficiency on the pathogenesis of experimental colitis but also of two other distinct sub-acute disease settings where the role of TREM-1 has so far not been addressed in vivo , i . e . inflammation induced by a parasitic and viral infection . Our findings demonstrate that Trem1−/− mice not only show a highly attenuated CD4+ T cell- and DSS-induced colitis but also display significantly reduced lesion size and diminished morbidity during infections with L . major and influenza virus , respectively . The substantial attenuation of illness and immune-mediated pathologies in Trem1−/− mice across these distinct models suggests a common mechanism by which TREM-1 signaling promotes inflammation irrespective of the original trigger . Several non-exclusive scenarios can be considered that may account for the attenuated disease in Trem1−/− mice: A priori reduced chemotactic recruitment of Trem1−/− neutrophils and monocytes , diminished pro-inflammatory activities , and/or reduced life-span of myeloid cell subsets . Although we observed significantly decreased numbers of distinct myeloid cell subsets in the LP of colitic Trem1−/− x Rag2−/− mice and at L . major-infected sites in Trem1−/− mice , we consider it unlikely that deficiency in TREM-1 causes an intrinsic primary defect in chemotaxis . When we analysed L . major infected sites at an early time-point , i . e . 3 days post infection , no differences in cellularity were detected between Trem1+/+ and Trem1−/− mice ( data not shown ) . Moreover , a recent study clearly demonstrated that TREM-1/3 proteins are not required for transendothelial migration of neutrophils [30] . Nonetheless , the markedly decreased expression of mRNA for monocyte- ( CCL2 ) , granulocyte- ( CXCL1 , CXCL2 , CXCL5 ) but also T cell ( CXCL9 ) chemoattractants in the LP of Trem1−/− x Rag2−/− mice will in a secondary manner certainly have contributed to the decreased accumulation of inflammatory cells . Besides the diminished expression of chemotactic mediators , the colonic LP of colitic Trem1−/− x Rag2−/− mice also exhibited substantially reduced mRNA levels of several innate cytokines , including IL-1β , IL-6 and TNF . The reduced expression of pro-inflammatory mediators in the entire colonic LP in Trem1−/− x Rag2−/− mice at the late stage of colitis induction certainly also mirrors the decreased cellular infiltration . However , considering the capacity of TREM-1 to augment the production of several chemotactic mediators either directly in infiltrating myeloid cells or likely also indirectly , in a paracrine manner ( e . g . through secretion of IL-1β ) [4] , [22] , [52] , we believe that TREM-1-amplified production of pro-inflammatory cytokines represents a key early pathogenic event that will ultimately determine the later disease course and may largely account for the attenuated disease in Trem1−/− mice . In this respect , it is noteworthy that the colonic LP of colitic Trem1−/− mice contained markedly fewer MHCIIint Ly6Chi cells or inflammatory macrophages with the capacity for expression of pro-inflammatory mediators [35] , [36] . As we have employed a CD4+ T cell-dependent colitis model and indeed observed considerably reduced CD4+ T cell numbers and correspondingly also mRNA levels for IFNγ and IL-17 in the colonic LP of transferred Trem1−/− x Rag2−/− mice , the question arises whether deficiency in TREM-1 may directly impact the priming of adaptive immune responses . Whereas in the colitis model we have not analysed CD4+ T cell responses in more detail , CD4+ T cells isolated from L . major-infected and CD8+ T cells retrieved from influenza virus-infected Trem1−/− mice , respectively , exhibited an unimpaired capacity for IFNγ production compared to T cells from Trem1+/+ mice . The deciding role of neutrophils in the L . major infection model [42] and the substantially decreased lesion size in Trem1−/− mice have prompted us to investigate the impact of TREM-1 on neutrophil-mediated functions in more detail . In particular , we were interested in the potential modulatory effect of TREM-1 ligation on neutrophil survival as delayed neutrophil apoptosis could also represent a critical pathogenic factor in intestinal inflammation [45] . In agreement with a previous report [30] , deficiency in TREM-1 caused no intrinsic predisposition for increased spontaneous neutrophil apoptosis . However , agonistic TREM-1 stimulation significantly promoted survival of SCF-condHoxb8 progenitor-derived neutrophils in vitro . We have also sought to assess the frequency of apoptotic neutrophils in Trem1+/+ versus Trem1−/− mice in situ by staining colonic tissue sections from colitic mice for cleaved caspase 3 ( data not shown ) . However , due to the likely very rapid clearance of apoptotic neutrophils by intestinal phagocytes , cleaved caspase 3 positive cells were rare in Trem1+/+ mice and even more scarce in Trem1−/− mice . Moreover , the substantially decreased cellular infiltrate in Trem1−/− mice further precluded an objective comparison and quantification of apoptotic cells in situ . While we cannot present definitive evidence that TREM-1 prolongs neutrophil survival in vivo , we still believe that the reduced presence of neutrophils in Trem1-deficient mice in the CD4 T cell induced colitis and L . major infection model relates to both: A reduced secondary recruitment based on the diminished expression of neutrophil chemotactic mediators and to a reduced life-span in the absence of TREM-1 ligation , with the former mechanism numerically perhaps being more relevant . Considering the various effector cell types and mechanisms by which TREM-1-mediated stimulation could contribute to disease , it may appear intriguing that deficiency in TREM-1 did not completely protect from disease . Accordingly , the degree of protection from colitis was not much higher in Trem1−/− mice compared to mice that were treated with the antagonistic LP17 peptide in our previous study [22] . We hypothesize that the absence of complete protection in Trem1−/− mice may primarily relate to the role of TREM-1 as an amplifier but not inducer of pro-inflammatory reactions [3] , [5] . While we cannot rule out a potential participation of TREM-3 , we believe that the protective effects seen in Trem1−/− mice are too substantial for a major involvement of TREM-3 in the inflammatory models analysed . One of the most striking findings of the present study was the observation that microbial control in the models analysed was apparently not impaired in Trem1−/− mice in spite of the blunted inflammatory responses . Hence , while Tnf−/− or anti-TNF-treated mice exhibit an aggrevated acute DSS-induced colitis [39] , [40] and also show enhanced parasite and bacterial burdens upon infection with L . major and L . pneumophila , respectively [41] , [53] , Trem1−/− mice appeared equally capable of controlling a parasitic , viral and bacterial infection as Trem1+/+ controls . This observation is in line with the main function of TREM-1 as an inflammatory fine-tuner which still allows for pro-inflammatory TLR or NOD-like receptor-induced reactions in its absence . Moreover , TREM-1 does not appear to be involved in phagocytic or direct antimicrobial activity of myeloid cells [30] , [54] , [55] . Still , conflicting data on the effect of a TREM-1 blockade on microbial control have been reported from various bacterial challenge models . Injection of a TREM-1/IgG fusion protein allowed for sufficient control of an E . coli-induced peritoneal infection and conferred protection [9] whereas maximal but not half-dose siRNA silencing of TREM-1 increased mortality in a fecal peritonitis model [29] . Similarly , administration of the antagonistic LP17 peptide protected rats from a P . aeruginosa-induced pneumonia [28] , whereas complete deficiency in TREM-1/3 led to markedly increased mortality in Pseudomonas aeruginosa-challenged mice due to defective transepithelial migration of neutrophils [30] . It has been proposed that the degree of TREM-1 blockade was a likely critical parameter to account for these disparant findings [29] , [30] . However , our findings demonstrate that microbial control must not necessarily be impaired in mice with a complete deficiency in TREM-1 , even when employing a L . pneumophila infection model where early neutrophil accumulation is also crucial for bacterial clearance [49] , [50] , [51] . Thus , we speculate that analogous to the divergent results obtained for endotoxin-challenged DAP12-deficient mice [56] , [57] , possibly the infection dose , the nature of the microbial agent and/or the kinetics of the infection may be critical parameters regarding the requirement for TREM-1 . Accordingly , in the presence of low abundance and/or low affinity TLR ligands or fast replicating agents , the inflammatory response may more stringently depend on TREM-1-amplified signaling . Alternatively , the expression of TREM-1 and its association with DAP12 may only preferentially be induced in situations of high abundance and/or high affinity TLR ligands whereas low level signaling would favour the engagement of TREM-2 . In summary , while the impact of TREM-1 on microbial control still needs further investigations across different experimental models , our extensive characterisation of Trem1−/− mice shows an unanticipated prominent role for TREM-1 in parasitic and viral infections . Our findings thus suggest that therapeutic targeting of TREM-1 holds considerable promise for distinct non-infectious and infectious inflammatory disorders and may bypass the increased risk for impaired microbial control which is associated with the general targeting of TNF .
Breedings and cohort maintenance were performed under SPF conditions in isolated ventilated cages in the central animal facility of the Medical School , University of Bern . All studies were conducted with age- and sex-matched animals . Trem1+/+ mice that were used as wildtype controls were originally derived from the interbreeding of Trem1+/− x Cre+/− mice , thus carrying the identical C57BL/6 genetic background and representing former littermates of Trem1−/− mice . To nonetheless adjust for potential differences in the microflora composition in Trem1−/− vs . Trem1+/+ mice , the bedding of Trem1+/+ and Trem1−/− mice was regularly exchanged between cages three weeks prior to the start of the experiments . All animal experiments were approved by the Veterinary Offices of the Cantons of Bern , Lausanne and Zurich and performed in compliance with Swiss laws for animal protection . The generation of Trem1-deficient mice was designed and carried out in collaboration with the TaconicArtemis GmbH ( Köln , Germany ) . To account for potential lethal effects of a total deletion of the Trem1 gene and to allow for a possible cell-specific ablation of Trem1 expression , a targeting vector was designed for conditional deletion of exon 2 , which encodes the extracellular part of Trem1 and also contains the putative ligand binding site [31] . As illustrated in the supplemental material Fig . S1 , the targeting vector was constructed on the basis of the cloning vector KS loxP ftr Neo BS to flank exon 2 with loxP sites , to comprise additional restriction sites ( AseI and AvaI ) and to contain PuroR ( flanked by F3 sites ) and Neomycin ( flanked by FRT sites ) positive selection marker cassettes to control for homologous recombination upstream and downstream of exon 2 , respectively . For counterselection , a Tk cassette was included downstream of exon 4 . As a template for the PCR reactions a BAC-based plasmid containing the entire genomic mouse Trem1 locus ( RP23-32N8 ) was obtained from BACPAC Resources Center BPRC ( Oakland , USA ) . The targeting vector was electroporated into a C57BL/6N . tac embryonic stem cell line ( TaconicArtemis ) . On day 2 , cells were selected with Puromycin and G418 and on day 8 counterselection with Gancyclovir was initiated . Isolated and expanded ES clones were screened for complete integration of the targeted allele by standard Southern blotting analyses with probes located upstream of exon 1 ( 5e2 ) or exon 3 ( ila1 ) ( Fig . S1 ) . Primer sequences for generation of the 5e2 Southern probes were: CGGATTTGACCAGGAATGACAG ( sense ) and CTTCCAGTTCATTCATGGACAGC ( antisense ) and for the ila1 Southern probe: AGCTCCTCTTGTCTGCCATTCAAGGC ( sense ) and GGCTACAACCTTGTTCTGCAG ( antisense ) . Eight positive clones could be identified and the ES clone A-A5 was subsequently injected into Balb/c derived blastocytes which were then transferred to pseudopregnant NMRI females . Chimeric offspring were bred to C57BL/6 female mice ( C57BL/6-Tg ( ACTB-Flpe ) 2Arte , TaconicArtemis ) transgenic for Flp recombinase to achieve deletion of the FRT and F3 flanked selection cassettes PuroR and Neomycin , respectively . Germline transmission of the targeted Trem1 allele was identified by coat color contribution and by PCR using oligo 1_sense ( GTGCTCAGAGAATGTCTTTGTATCC ) and oligo 4_antisense ( CCCTGGTCAGACCATTTACC ) which either yield a 1 . 3 kb fragment for the wildtype ( WT ) allele or a 1 . 6 kb fragment for the conditional Trem1flox allele . ( Fig . S1 ) . Cycling conditions were: 5′ at 95°C followed by 35 cycles consisting of 30″ at 95°C , 30″ at 60°C , 1′ at 72°C , followed by 10′ at 72°C . Thus identified Trem1+/flox mice ( C57BL/6-TREM-1tm1821_33 . 1Arte ) were mated with male mice carrying the Cre recombinase under the control of the ROSA26 locus ( C57BL/6-Gt ( ROSA ) 26Sortm16 ( Cre ) arte , TaconicArtemis ) to obtain systemic deletion of one Trem1 allele ( Trem1+/− ) . Trem1+/− x Cre+/− mice were interbred to achieve deletion of Cre and to obtain wildtype controls ( Trem1+/+ ) and heterozygous ( Trem1+/− ) and homozygous ( Trem1−/− ) Trem1-deficient mice . Deletion of exon 2 in Trem1+/− and Trem1−/− mice was assessed by the genotyping PCR strategy described above and depicted in Fig . S1 . Trem1+/+ and Trem1−/− mice were subsequently expanded for experiments . For the CD4 adoptive transfer model of colitis , Trem1−/− x Rag2−/− mice were generated by crossing Trem1−/− mice with Helicobacter+ Rag2−/− mice and interbreeding of the F1 offspring . The Helicobacter+ status of the Trem-1−/− x Rag2−/− offspring was confirmed by PCR testing ( Microbios GmbH , Reinach , Switzerland ) . The following mAbs were used: anti-mouse CD11b-Pacific Blue ( M1/70 ) , CD45-Pacific Blue ( 30-F11 ) , CD45-Brilliant Violet570 ( 30-F11 ) , CD4-APC-Cy7 ( RM4-5 ) , Gr1-PE ( RB6-8C5 ) , NK1 . 1-PE-Cy7 ( PK136 ) , Ly6G-APC-Cy7 ( 1A8 ) and F4/80-biotin ( BM8 ) IL-7Rα-biotin ( A7R34 ) , CD3ε-biotin ( 145-2C11 ) CD19-biotin ( 6D11 ) , Gr1-biotin ( RB6-8C5 ) and Ter119-biotin ( TER119 ) , all purchased from Biolegend; anti-mouse CD115-PE ( AFS98 ) , Gr1-APC ( RB6-8C5 ) , CD45-eFluor605 ( 30-F11 ) , CD45 . 1-PE ( A20 ) , MHCII-APC ( M5/114 . 15 . 2 ) , CD11c-PE ( N418 ) , CD11b-eFluor450 ( M1/70 ) , Streptavidin-PE-Cy7 , were purchased from eBioscience ( San Diego , USA ) ; anti-mouse Ly6C-FITC ( AL-21 ) from BD Pharmingen ( San Diego , USA ) and anti-mouse TREM-1-APC ( 174031 ) from R&D Systems . DAPI ( Invitrogen ) was used in a final concentration of 0 . 5 µg/ml to exclude dead cells . Prior to FACS stainings , Fc receptors were blocked using supernatant from the hybridoma 2 . 4G2 . Cells were acquired on a LSRII SORP ( BD Biosciences , San Diego , USA ) and analysed using FlowJo cytometric analysis program ( Tree Star , Ashland , USA ) . To assess the presence of histopathological alterations on formalin-fixed , paraffin-embedded and hematoxylin-eosin-stained colonic tissue sections , a scoring system ranging from 0 ( no alterations ) to 15 ( most severe signs of colitis ) was established , including the following parameters: cellular infiltration ( 0–3 ) , loss of goblet cells ( 0–3 ) , crypt abscesses ( 0–3 ) , epithelial erosions ( 0–1 ) , hyperemia ( 0–2 ) , thickness of the colonic mucosa ( 0–3 ) . Histological scoring was performed by a pathologist ( V . G . ) blinded to sample identity . RNA was isolated using RNA isolation reagent ( Tri-Reagent , Molecular Research Center ) . DNA was digested using DNase I ( Ambion ) , and cDNA was generated using High Capacity cDNA Reverse Transcription Kit ( Applied Biosystems ) . Expression of genes was analysed using Qiagen Quantitect Primer Assays on a 7500 Real-time PCR System ( AB Biosystems ) . The house keeping gene GAPDH was used for normalization of gene expression . Cell-free supernatants derived from SCF−cond Hoxb8 neutrophils and BAL fluid from influenza virus infected mice were analysed by ELISA ( Biolegend ) . All data were analysed with GraphPad Prism software using the Student's t-test or 2-way ANOVA .
|
Triggering receptor expressed on myeloid cells-1 ( TREM-1 ) is an immune receptor expressed by myeloid cells that has the capacity to augment pro-inflammatory responses in the context of a microbial infection . While a TREM-1-amplified response likely serves the efficient clearance of pathogens , it also bears the potential to cause substantial tissue damage or even death . Hence , TREM-1 appears a possible therapeutic target for tempering deleterious host-pathogen interactions . However , in models of bacterial sepsis controversial findings have been obtained regarding the requirement of TREM-1 for bacterial control - depending on the overall degree of the TREM-1 blockade that was achieved . In order to conclusively investigate harmful versus essential functions of TREM-1 in vivo , we have generated mice deficient in Trem1 . Trem1−/− mice were subjected to experimentally-induced intestinal inflammation ( as a model of a non-infectious , yet microbial-driven disease ) and also analysed following infections with Leishmania major , influenza virus and Legionella pneumophila . Across all models analysed , Trem1−/− mice showed substantially reduced immune-associated disease . We thus describe a previously unanticipated pathogenic role for TREM-1 also during a parasitic and viral infection . Importantly , our data suggest that in certain diseases microbial control can be achieved in the context of blunted inflammation in the absence of TREM-1 .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"immunopathology",
"pathogenesis",
"inflammation",
"immune",
"cells",
"cytokines",
"monocytes",
"immunity",
"immune",
"activation",
"t",
"cells",
"immune",
"defense",
"immunology",
"immunity",
"to",
"infections",
"biology",
"microbiology",
"host-pathogen",
"interaction",
"microbial",
"control",
"immune",
"response",
"immune",
"system"
] |
2014
|
TREM-1 Deficiency Can Attenuate Disease Severity without Affecting Pathogen Clearance
|
The slime mold Dictyostelium discoideum is one of the model systems of biological pattern formation . One of the most successful answers to the challenge of establishing a spiral wave pattern in a colony of homogeneously distributed D . discoideum cells has been the suggestion of a developmental path the cells follow ( Lauzeral and coworkers ) . This is a well-defined change in properties each cell undergoes on a longer time scale than the typical dynamics of the cell . Here we show that this concept leads to an inhomogeneous and systematic spatial distribution of spiral waves , which can be predicted from the distribution of cells on the developmental path . We propose specific experiments for checking whether such systematics are also found in data and thus , indirectly , provide evidence of a developmental path .
The slime mold Dictyostelium discoideum is a model organism for the study of pattern formation and excitable media dynamics in biological systems [1] . Several stages of its life cycle exhibit self-organized formation of structures successively building up on one another . Here we are exclusively concerned with the starvation-induced passage from a colony of chemotactically quiescent single cells to the cAMP signaling stage prior to the onset of aggregation and slug formation . Starting from the spontaneous emissions of a few starving cells , the whole colony enters a regime of excitable media dynamics [2]–[4] , where a local supra-threshold concentration of cAMP causes cells to produce and release more cAMP , which then diffuses to neighboring cells . The behavior of the single cells gives rise to a macroscopic dynamics exhibiting typical excitable media attractor states , namely circular radially growing ‘target waves’ caused by periodic oscillation of a central pacemaker element and self-sustained spiral waves . There are several mathematical models describing this transition [5]–[7] , making different assumptions about the exact nature of the underlying biological processes . Single D . discoideum cells have recently been shown experimentally to have distinct and persistent reactions to external stimuli [8] . In particular , the response to repeated stimuli varied substantially less for an individual cell under repeated stimuli , compared to the ensemble variation , indicating that the average response is indeed a cell property , varying across the cell population but rather fixed in time . As an analogy to physical systems , we like to interpret the arising situation as a ‘jagged potential landscape’; a hypothetical basic process corresponds to a smooth potential , where an injected particle will almost certainly come to rest in some sink of the landscape , and , in the case of several stable conformations emerging under variation of some control parameter and separated by unstable equilibrium positions ( the typical scenario for a second-order phase transition ) , random temporal fluctuations such as thermal noise decide the result ( see , e . g . [9] for the relation between phase transitions and self-organized processes ) . However , the biological variability of a real system combined with the finite number of constituents adds a layer of static roughness to the potential landscape , so that the influence of small ‘bumps’ may well outrank thermal noise , leaving a distinct fingerprint of the cell configuration in an ensemble of experiments and thus systematically biasing the asymptotic configuration of the system . We believe that in principle the result of the self-organized signaling , namely the spatial layout of the spiral wave pattern , can be predicted from the location and the properties of some cells playing key roles in triggering certain phases of cAMP communication . We recently succeeded in demonstrating this in a rather detailed fashion [10] for the model developed in [6] , [11] , which was also used to draw a connection between the macroscopic spiral wave density and the genetic feedback strength of the cAMP dynamics [12] . A key finding of [10] is the pronounced anticorrelation between the location of pacemaker cells ( which are explicitly included in that model ) and spiral occupancy , which enabled us to identify ( and model geometrically ) the most relevant microscopic mechanism of spiral formation , leading to a quantitatively successful prediction scheme for the spiral tip probability based only on pacemaker cell locations . If for several mathematical models of one real system the mechanisms and rules can be identified that map single element properties to emerging patterns , one can check these for agreements and differences which may open a route to experimental testing of macroscopic predictions , thus testing the assumptions regarding microscopic processes that are not easily accessible to direct observation . Therefore we here explore the deep link between cell properties and pattern features in a realistic mathematical model of Dictyostelium signaling behavior , namely the developmental path model [5] , which is motivated by experimental evidence for a slow variation of the cells' kinetic properties [13]–[15] . The Methods section describes the model we analyze here , as well as giving a brief overview over the methods we used to detect significant events in the spatio-temporal evolution of the system and the analysis of point processes , which we used to relate our numerical findings to experimental data . Using these we were able to identify effective pacemaker cells and spiral creation statistics , as described in the results section , enabling us to find qualitative differences in the mapping of pacemaker positions to macroscopic spiral probabilities in two models of D . discoideum .
Although the model used here does not explicitly contain pacemaker cells like some other models of D . discoideum ( [6] , [12] ) , we can expect that , since signal transduction is enabled rather suddenly , when almost all cells have crossed into the excitable regime , only the cells that are in the oscillatory regime at that exact moment in time can be the ones to initiate the first generation of target waves . These cells effectively assume the role of pacemakers in the early shaping of the emerging patterns . Figure 1 shows target wave events detected with the algorithm summarized in the Methods section over a density plot of cells that are in the oscillatory regime when signaling begins at . Two regimes are discernible: early locally repeating target waves growing in ‘stems’ mostly from clusters of oscillatory cells ( dark areas ) and a regime where several target wave centers drift apart slowly in branch-like structures , starting at about , when most cells have entered the autonomously oscillatory phase of the developmental path . In the early stage , clusters of oscillatory cells have a high chance of creating a repeating target wave center before being entrained by target waves emanating from single pacemakers in the neighborhood , since any of them can initially trigger the surrounding excitable cells; the reduced expected ‘time to next excitation’ offers a selection advantage compared to solitary oscillatory cells . The slower pacemakers of the cluster are in this case enslaved by the surrounding pattern , quickly leading to their synchronization with the initially triggering pacemaker . The behavior of predominantly oscillatory cells in the latter stage ( ) is a collective oscillatory dynamics , as opposed to the excitable behavior observed in the early and late stages of the developmental path employed here . The dynamical behavior of single cells changes qualitatively when the bulk of cells in the system crosses over into the oscillatory regime: Cells no longer react to their neighbors activating them , but oscillate autonomously . The target wave pattern established in the early phase ( ) remains imprinted on the system and persists for a while , but is no longer a completely stable attractor of the collective system . The diffusive term in the extracellular cAMP concentration coupled with the degradation term act as a phase damper , favoring synchronous bulk oscillations . We believe this amplifies the irregularization of the amplitudes along the circumference of a target wave ( cf . Figures 2 and 3 ) , simplifying the breakup of waves and finally spiral formation , as described below . Depending on the desynchronization parameter , a varying number of cells is eligible to produce target waves . Since signal transduction ( at the parameter setting used here ) is enabled at and the oscillatory regime starts at and ends at ( cf . Figure 2 ) , we can find this number by interpreting Eq . ( 7 ) as a function of , ( 1 ) which is a monotonically rising function for . So , for large enough to produce signaling patterns at all , and growing within reasonable bounds , one expects a growing number of potential pacemaker cells and hence a decreasing correlation between the ( binary ) spatial distribution of these cells and the time-averaged plane of detected target wave events because of the rising number of pacemaker cells that are entrained without leaving a strong fingerprint . Numerical experiments qualitatively confirmed this expectation: For , the correlation coefficient reaches about 40% and then decays rapidly with growing ( data not shown ) . Combining the results for the detection of spiral- and target waves , we arrive at a fairly detailed picture of the behavior of the system ( Figures 4 and 5 ) . A typical repeatedly pulsing target wave ( ) fractures ( ) into several small active centers that drift apart in the oscillatory regime ( see also Figure 1 ) and result in pairs of counter-rotating spirals ( ) in an apparently typical distance from the original target wave center . The spirals in a pair repel each other ( the target wave shrinks at both open wave ends , cf . [16] ) , drifting apart primarily transversally relative to the original target wave emitter and either annihilate with opposite-handedness spirals from an adjacent pair ( , the wave segment vanishes ) or come to rest and persist indefinitely in a dynamic steady state ( ) . After spiral annihilation events , oscillatory cells in the vicinity can easily ‘hijack’ the location , since it has last experienced a very target-wave-like cAMP pulse , and initialize a new train of target waves ( left , corresponding to the target wave visible in the lower left of Figure 3 at ) ; these persist at most until the last cells have left the oscillatory regime . It is noteworthy that , comparing the raw system data with the phase data in Figure 3 , phase singularities are often recognized by the algorithm long before the corresponding spirals in the raw data seem to come into existence; one could hence question the validity of this identification . However , the very continuous nature of subsequently identified phase singularities as shown in Figure 4 , that in later stages also coincides perfectly with spirals apparent in the raw data , leads us to conclude that a creation of a pair of opposite-handedness phase singularities indeed constitutes the ‘birth’ of a counter-rotating spiral pair , even though it might not yet be discernible as such , looking only at the raw system data . Do the sites of spiral pair creation events differ systematically from the rest of the cell population ? We detected the sites of spiral pair creation events in many differently initialized runs of the model presented here , in search of another specific fraction of cells that is responsible for the breaking of target waves and hence for the creation of spiral wave pairs . To our astonishment , these events seem not be related with the position of a cell on the developmental path: The cell age offset values at the sites of spiral creation events exactly mimic the offset distribution , Eq . ( 5 ) , of the whole population ( data not shown ) . It thus seems that the locations of spiral pair creation are at least not directly influenced by cell properties , as we expected from looking into the previous works [3] , [17] . Is there a typical radius for spiral creation events , i . e . a typical distance from the target wave events around which the spiral waves organize themselves ? We tested this hypothesis by running several hundred simulations in which the cell age distribution was randomized , but we placed 3×3 clusters of pacemaker cells ( offset 60 . 0 ) at fixed positions before running the model . We chose completely synchronized clusters of pacemakers for simplicity , preempting their dynamical synchronization to the fastest pacemaker of the cluster , and at the same time allowing complete control over the relative phase of several pacemaker clusters . In fact , ring-like structures centered on the manually placed pacemaker clusters are visible in the average spiral wave occupancy ( Figure 6 ) , but their radius and crispness ( i . e . , the visual clarity ) depend on the distance between the pacemaker clusters . For small distances up to 20 grid points , no separate rings are discernible and the localization of the ring is very poor , indicating a wide spread in spiral creation radii . For medium distances ( 30–60 ) there are clearly separated halos that are much more localized . For larger distances the appearance seems to revert to two separate instances of the weakly localized rings observed for very small distances . These effects can be explained in an artificially created minimal situation . In order to reduce the observed average patterns to contributions of the manually placed pacemaker clusters , we changed the cell age offset distribution , Eq . ( 5 ) , to not contain random pacemaker cells . To achieve this , when creating the concrete cell age offset distribution for a given run , cells with an age offset between 20 and 110 minutes ( corresponding to pacemaker candidates when the medium becomes active at about , plus a head start of ten minutes for manually placed pacemakers to dominate the system ) had their values re-randomized according to Eq . ( 5 ) until they ended up outside of this interval . In Figure 7 we demonstrate that this modified cell age distribution does not qualitatively change the temporal evolution of the system ( apart from removing ‘noise’ ) , and that the few manually placed pacemaker clusters are sufficient to generate the initial stage target waves as well as spiral waves later on . The resulting spiral formation and -occupancy statistics of this modified cell age distribution , however , are vastly clearer than before ( Figure 8 ) . It is now readily discernible that the formation of spiral wave pairs ( omitting boundary effects ) happens almost exclusively on lines parallel to and in a fixed distance from the Voronoi diagram of the pacemaker clusters ( i . e . the partition of the plane into areas that have the nearest pacemaker in common ) . Spiral pairs are formed in front of areas where waves collide and annihilate , probably based on a prolonged refractory time ( or oscillation period ) in these areas , due to the large peak amount of extracellular cAMP deposited there . In contrast to common expectation , we could not observe this process as a sudden breaking of target waves , it rather seems to be a gradual build-up of phase lag during the time when most cells are oscillatory ( cf . in Figure 3 ) . The very exact localization near a specific distance from the Voronoi lines is due to the synchronization of our manually placed pacemakers – since their age offset is exactly identical , they fire in phase and hence the target waves emanating from pairs of adjacent pacemakers always meet exactly at half their distance to each other . Also , consistent with the reason for spiral formation stated above , the probability of spiral pair formation is higher near vertices of the Voronoi diagram , representing points where target waves from three or more pacemakers meet and annihilate , and on direct connections between adjacent pacemakers , where target waves collide head-on . Removing the artificial constraint that all manually placed pacemaker clusters oscillate in phase does not qualitatively change this result . Since the pacemakers already are in the oscillatory regime when signaling is enabled , their phase difference can at most be , corresponding to one oscillation period . We selected the starting time offset of all pacemaker clusters uniformly from 60 to 66 minutes ( ) . The phase difference distribution of adjacent pacemaker clusters is then triangular , peaking at zero and stretching to very unlikely maximum differences of . Hence , the positions of target wave collisions are still strongly centered on half the corresponding pacemaker cluster distances and the resulting distributions corresponding to Figure 8 are slightly washed out but qualitatively unchanged; the Voronoi diagram of the pacemaker clusters is still clearly discernible as the entity governing the statistics of spiral pair formation and consequently spiral occupation . We hence conclude that the locations of spiral pair creation do not directly depend on the cellular properties at these sites , but on geometrical constraints which are an indirect consequence of the heterogeneity of cell properties given by their positions on the developmental path . Note that the Voronoi diagram of the pacemaker cells arises here by the dynamical exploration of the ensemble of possible points for spiral formation over many numerical runs . It is not connected to the fact that in D . discoideum cell streaming experiments , the initially homogeneous plane is separated into several basins of attraction , which very nearly correspond to the Voronoi cells of the spiral cores ( again , subject to phase differences ) . The Voronoi pattern in Figure 8 is a summary of the geometrical constraints arising implicitly from the distribution of cells on the developmental path , while the explicit partitioning into Voronoi cells during aggregation is simply a consequence of each D . discoideum cell moving to the nearest spiral core under the influence of the chemotactic signal . The mechanism of spiral formation outlined above was not discernible in similar clarity , employing the unmodified cell age distribution , because of interference from randomly emerging pacemakers . The varying crispness in halo appearance in Figure 6 can also be explained in these terms . For very small distances , the two clusters effectively act as one pacemaker , since no fully formed waves are established between them , and spiral pairs are formed only by interactions with random pacemakers , which emerge at different positions , radii and relative phases in every numerical run , leading to a very fuzzy average image . Once the distance between the two manually placed clusters is large enough so that fully developed target waves are created emerging from each of them , this pair constitutes the most predominant ( and consistent ) cause of spiral formation , giving rise to fairly clear halos , especially directly between them . The further these clusters are separated , the higher is the probability of interference from random pacemakers , until at the maximum considered distance one effectively has two instances of a single consistent pacemaker cluster , interacting only with their respective varying neighborhoods of randomly emerging pacemakers . Translating these insights back to the original setup of the model is straightforward in principle , but tricky in detail . Since for a desynchronization a large fraction ( ≈29% ) of all cells has the potential to act as pacemaker cells , but few actually emerge because of the system's limited carrying capacity for sustained target waves per area , one needs a good scheme to predict the formation sites of initial target waves . The spatial density of pacemaker candidates is a promising start , but not quite sufficient , as Figure 1 shows . This difficulty can be reduced by choosing smaller , just large enough to give the randomly emerging pacemakers time to establish target waves before all cells enter the oscillatory regime , but the number of pacemaker candidates ( e . g . ≈1350 for and a 100×100 grid , cf . Eq . ( 1 ) ) is still substantially larger than the number of persistent target waves the system can sustain ( few tens ) . Alternatively and more realistically , regarding a possible application to experimental data , one can start from an ignorance of detailed cell properties and predict the spiral positions only from observed target wave positions . Since one cannot expect single experiments to yield statistically significant agreement with an ensemble average of many idealized repetitions we performed in silico , it might be most instructive to only compare observed spiral tip density per unit area in two categories , namely ‘high’ and ‘low’ expected spiral density , computed from the distribution of early target wave emitters . An experimental setup with a high degree of similarity to our manipulated cell age distribution might be attainable by placing a few ( possibly fluorescence-marked ) cells from an older population already in the target wave phase into a younger colony where target wave signaling has not yet been established . The geometrical spatial systematics may very well serve as an experimentally accessible evidence for such a developmental path in the real system . Other researchers have strived to explore the spatial correlations between target and spiral waves in experimental data [14] , but so far the analyses proved difficult . We here for the first time apply methods from point process statistics [18] to the analysis of Dictyostelium signaling patterns . As outlined in the Methods section , this allows a quantification of the over- or underrepresentation of pairs of events at specific distances , resulting in typical ‘correlation profiles’ between target waves and spiral cores over distance . This method can at most very indirectly capture more detailed geometric correlation patterns ( such as demonstrated in the previous Section ) , but allows the quantification of pair correlations in data sets that appear unordered to the naked eye . In addition , this method allows in principle to perform a cumulative analysis of many experimental runs , even despite their larger diversity when compared to numerical simulations . We digitalized the data points in Figure 2a of [14] and compared the resulting curves to curves extracted from the model from [6] ( ‘Levine model’ ) , the developmental path model analzyed here ( with , corresponding to 29% pacemaker candidates ) and additional experimental data sets kindly provided by Christiane Hilgardt ( University of Magdeburg ) and Satoshi Sawai ( University of Tokyo , [19] ) . We show here only one curve per ( experimental ) source , where the quality of target and spiral wave detection was highest . Apart from detection artifacts , all experimental curves exhibited the same qualitative behavior . Figures 9–11 show the reduced partial pair correlation functions ( see Methods ) for target-spiral , spiral-spiral and target-target comparison , respectively . The interplay between target and spiral waves ( Figure 9 ) is dominated by an underrepresentation ( suppression ) of these pairs for short distances , which is the qualitative anticorrelation we found in [10] as well as in our present work . Note that all experimental curves we analyzed also exhibit this feature qualitatively . In some cases it can be somewhat obscured by crosstalk between target and spiral wave recognition , which is problematic mainly for experimental data sets ( dark blue curve ) . Figure 10 shows a strong suppression of spiral-spiral pairs below a minimum distance . This suppression typically has a much longer range than the minimum distance we manually imposed to prevent the double recognition of a single spiral signal as two points ( the typical diameter of a spiral peak ) . Following this suppression there is a regime of overrepresentation that has its main root in the existence of stable pairs of counterrotating spirals . Note that this peak is apparently shifted towards higher distances for the model from [6] , but this is mainly due to the smaller length scale this model is simulated at . The most robust feature of Figure 11 is a significant overrepresentation of pairs of target waves in very short distances , indicating clustering . In the Levine model , this effect is much less distinct than in the developmental path model , due to the much smaller number of repeated firing events before a spiral pattern is established . The longer spatial scale in the developmental path model stems from the fracturing of target waves , resulting in large target wave clusters . One possible interpretation of this clustering is a correlation between the locations of specific ‘pacemaker’ cells and target waves ( per construction , for the mathematical models ) . On the other hand , a clustering of target waves can also occur in scenarios where every quiescent cell has the potential to fire spontaneously ( the original setup of [6] ) ; the higher firing probability near the source of the previous target wave center is then based only on the longer time since the last firing event . The expected degree of clustering in such a setup depends on the typical time scale until newly quiescent cells fire , compared to the wave speed . There is a possible objection to the anticorrelation hypothesis between spiral and target waves: One can argue that for the curves corresponding to spiral-spiral and spiral-target pairs there is a competition which trivially blocks these events from occuring in close proximity , whereas for the target-target curve there is no such competition since they occur sequentially instead of at the same time . This is absolutely true for the spiral-spiral case , we in fact assume that the length of the repulsive plateau is an indicator for the shortest length scale at which stable spiral pairs can coexist . However , if one accepts that spiral waves in some form or other result from target waves , which are the fundamentally simpler patterns that can easily arise from spontaneous firing events of few cells , one has to accept that target waves first occur prior to spiral waves , when there is no coexistence and hence competition . This is rather clearly the case for both mathematical models considered here , were we observe a relatively sharp transition from target to spiral waves . The aforementioned objection hence only holds for target-spiral pairs in the late stage of signaling when spirals have been established . Furthermore , since both mathematical models considered so far qualitatively predict a target-spiral anticorrelation for short distances , this point does not contribute to the ultimate question of which model better captures the experimental evidence . It should be noted that , ideally , one would analyze correlations between cell properties and pattern features . At the moment , however , to our knowledge such data do not exist for Dictyostelium . Exact length scales for experimental data would also be valuable in comparing several curves over real-space distances .
In this report we employed a technique for the identification of spatio-temporal target waves and used it in conjunction with spiral tip recognition , based on the established phase singularity technique , to identify typical temporal motifs of events in a developmental path model for the social amoeba Dictyostelium discoideum . This analysis follows up on our earlier investigation into the predictability of spiral patterns from the knowledge of cell properties conducted in a more schematic model of D . discoideum [10] . Not surprisingly , the more complex and more closely biologically motivated model examined here exhibits a more complex statistical dependence of the resulting spiral patterns on the cell properties . Nevertheless , we were able to identify a specific fraction of cells that function as effective pacemakers , as well as the dominant mechanism of spiral formation , and employ this knowledge to engineer the spatial statistics of target wave and spiral creation by manipulation of the pair correlations of these cells . Similarly to our findings in [10] , one observes an anti-correlation between ( now dynamically generated , effective ) pacemaker cell locations and the locations of spiral formation and asymptotic spiral position . However , the structure of spiral creation and meandering in the region between pacemaker locations is more complex; spiral tips are formed at a specific distance from lines of the Voronoi diagram of the pacemakers , and meander on roughly circular orbits around pacemakers , without intruding into the halos of adjacent pacemakers . Also , in strong contrast to the ‘simple’ anticorrelation we found in [10] , the area near the lines of the Voronoi diagram ( the area right between pacemakers ) is expected to hold a strongly reduced amount of spiral tips . The results presented here provide further evidence supporting our general hypothesis that single element properties are systematically mapped onto patterns and thus conserved through processes of self-organization ( as opposed to enslaved and deleted by the collective ) , as outlined in [10] and the introduction of this paper . Furthermore , we have now presented mapping schemes for two numeric models of D . discoideum , yielding different predictions regarding the relationship between pacemaker positions and spiral wave tip statistics . We compared both of these models to experimental data from [14] , introducing the statistical tool of point processes to the Dictyostelium signaling debate . We were able to demonstrate that the spiral and target waves in the data from [14] are not uncorrelated , as claimed there , but that they follow qualitatively identical systematics as other experimental data as well as both considered mathematical models , including the anticorrelation between target wave centers and asymptotic spiral core positions . We have so far not been able to conclusively distinguish which mathematical model better captures the real system , mainly due to the increased noise level and the large variation one typically observes from run to run in experiments .
The D . discoideum model considered here has been formulated in [20] and extended to include a developmental path in [5] , with some additional discussion in [17] . It is given by three coupled differential equations for the total fraction of active cAMP receptor ( ) and the normalized concentrations of intracellular ( ) and extracellular ( ) cAMP , respectively , ( 2 ) with ( 3 ) The biological meaning of the main terms in these equations are the following: The cAMP receptors on the cell surface are de- and resensitized depending on the extracellular cAMP concentration and the currently active receptor fraction . Intracellular cAMP is produced depending ( among other factors ) on the current activity of the adenylate cyclase ( cf . below ) . Upon diffusing to the extracellular area it is degraded by the action of phosphodiesterase ( PDE , both bound to the cell membrane and extracellular PDE ) and otherwise diffuses freely . The exact forms of the nonlinearities stem from the reduction from nine to three dynamic variables performed in [20] . For clarity , we use exactly the notation from [5] . This model has been studied in great detail in terms of its dynamical regimes as a function of the position in parameter space ( cf . [20] and references therein ) . Here we do not consider a wide range of parameter constellations and instead focus on the dynamical processes leading to target and spiral wave formation . Throughout this paper we use the parameter setting discussed in [5] , i . e . , , , , , , , , , , , , , . The grid spacing is 100 µm , a system time unit is identified with a minute [5] . We integrated these equations using an explicit Euler scheme with a fixed step size , again as in [5] . The simulations were performed in custom software in C++ . These equations and parameter settings give rise to several dynamic regimes , including most importantly a steady-state regime where external cAMP stimuli do not trigger a reaction , an excitable regime where external stimuli are followed by a sharp increase of cAMP production with a subsequent recovery period , and finally an oscillatory regime where the cells autonomously oscillate between phases of cAMP production and quiescence ( Figure 12 ) . Lauzeral et al . [5] proposed that the maturation of cells has the effect of modifying their behavior ( described in this model by the maximum activity of adenylate cyclase and the extracellular phosphodiesterase rate constant ) , transporting them through these regimes along a fixed predetermined developmental path , thus giving rise to macroscopic cAMP patterns ( cf . Figure 2 ) which then give the cue for cell aggregation and finally lead to the following stages of the cell cycle . In order to introduce the initial heterogeneity needed for the formation of spiral waves , it is assumed that cells have different properties at the onset of starvation , for example different stages of their cell cycle , which are represented as differing starting position offsets on this path . Here we consider only the developmental path 3 of [5] , given by a combined sigmoidal variation of and , ( 4 ) where again we follow [5] closely , both in notation and in the parameter values , i . e . , , , , , , , . Figure 12 shows this developmental path in the parameter plane of a single oscillator ( after [5] ) . A total number of cells arranged on a regular spatial grid is placed on this path with varying starting time offsets according to an exponential probability density , ( 5 ) Note that we refer to each grid point as a ‘cell’ for practical purposes , although it actually represents a cluster of about ten cells with synchronized properties [5] . Throughout this paper we use grids of 100×100 cells and unless explicitly noted otherwise . It is noteworthy that the complete dynamics of this model depends only on the concrete choice of starting time offsets and contains no other random elements . The exponential distribution implies that the number of cells with starting time in the interval is approximately ( 6 ) After a time has passed , the number of cells in the interval is ( 7 ) We used algorithms to detect target and spiral wave configurations , which are well-known attractor states for excitable media dynamics , and which play important roles in the shaping of the self-organized cAMP communication process . In order to detect spiral waves , we used the phase singularity method introduced by Gray et al . and Bray et al . [21] , [22] in the context of heart tissue dynamics , which was to our knowledge first applied to D . discoideum data in [12] . We observed an influence of the sample size entering the specific average in the underlying embedding process ( cf . [22] ) on the exact recognized phase singularity position: Using a global average over all raw values ( after the end of the quiescent period ) for each grid point caused phase singularities of apparently pinned spiral waves to circle around an empty core on a decaying helix trajectory in space-time . While this did mimic the experimentally observed behavior , it was an inconvenience when trying to visually trace spiral cores . Using a gliding time average over about three signal periods ( 20 minutes ) removed this meandering and yielded the spatially fixed phase singularities shown in this paper . Figure 3 shows a time series of raw system data contrasted with the corresponding phase data and detected phase singularities . In order to detect target waves we developed a 3D fitting algorithm based on the already calculated smooth and amplitude-insensitive phase data ( article in preparation ) . At its core , it fits cones ( the spatio-temporal evolution of target waves , neglecting curvature effects on wave speed ) to connected voxel segments , representing contour shells extracted from the spatio-temporal phase data and subjected to causal consistency constraints ( maximum observed wave velocity ) . The tip of a successfully fitted cone corresponds to the point in space-time where an observed target wave was created; we call these points target wave events . An analogous spiral fitting algorithm was also developed , but not employed because of high computational cost and inferior performance compared to the phase singularity technique , given the relatively low-noise environment of the computational model discussed here . We used the mathematical concept of point processes to quantify correlations between temporal projections of spiral and target wave events in the developmental path model [5] discussed here , the more phenomenological excitability model from [6] and experimental data from [14] . The core idea of point processes is to take a given distribution of possibly several types of points ( marked point processes , here: locations of target wave events and asymptotic positions of spiral waves ) and calculate a variety of measures comparing e . g . the observed frequency of point pairs in specific distances to the expected frequency if the points were randomly distributed ( see e . g . [18] ) . We found the partial pair correlation function to be the most distinct and at the same time straightforward quantifier of the relevant system statistics . For the two dimensional case used here it is defined as [18] ( 8 ) Here and are the sets of all points of types and , respectively , and are the respective intensities ( expected number of points of type or per unit area ) , is the box kernel function with bandwidth and is the area of the intersection between the sampling window W shifted to and . The latter term is intended to correct for edge effects . The bandwidth quantifies the width of the band around from which one accepts point pairs contributing to the value of and should be chosen separately for each type of data set . Larger values increase the number of point pairs taken into consideration and thus improves the statistics but at the same time reduces the achievable resolution in ; we increased in steps of . We used for data from numerical simulations , for the data from [14] , for the data by Christiane Hilgardt ( 300×500 pixel areas from digital photographies of dark-field experiments , downsampled to 150×250 , to which the bandwidth refers ) as well as for the data by Satoshi Sawai ( [19] , 640×480 downsampled to 320×200 ) . The partial pair correlation function quantifies the probability of simultaneously finding a point of type and another point of type in infinitesimal volumes at distance , normalized by the expected probabilities . For complete spatial randomness one expects a constant value of one . Values greater than one indicate an overrepresentation of these pairs at distance and correspondingly values of less than one correspond to underrepresentation . For large distances between points one expects an asymptotic behavior tending towards one , where the distances become so great that points do not influence each other significantly; by definition , . Since the spatial resolutions and image sizes of the data we want to compare are different ( and in some instances unknown ) , we renormalized distances to the maximum diameter of the sampling windows , i . e . the image diagonal . Length scales are then remapped to fractions of the image diameter and the curves become more easily comparable . One should keep in mind , though , that the does not represent the same real-space distance for all curves , e . g . for the simulations we used identical 180×180 grids , but a single grid distance corresponds to 0 . 6 mm in the model from [6] ( as used in [12] ) and to 1 . 0 mm in the developmental path model analyzed in-depth here . We found that edge effects are still visible despite the edge correction term , so we always plot , which we call the reduced partial pair correlation function , where is the average curve from hundred realizations of randomly distributed points in the same sampling window , keeping the numbers of target wave events and spirals fixed to the original amount found in the respective source data . Values above or below zero thus correspond to over- or underrepresentation compared to the null model of completely random point distributions , in units of the expected probability based on the intensities . It should be clear that given the rather finite-sized sampling windows and the differences between the considered data sets , our comparisons based on this technique should be used mainly as qualitative indicators .
|
Spatio-temporal pattern formation is a core discipline of theoretical biology . Formation of large-scale patterns from local interactions can very prominently be observed in the swarming behavior of fish and birds , in animal markings or bacterial growth patterns . It also plays a critical role in the life cycle of the social amoeba Dictyostelium discoideum . A homogeneous colony of amoebae is partitioned into subgroups that will form migrating slugs by a collective phase of chemotactic signaling , exhibiting typical and well-known patterns for this sort of excitable dynamics ( circular and spiral waves ) . The mechanism of spatial localization of aggregation centers ( that is , the centers of periodic circular and spiral waves ) is unclear , despite its crucial role to the organism's procreation . Here we demonstrate for an established computational model of D . discoideum that the initial properties of potentially very few cells have a driving influence on the resulting asymptotic collective state of the colony . Analogous processes take place in diverse situations such as , e . g . , heart cells ( where spiral waves occur in potentially fatal ventricular fibrillation ) , so that a deeper understanding of this additional layer of self-organized pattern formation would be beneficial to a wide range of applications .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"biophysics/cell",
"signaling",
"and",
"trafficking",
"structures",
"biophysics/theory",
"and",
"simulation",
"computational",
"biology/systems",
"biology"
] |
2009
|
Predicting the Distribution of Spiral Waves from Cell Properties in a Developmental-Path Model of Dictyostelium Pattern Formation
|
Advances in sequencing have generated a large number of complete genomes . Traditionally , phylogenetic analysis relies on alignments of orthologs , but defining orthologs and separating them from paralogs is a complex task that may not always be suited to the large datasets of the future . An alternative to traditional , alignment-based approaches are whole-genome , alignment-free methods . These methods are scalable and require minimal manual intervention . We developed SlopeTree , a new alignment-free method that estimates evolutionary distances by measuring the decay of exact substring matches as a function of match length . SlopeTree corrects for horizontal gene transfer , for composition variation and low complexity sequences , and for branch-length nonlinearity caused by multiple mutations at the same site . We tested SlopeTree on 495 bacteria , 73 archaea , and 72 strains of Escherichia coli and Shigella . We compared our trees to the NCBI taxonomy , to trees based on concatenated alignments , and to trees produced by other alignment-free methods . The results were consistent with current knowledge about prokaryotic evolution . We assessed differences in tree topology over different methods and settings and found that the majority of bacteria and archaea have a core set of proteins that evolves by descent . In trees built from complete genomes rather than sets of core genes , we observed some grouping by phenotype rather than phylogeny , for instance with a cluster of sulfur-reducing thermophilic bacteria coming together irrespective of their phyla . The source-code for SlopeTree is available at: http://prodata . swmed . edu/download/pub/slopetree_v1/slopetree . tar . gz .
Learning how to obtain complete genomes was a critical step to understanding biology and was achieved as early as 1977 for the genome of bacteriophage ɸX174 [1] . Since then , methods for obtaining full genome sequences have advanced tremendously [2–4] , leading to a second critical transition , when the number of genome sequences became too large for traditional , alignment-based , phylogenetics [5–8] . Even during the time of Sanger sequencing , the number of bacterial genomes began to cross this threshold [9] . With the development of next generation sequencing technology , we are experiencing a flood of complete genomes and metagenomes [10] . Molecular phylogenetics enabled the classification of prokaryotic organisms . In 1977 , a multiple sequence alignment ( MSA ) of the small subunit ( SSU ) 16S rRNA gene revealed the existence of the three domains of life [11] , making the SSU rRNA the gold standard for phylogenetics [12–14] . As more sequences became available , additional genes were used as phylogenetic markers , including protein elongation factors EF-α/Tu and EF-2 [15–17] , chaperones Hsp60 and Hsp70 [18 , 19] , the largest subunits of the RNA polymerase [20 , 21] , RecA [22] , a variety of aminoacyl-tRNA synthetases [23] and others . Approaches using single genes originally generated a wealth of phylogenetic insight , but these trees were frequently incongruent with one another [24 , 25] . To improve the accuracy of phylogenetic methods , phylogeneticists began to concatenate multiple conserved genes to produce larger MSAs and therefore better resolved trees [25–28] . The size and functional diversity of these gene groups is largely dependent on the number and diversity of taxa [29] . For instance , in the recent work of Lang and Eisen [25] , an analysis of ~900 diverse prokaryotes from both bacteria and archaea identified only 24 suitable ( i . e . paralog-free ) genes . These consisted of a subset of ribosomal proteins , two translation factors that both interact with the ribosome , and the alpha subunit of a phenylalanyl-tRNA synthetase which was the only protein in the set not interacting with the ribosome and which contributed only ~5% of the overall alignment used to generate phylogeny . A similar situation was seen by Ciccarelli et al . , in which for a group of 191 organisms , the set of 31 genes used in the final alignment consisted of 23 ribosomal proteins [30] . Widespread horizontal gene transfer ( HGT ) also interferes with a straightforward definition of evolution by descent [31–34] . Therefore , we are still making our way to a consensus to a definition of prokaryotic evolution by descent . In contrast to the majority of traditional MSA-approaches , which often require extensive curation to produce high quality alignments of orthologs , alignment-free methods are scalable and require minimum manual intervention [35–38] . The idea of using complete genomes to perform phylogeny has a long history [39] , but lay dormant until enough complete genomes became available . The rate at which these methods are now appearing reflects the pressing need for unsupervised , scalable methods . An additional advantage is that because they use complete genomes , they may provide a more sound approximation for organismal phylogeny [40] . Alignment-free methods compute similarity or distance metrics using a variety of statistical properties belonging to k-mers ( fixed-length substrings , also called n-grams , n-mers , k-tuples , and k-words ) in genomes . These methods are often divided into two main classes: methods using fixed-length word counts and methods using match lengths . Word count methods relying on exact word matches include Composition Vector Trees ( CVTrees ) [41 , 42] , Feature Frequency Profiles ( FFP ) [43] , and D2 statistics [44–46] . Each of these methods relies on different properties of counting exact matches fixed-length k-mers . CVTree calculates the frequency of all length-k k-mers in all proteomes; these frequency or composition vectors , after a background subtraction correcting for random neutral mutations , are then compared to one another and a correlation is calculated by means of the cosine of the angle between them , which is normalized to produce the final value . FFP tabulates the counts for all possible features in the genome of fixed length k , which as in CVTree also are passed through a normalization procedure to form a probability distribution vector ( i . e . an FFP ) ; distances are then calculated using Jensen-Shannon Divergence [47] . D2 measures sequence dissimilarity by the logarithm of the ratio between conserved and non-conserved k-mers . Word count methods employing inexact matches include Co-phylog [48] and Spaced Word Frequencies ( SWF ) [49 , 50] . Co-phylog identifies seed alignments ( exact or approximate matches ) between the query and subject sequences and then extends them into longer alignments ( i . e . ‘micro-alignments’ ) ; this method has an additional advantage in that it runs on raw next-generation sequencing data . SWF is similar to Co-phylog , using a mask consisting of positions that are either match or don’t care , and using the frequencies of these spaced words , with the don’t care positions ignored according to the specified pattern . Match length methods can also be divided into those allowing zero mismatches or those allowing some number of mismatches . Exact match length methods include Average Common Substring ( ACS ) [51] , Kr [52] , and Underlying Approach ( UA ) [53] . Conceptually , the ACS method is the most similar to the method we present , and calculates its distance metric by means of variable length , exact matches between genomes or proteomes . For every position in one genome or proteome , ACS finds the longest length match in the other . This list of matches is then averaged , normalized , and a correction applied that transforms it from a similarity measure to a distance . The Kr method is closely related to the ACS method; taking two unaligned DNA sequences , Kr estimates the number of substitutions per site by determining for every suffix present in the one sequence the shortest prefix that is absent from the other ( called shustrings ) . UA uses a scoring function on matching statistics between unique , independent subwords . k-Mismatch ACS ( kmacs ) [49] is an extension of the ACS approach which approximates the number of substring matches with up to k mismatches . Another , more recent extension of the ACS approach is ALFRED-G , also capable of computing lengths of shared sequences with mismatches allowed [54] . ACS-like methods which allow for k>1 mismatches can be highly costly in computational terms , but there has recently been some headway in improving their efficiency [55] . We present SlopeTree , a new alignment-free method which measures evolutionary distance by quantifying how quickly the number of matching sequences between two proteomes decays as a function of sequence length . The sequences that we employ to this end are k-mers , i . e . substrings of length k . The method considers uneven composition of amino acids , the possibility of backwards mutations , a background of coincidental matches over short k-mer lengths , and the issue of HGT . HGT is highly relevant for alignment-free methods because it adds a spurious contribution of similarity between genomes [56 , 57] . There are multiple possible signatures of horizontally transferred proteins , for instance unusual codon usage [58–60] . We identified a novel signature based on analysis of multiple copies of almost identical protein sequences in a genome , and those multiple copies almost invariably belonged to one of two categories: one category was of EF-Tu translation factor , which is frequently present in multiple copies; and the second was of mobile elements , as inferred from a very narrow or scattered phylogenetic footprint , even within a single species . When annotated , these mobile elements consisted primarily of parasitic proteins resulting from phage infections . Another level of filtering is done by means of a dual evolutionary stability index indicating conservation and lack of stability , i . e . a paralogy score , with a large instability value representing very likely cases of HGT . A mobile element ( ME ) filter and a separate , conservation filter were built into SlopeTree using the earlier mentioned novel signature and the paralogy score . To measure the similarity between two proteomes , SlopeTree yields a slope ( explained in the Algorithms section ) ; a third HGT correction is based on the curvature of this slope . Therefore , SlopeTree is unique in that it is not only robust to HGT , but it explicitly identifies and corrects for HGT at multiple stages of the analysis . By subtracting the background of short length , coincidental matches and restricting itself to a range of longer lengths ( ~7 or more amino acids ) , SlopeTree is able to follow the evolution of the highly conserved segments of proteins , using approximately 10 , 000 to 40 , 000 amino acids per genome pair . The highly conserved regions that SlopeTree targets correspond to the alignable regions in an MSA . For 72 strains of Escherichia coli and Shigella , 73 archaea , and 495 bacteria , we built trees using different degrees of HGT-correction . We compared these trees to trees based on phylogenetically broad concatenated alignments from the literature [25] , in which supermatrices were constructed from 24 single-copy , ubiquitous genes and then passed to a Maximum Likelihood ( ML ) routine for tree-building . These comparisons were performed to assess the accuracy of our method and to identify potential biological sources for differences . The SlopeTree strain-level trees were remarkably stable for different inputs . Even when only mobile elements together with proteins that are not part of the core were considered , the tree topology was highly similar . The archaeal trees were more fluid upon restricting the method to the most conserved proteins , but the majority of clades and relationships between deep branches remained the same . The deep , short branches in the bacterial trees were the most unstable , which is related to a generic problem of defining phylogenetic relationships in evolutionary radiation . For archaea and bacteria , we calculated the symmetric difference distance [61] to the trees built from supermatrices for trees built by SlopeTree , ACS , CVTree , D2 , kmacs , Spaced Words and ALFRED-G . By applying our ME filter and conservation filter to the data prior to running the main SlopeTree routines , we were able to significantly reduce the distances to the trees built from supermatrices not only for SlopeTree but for all other alignment-free methods . We observed approximately 20 bacteria whose placement on the phylogenetic trees frequently disagreed between alignment-free methods and the current NCBI classification . The consistency of these alternative placements for these bacteria when applying alignment-free methods suggests that their classification may require revision , or at the very least have complex histories . This is further supported by the fact that several of these bacteria had similar disagreements between the trees built from supermatrices and the NCBI classification .
Our method is based on k-mers that are substrings of length k . The SlopeTree package includes both the main SlopeTree algorithm , which estimates evolutionary distance by quantifying how quickly the number of matching sequences between two proteomes decays as a function of sequence length , and several independent modules for filtering mobile elements and less-conserved proteins out of the data and recalculating distances for pairs still exhibiting significant HGT even after the earlier filtering steps . Altogether , the method consists of the following four modules: ( 1 ) a Mobile Element Filter , ( 2 ) a Conservation and Stability Filter , ( 3 ) the SlopeTree Main Algorithm and ( 4 ) a Pair-Wise Horizontal Gene Transfer ( HGT ) Correction . A flowchart is provided in S1 Fig . The Mobile Element Filter exploits a novel signature which is based on analysis of multiple copies of almost identical protein sequences in a genome . These highly repetitive proteins proved almost always to be mobile elements . The Conservation and Stability Filter calculates for each protein a value , which we call a paralogy score , from the ratio of the sum of how many genes each of the protein’s k-mers has a match with in other genomes to the sum of the total number of genomes the protein’s k-mers have matches with . This ratio effectively separated orthologous proteins evolving by descent , which typically have a gene to genome ratio of one and therefore had paralogy scores of approximately one . Mobile elements on the other hand , have paralogy scores frequently much greater than one because their presence , absence , and copy number are much more unstable , while unconserved proteins which simply have no k-mer matches with any other proteins in the input have scores of 0 . The SlopeTree Main Algorithm estimates a distance for every pair of organisms from the decay in the number of exact sequence matches as a function of match length . The Pair-Wise HGT Correction assesses the slopes produced by the SlopeTree Main Algorithm and identifies pairs of organisms that appear to have shared significant horizontal transfers; it runs the SlopeTree Main Algorithm on these pairs combined with a reference set to identify proteins that the pair shares but that are absent from the reference , and then it re-runs the SlopeTree Main Algorithm on just the pair , with the flagged proteins removed . The four modules are not necessarily run together; for instance , the SlopeTree Main Algorithm can be run on unfiltered data or data passed through only one of the filters . Input . A set S of n proteomes 〈S1 , S2 , … , Sn〉 and a set T = 〈T1 , T2 , … , Tl〉 , with T taken from l taxonomically diverse organisms where Ti consists solely of the highly conserved proteins of the organism i . In practice , l is generally much smaller than n , but this is not required . Output . A set V = 〈V1 , V2 , … , Vn〉 where each Vi consists of all proteins in Si , minus the mobile elements . Algorithm . Let pij be the jth protein in Si , and let pkij[h] be a k-mer from pij of length k , starting at index h , where 0≤h<f given that pij has length f . For those k-mers at the end of each protein where h+k>f , the suffix is expanded by the necessary number of empty characters to fill the remainder of the k-mer . Each k-mer is stored as a 2-tuple consisting of the k-mer and the gene ID ( j ) . Let Ai be the alphabetically sorted list of all 2-tuples from Si . For every protein pij , there is a pair of integers , rij and cij , both initialized to 0 . Starting from the first k-mer in Ai , we pass down the list until a k-mer with more than u mismatches with this first k-mer is found . For all proteins with k-mers in this block , rij is incremented . This process is repeated until the end of Ai is reached , always starting from the first k-mer to not be a member of the current block of matches . Separately , we repeat the k-mer compilation process described above on T to generate a single , alphabetically sorted list of 2-tuples across all proteomes in T . Duplicates are removed from this list to make a new list B consisting of each k-mer and the number of times it appears in T . Those k-mers appearing only once are given a count of 1 . Then for every k-mer in Aj , we query B; the value of cij is increased by the count stored in B for every exact match between B and any k-mer in any protein pij . Having set all rij and cij for all pij in Si , we define a linear function such that all pij with rij≥acij+b are removed from proteome Pi and the reduced proteome we call Vi . Computational complexity . For n organisms and m amino acids in S , let m = m1+m2+…+mn . For l organisms and k amino acids in T , let k = k1+k2+…+kl . The compilation of Ai is done in O ( m ) time , and the time required for sorting each Ai is O ( mi log mi ) , which summed over all n organisms is O ( m log m ) . Similarly , the time to compile all k-mers in T is O ( k ) and to sort them requires O ( k log k ) time . The order of the algorithm is dominated by the sorting , and therefore the computational complexity of the filter is O ( m log m + k log k ) . Input . A set W of n+k proteomes consisting of two sets of proteomes: a set V of n proteomes 〈V1 , V2 , … , Vn〉 and a set U of z proteomes 〈U1 , U2 , … , Uz〉 , with U taken from taxonomically diverse organisms . Output . A set H = 〈H1 , H2 , … , Hn+k〉 where Hi is the subset of Wi containing conserved proteins with stable copy number . Algorithm . Let pij be the jth protein in Wi , and let pkij[h] be a k-mer from pij of length k , starting at index h , where 0≤h<f given that pij has length f . For those k-mers at the end of each protein where h+k>f , the suffix is expanded by the necessary number of empty characters to fill the remainder of the k-mer . Each k-mer is stored as a 3-tuple consisting of the k-mer , the proteome ID ( i ) , and the gene ID ( j ) . Let D be the alphabetically sorted list of all 3-tuples from both V and U . We define a k-mer cluster to be a block of adjacent k-mers in D in which no k-mer has more than u mismatches with the previous k-mer . Starting from the first k-mer in D , we compare adjacent k-mers to identify all clusters in D . At the end of this process , the k-mers in adjacent clusters are checked against one another and merged by the same rule of no more than u mismatches , a step which circumvents the frequent problem of stray k-mers interrupting what would otherwise be a single block of matches . We call this final set of clusters C . Every protein in pij from W is assigned a pair of integer arrays , Gij and Fij each initialized at every index to 0 ( default size = 10 ) . For each cluster in C , let g be the number of organisms from U with at least one k-mer in the cluster , and let f be the number of total 3-tuples in the cluster with k-mers from U , including repeats . We use Gij and Fij to accumulate the sums of f and g , respectively , for each cluster; the index of the array for a given cluster is selected by a function of the fraction of the total proteomes in U with hits in the cluster . If y is the number of proteomes in U with hits in the cluster , o = ⌊10y/z⌋ . For every protein pij with a k-mer in a given cluster from C , let g and f be added to the values of Gij and Fij at index o , respectively . After passing through all clusters in C , we assign a paralogy score for every protein pij , for each possible value of o , where we define a paralogy score Xijo for each value of o as Xijo= ∑k=ok<10Gij[ o ]/∑k=0k<10Fij[o] . H consists of all proteomes in V and U , where only proteins that have 0<Xijo≤ orthology cutoff ( default = 1 . 3 ) retained . How conserved the final set H is depends on the user’s selection of o . The reference set U is not mandatory . When a reference set is absent , the whole set V is treated as the reference by the algorithm . Computational complexity . As in Algorithm 1 , the time to compile the sorted list of k-mers is O ( m log m ) , where m is the total number of amino acids in W . The clustering is performed in O ( m ) time , and the calculation of final scores is performed in O ( n+k ) time . Therefore , the computational complexity of the filter is O ( m log m ) . Input . A set H of n proteomes 〈H1 , H2 , … , Hn〉 . Output . A distance matrix D of SlopeTree evolutionary distances between all pairs in H , such that Dij is the SlopeTree distance between proteomes Hi and Hj . Algorithm . Let pij be the jth protein in Hi , and let pkij[h] be a k-mer from pij of length k , starting at index h , where 0≤h<f given that pij has length f . For those k-mers at the end of each protein where h+k>f , the suffix is expanded by the necessary number of empty characters to fill the remainder of the k-mer . Each k-mer is stored as a 3-tuple consisting of the k-mer , the proteome ID ( i ) , and the gene ID ( j ) . Let L be the alphabetically sorted list of all 3-tuples . Let mrxy be an exact sequence match of length r , where 1≤r≤k for proteomes Px and Py , where each match is counted exactly once . Let Mrxy be count of all mrxy , where the same sequence is only counted once . For all r in the evolutionarily relevant range , ~r>8 amino acids , we define Dxy as an estimate of the evolutionary distance between proteomes Px and Py , where Dxy is the decay in the histogram of ln ( Mrxy ) as a function of r . Computational complexity . For n organisms and m amino acids , let m = m1+m2+…+mn . The compilation of L is done in O ( m ) , and the sort within all organisms is equal to ΣO ( mi log mi ) which is equal to O ( m log m ) . The match-counting algorithm then requires O ( m log m + n2 ) time . Thus , the time complexity is O ( m log m + n2 ) , with m>>n . We treat the alphabet size as a constant here . Input . A previously calculated SlopeTree distance matrix D ( defined in Algorithm 3 ) , a list Q of proteome pairs flagged as requiring additional correction , and a set R of proteomes , with R taken from taxonomically diverse organisms . Output . A new distance matrix D`identical to D except for the distances between all pairs in Q , which have been recalculated . Algorithm . Let pij be the jth protein in Ri , and let pkij[h] be a k-mer from pij of length k , starting at index h , where 0≤h<f given that pij has length f . For those k-mers at the end of each protein where h+k>f , the suffix is expanded by the necessary number of empty characters to fill the remainder of the k-mer . Each k-mer is stored as a 3-tuple consisting of the k-mer , the proteome ID , and the gene ID . Let S be the alphabetically sorted list of all 3-tuples from R . Let v and w be a pair in Q . Then for this pair , we compile an alphabetically sorted list of 3-tuples and call this list P . Let S and P be merged and this list passed to Algorithm 3 , i . e . the SlopeTree Main Algorithm for counting matches . During the match-counting , let any protein pij contributing a match between v and w with a nit-score ( proportional to the length of the match , described in Implementation ) higher than some cutoff x , and with fewer than y hits among the reference set , be marked . Having reached the end of the merged list of S and P , and having marked all proteins from v and w , we rerun Algorithm 3 on P , but ignoring matches from the marked proteins , to produce a new distance , D`vw . Let the original distance Dvw be replaced by the new distance D`vw , and the matrix D`be the matrix in which every element has been updated in this way for all pairs in Q . Computational complexity . Compiling the alphabetically sorted list S takes O ( r log r ) time , where r is the total number of amino acids in R . Similarly , compiling P takes O ( p log p ) time , where p is the total number of amino acids in v and w . Each first iteration of the SlopeTree main algorithm then requires O ( r log r + p log p ) time , and running the pair requires O ( p log p ) time . This must be repeated for every pair in Q . For a total of n organisms , i . e . a distance matrix to recalculate that is n by n , the worst case scenario is that every pair has been flagged , requiring that n2/2 distances be recalculated , but in practice , and especially after having applied the filters described in Algorithms 1 and 2 , the number of pairs in Q is much smaller . The algorithms behind the four main modules of the SlopeTree package ( S1 Fig ) were described in the Algorithms section . Here we present some important details regarding their implementation , including how the methods address uneven composition of amino acids , the possibility of backwards mutations , and the background of coincidental matches over short k-mer lengths . The source-code for SlopeTree is available at http://prodata . swmed . edu/download/pub/slopetree_v1/slopetree . tar . gz . We observed occasional curvature in the SlopeTree histograms ( Fig 1F ) . The linear fit was inadequate for plots exhibiting this curvature . Manual inspection of the proteins associated with long length matches between organisms with unexpectedly close distances identified several cases of horizontal gene transfer ( HGT ) . We implemented a quadratic fit to address this , which produced better slopes for a number of cases . However , the quadratic fit also performed poorly when it came to large-scale HGT , e . g . cases involving single copy phages . For this reason , we developed the two filters and the final HGT correction ( Algorithms 1–2 , 3 ) . Mobile elements are often present in multiple copies in a single genome , with their k-mers therefore also being present in multiple copies; we used this feature of mobile element k-mer copy number to identify and remove these proteins . This criteria removed an average of 118 proteins from each archaea ( stdev = 116 ) and 162 proteins from each bacteria ( stdev = 246 ) . The archaea with the most mobile elements removed was Methanosarcina acetivorans C2A , which had 744 proteins removed out of a total 4540 . The bacteria with the most mobile elements removed , and which did not show issues with data quality , was Arthrospira platensis NIES-39 , which had 2143 proteins removed out of a total 6630 . The effect this filtering had on the distance to the Eisen-trees was variable; SlopeTree and CVTree show negligible difference before and after the application of the filter; ACS and kmacs showed a small reduction in distance to the Eisen-trees; and D2 and Spaced Words showed a significant reduction in distance to the Eisen-trees . The conservation filter used a taxonomically diverse reference set of organisms to identify proteins with k-mers that had hits for a minimum fraction ( ~o ) of the reference set , and calculated paralogy scores that provided an estimate of a protein’s copy number profile across the entire reference set . This filter was applied to the majority of the ST-trees , in conjunction with the ME-filter . The purpose was to observe how the phylogenetic trees might change as the input was reduced to an increasingly conserved core , and to assess whether these automatic filters could help produce higher quality trees while keeping the methods completely unsupervised . As a validation , we generated histograms from the paralogy scores for proteins with specific keywords in their annotations , with for example ‘ribosomal’ as an instance of a core protein and ‘chemotaxis’ as an instance of an unstable , often horizontally transferred protein ( S3 Fig ) . The former has a sharp peak at the paralogy score of 1 which decreased but does not disappear for increasing o . The latter has two peaks at 0 and 5 , with all paralogy scores of 1 disappearing by o = 2 , indicating that chemotaxis proteins are frequently absent or present in multiple copies . Proteins with paralogy scores less than 1 and greater than 1 . 3 are filtered out; therefore , as o is raised , chemotaxis and other similar proteins are gradually eliminated while the majority of ribosomal proteins and other stable , conserved proteins are retained . For every method , this filtering steadily reduced the distance to the Eisen-trees ( Table 1 ) and organisms that were misplaced ( according to the NCBI taxonomy ) in the unfiltered trees were frequently placed correctly in the more filtered trees . To be valid inputs to SlopeTree , proteomes cannot be filtered beyond a certain level . This is because SlopeTree distances are derived from the decay of k-mers as a function of match length , and when the average proteome size drops below ~100–200 proteins , the algorithm begins to encounter pairs that no longer have measurable or informative slopes . This defines a filtering limit for SlopeTree in the vicinity of o = 8 or o = 9 , but not all alignment-free methods have this constraint . The pair-wise HGT correction was designed to correct very occasional but serious error when a single copy phage was transferred between distal organisms . The mobile element filter is not designed to identify single copy phages , which represent a rare category of phages . The ME filter , conservation filter , and pair-wise HGT correction are separate modules in SlopeTree that are applied at different times and address slightly different issues in the data . However , they overlap in many of the proteins they remove; for instance , the mobile element filter and conservation filter both remove many proteins that the HGT filter would remove , were the conservation filter not applied . In general , we found that by o = 3 or o = 5 , most problematic proteins were already removed and the HGT filter had little impact on the final trees . A series of ST-trees was built for 62 E . coli and 10 Shigella ( Fig 3 and S4 Fig ) , which were all the complete proteomes available for these species at the time of this writing . This was to test the range at which SlopeTree could still resolve sensible evolutionary distances . Escherichia fergusonii and Escherichia blattae were included in the run as outgroups to root the trees , but were removed from the final distance matrices prior to tree-building because their presence excessively compressed the other distances . To assess whether longer k-mers might produce more accurate distances at the strain-level , we built a tree using 20-mers ( Fig 3 ) and another using 40-mers . We did not observe an improvement; the 20-mer and 40-mer trees were in very close agreement , with topological differences arising from short branches mainly in the B2 phylogroup . We built additional trees using proteomes filtered for mobile elements , and also proteomes filtered for stability and conservation , in which the reference set for the conservation filter was simply the entire input . The average number of proteins per proteome for the 72 E . coli and Shigella , prior to filtering , was 4730 ( stdev = 485 ) . When the set was filtered just for mobile elements , the average size was reduced to an average of 4282 proteins ( stdev = 402 ) . This set , with mobile elements removed , was filtered against itself for the smallest possible filtering parameter ( o = 0 ) , reduced the average proteome size to 4071 ( stdev = 362 ) ; for self-filtering on o = 5 , the average size was then 3465 ( stdev = 209 ) ; and for o = 10 , the average size was 1290 ( stdev = 9 ) . For all trees , the trees were highly similar to the unfiltered trees . We performed more aggressive conservation filtering against a reference set of 30 diverse bacteria ( o = 3 ) , leaving an average of 343 ( stdev = 41 ) proteins per proteome . This was done to investigate whether the trees built from the most conserved genes across the entire domain of bacteria matched those built without filtering and those built with loose filtering . Again , we observed only minor changes in topology , mostly involving short branches . As an additional validation , we reduced the unfiltered 20-mer tree to the set considered in Touchon et al . [65] which was used as a reference for another alignment-free method in Sims et al . [66]; these two topologies were also found to be in agreement . The ST strain-level topology also agreed with current phylogroups of E . coli and Shigella . There are different means for determining phylogroups , with some assignments varying between approaches [67 , 68]; SlopeTree supports the grouping of E . coli IAI39 uid59381with phylogroup D and E . coli APEC O78 uid187277 with phylogroup C . Pathotypes do not follow phylogeny [69] and when they were mapped the trees , their placement was scattered . The genes responsible for pathogenicity are frequently mobile elements [56 , 70 , 71] , so we constructed an ST-tree from mobile elements and less conserved proteins removed during filtering on o = 0 , to investigate whether strains of the same pathotype would cluster . We did not see this effect; not surprisingly , this tree differed from the other trees in several placements , but nevertheless held many groupings in common , particularly between the more closely related strains ( S4 Fig ) . When strains differ by very few mutations in DNA , most of these will not cause changes in coding sequence . For such cases , performing phylogenetic analyses by following the easily identifiable mutations at the DNA level is the more accurate and practical approach . A series of ST-trees was constructed for 73 archaea ( Fig 4 and S5 Fig ) . These 73 were all the archaea in Lang et al . [25] that had available proteomes in NCBI . Two archaea were pruned from the distance matrix prior to building the trees: Candidatus Korarchaeum cryptofilum OPF8 uid58601 , and Nanoarchaeum equitans Kin4 M uid58009 . Both were automatically flagged by SlopeTree for having an unusually low number of conserved genes compared to the rest of the set . As with the strain-level analysis , we generated both unfiltered ST-trees and also filtered ST-trees , and also applied our pair-wise HGT correction . These trees were compared to the Eisen-73 and Eisen-71 trees . Differences in filtering parameters produced some changes in topology , with distances to the Eisen-73 tree generally decreasing as filtering increased . For instance , without filtering ( but with pruning ) , the symmetric difference distance was 52 , compared to 38 for filtering on o = 5 . For the purpose of comparison , we also built trees on unfiltered and filtered data using five other alignment-free methods: ACS ( S6 Fig ) , CVTree ( S7 Fig ) , D2 ( S8 Fig ) , kmacs ( S9 Fig ) , and Spaced Words ( S10 Fig ) . A smaller set of trees , due to the long run-time of the program , was calculated for ALFRED-G ( S11 Fig ) . The symmetric difference distances to the Eisen-73 and Eisen-71 trees are shown in Table 1 , with more distances available in S1 Table . We built a series of ST-trees for 495 bacteria on unfiltered data , filtered data ( varying the value of o ) , and with and without the final pair-wise HGT correction ( Fig 5 and S12 Fig ) . As the root , we chose the division between the gram-negative and gram-positive bacteria . Organisms identified by SlopeTree as problematic ( e . g . unusual number of conserved genes , reduced genomes , significantly fragmented assemblies , candidate division , etc . ) were retained throughout the entire SlopeTree run , but pruned from the majority of the final trees ( S1 Text ) . Mobile element and conservation filtering reduced the distance to the Eisen-495 tree for all methods , fixing several misplacements of individual organisms as well as shifting whole branches to locations more in keeping with the current NCBI classifications . By ‘misplacement’ we mean a disagreement with the current NCBI classification . For the purpose of comparison , we built trees on full and filtered data using ACS ( S13 Fig ) , CVTree ( S14 Fig ) , D2 ( S15 Fig ) , kmacs ( S16 Fig ) , and Spaced Words ( S17 Fig ) . We also built trees using ALFRED-G , but could only test the o = 5 and o = 7 inputs due to the long run-time of the program ( S18 Fig ) . The ALFRED-G distances are included in S1 Table . There is no consensus regarding the positions of the deep branches of phylogenetic trees . Even the attempt to root the tree on the division between gram-positive and gram-negative bacteria could not be done cleanly , with the Chlamydiae , Cyanobacteria and Spirochaetes moving between these two groups for different levels of filtering . Not just SlopeTree , but all alignment-free methods have changes in their tree topologies as the inputs are filtered more aggressively . Nevertheless , we observed some stable features in the ST-trees that are stable for the other methods as well . These include a clade consisting of the Gammaproteobacteria , Betaproteobacteria , and Alphapoteobacteria . The Bacteroidetes , Chlorobi , and Gemmatimonadetes form another stable clade , typically neighboring a group consisting of the Spirochaetes and some subset of the Planctomycetes-Verrucomicrobia-Chlamydia ( PVC ) superphylum [72 , 73] . These features are consistent with the Eisen-495 tree . The Deltaproteobacteria however are almost always polyphyletic or paraphyletic . The position of the Acidobacteria is also variable , grouping with the Proteobacteria ( mainly the Deltaproteobacteria ) or the PVC group . The Epsilonproteobacteria are consistently monophyletic , but they group with the Proteobacteria for raw and less-filtered trees ( up to o = 3 ) and the Aquificae or PVC group for more filtered trees ( o = 5 or more ) . SlopeTree usually places the Aquificae and a diverse , sulfur-reducing thermophilic group with the gram-negative bacteria , close to a group of Deltaproteobacteria . Filtering and the pair-wise HGT correction move this clade to an area that is separate from the majority of the gram-negative bacteria ( Proteobacteria , Bacteroidetes , Chlorobi , Verrucomicrobia , Planctomycetes , etc . ) and the gram-positive bacteria ( Actinobacteria , Firmicutes ) alike . The Cyanobacteria are also often found in this area; they are typically on a short , deep branch and in the filtered trees , they neighbor the Deinococcus-Thermus . In the unfiltered ST-tree in which the pair-wise HGT correction was not performed , the Cyanobacteria are grouped with the Proteobacteria , which agrees with the Eisen-495 tree . However , a cursory investigation of the prospective HGT pairs for the members of Cyanobacteria present in the analysis revealed numerous possible transfers with the Proteobacteria , and the pair-wise HGT correction alone , even with no filtering , moved the Cyanobacteria away from the gram-negative bacteria and into the neutral area . This area also often includes a clade consisting of the Thermotogae and Synergistetes , another stable group whose placement in the trees varies between this area and a placement deep within the gram-positive bacteria . The remainder of the tree consists predominantly of gram-positive bacteria . The Firmicutes and Actinobacteria typically share a common root , in agreement with the Eisen-495 tree . The Firmicutes are polyphyletic in all ST-trees , with the Tenericutes branching from within them . Whether the Tenericutes are their own phylum or belong within the Firmicutes is debated [74]; SlopeTree consistently groups them within the Firmicutes , matching the Eisen-495 tree . The occasional presence of the Thermotogae within the Firmicutes is at least in part due to a clear instance of HGT discussed later , but it has been observed that the Thermotogae and Firmicutes , in particular Clostridia , show similarity at the whole-genome level [75 , 76] . The Fusobacteria are also in this clade , first nested within the Firmicutes but then more and more basal as filtering increases . The placement of the Fusobacteria with the gram-positive bacteria , despite their being gram negative , has support [76 , 77] . This generally gram-positive clade also often included the Chloroflexi . Like the Thermotogae , the Chloroflexi mostly stain Gram negative , but are monoderms [78] . This placement is seen in the majority of trees produced by the other alignment-free methods and is also seen in the Eisen-495 tree . It is to be expected that different phylogenetic methods will produce different phylogenetic trees . However , the set of organisms that is misplaced in the trees according to the current NCBI taxonomy is remarkably consistent between all alignment-free methods and many of these misplacements were present in the supermatrix tree and specifically discussed in Lang et al . [25] . We discuss some of them below . We observed two main classes of HGT for the pair-wise HGT correction . The first was associated with single copy phages . D . lykanthroporepellens and both Syntrophobacter fumaroxidans and Desulfarculus baarsi serve as an example of this . The second was related to adaptation-associated proteins . Petrotoga mobilis and Mahella australiensis , which shared a transfer of proteins associated with resistance to a toxic environment , are an example . Both were addressed by means of a combination of mobile element filtering and a sufficiently high value for o , and in general this was our preferred approach because it is significantly more efficient than running the HGT filter on a large number of pairs . However , we did observe that for a very low o , or when filtering was not applied , the pair-wise HGT correction was able to correct the placement of D . lykothroporepellens , D . mccartyi , and P . mobilis ( S19 Fig ) . In addition , it amended the placement of Leptospira biflexa serovar Patoc and Leptospira interrogans serovar Lai , two Spirochaetes which every alignment-free method misplaced unless using a very high level of conservation filtering . Rhodothermus marinus and Salinibacter ruber M8 , classified as Bacteroidetes , were also moved from the Chlorobi back to the Bacteroidetes . The correction also caused some substantial reordering of the deeper branches . The Gammaproteobacteria , which are completely monophyletic in the uncorrected tree , are split into two groups in the HGT-corrected tree , in both cases forming a monophyletic clade with the Betaproteobacteria; this split is often seen in the other alignment-free methods and may be an indication of a missing “eta” class for the Proteobacteria [25 , 84] . The pair-wise HGT correction also removed the Cyanobacteria from the Proteobacteria , placing them close to the root alongside the Deinococcus-Thermus which were also shifted out of the Firmicutes . The Spirochaetes and Chlamydiae were also moved from the gram-positive bacteria to the gram-negative bacteria . The symmetric difference distance [61] was calculated between all alignment-free trees and the Eisen-trees , using the treedist program in PHYLIP [87] . However , we note that the Eisen-trees are only approximations of the real evolutionary history , and that the methods should not be judged as “better” or “worse” purely according to their distances to these approximations . The kmacs method , with mobile element filtering and conservation filtering on o = 7 , achieved the closest tree to the Eisen-tree for both bacteria and archaea , with a symmetric difference distance of 350 and 32 . D2 also achieved a distance of 32 to the Eisen-71 tree . For bacteria and archaea , SlopeTree achieved 384 and 38 , both at o = 5 . Filtering lessened the distance to the Eisen-trees for all methods ( Fig 6B and Table 1 ) . We observed a distinct difference in the nature of the branch lengths between different methods; D2 , SlopeTree and Spaced Words fall into one group , having a wider range of branch lengths , while ACS , CVTree , kmacs , and ALFRED-G have branch lengths that are restricted to a more narrow range ( Fig 6A , 6C and 6D ) . ACS appears to be the most restricted in this regard , and we found that by applying the conservation filter , the range for a given method’s distances was somewhat widened .
Different measures of evolution , for instance different alignment-free methods , will produce different trees . Generally , these measures are correlated , generating highly concordant trees . Each alignment-free method defines similarity between organisms in its own units , but it still needs to be established how each of these measures can be transformed into units of accumulation-of-mutations and with what level of accuracy . SlopeTree was designed to provide a measure with a close relationship to the accumulation of mutations . In the absence of selection , this relationship would be given by a simple formula , but at larger evolutionary distances , the slope is defined by slowly evolving protein segments subject to strong negative selection . At the domain level , the relationship becomes nonlinear and requires calibration between the slope and the number of accumulated mutations . At very large distances , such as those between domains , the slope loses its relationship to evolutionary distance entirely . However , this is only significant for rooting archaeal and bacterial phylogenies . The uniformity of the branch lengths from the “root” to the tips in the SlopeTree trees is not an artifact of the distance measure being nonlinear or saturating at some value . It may be a consequence of looking at a large number of conserved sites and if a particular locus evolved faster for a particular genome pair , its contribution becomes much smaller . Heterotachy , which is variable between positions in an alignment , has very different consequences in terms of branch length estimation for alignment-based methods and current alignment-free methods . Considering that there is much larger variability in branch lengths by alignment-based methods , it appears that more uniform branch lengths are a consequence of two factors: averaging between more proteins and potentially smaller sensitivity to heterotachy which is variable between positions in an alignment . SlopeTree includes a filter for mobile elements and a conservation filter which is applied to all proteomes prior to the main run . A conservation filter follows , which is adjustable . As the level of filtering increased , the distances between the ST-trees and the Eisen-73 or the Eisen-495 trees decreased . All other alignment-free methods that we tested also benefited from filtering the data prior to running , at least in terms of their distances becoming closer to the Eisen trees . An additional benefit to this is that filtering the data beforehand decreases the run-times . The number of matches contributing to the assessment of evolutionary distances can be limited for longer distances or small genomes . Including mismatches adds a substantial number of informative , i . e . non-random , matches to the analysis . As can be seen with kmacs , the inclusion of mismatches can greatly improve phylogenetic distances . SlopeTree is essentially a type of survival analysis; therefore , it can apply to partial matches just as well as to those that are exact , and it is our expectation that such extension will produce even better results .
The archive all . faa . tar . gz was downloaded from the NCBI ftp website ( ftp://ftp . ncbi . nih . gov/genomes/Bacteria/ ) in May 2015 . The archive taxdump . tar . gz was downloaded from the NCBI taxonomy website ( ftp://ftp . ncbi . nlm . nih . gov/pub/taxonomy/ ) also in May 2015 . In the NCBI taxonomy , the root nodes for bacteria and archaea are 2 and 2157 , respectively , and out of the 2774 organisms in the FASTA archive , 165 were identified as archaea and 2607 as bacteria . The Maximum Likelihood trees , S1 and S4 files from Lang et al . [25] , built from the concatenations of 24 conserved proteins , were downloaded and organisms compared to those present in the FASTA archive . Allowing for some imperfect matches ( e . g . Haliangium ochraceum SMP 2 DSM 14365 in the ML tree , opposed to Haliangium ochraceum DSM 14365 in the archive ) and some differences in strains ( e . g . Eubacterium siraeum DSM 15702 uid54603 in the ML tree , opposed to Eubacterium siraeum uid197160 in the archive ) , 73 archaea and 495 bacteria were found in common between the ML trees and the archive . Two lists were compiled of organisms to remove from the ML trees and these lists and trees were given as input to the program nw_prune , from the package newick-utils ( version 1 . 6 ) [90]: . /nw_prune Eisen_newick_ML_journal . pone . 0062510 . s008 . txt $ ( cat pruning_bacteria . txt ) > eisen_495_tree_bacteria_newick . txt . /nw_prune Eisen_ML_841_journal . pone . 0062510 . s011 . txt $ ( cat pruning_archaea . txt ) > eisen_73__tree_archaea_newick . txt These two supermatrix-derived trees are referred to as the Eisen-73 tree and the Eisen-495 tree and were produced for comparison purposes ( S2 Fig ) . For all distance matrices produced by SlopeTree and the other methods discussed here , we used rapidNJ version 2 . 0 . 1 [91] to construct the trees . . /rapidnj distance_matrix . txt > distance_matrix_tree . txt A raw tree consisting of the full sets bacteria and archaea is available for each method ( SlopeTree and alternative alignment-free methods ) . The remaining trees were pruned of the organisms that SlopeTree automatically flagged as problematic , 2 for archaea and 50 for bacteria ( S1 Text ) . The distance matrices were pruned of the flagged organisms before being passed to rapidNJ . Pruned versions of the Eisen-trees were also created ( S2 Fig ) , using nw_prune as described above with the organisms flagged by SlopeTree added to the file of organisms to prune . This was necessary for the pruned trees to be comparable to the Eisen-trees . The scripts we refer to in this section are included in the SlopeTree package ( http://prodata . swmed . edu/download/pub/slopetree_v1/slopetree . tar . gz ) . The figures for all of the trees in this manuscript were generated using the ITOL web-server [92 , 93] . All bacterial proteomes were moved to the directory FAA within the directory Bacteria . All archaeal proteomes were moved to the directory FAA within the directory Archaea . All proteomes for the strain-level analysis were moved to the directory FAA within the directory Ecoli . The distance matrices for these two sets were then generated with the following two scripts: bash dSTm . sh Bacteria/ 20 B . . /Taxonomy/ bash dSTm . sh Archaea/ 20 A . . /Taxonomy/ bash dSTm . sh Ecoli/ 20 B . . /Taxonomy/ The distance matrices were then passed to rapidNJ . We refer to these trees as the “raw” trees . We manually selected thirty diverse bacteria from the raw ST-tree as our reference set for the bacterial runs . Similarly , we manually selected ten diverse archaea for the archaeal runs . The specific organisms selected are listed in S2 Text . The reference sets for bacteria and archaea were moved to Bacteria_ref/FAA and Archaea_ref/FAA , respectively . We then filtered them for conservation , using our conservation filter , for the parameter of o = 7: For bacteria: bash pFilt . sh Bacteria_ref/ 20 . /fpwrite Bacteria_ref/–f 10 –o 7 For archaea: bash pFilt . sh Archaea_ref/ 20 . /fpwrite Archaea_ref/–f 10 –o 7 These commands generated proteomes that had been reduced to their core proteins . These reduced proteomes were moved to new directories Bacteria_ref_10_7/FAA and Archaea_ref_10_7/FAA and the list of merged and sorted 20-mers generated for each of them: bash dMT . sh Bacteria_ref_10_7/ 20 B bash dMT . sh Archaea_ref_10_7/ 20 A This created a set of sorted 20-mers from conserved proteins from a diverse reference set for bacteria and for archaea . These sets were used as the reference for the mobile element filtering: . /mef Bacteria/ Bacteria_ref_10_7/MERGED_TAGS/ . /mef Archaea/ Archaea_ref_10_7/MERGED_TAGS/ . /mef Ecoli/ Bacteria_ref_10_7/MERGED_TAGS/ This produced , for bacteria , archaea and our set of E . coli , a set of proteomes in which the mobile elements were eliminated . These reduced proteomes were automatically written out to Bacteria/FAA_mobelim , Archaea/FAA_mobelim and Ecoli/FAA_mobelim . We moved these reduced proteomes to Bacteria_MEF/FAA , Archaea_MEF/FAA and Ecoli_MEF/FAA and moved the organisms that had been chosen for the reference sets to FAA_ref directories within each main directory . We then ran the main SlopeTree script to produce the final distance matrices: bash dSTm . sh Bacteria_MEF/ 20 B . . /Taxonomy/ bash dSTm . sh Archaea_MEF/ 20 A . . /Taxonomy/ bash dSTm . sh Ecoli_MEF/ 20 B . . /Taxonomy/ Trees were then built using rapidnj . The FAA and FAA_ref directories from Bacteria_MEF/ and Archaea_MEF/ , and the FAA directory for Ecoli_MEF , were copied to Bacteria_MEF_CF , Archaea_MEF_CF , and ECOLI_MEF_CF , respectively . We then ran the filtering code: bash pFilt . sh Bacteria_MEF_CF/ 20 B bash pFilt . sh Archaea_MEF_CF/ 20 A bash pFilt . sh Ecoli_MEF_CF/ 20 B For bacteria and archaea separately , we generated five sets of proteomes filtered on o = 0 , o = 1 , o = 3 , o = 5 and o = 7 . The following two commands use o = 3 as an example: . /fpwrite Bacteria_MEF_CF/–f 10 –o 3 . /fpwrite Archaea_MEF_CF/–f 10 –o 3 This command generated filtered proteomes , still divided into main set and reference set , for both bacteria and archaea . These filtered proteomes were moved to their own directories , Bacteria_MEF_CF_10_3 and Archaea_MEF_CF_10_3 for the case of o = 3 and so on for other values of o . Finally , each of these new directories , which contained an FAA and FAA_ref that had been reduced for both mobile elements and also less conserved proteins , was passed to the main SlopeTree script: bash dSTm . sh Bacteria_MEF_CF_10_3/ 20 B . . /Taxonomy/ bash dSTm . sh Archaea_MEF_CF_10_3/ 20 A . . /Taxonomy/ Similar steps were followed to generate the filtered proteomes for our set of E . coli , using the same set of 30 bacteria in FAA_ref for the more aggressive filtering . In addition , E . coli was filtered against itself , i . e . no reference set . All that was required for this self-filtering was to not provide an FAA_ref directory when pFilt . sh was run . Trees were then built with rapidNJ . The correction for HGT was applied only to proteomes already filtered of mobile elements and filtered on o = 3 , o = 5 , and o = 7 . For o = 3 , the command was: . /fh Bacteria_MEF_CF_10_3/ For each data set , this command produced new distance matrix which was then passed to rapidNJ . Trees were built using several other , alignment-free methods: ACS , CVTree , D2 , kmacs , Spaced Words , and ALFRED-G . Each method was run on the 495 bacteria and 73 archaea for: a ) raw proteomes , b ) proteomes filtered of mobile elements , and c ) proteomes filtered of mobile elements and also filtered for conservation on o = 0 , 1 , 3 , 5 , and 7 . The final pair-wise HGT-correction which was applied to the SlopeTree runs for o = 3 , 5 , and 7 was not applied to these alternative methods because unlike the mobile element filter and conservation filter , the pair-wise HGT correction currently cannot be run independently of SlopeTree . For the matrices produced by these alternative methods , we built trees using rapidNJ . Version 1 . 2 of the ACS code was used to build the ACS trees with the following command: . /ACS -a <path to ACS directory>/ACS_input_file—o distance_matrix . txt—A -A ACS_matrix . txt Trees were built using rapidnj on the file written out by the -o option . Version 4 . 2 of CVTree was used . The commands to build the matrices were the following: . /cvtree—i -i cvtree_input_file . txt -d FAA/ -k 6 -t aa -c out/ . /batch_dist . pl 1 . 5 cvtree_input_file . txt out/ out_matrix_k6 . txt Version 1 . 0 of D2 was used . The command to build the matrices was the following: java -Xmx126g -jar jD2Stat_1 . 0 . jar -a aa -i input . faa -o matrix We ran kmacs with k = 14: . /kmacs input . faa 14 We ran Spaced Words with k = 12 and Euclidean distances . Evolutionary distances were not available for amino acid sequences: . /spaced—k -k 12 –d EU input_file . faa We ran ALFRED-G with k = 6 and x = 1 . build/alfred . x -f input . fas -o output . txt -k 6 -x 1 All trees were compared to the Eisen-trees using the treedist tool from PHYLIP [87] for the symmetric difference distance . Using a keys file generated for the purpose of finding matches between the original FASTA archive and the Eisen-trees , we renamed the nodes of the Eisen-trees and alignment-free trees so that they were identical and renamed the two tree files intree and intree2 for treedist .
|
Due to their lack of distinct morphological features , bacteria and archaea were extremely difficult to classify until technology was developed to obtain their DNA sequences; these sequences could then be compared to estimate evolutionary relationships . Now , due to technological advances , there is a flood of available sequences from a wide variety of organisms . These advances have spurred the development of algorithms which can estimate evolutionary relationships using whole genomes , in contrast to the more traditional methods which used single genes earlier and now typically use groups of conserved genes . However , there are many challenges when attempting to infer evolutionary relationships , in particular horizontal gene transfer , where DNA is transferred from one organism to another , resulting in an organism’s genome containing DNA that does not reflect its evolution by descent . We developed a new whole-genome method for estimating evolutionary distances which identifies and corrects for horizontal transfer . We found that for SlopeTree and all other whole-genome methods we applied , horizontal transfer causes some evolutionary distances to be grossly underestimated , and that our correction corrects for this .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"sequencing",
"techniques",
"taxonomy",
"horizontal",
"gene",
"transfer",
"split-decomposition",
"method",
"microbiology",
"gene",
"transfer",
"multiple",
"alignment",
"calculation",
"phylogenetics",
"data",
"management",
"phylogenetic",
"analysis",
"genetic",
"elements",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"mobile",
"genetic",
"elements",
"computer",
"and",
"information",
"sciences",
"sequence",
"alignment",
"proteins",
"evolutionary",
"systematics",
"molecular",
"biology",
"molecular",
"biology",
"assays",
"and",
"analysis",
"techniques",
"archaean",
"biology",
"biochemistry",
"computational",
"techniques",
"proteomes",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"evolutionary",
"biology",
"evolutionary",
"processes"
] |
2016
|
Phylogeny Reconstruction with Alignment-Free Method That Corrects for Horizontal Gene Transfer
|
Methamphetamine ( Meth ) is abused by over 35 million people worldwide . Chronic Meth abuse may be particularly devastating in individuals who engage in unprotected sex with multiple partners because it is associated with a 2-fold higher risk for obtaining HIV and associated secondary infections . We report the first specific evidence that Meth at pharmacological concentrations exerts a direct immunosuppressive effect on dendritic cells and macrophages . As a weak base , Meth collapses the pH gradient across acidic organelles , including lysosomes and associated autophagic organelles . This in turn inhibits receptor-mediated phagocytosis of antibody-coated particles , MHC class II antigen processing by the endosomal–lysosomal pathway , and antigen presentation to splenic T cells by dendritic cells . More importantly Meth facilitates intracellular replication and inhibits intracellular killing of Candida albicans and Cryptococcus neoformans , two major AIDS-related pathogens . Meth exerts previously unreported direct immunosuppressive effects that contribute to increased risk of infection and exacerbate AIDS pathology .
Chronic methamphetamine ( Meth ) abuse has reached epidemic proportion throughout the United States , where a 2003 survey indicated that approximately 5% of the population over 12 years of age has tried Meth and the rate of treatment admissions for primary Meth abuse increased over 3-fold ( The DASIS Report , http://www . oas . samhsa . gov/2k6/methTX/methTX . htm ) [1 , 2] . In particular , among gay and bisexual men [3] Meth it is associated with high-risk sexual behavior , HIV viral infection , and a high incidence of AIDS [4] . Meth exacerbates AIDS pathology , including cognitive deficits [5 , 6] , and is strongly suspected to inhibit normal immunological response to secondary infections such as hepatitis C , which is prevalent in those who smoke or take Meth intranasally [7] . It has recently been suggested that Meth could contribute to a particularly rapid progression of AIDS in individuals exposed to a strain of HIV that is resistant to drug treatment [8 , 9] . Indeed , animal studies clearly demonstrate that Meth suppresses both innate and adaptive immunity [10 , 11] , enhances cytokine production in combination with HIV TAT protein [12] , and alters gene expression in cells of the immune system [13] . However , the molecular basis for Meth's immune suppression is unknown . Here we have examined the relationship between Meth and the impairment of specific immune cell functions . In the clinical setting , Meth abusing individuals who present with opportunistic infections possess high blood and tissue levels of the drug . Meth is generally self-administered in this population in binges of 3–4 grams ingested over a six day interval [14] , with an initial ingestion of ∼0 . 5 gram ( Judith Rabkin , Columbia University , personal communication ) and a total mean level of 2 . 2 grams ingested during the first day [15] . Such high levels of administration result in a blood concentration of ∼10–50 μM Meth , and levels in the hundreds of micromolar range in organs including brain and the spleen [16] ( see Results ) . Pathogens are processed and displayed by antigen presenting cells ( APCs ) for T cell recognition . Antigen presentation in tissue resident macrophages , as well as dendritic cells , involves fluid phase or receptor-mediated endocytosis followed by fusion of the phagosome with specialized lysosomes known as MHC class II compartments ( MIIC ) [17] . The foreign antigens are partially degraded by lysosomal hydrolases and the resulting peptides are loaded on MHC class II molecules and transported to the cell membrane to be presented to T cells . Both pathogen transport to the MIIC and its degradation into immunogenic peptides are functions that require an acidic endosomal pH . Endogenous antigens and viruses can be delivered to the MIIC upon autophagosome fusion [18] . Autophagy also mediates single-stranded RNA virus detection and consequent interferon-α release in plasmacytoid dendritic cells activating anti-viral cellular defense mechanisms in uninfected cells [19] . Further evidence shows the direct role of autophagosomal-lysosomal degradation in elimination of intracellular HSV-1 viral particles [20] and that the HSV-1 has alternative methods to counteract cellular autophagy [21] . In this study , we have identified a novel molecular mechanism that explains the immunosuppressive effects of Meth via the alkalization of acidic organelles in dendritic cells and macrophages that are critical for the immunological function of these APCs .
To model effects of Meth on the immune system , we estimated Meth levels in drug abusers . Meth is self-administered intravenously , by nasal inhalation , anally , and orally , in typical doses of 250–500 mg by occasional users to levels as high as 1 g by chronic abusers ( personal communication , Perry N . Halkitis , New York University ) . Meth blood levels measured in individuals detained by police in California were 2 . 0 μM on average but as high as 11 . 1 μM [22] . Controlled studies indicate that a single 260 mg dose reaches a level of 7 . 5 μM [22] . Thus , a single dose of 260 mg – 1 g would be expected to produce 7 . 5 – 28 . 8 μM blood Meth levels . The abusers however tend to self-administer METH in binges , and as the drug exhibits a half-life of 11 . 4 - 12 h [23 , 24] , this leads to higher levels . Recently published studies modeling binge pattern of use in individuals show that after the fourth administration of 260 mg during a single day produces maximum blood levels of 17 μM , reaching 20 μM on the second day of such a binge [22] . Thus , binge doses of 260 mg – 1 g would produce 17 - 80 μM blood Meth levels . The estimates appear consistent with blood levels detected after fatalities [25–27] , ranging as high as 84 μM for an individual for whom Meth intoxication was determined as the cause of death ∼16 hours after ingestion . It is also important to estimate how Meth is distributed from blood to other tissues involved in immune response , particularly the spleen , which houses high numbers of dendritic cells . Tissue-to-serum Meth ratios in rat are: brain , 9 . 7; kidney , 35 . 3; spleen , 14 . 3 [16] . Thus , relevant levels in spleen , the organ critical for immune response , after administration of 250 mg – 1 g as a single dose is 100 – 400 μM , and during binges between 240 – 1144 μM . Dendritic cell MIIC processing organelles are characterized by acidic pH [28] , a limiting membrane enclosing internal vesicles or lamellae [17] , the presence of proteases [29] , internal expression of LAMP [30] and MHC II proteins [31] . The morphological and functional integrity of the MIIC depends on the maintenance of an acidic pH [32] due to the action of the V-ATPase , which may also regulate transport from early to late endosomes and lysosomes [33] . Meth and its metabolite , amphetamine , are membranophilic weak bases that collapse intracellular organelle pH gradients in neurons [34 , 35] . We therefore tested whether Meth collapsed intracellular organelle pH gradients in dendritic cells by monitoring quenching of acridine orange , a weak base vital dye that accumulates in acidic organelles including endosomes and lysosomes [35] . We found that Meth concentrations of 50 μM or higher rapidly collapsed acidic organelle pH gradients in dendritic cells ( Figure 1A–1B ) . LysoSensor Yellow/Blue ratiometry is a generally accepted method to measure organellar pH in live cells , and was used to measure average pH in acidic organelles using LysoSensor Yellow/Blue fluorescent dye . After 10 min treatment with Meth ( 50 or 100 μM ) or chloroquine ( 10 μM ) acidic organellar pH was elevated ( pH 6 . 5 , 6 . 9 and 6 . 4 respectively ) significantly above the levels in untreated control cells ( pH 4 . 8 ) ( Figure 1C ) . As predicted with alkalizing agents , Meth also disrupted MIIC structure [36] , producing large organelles ( > 1 μm diameter ) devoid of internal vesicles ( right panels ) with LAMP-1 and MHC II staining confined to the limiting membrane ( Figure 1D and 1E ) . Chloroquine ( Clq ) [10 or 20 μM] , another weak base was used as a control showed similar effects to Meth as described above ( Figure 1A–1E ) . At the used concentrations ( 20 and 100 μM ) Meth did not affect cell viability as determined by flow cytometric analysis ( Figure S1 ) . To investigate whether Meth-induced endosomal alkalization blocks antigen processing by impairing dendritic cell lysosomal proteolytic degradation of foreign proteins , we exposed cells to the fluorescently labeled MHC II antigens [bovine serum albumin ( BSA ) , casein , and ovalbumin ( OVA ) ] , and measured the degradation of each protein by western blot . In untreated cells , each antigen was proteolytically degraded , while antigen degradation was blocked in Meth or Clq treated cells ( Figure 2A and 2B ) . Meth ( 10 , 50 , 100 μM ) and Clq ( 10 , 20 μM ) effectively inhibited processing of antigens previously taken up by dendritic cells ( Figure 2A ) . Control experiments showed that Meth did not block endocytosis of 2 μm fluorescent dextran beads by dendritic cells ( Figure S2 ) , indicating that the drug does not inhibit nonspecific phagocytosis [37] . To determine the stage at which Meth compromised post-endocytic proteolytic antigen processing , we prepared fractions of early and late endosomes and lysosomes from dendritic cells exposed to BSA and casein antigens . Each fraction was examined for β−hexosaminidase to identify lysosomes and late endosomes , the transferrin receptor ( TrfR ) to identify early endosomes , and LAMP-1 to identify late endosomes and lysosomes ( Figure 2C and 2D ) . No detectable antigen remained in early endosomes under any of these conditions . There were , however , much higher levels of BSA and casein in lysosomes and late endosomes of Meth- and Clq-treated cells than in controls . In particular , casein was completely degraded in untreated cells but relatively unprocessed in dendritic cells treated with Meth or Clq ( Figure 2E ) . Thus , Meth inhibited antigen proteolysis within late endosomal/lysosomal compartments . Similarly , processing of invariant chain was compromized after Meth treatment as indicated by increased levels of p25/28 and p10 fragments during the chase time-point ( Figure 2F ) . MHC II antigen presentation stimulates T cell proliferation , providing a means to measure effects of Meth on antigen presentation . We prepared cultures of immature bone marrow-derived dendritic cells and splenic purified T cells , both from OTII transgenic mice [38] . Using OVA as an antigen , we assayed cellular proliferation of T cells by radiolabeled thymidine uptake and incorporation into DNA . At all levels tested , Meth decreased the T cell proliferative response to the intact antigen ( Figure 2G ) but not the pre-processed OVA-323–339 peptide ( Figure 2H ) . In addition to blocking lysosomal antigen degradation , Meth could disrupt processing and presentation of antigens by inhibiting autophagosome formation , thereby halting antigen delivery to MIIC . To test this possibility , we prepared dendritic cells from a transgenic mouse expressing a GFP-fused autophagosome-associated protein LC3 ( GFP-LC3 ) which has been used as an in vivo autophagosomal marker [39] . Consistent with previous reports in neurons [40] , 50–500 μM Meth induced autophagosome accumulation in dendritic cells ( Figure 3A–3C ) . Our data show that Meth does not block autophagosome formation , but rather impairs lysosomal-autophagosomal degradation , resulting in the accumulation of autophagosomes , and impaired intracellular antigen proteolysis . Together with the evidence from endosomal/lysosomal fractions ( Figure 2C–2E ) , these results strongly suggest that Meth inhibits degradative antigen processing by disrupting pH gradients . As with dendritic cells , we also found that Meth and Clq effectively collapsed the intracellular pH gradients within macrophages ( Figure S3A–S3C ) and blocked autophagosome degradation , resulting in the accumulation of GFP-LC3 labeled autophagosomes ( Figure S4A–S4D ) . Similar results were obtained with the fluorescent dye monodansylcadaverine , a lipophilic weak base that accumulates in lysosomes and autophagosomes ( Figure S4E ) . Since Meth users can also present with bacterial infections [41] , we examined whether Meth disrupts macrophage phagocytosis , a primary mechanism for clearance of these extracellular pathogens . Murine peritoneal-derived macrophages were incubated in the presence or absence of Meth . After 2 h , IgG-antibody coated erythrocytes [E ( IgG ) ] were added to each of the conditions and the number of ingested erythrocytes counted . We found that Meth ( 50 and 250 μM ) inhibited E ( IgG ) phagocytosis ( Figure 4A and 4B ) by 20 and 45% , respectively . Meth did not inhibit the receptor-independent endocytosis of Lucifer Yellow ( Table S1 ) . Similar effects were observed with Clq ( Figure 4A and 4B ) These results demonstrate that Meth inhibits Fcγ-mediated phagocytosis in macrophages and are consistent with the observations that the macrophage Fcγ receptors are continuously recycled between phagosomes and the plasma membrane [42] , a process that requires appropriate acidification of secretory vesicles and tubules for trafficking [43] . These results demonstrate that Meth inhibits Fcγ-mediated phagocytosis in macrophages . As a next step we analyzed the effects of Meth on Candida albicans ( Ca ) and Cryptococcus neoformans ( Cn ) phagocytosis and killing by murine macrophages since these two organisms are the most commonly isolated fungi in individuals infected with HIV [44] . Clq and Meth ( 10 and 50 μM ) inhibited phagocytosis of Ca and Cn by macrophages by 40% ( Figure 4C ) . Moreover , Meth enhanced the proliferation of fungi within macrophages ( Figure 4D ) , indicating that intracellular replication of both yeast was facilitated by Meth . In contrast , Clq had no significant , or slightly reducing , effect on Ca ( p= 0 . 056 ) and Cn ( p = 0 . 060 ) CFU numbers in macrophages . Control experiments showed that in the absence of macrophages , Ca and Cn proliferation was unaffected by the addition of Clq or Meth to the BHI medium ( data not shown ) . To further examine why chronic Meth abuse has been recently associated with the rapid development of immune deficiency among gay and bisexual men [3] , we studied the effect of Meth on HIV proliferation in macrophages from HIV-transgenic mice . These JR-CSF/huCycT1 double transgenic mice express HIV-1 JR-CSF , which is a full length R5 HIV-1 provirus regulated by the endogenous HIV-1 LTR , as well as the human cyclin T1 controlled by a murine CD4 expression cassette [45] . These mice have constitutive HIV production in CD4 T lymphocytes and monocytes . GM-CSF differentiated bone marrow cells were either left untreated or treated with a range of Meth levels ( 10 , 50 , 150 μM ) or NH4Cl ( 10 mM ) , another weak base , for 7 days . Cell supernatants were then collected and p24 , a secreted HIV-specific protein , was quantified by Elisa . Cells from JR-CSF/huCyc T1 mice treated with Meth for 7 or 9 days exhibited a 30–60% increase ( Figure S5A ) of p24 antigen production . Clq does not provide a positive control , since its well-established inhibition of HIV production is probably due to inhibition of viral capsid protein glycosylation in the Golgi [46] . Thus , in this experiment NH4Cl was used as positive control ( Figure S5A ) . To determine whether HIV replication was affected by Meth in vivo , we studied the effect of Meth on HIV virus proliferation in the JR-CSF/huCycT1 double transgenic mice [45] . Mice were treated with increasing concentration of Meth over a 7 day-period . One group of animals received 5 mg/kg of Meth at day 0 , 2 and 4 and was sacrificed at day 6 ( low Meth ) . Another group received 6 mg/kg at day 0; 7 mg/kg at day 2 and 7 . 5 mg/kg at day 4 ( high Meth ) and was also sacrificed at day 6 . Copy number of HIV-1 RNA was quantified in the serum of each mouse by RT-PCR using primers spanning the highly conserved region of the HIV-1 gag gene . No statistically significant differences were observed between the untreated and the Meth-treated mice ( Figure S5B ) . These data suggest that Meth treated HIV-1 transgenic mice do not exhibit an increase in HIV viral load .
We find that the widely abused addictive psychostimulant , Meth , at pharmacologically relevant levels acts as an immunosuppressive agent , due to its inhibition of endosomal acidification . These actions result in Meth's inhibition of antigen presentation and phagocytosis . Maintenance of low endosomal and lysosomal pH serves many functions , including regulation of protein degradation , pathogen inactivation , and regulation of the amount of several surface receptors . All of these functions require active transport via the endocytic pathway and fusion with lysosomal compartments . First , via alkalization , METH-inhibited lysosomal-autophagosomal degradative function for both exogenously and endogenously internalized antigens resulting in accumulation of proteins entering the endocytic pathway through phagocytosis , as well as autophagic vacuoles . Chaperone-mediated autophagy and macroautophagy in lysosomes have been described as major pathways for endogenous antigen processing in MHC class II compartments [18 , 47 , 48] and as a means to directly degrade intracellular virus particles [20 , 21] . Thus , through its alkalizing effect , Meth blocks normal antigen processing and presentation . Progression of internalized antigens along the endocytic pathway rely on the progressive maturation of early endosomes into late endosomes and ultimately to lysosomes . Maintenance of an acidic internal pH and a pH gradient in these compartments is important for the cargo progression . Endosomal acidification is accomplished by H+ transport across the endosomal limiting membrane by the proton pump vacuolar ATPase ( V-ATPase ) . The recent discovery that V-ATPase interacts with components of the endocytic transport machinery indicates that V-ATPase is also a pH sensor that regulates early to late endosomal transport [33] . This would explain why endosomal alkalization by Meth not only disrupts antigen processing but phagocytosis and cargo progression along the endosomal pathway . Second , endosomal alkalization by Meth inhibited Ca and Cn phagocytosis and killing by macrophages . These effects are expected to be particularly devastating in AIDS-related disorders , since Ca and Cn are the two most commonly isolated fungi from sterile body fluids obtained from HIV infected individuals [44] . Also , these results could explain recent reports of rapid insurgence of AIDS in Meth-addicted individuals soon after infection with HIV . Thus , Meth also blocks pathogen killing by macrophages . Exposure to Clq and other weak bases was already been shown to inhibit growth of Cn in macrophages [49 , 50] even though Clq has no direct toxicity to Cn [50] . Not all basic compounds are equivalent in their fungicidal activity; for example , ammonium chloride also enhances the anti-cryptococcal activity of macrophages , yet the potency of ammonium chloride is less than that of Clq , which may be a result of the ability of ammonium chloride to inhibit phagolysosomal fusion [51] . Hence , Meth apparently has complex effects on macrophages that result in an intracellular milieu that enables the replication of the examined pathogenic fungi . Third , even though an increase in p24 secretion has been observed in Meth-treated dendritic cells and macrophages , a direct effect of Meth on HIV viral load could not be demonstrated . It is likely that viral proteins enter late endosomal and lysosomal compartments independently of viral assembly , and due to inhibition of processing , more viral proteins are secreted in the extracellular milieu . Within the infected macrophages , HIV has been previously shown to assemble in compartments with characteristics of multivesicular late endosomes ( CD63 , Lamp-1 , CD81 and CD82 positive ) [52] . It now appears that the HIV particles present in multivesicular endosomes are the results of endocytosis , and the site of viral assembly is at the invagination of the plasma membrane particularly enriched in tetraspannin proteins [53] . These results would explain why , by compromising endosomal pH , an increase in HIV viral load has not been observed . In conclusion , the immunosuppressive effects of Meth are consistent with reports that Meth-treated mice demonstrate decreased immunity [10 , 11] . To be noted is that even though there is a linear concentration-dependent response between Meth concentration and pH disruption such an effect is less evident at the biological level . In all assays ( antigen processing , Ca or Cn killing and p24 production ) different Meth dosages behave very similarly . Likely a small disruption in the endosomal pH is sufficient to alter the microenvironment and endosomal-related functions . This is consistent with the effects of low concentrations of Meth tested on lysosomal pH in Figure 1C , with alkalinization of > 1 pH unit , which would effectively inhibit lysosomal protease activity . The collapse of endosomal pH by Meth and the resulting decrease of normal immune response provide an explanation for the compromised immunity and exacerbate infections occurring in Meth abusers . In fact , Meth is strongly suspected to more dramatically inhibit normal immune responses than other drugs of abuse since Meth users often present with skin lesion and “Meth mouth” , a devastating periodontal disease ( http://www . drugfree . org/ ) [54] and [1 , 2] . A similar immunosuppressive activity has also been shown for chloroquine , also a well known inhibitor of endosomal acidification [36] . In particular , Meth immunosuppression may underlie the mechanism of the recently reported extremely rapid development of immune deficiency , with devastating effects in AIDS-related disorders in Meth abusers that had contracted HIV . In particular , there is evidence suggesting the presence of a new population of HIV+ positive men who are developing AIDS over months rather than over 10 or more years as is typical . The most widely reported individual was documented by Dr . Martin Markowitz of the Aaron Diamond AIDS Research Center in New York in 2004 . In this individual , a gay man who was also a Methabuser tested HIV negative in May , 2003 , then likely contracted HIV during unprotected sex in mid-October 2004 , displayed acute retroviral syndrome in November , 2004 and 3-drug-class-resistant HIV-1 ( 3DCR HIV ) with apparently rapid progression to AIDS by December , 2004 ( CDC , MMWR July 28 , 2006/55 ( 29 ) ;793–796 ) . Meth self-administration by HIV+ individuals during the acquisition of sexually transmitted pathogens appears likely to interfere with immunological resistance and lead to AIDS progression .
Mouse femur hematopoietic stem cells from bone marrow were harvested from the hind legs of 8 to 12-week old male wild-type C57BL/6J ( The Jackson Laboratory , Bar Harbor , Maine ) or GFP-LC3 transgenic mice [39] , and plated at 2 × 106 cells/ml density in DMEM supplemented with 10% FBS , 1x non-essential amino acids ( Gibco , Carlsbad , California ) , 2 mM L-glutamine , 1 mM sodium pyruvate and 20 mM HEPES . For differentiation to dendritic cells or macrophages , 10 ng/ml of recombinant mouse GM-CSF ( Biosource , Carlsbad , California ) or 10 ng/ml of recombinant mouse M-CSF ( R&D Systems , Minneapolis , MN ) was added to media , respectively . Cells were fed every 2 days with fresh DMEM containing the appropriate macrophage colony stimulating factor . Cells were trypsinized after 8 days and , unless otherwise noted , plated at 4 × 105 cells/cm2 density to be used for experiments the following day . Resident mouse peritoneal macrophages were isolated from 8 to 12 week-old female wild-type C57BL/6J as described [55] and plated on 12 mm-diameter glass coverslips in 24-well tissue culture plates at 3 × 105 cells/well density in RPMI media supplemented with 10% FBS with streptomycin and penicillin . Non-adherent cells were removed by washing 2 h after plating . The remaining adherent cells were over 95% macrophages as assessed by esterase staining [55] . Cells were incubated overnight prior to experiments . Cells were stained with 10 mM acridine orange in phenol red-free media for 1 h . Images of stained cells were acquired using fluorescence microscopy as described above . Images of 10 fields with 5 to 10 cells for each image were taken using multiple stage positions in Multidimensional Acquisition mode under conditions of no photobleaching ( ND2 filter , 1000 msec exposure ) that enabled the acquisition of multiple images of the same cells . Stage position for each image was stored so images of the same fields and cells could be taken before and after treatment . Phenol red-free media containing 10 mM acridine orange in the presence or absence of Meth or Clq was added to the cells . At the end of the incubation time , images of treated cells were taken and used for morphometric analysis . Change in acridine orange intensity was measured as a change in average pixel intensity in the cytoplasmic area using MetaMorph Version 6 . 1r6 image analysis software ( Molecular Devices , Sunnyvale , CA ) . Background mean pixel intensity was measured in nuclear area and subtracted . Dendritic cells were stained with 5 μM LysoSensor Yellow/Blue ( Invitrogen , Carlsbad , CA ) for 5 min before Meth ( 10 , 50 , 100 μM ) or Clq ( 10μM ) was added for an additional 10 min incubation followed by washing with phosphate-buffered saline ( PBS ) ( pH 7 . 4 ) . Fluorescent images were taken of the same cells using Olympus IX81 microscope with Photometrics CoolSNAP HQ cooled camera , MetaMorph Version 6 . 1r6 imaging software ( Molecular Devices , Sunnyvale , CA ) , Olympus PlanApo 40x/1 . 4 Oil objective , equipped with fluorescent yellow customized Chroma ( D350/50 excitation , 400DCLP dichroic splitter and D535/40m emission ) ( Chroma technology Corp . , Rockingham , VT ) and blue Chroma 31000v2 ( D350/50 excitation , 400DCLP dichroic splitter and D460/50m emission ) filter sets . Average pixel intensity was measured in the cytoplasm of the cells excluding the nucleus using MetaMorph software , and the ratio of yellow to blue intensity was compared to a pH calibration curve to determine pH values . For the calibration curve , cells were stained with 5 μM LysoSensor Yellow/Blue for 20 min , washed with PBS and incubated in buffer of known pH ( 4 . 0 to 7 . 4 ) containing 10 μM monensin and 10 μM nigericin [56] before images were taken and processed as above . Dendritic cells were derived and incubated for 4 h with or without Clq ( 20 μM ) or Meth ( 100 μM ) . Cells were fixed in 2% paraformaldehyde and 2 . 5% glutaraldehyde in 0 . 1 M sodium cacodylate buffer , postfixed with 1% osmium tetroxide followed by 1% uranyl acetate , dehydrated through a graded series of ethanol and embedded in LX112 resin ( LADD Research Industries , Williston , VT ) . Ultrathin ( 80 nm ) sections were cut on a Reichert Ultracut UCT , stained with uranyl acetate followed by lead citrate and viewed on a JEOL 1200EX transmission electron microscope at 80 kV . For immunogold labeling , cells treated as described above were fixed in 2% paraformaldehyde and 4% polyvinylpyrolodone in phosphate buffer 0 . 2 M ( pH 7 . 4 ) at 4 °C . Fixed cells were processed for ultrathin cryosectioning as previously described [14] . Immunogold labeling was performed using LAMP-1 antibody ( clone 1D4B , BD Pharmingen , San Diego , CA ) followed by anti rat Ig-G coupled with 10 nm gold particles and biotinylated anti-MHC II ( clone M5/114 . 15 . 2 , BD Pharmingen , San Diego , CA ) followed by streptavidin gold ( 15 nm ) . Contrast was obtained with a mixture of 2% methylcellulose ( Sigma , St . Louis , MO ) and 0 . 4% uranyl acetate pH 4 ( EMS , Hatfield , PA ) . FITC-coupled BSA , ovalbumin or casein were fed to immature bone marrow-derived dendritic cells ( between 1 to 3 × 107 cells for each condition ) at a concentration of 100 μg/ml . Cells were untreated or treated with Meth or Clq . After overnight incubation , cells were washed twice in PBS and lysed in 150 mM NaCl , 50 mM Tris-HCl and 1% NP40 supplemented with protease inhibitor cocktail . Post-nuclear supernatants were normalized for protein content and 80 μg of total protein was run on SDS-PAGE gel . Membrane blots were probed with the anti FITC mAb or β-tubulin mAb ( Sigma , St . Louis , MO ) . The immature bone marrow-derived dendritic cell line JAWS ( ATCC ) was used for sub-cellular fractionation . One hundred million cells were used for each experimental condition . Cells ( 4 × 106 cells/ml ) were incubated overnight with 80 μg/ml of FITC-labeled BSA or casein in presence or absence of Clq ( 20 μM ) or Meth ( 50–100 μM ) . Cells were then lysed in 250 mM sucrose , 1 mM EDTA pH 7 . 4 . Early and late endosomes and lysosomes were prepared over consecutive Percoll gradients ( 27% and 10% ) from cells treated as reported above [57] . Each fraction was tested for β−hexosaminidase to locate lysosomes and late endosomes . The late endocytic marker Lamp-1 ( clone 1D4B , BD Pharmingen , San Diego , CA ) and the early endosomes/plasma membrane marker transferrin receptor ( TrfR ) ( clone M-A712 , BD Pharmingen , San Diego , CA ) were also used to assess the purity of the endosomal preparations . Pulled fractions 3–6 from the 27% Percoll gradient ( lysosomes ) , 2–5 from the 10% Percoll gradient ( late endosomes ) and 7–10 from the 10% Percoll gradient ( early endosomes ) were run on SDS-PAGE and blotted membranes analyzed for FITC-labeled antigens as reported above . GM-CSF differentiated bone marrow dendritic cells were cultured in methionine- and cysteine-free medium complete DMEM media containing 5% dialyzed serum for 1 h . Cells were then labeled with 0 . 2 mCi/ml [35S]-methionine ( Perkin Elmer , Waltham , MA ) for 30 minutes ( pulse ) . Cells were then washed three times and incubated in complete DMEM media supplemented with 10X cold methionine for 4 hours ( chase ) in the presence or absence of 50 μM Meth . Cells were subsequently lysed in 1% NP40 , 150 mM NaCl , 50 mM Tris containing a cocktail of protease inhibitors ( Complete Mini , Roche Diagnostics , Indianapolis , IN ) for 30 min on ice , spun at 14000 rpm for 30 minutes to remove cell nuclei and debris . The amount of incorporated radioactivity in each sample was determined by precipitating 10 μl of the post-nuclear supernatants with 10% trichloroacetic acid ( TCA ) . Equivalent amounts of radioactive lysates were pre-cleared with rat serum adsorbed to Prot G beads followed by Protein G beads alone , for 2 hours at 4°C . Immunoprecipitation was performed using 10 μg of anti CD74 ( clone In-1 , Pharmingen , San Diego , CA ) bound to Protein G beads . The beads were washed 3 times with lysis buffer and eluted with sample buffer . The elute was boiled and resolved by SDS-PAGE . The gel was subsequently dried and exposed in autoradiography . Bone marrow-derived dendritic cells from OT II transgenic mice ( Jackson Laboratory , Bar Harbor , MN ) were grown in 10 ng/ml of mouse GM-CSF for 10–12 days . Splenic T cells were purified using the pan-T cell isolation kit ( Miltenyi Biotec , Auburn , CA ) according to the manufacturer's suggestions . One hundred thousand dendritic cells were cultured with 4 × 105 T cells in the presence or absence of 0 , 3 , 10 , and 30 μM OVA protein for 3 days at 37°C . In some experiments Meth or Clq was added on the first day of culture until the end of the proliferative response . In other experiments , dendritic cells were pretreated with Meth or Clq for 4 h at 37°C , washed and fixed in 1% paraformaldehyde before adding OVA 323–339 peptide and the T cells . In all experiments , [3H]-thymidine ( 1 μCi/well ) was added during the last 18 h of incubation to assay T cell proliferation . Plates were harvested and the DNA [3H]-thymidine incorporation was monitored using a Wallac liquid scintillation counter ( Perkin Elmer , Waltham , MA ) . Bone marrow-derived dendritic cells and macrophages from GFP-LC3 animals were plated at 4 × 105 cells/cm2 density in glass bottom dishes and incubated overnight . Meth or Clq was added to the media , and the cells incubated for 2 h and 24 h before fluorescence microscopy using an Olympus IX81 microscope with Photometrics CoolSNAP HQ cooled camera , MetaMorph Version 6 . 1r6 imaging software ( Molecular Devices , Sunnyvale , CA ) , TC-324B Automatic Temperature Controller ( Warner Instrument Corporation , Hamden , CT ) , Olympus PlanApo 60x/1 . 4 Oil objective , and a Chroma FITC 41001 ( HQ480/40x excitation , Q505LP dichroic splitter and HQ535/50m emission ) ( Olympus , Center Valley , PA ) to determine GFP-LC3 puncta formation in the cells . At least 100 cell profiles per dish were assayed in triplicate for GFP-LC3 puncta . Mouse peritoneal macrophages at 3 . 5 × 105 cells/cm2 density on cover slips were incubated at 37°C with 20 μl of IgG opsinized sheep erythrocytes in 520 μl volume for 90 min . Uningested erythrocytes were lysed by sequential washing with PBS , water , and PBS . The phagocytic index was quantified by measuring the number of erythrocytes phagocytosed per 100 macrophages using bright field microscopy and 20x magnification , and inhibition is identified as percent of control . Data were collected from 4–7 independent experiments . The average control phagocytic index was 433 . To determine the phagocytic index for Candida albicans ( Ca ) and Cryptococcus neoformans ( Cn ) the macrophage like-cell line J774 . 16 ( cultured in DMEM with 10% heat-inactivated FCS , 10% NCTC-109 medium , and 1% nonessential amino acids ) was treated with 10 or 50 μM Methfor 2 h and then washed three times in media . As controls , J774 . 16 cells were incubated in medium alone or in the presence of 20 μM chloroquine . Ca strain SC5314 yeast cells were grown in brain heart infusion medium ( BHI ) at 37°C for 24 h then washed three times in PBS prior to application to macrophage . Cn yeast strain H99 was grown for 24 h in BHI at 37°C , incubated with capsule-specific monoclonal antibody ( mAb ) 18B7 [58] as an opsonin at 10 μg/ml for 2 h and then washed three times in PBS prior to incubation with macrophage . Ca and Cn cells were added to the macrophage monolayer at an effector to target ratio 1:1 , and the suspension was incubated at 37°C for 30 min with Ca or 1 h with Cn . After incubation , remaining extracellular yeast cells were removed with three washes of PBS . The phagocytic index was determined by microscopic examination . For each experiment , five fields in each well were counted , and at least 100 macrophages were analyzed in each well . Wells were performed in triplicate for each condition examined . Colony counts were made to determine the number of viable Ca and Cn yeast cells after phagocytosis . For the colony forming unit ( CFU ) determination , J774 . 16 macrophages were treated with or without Methor chloroquine and infected with Ca or Cn as described . The cultures were washed after 2 h to remove extracellular yeast and then incubated for an additional 22 h . After the 24 h total incubation , macrophage cells were lysed by forcibly pulling the culture through a 27-gauge needle 5 times . The lysates were serially diluted , and plated on Sabouraud dextrose agar at 37°C . CFU determinations were made after 72 h . Controls also consisted of yeast grown without macrophage , but in the presence of chloroquine or methamphetamine . All tests were preformed in triplicate .
|
There is a new population of HIV+ men who are developing AIDS over months instead of years as typical . It has recently become popular among gay and bisexual men to consume very high levels of Meth . Unsafe sex together with Meth abuse has been suspected to lead to rapid disease progression . While studies show exacerbated AIDS symptoms and disease progression in HIV+ Meth abusers , the molecular mechanism is yet unknown . It was postulated , yet unproven , that the rapid disease progression might be due to a mutant “superstrain” of HIV that was extremely virulent . It was also assumed that the effects of the drug on behavior may lead to unsafe sex , although this would not explain the more rapid time course of the disease . We now demonstrate the first direct evidence that Meth is an immunosuppressive agent , and that the molecular mechanism of this immunosuppression is due to the collapse of acidic organelle pH in cells of the immune system , inhibiting the functions of antigen presentation , as well as phagocytosis . These effects compromise the immune response to opportunistic infections and HIV . These findings could have a major impact on public health , as there are over 35 million Meth abusers worldwide
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"viruses",
"infectious",
"diseases",
"public",
"health",
"and",
"epidemiology",
"virology",
"yeast",
"and",
"fungi",
"in",
"vitro",
"immunology",
"mus",
"(mouse)",
"homo",
"(human)"
] |
2008
|
Methamphetamine Inhibits Antigen Processing, Presentation, and Phagocytosis
|
Trachoma is the leading infectious cause of blindness in the world , and for endemic communities , mass treatment with azithromycin reduces the pool of infection . High coverage is essential , especially in children as they are the infectious reservoir . However , infection remains post-mass treatment . We sought to determine risk factors for infection in children post-mass treatment . All children under 9 years in 4 villages in Tanzania were followed from baseline pre-mass treatment to six months post treatment . 1 , 991 children under nine years were enrolled in the longitudinal study and data on individual and household characteristics was collected at baseline . Clinical trachoma was determined by an ocular exam and infection detected by PCR of an eyelid swab . Azithromycin was offered and infection was reassessed at 6 months . A multilevel logistic regression model was used , accounting for household clustering of children for analysis . Baseline infection was 23 . 7% and at 6 months was 10 . 4% , despite 95% coverage . Infection at baseline was positively associated with infection at 6 months ( OR = 3 . 31 , 95%CI 2 . 40–4 . 56 ) and treatment had a protective effect ( OR = 0 . 45 , 95%CI 0 . 25–0 . 80 ) . The age group 2–4 years had an increased risk of infection at 6 months . The household characteristics predictive of infection at 6 months were increasing number of children infected in the household at baseline and increasing number of untreated children in the household . While one round of mass treatment with high coverage did decrease infection by over 50% , it appears that it is not sufficient to eliminate infection . Findings that young children ( ages 2–4 years ) and households with increasing numbers of infected and untreated children have a positive association with infection at 6 months suggest that such households could be targeted for more intensive follow up .
Trachoma is a blinding chronic conjunctivitis caused by repeated episodes of ocular Chlamydia trachomatis infection . It continues to be the leading infectious cause of blindness in the world , largely in the most resource poor countries [1] . The World Health Organization ( WHO ) has made a commitment to eliminate blinding trachoma by the year 2020 ( Global Elimination of Blinding Trachoma ) [2] . The strategy designed for this initiative can be summarized with the acronym , SAFE , Surgery for trichiasis , Antibiotic to reduce the community pool of C . trachomatis infection , Facial cleanliness , and Environmental change to reduce transmission . In 1999 , a study by Schachter et al . demonstrated that azithromycin could be effective in a mass treatment campaign to reduce the infection in endemic villages . The manufacturer has since provided the drug free of charge to trachoma endemic countries that apply for donation [3] . Eliminating infection with a single dose of antibiotic is very appealing , because of improved compliance . In theory , infection could potentially be eliminated from a population in a single effort . This has been reported in some villages that have low trachoma prevalence [4] . However , in hyper-endemic villages , infection is still found in treated children following mass treatment . The biggest reservoir of active disease and infection is seen in children [5] . Two months after mass treatment of a hyper-endemic village in Tanzania , more than 30% of children with high bacterial loads at baseline continued to be infected [6] . In hyper-endemic villages , where active trachoma may be as high as 50% in children in cross sectional surveys , it has been suggested that up to 6 rounds of mass treatment may be necessary to eliminate infection [7] . In order to further investigate infection following mass treatment , we sought to determine the household and child factors that predict infection post treatment . Identifying such factors could help determine how to implement mass treatment .
At the beginning of 2009 a household census was performed in 4 villages in Kongwa district , Tanzania . Children less than 9 years of age were identified and their parents were invited to have their children participate in a longitudinal study of trachoma and infection over a three year period . Written informed consent was obtained from all parents or guardians of children in this study . All procedures for the study were approved by the Johns Hopkins University Institutional Review Board ( JHU IRB ) and the National Institute of Medical Research ( NIMR ) in Tanzania . Data on the potential risk factors at the individual and household levels were collected at baseline . The individual characteristics included age , gender , active trachoma and presence of Chlamydia trachomatis at baseline , and facial cleanliness in children 5 years of age and younger . The household characteristics were: distance to the nearest source of water ( measured as self report of the time to walk one way ) , presence of a latrine as observed by the interviewer at the household , and average years of education of the head of the household based on self report . Facial cleanliness was assessed on children ages five years and younger by direct observation , using three elements: presence of ocular or nasal crusting , and observation of one or more flies on the face in a three second window , as measured at the time of the exam [8] . Active trachoma and presence of C . trachomatis were also determined at 6 months . Trained trachoma graders performed an ocular exam using 2 . 5x loupes and a torch to determine active trachoma . The trachoma graders were standardized during a two week training exercise prior to the start of the study . Inter-observer agreement , and agreement with a senior grader had to be above kappa = 0 . 6 for TF and TI . Graders are monitored every six months by comparison of grades against the senior grader using a series of trachoma images [9] . Active trachoma was graded using the World Health Organization simple grading scheme [10] . An eyelid-conjunctival swab was obtained to assess presence of C . trachomatis DNA . These swabs were taken using strict protocols to avoid contamination . Both the trachoma grader and eyelid flipper used gloves and changed them between each child examined . In addition , “air control” swabs were taken on a random sample of 5% of children . The swabs were waved in the air above the child and not touched to the eyelid , marked as per any other sample and sent for processing . If positive , further investigation were carried out to determine the source of contamination . The samples were kept at −20°C until they were sent to the International Chlamydia Laboratory at Johns Hopkins University for processing . The samples were analyzed using real-time Polymerase Chain Reaction ( Amplicor: Roche Molecular Systems , Indianapolis , IN ) according to manufacturer's directions . Optical density was used to identify positive results and defined as >0 . 7 . Optical densities between 0 . 2 and 0 . 7 were considered equivocal and were retested . Dummy swabs were performed to test for contamination in the field and laboratory . Results of dummy swabs showed evidence of lab contamination that affected 34 samples at baseline . Contaminated samples were excluded; all further tests for contamination were negative at baseline and follow up . A single dose of azithromycin at a dose of 20 mg/kg up to 1 g was offered to all residents 6 months of age and older in each of the 4 villages by community treatment assistants , who observed and recorded treatment . Infants under 6 months of age were treated using topical tetracycline ( 6 weeks of 1% tetracycline topical ointment twice daily ) . Treatment was immediately after the time of the baseline assessment . Contingency table analysis was used to identify factors associated with infection at six months , corresponding p-values were obtained from a logistic model with infection at 6 months as the outcome , accounting for clustering at the household level . In the multivariate stage , logistic regression was used to model presence of infection at six months as a function of the factors found to be associated at a 0 . 15 level in the univariate analysis , a backward elimination strategy was used to find a parsimonious model . Among variables whose p-value was greater than 0 . 05 , we deleted the one that had the highest p-value . We proceeded iteratively until all variables in the model had associated p-values less than 0 . 05 . The generalized estimating equation ( GEE ) approach was used to correct the standard errors to account for the correlation among children members of the same house ( procedure GENMOD in SAS , binomial distribution , logit link function , exchangeable correlation structure ) .
There were a total of 2201 children less than 9 years of age in the four villages at the time of census ( Figure 1 ) . A total of 49 ( 2 . 2% ) of the children did not participate in the study at baseline . An additional 161 ( 7 . 5% ) did not return at 6 months for the ocular exam . Thus , our study population consisted of 1991 children with collected data at baseline and 6 months . A comparison of baseline characteristics of the study population and the 161 children who missed the 6-month ocular exam revealed no significant differences for all characteristics except treatment ( Table 1 ) . The study population had antibiotic coverage of 95 . 0% while the children with missing 6 month data had 87 . 5% coverage , reflecting the fact that some of them left the study even prior to mass treatment . Of note , the baseline prevalence of infection in study children was 24% compared to 22% in those who did not participate at 6 months , a non-significant difference . Overall the prevalence of Chlamydia infection and active trachoma respectively were 23 . 7% and 27 . 8% in the study population at baseline . At 6 months , 203 children were infected ( 10 . 4% ) , down from 23 . 7% at baseline . Characteristics associated with infection at 6 months are shown in table 2 . There was no difference in infection at 6 months by gender or those with clean or unclean faces at baseline or the available household characteristics ( education , distance to water , presence of a latrine ) . Prevalence of infection at 6 months were higher in those 2–4 years of age , with baseline infection , and in the untreated group ( Table 2 ) . Children were more likely to be infected at 6 months if other children in the household were infected at baseline ( 7% , 15% , 28% for 0 , 1 , 2+ infected children , test for trend p = 0 . 001 ) . In addition the risk of infection increased with the number of untreated children in the same household ( test for trend p = 0 . 02 ) . The effect of treatment seemed to be different depending on the infection status at baseline ( Table 3 ) . Among infected children , children who were treated at baseline had 6–month prevalence of infection of 22 . 0% , compared to 5 . 9% in uninfected children who were treated . Children who were untreated and infected at baseline had a 3 . 1-fold higher chance of being infected at 6 months compared to untreated and uninfected children , 50 . 0% vs . 16 . 0% . However , after adjusting for other factors the interaction between infection at baseline and the effect of treatment was not statistically significant ( p = 0 . 59 ) . Of the sub group of 1493 children without infection at baseline , 6 . 4% were infected at 6 months . In this sub group , treatment was the only explanatory factor comparing those who stayed free of infection with those 96 who developed infection at 6 months . Among those free of infection , 95 . 1% had been treated compared to 86 . 4% of those who developed infection . We found no difference in gender , mean age , percentage of children under 5 years of age with unclean faces , or whether at baseline there was another infected child in the house ( Table 4 ) . Our final model predicted infection at 6 months , adjusted for multiple variables and household clustering ( Table 5 ) . Factors that independently increased the odds of having infection at 6 months included being in the age group 2–4 years old ( OR = 1 . 41 ) , having infection at baseline ( OR = 3 . 31 ) , and living in a household with other infected children ( OR = 1 . 39 per additional infected child ) . Treatment significantly reduced the odds of infection at 6 months ( OR = 0 . 45 ) , and other untreated children in the household were associated with increased risk ( OR = 1 . 58 per additional untreated child ) .
In these four endemic communities we found an overall prevalence of Chlamydia infection at baseline of 23 . 7% . Mass antibiotic treatment coverage reached 95% of the children in the population , which is close to a goal of full coverage of all children . Despite such high coverage , the infection at 6 months was 10 . 4% . One factor that predicted infection at 6 months was infection at baseline . This is consistent with a similar finding in a study from another village with similar endemicity in Tanzania [6] [11] . At six months , these infections may be due to lack of treatment , although with high coverage these are few case; the infections may also be due to high bacterial loads in those treated , which made them less likely to be cured after a single dose . West et al found previously that those children with the highest bacterial load pre-treatment were most likely to still have infection 2 months after treatment [6] . Potentially , the dose of azithromycin could be increased for these children , but a randomized trial of 30 mg/kg versus 20 mg/kg dose in children with severe trachoma did not demonstrate any difference in outcome [12] . Recent data from Ethiopia suggests that more frequent treatment , perhaps every six months , would be more effective , although this needs to be confirmed [7] . The feasibility for programs of implementing treatment even every six months to families of young children would need careful study of the cost-effectiveness of such a strategy , as it would likely double the yearly costs of mass treatment as well as demand for azithromycin . Another factor associated with infection at 6 months , independent of self-infection at baseline , was the number of children infected in the household at baseline , and number of untreated children in the household . This finding is not unexpected , as some of the infected siblings may well be the source of reinfection [13] . To the extent that not all children are treated in the household leaves a risk for other members , as intra-household transmission appears to be relatively fast [11] , [13] . Another study has shown that missing treatment does not occur at random , but clusters in households , suggesting that finding households that are non participants in mass treatment , and determining strategies to improve coverage in that group may be needed 14 . Finally , the 2–4 year age group was at increased risk of infection post-mass treatment , independently of treatment status or infection status . We did not collect data on bacterial load and it is possible that this particular age group has the highest load of infection , thus most likely to remain infected post treatment . We expected the highest prevalence of infection at 6 months would be the infants who had to be treated with topical tetracycline , and for whom compliance with a six-week course could not be assured . However , infection post treatment was greatest in the 2–4 year olds , and there was no association of 6 month infection with having an infant or infected infant in the household . One limitation to our study is that we were not able to measure chlamydial load . Thus , we cannot be certain if this is the reason that the children ages 2–4 had higher odds of infection at 6 months . Another potential limitation is loss to follow-up of children in the study . There are two points at which we lost subjects from our original eligible population . The first was at the start of the study , where 49 ( 2 . 1% ) of children did not participate . This is a small number compared to the total size of the population and is unlikely to bias our results . The second point was at the 6-month follow-up visit , where 161 ( 7 . 3% ) children did not return for a follow up exam . There were no differences in baseline characteristics between the 1991 children in the study sample and the 161 children missing 6-month data , except for the percentage receiving antibiotics . Thus , they may have been more likely to have infection at 6 months , but this would not be expected to bias our findings as they were no more likely to have infection at baseline . In addition , 34 of the 1991 samples ( 1 . 7% ) at baseline could not be used . It is unlikely that this loss contributed to bias as the number lost was so small , and was the result of error in the laboratory , which is masked to any clinical or treatment data . In summary , although we found that village coverage with azithromycin decreased infection at 6 months in children ages 9 and under , it did not eliminate the pool of infection . Clearly , one round of mass treatment in these communities with infection above 20% is not enough . Based on these findings , which support our previous findings in a single community , mass treatment of communities should be continued as part of the SAFE strategy to eliminate blinding trachoma . The World Health Organization recommends at least three annual mass treatment interventions as part of SAFE , and our study will provide data in the future on potential re-emergence following a second and third round of mass treatment in line with those recommendations 15 .
|
Trachoma control programs aim for high coverage of endemic communities with oral azithromycin to reduce the pool of infection with Chlamydia trachomatis . However , even with high coverage , infection is seen following treatment . In four communities in Tanzania , we followed every child aged under ten years from baseline through treatment to six months post-treatment . We determined who had infection at baseline and who still had or developed infection six months later . Coverage was over 95% in children in these communities , and infection in these children decreased by over 50% at six months . The study found that , at baseline , uninfected children who were treated had prevalence of infection at 6 months of 6% , but infected children who were treated had prevalence of infection of 22% at 6 months . Other risk factors for infection at 6 months included living in a household with other infected children , and living in a household with untreated children . Our data suggest that households with untreated children might be targeted for more intensive follow up to increase coverage and reduce subsequent infection in the community .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases/neglected",
"tropical",
"diseases"
] |
2011
|
Risk Factors for Ocular Infection with Chlamydia trachomatis in Children 6 Months following Mass Treatment in Tanzania
|
Nuclear receptor ligand binding domains ( LBDs ) convert ligand binding events into changes in gene expression by recruiting transcriptional coregulators to a conserved activation function-2 ( AF-2 ) surface . While most nuclear receptor LBDs form homo- or heterodimers , the human nuclear receptor pregnane X receptor ( PXR ) forms a unique and essential homodimer and is proposed to assemble into a functional heterotetramer with the retinoid X receptor ( RXR ) . How the homodimer interface , which is located 30 Å from the AF-2 , would affect function at this critical surface has remained unclear . By using 20- to 30-ns molecular dynamics simulations on PXR in various oligomerization states , we observed a remarkably high degree of correlated motion in the PXR–RXR heterotetramer , most notably in the four helices that create the AF-2 domain . The function of such correlation may be to create “active-capable” receptor complexes that are ready to bind to transcriptional coactivators . Indeed , we found in additional simulations that active-capable receptor complexes involving other orphan or steroid nuclear receptors also exhibit highly correlated AF-2 domain motions . We further propose a mechanism for the transmission of long-range motions through the nuclear receptor LBD to the AF-2 surface . Taken together , our findings indicate that long-range motions within the LBD scaffold are critical to nuclear receptor function by promoting a mobile AF-2 state ready to bind coactivators .
The nuclear receptor ( NR ) superfamily of ligand-regulated transcription factors controls the expression of genes essential to metabolism , development and systemic homeostasis [1]–[3] . NRs are modular proteins typically composed of a conserved N- terminal Zn-module DNA binding domain ( DBD ) that targets specific response elements , a variable hinge region , and a C-terminal ligand binding domain ( LBD ) capable in most cases of responding to specific small molecule ligands [4] . NR LBDs contain a shallow activation function 2 ( AF-2 ) surface formed by helices α3 , α3′ , α4 and αAF that is essential for ligand-dependent interactions with transcriptional coregulators . The AF-2 surface complexes with LxxLL-containing transcriptional coactivators in the presence of agonist ligands , and with distinct leucine-rich corepressor motifs in the presence of antagonists or in the absence of ligand [4] , [5] . The pregnane X receptor ( PXR ) controls the expression of a wide range of gene products involved in xenobiotic metabolism and endobiotic homeostasis [6]–[8] , and is unusual in the NR superfamily in several respects . First , PXR responds promiscuously to a wide range of chemically-distinct ligands from small lipophilic phenobarbital ( 232 Da ) to the large macrolide antibiotic rifampicin ( 823 Da ) ; in contrast , most NRs are highly specific for their cognate ligands [9]–[11] . Second , the PXRs of known sequence contain a 50–60 residue insert that , as observed in human [12]–[14] , creates a unique β-turn-β motif and novel PXR homodimer interface . All NR LBDs fold into a three-layer α-helical sandwich in which α10 forms standard homodimerization interactions ( e . g . , for steroid receptors like the estrogen receptor-α , ERα ) or heterodimerization interactions ( e . g . , with RXR for orphan receptors like PXR ) [2] , [15] , [16] . The PXR LBD , in contrast , contains a second oligomerization interface at the novel β-turn-β motif in which intercalating tryptophan and tyrosine residues ( Trp-223/Tyr-225 ) lock across the dimer to form an aromatic zipper [4] , [5] , [12] ( Figure 1A ) . It has been shown that this dimer interface is essential to PXR function , and that the specific disruption of homodimerization eliminates the ability of the receptor to interact with transcriptional coactivators like steroid receptor coactivator 1 ( SRC-1 ) , but does not impact PXR's subcellular localization or its association with DNA , RXR , or activating ligands [12] . This work led to the proposal of a PXR-RXR heterotetramer as the functional unit [12] ( Figures 1A , 1B ) . The unique PXR homodimer interface , however , is located more than 30 Å from the coactivator binding site at the receptor's AF-2 surface ( Figure 1A ) . Thus , we hypothesize that long-range motions within the PXR LBD are essential for communicating the stabilizing effect of PXR homodimerization to the AF-2 domain . To test this hypothesis , we performed all-atom molecular dynamics ( MD ) simulations on both the PXR LBD , as well as two other nuclear receptor LBDs , in various states ( Table 1 ) . The former orphan peroxisome proliferator-activated receptor-γ ( PPARγ ) is functional as a heterodimer with RXR , while the steroid estrogen receptor-α ( ERα ) is active as an analogous homodimer ( Figure 1B ) . We examined LBDs in inactive states ( e . g . , monomers or mutants ) , as well as those in the proper functional states ( e . g . , homo- or heterodimers , or as a heterotetramer for RXR ) we have termed “active-capable . ” Our results support the conclusion that the NR LBD provides a scaffold for long-range motions that prepare the AF-2 surface for binding to transcriptional coactivators .
Six all-atom molecular dynamics ( MD ) runs were performed for 20–30 ns on three nuclear receptor LBDs ( Table 1 ) . PXR was examined both as a heterodimer with RXR and as a heterotetramer with RXR ( 30 ns simulations ) . Wild-type PPARγ was examined as a heterodimer with RXR , and the inactive PPARγ P467L mutant was also examined as a heterodimer with RXR ( 20 ns simulations ) . Finally , ERα was examined both in its inactive monomeric state ( 20 ns ) , and as an ERα homodimer ( 25 ns ) . All six trajectories were judged as stable by two criteria . First , the total energy of each system , calculated as the sum of kinetic and potential energy at each time point , was found to be essentially constant after the first 2–3 ns ( Figure 2 , Figure S1 ) . These results indicate that after a short period of equilibration , each simulation was sampling an energetically stable conformational ensemble . Second , all trajectories were analyzed in terms of moving average all-atom root mean square deviations ( RMSDs ) from starting crystal structures over the simulation time course ( Figures S2 , S3 ) . The PXR-RXR trajectories exhibited RMSD values of 0 . 7–5 . 0 Å ( Figure S2 ) , while the PPARγ- and ERα-containing trajectories exhibited values of 1 . 7–3 . 5 Å ( Figure S3 ) . Such deviations were considered low for systems of this size ( e . g . , 1044 residues for the PXR-RXR heterotetramer ) . The RMSD results indicate that all simulations were stable for at least the last 10 ns of each trajectory ( Figures S2 , S3 ) . Thus , the final 10 ns section of each simulation was used for subsequent analysis . The PXR-RXR heterotetramer is expected to have distinct functional dynamics relative to the heterodimer because the heterotetramer contains the unique PXR homodimer interface shown to be essential for receptor activity [12] ( Figure 1 ) . Thus , we examined the PXR LBDs in both the PXR-RXR heterodimer and PXR-RXR heterotetramer simulations over the last 10 ns of each trajectory using essential dynamics analysis . Essential dynamics discriminates between concerted motions of residue clusters within a protein and uncorrelated residue fluctuations [17] . We computed normalized covariance matrices [18] to classify the relationships between all possible residue pairs in the protein ( Figure 3A ) . In this analysis , correlation ( two residues moving in the same direction ) is indicated by residue-residue correlation coefficients approaching +1 , while correlation coefficients approaching −1 indicate anticorrelation ( residues moving in opposite directions ) . Correlation coefficients near zero , in contrast , are associated with residue pairs that lack a dynamic relationship . The PXR LBDs in the PXR-RXR heterotetramer exhibit significantly more residue-residue correlation relative to the PXR LBD in the PXR-RXR heterodimer ( Figure 3A ) . Indeed , the distribution of correlation coefficients for the PXR LBD in the PXR-RXR heterodimer has one peak centered close to zero , indicating the majority of residue-pairs are not correlated ( data not shown ) . In contrast , the correlation coefficient distribution for the PXR LBDs in the PXR-RXR heterotetramer has two distinct peaks , one positive and one negative , indicating both residue-residue correlation and anticorrelation ( data not shown ) . Clusters of correlated PXR residues from the PXR-RXR heterodimer and heterotetramer that exhibited concerted motion were then examined for the strength of their residue-residue correlation coefficients and the biological significance of those dynamics . Clustering the PXR LBDs from the PXR-RXR heterotetramer at a correlation coefficient less than 0 . 6 produced a single cluster containing the complete PXR-RXR homodimer , while clustering at a correlation coefficient above 0 . 8 resulted in clusters comprised of only 2–5 residues . Neither coefficient cutoff alone could interrogate the biological relevance of the concerted motions; thus , we classified clusters using three correlation coefficient cutoffs ( Figure 3B , C ) . Such cutoffs discriminate between weak ( 0 . 6 or less ) , medium ( between 0 . 6 and 0 . 7 ) , and strong ( between 0 . 7 and 0 . 8 ) correlations between PXR residues . The same cutoffs set in the heterotetramer were used , for consistency , to cluster the relatively weak correlated motion observed in the PXR-RXR heterodimer . Indeed , the PXR LBD from the heterodimer exhibited only five small correlated clusters , with smaller regions of these weakly correlated clusters remaining at a medium strength correlation coefficient , and only one group of a few residues identifiable at a strong correlation coefficient ( Figure 3B ) . In distinct contrast , however , the PXR LBDs in the PXR-RXR heterotetramer form a strongly correlated unit ( Figure 3C ) . The β-sheet region involved in the PXR-PXR homodimerization interface ( β1 , β1′ , β3 , β4 ) , together with α-helices 1 , 3 , 3′ , 4 and 9 , exhibit the strongest degree of correlation; the residues of these β-sheets and α-helices are all clustered together with a correlation coefficient of 0 . 8 . The neighboring helices , including αAF , also exhibit highly correlated motion with correlation coefficients >0 . 6 ( Figure 3C ) . The strength of residue-residue correlations throughout this region suggest that α3 forms a critical conduit through which the stabilizing effects of the homodimer interface involving β1 , β1′ , β3 , and β4 are communicated to helices 3 , 3′ , 4 and AF of the AF-2 surface . In the PXR-RXR heterodimer , however , the same β-sheet region is anticorrelated with the AF-2 domain ( Figure 3B ) . We next examined the motions in the four helices that create the AF-2 coactivator binding surface on PXR: α3 , α3′ , α4 , and αAF . The concerted motion of this surface was compared between the PXR LBDs in the PXR-RXR heterodimer and heterotetramer trajectories , and was examined using both quasiharmonic analysis ( QHA ) and normal mode analysis ( NMA ) . Both methods have benefits and limitations . For quasiharmonic analysis , its benefits are all-atom resolution and the use of explicit solvent , but it is limited by the time constraints of all-atom MD . Normal mode analysis has the benefit of observing motions on a longer timescale than available with QHA , but is limited to analyses based upon the coarse grained model solely of the macromolecule . Our results agree with others that it takes more NMA modes than QHA modes to describe the same motions [19] . Thus , we employed the first two modes from QHA and first 14 nontrivial modes from NMA ( see Methods ) . Eigenvectors from these analyses are associated with the magnitude and direction of motion , and these eigenvectors can be used to create visuals of the NR's motion . After examining the vectors describing the primary modes of motion derived from QHA for each α-carbon position in the PXR LBDs of the PXR-RXR simulations , a single average vector was calculated to describe the motion of seven of the eleven α-helices in the LBD . The remaining four helices , α3 , α4 , αAF and α10 , displayed distinct motions at their termini; thus , for these helices , two average vectors were employed . The results of this analysis show that the PXR LBD helices from the PXR-RXR heterotetramer move as a single unit , and in one direction ( Figure 4A ) . This correlation is especially evident in the AF-2 surface , as α3 , α3′ , α4 and αAF all move together in the same direction ( Figure 4A inset ) . In contrast , the PXR LBD from the PXR-RXR heterodimer exhibited relatively small , disjointed motions ( Figure 4B ) . This lack of helix-helix correlation includes the AF-2 surface helices α3 , α3′ , α4 and αAF ( Figure 4B inset ) . AF-2 mobility identified by QHA was also assessed by examining the angles between the directions of motion as defined by the eigenvectors for α-carbons of residues important to coactivator binding ( Table 2 , Methods ) . As such , if two residues in the AF-2 surface are moving together , the angle between them is small ( see Methods , Equation 1 ) . The average angle from the sum of motion vectors ( modes ) 1 and 2 between AF-2 domain residues in the PXR-RXR heterodimer simulation was 71 . 6° . In contrast , the average angle for the same residue pairs in the PXR-RXR heterotetramer simulation was 31 . 5° ( Table 2 ) . Taken together , these QHA results support the conclusion that the intramolecular β-sheet formed by the PXR homodimer interface produces highly correlated AF-2 surface motions in the PXR-RXR heterotetramer complex . In a second analysis , modes of motion of the AF-2 surface of the PXR LBD were examined from both the heterodimer and heterotetramer trajectories using NMA . Similar to the QHA study above , angles between the directions of motion as defined by the eigenvectors for α-carbons of residues important to coactivator binding were calculated ( Table 2 , Methods ) . The average angle observed in the AF-2 surface in the PXR-RXR heterotetramer was 13 . 5° using NMA , even smaller than the average angle found using QHA ( Table 2 ) . In contrast , the average angle for the same PXR AF-2 residues in the PXR-RXR heterodimer was 70 . 1° , nearly identical to the value found using QHA ( Table 2 ) . Thus , these data support the conclusions of the QHA study , and indicate that a high degree of helix-helix correlation is present in the AF-2 surface of the PXR-RXR heterotetramer relative to the heterodimer . Similarities between the QHA and NMA results strengthen this collective conclusion , particularly because QHA is based on shorter dynamic movements of all atoms , while NMA examines harmonic oscillations that occur on longer time scales . Plots of the angles between the vectors of motion of all possible PXR LBD residue pairs from both the heterodimer and heterotetramer simulations for the QHA and NMA studies are shown in Figures 5A and 5B , respectively . Areas in green represent angle values close to zero ( vectors moving in the same direction , or correlated ) , while areas in yellow indicate vectors with angles close to 180° ( vectors moving in the opposite direction , or anticorrelated ) . In both plots , a high degree of correlated motion is observed for the PXR LBD in the PXR-RXR heterotetramer , while significantly less correlation is observed for the LBD in the heterodimer ( Figures 5A , B ) . The similarity between Figures 5A and B , from selected modes of QHA- and NMA-identified motion , and Figure 3A , from all modes of motion , indicates that enough modes were chosen in both QHA and NMA to represent the motion of each LBD ( Methods ) . In addition , both the QHA and NMA plots for the heterotetramer indicate similar correlated structural elements . For example , the PXR β-sheet moves in a more correlated manner with respect to αAF in the heterotetramer relative to the heterodimer ( Figures 5A , B ) . In summary , long-range motions impacted by the oligomeric state of PXR play a central role in the function of this nuclear xenobiotic receptor . We next examined whether the unliganded LBDs of other members of the NR superfamily would also exhibit correlated AF-2 surface motions . As stated above , 20–25 ns MD simulations were performed on two inactive NR states , the ERα monomer and the PPARγ P467L-RXR heterodimer complex , and on two “active-capable” states , the ERα homodimer and the wild-type PPARγ-RXR heterodimer . A P467L mutation has been shown to inactivate PPARγ [20] . Only moderate levels of residue-residue correlation and anticorrelation were observed for both states of ERα and PPARγ ( Figures S4A , B ) . Examination of correlation coefficient distributions in these simulations reveals that all remain close to zero , indicating relatively non-correlated motion ( data not shown ) . In spite of their relatively limited overall correlation , however , the active-capable forms of ERα and PPARγ-RXR exhibited correlated AF-2 domain motions . Similar to the analysis of the PXR trajectories , both QHA and NMA were employed to examine these ERα and PPARγ simulations . Results from QHA studies reveal that the active-capable forms of ERα , and PPARγ exhibit more correlated AF-2 motions than their inactive counterparts ( Figure 6 ) . Angles between the vectors describing AF-2 surface helix motions in PPARγ and ERα states using both QHA and NMA further support the overall conclusion that active-capable states exhibit correlated AF-2 surfaces ( Table 3 , 4 ) . For example , the average angles for ERα homodimer and wild type PPARγ-RXR determined using NMA are 41 . 0° and 48 . 8° , respectively , while those for the inactive ERα monomer and the PPARγ P467L mutant are 63 . 1° and 58 . 3° . Again , the AF-2 correlation in motion observed using the shorter time scales of all-atom molecular dynamics simulations and QHA are also seen in the longer harmonic oscillations of NMA . In summary , correlated motion appears to be a consistent feature in the AF-2 domains of active-capable nuclear receptor LBDs .
The differences in human PXR LBD motion between two oligomeric states of the receptor ( as a heterodimer and a heterotetramer with RXR ) were examined using molecular dynamics trajectories , essential dynamics , quasiharmonic , and normal mode analyses . It was hypothesized that the PXR heterotetramer , in which PXR LBD monomers form a unique homodimer shown to be critical for transcriptional regulation [12] , would exhibit functionally-relevant motion . Indeed , we find that this “active-capable” form of PXR exhibits not only significantly more overall motion and more correlated motion relative to the heterodimer , but also highly correlated motion in the AF-2 surface responsible for functionally-essential contacts with transcriptional coactivators ( Figures 4 , 5 ) . These data suggest that a high degree of motion promotes the proper function of this nuclear receptor , provided that the motion is correlated to preserve the state of the receptor ready to bind to leucine-rich coactivator motifs . In addition , these results indicate that long-range motions are critical to the function of the xenobiotic receptor PXR . The homodimer interface unique to the PXR LBD is located approximately 30–35 Å from the AF-2 surface ( Figure 1 ) . Essential dynamics have revealed that the β-sheet and six α-helices in PXR ( 1 , 3 , 3′ , 4 , 9 , AF ) , including those that comprise the AF-2 surface , move as a single unit in the heterotetramer trajectory ( Figure 3 ) . This suggests a structural mechanism by which PXR homodimerization creates a ten-stranded intermolecular β-sheet ( Figure 1 ) that positively impacts AF-2 domain motion . The N-terminal portion of α3 appears to serve as a critical bridge between the PXR β-sheet and the AF-2 helices , such that correlated β1-β4 motion is “communicated” to α3-α4 and αAF ( Figure 4 ) . This relationship explains how the obligate PXR monomer mutant Trp-223-Ala/Tyr-225-Ala , in which the interlocking aromatic residues at the homodimer interface are eliminated , is still able to bind to ligand , DNA and RXR , but not to transcriptional coactivators at the AF-2 surface [12] . This hypothesized path of “communication through motion” mediated by α3 and involving several β-strands , as well as α1 and α9 , correlates well with existing PXR structure-function data . First , Met-243 , located in the N-terminal portion of α3 , is contacted by ligands in all reported PXR LBD crystal structures [4] , [21] , [22] . Thus , they appear critical for the ligand-enhanced transcriptional activity exhibited by PXR . Second , single mutations in either α3 or α3′ , such as Thr-248-Glu , Lys-277-Gln and Pro-268-His , result in a loss of PXR activity [23] , [24] . In addition , although the α3 double-mutant Lys-277-Gln/Thr-248-Glu restores transcriptional activation , it abolishes the antagonism of ketoconazole , hypothesized to function by binding the AF-2 surface [24] , [25] . Third , the α1 and α9 mutants Asp-163-Gly and Ala-370-Thr , respectively , represent a class of PXR variants that are distantly located from the AF-2 domain but result in reduced transcriptional activity [26] . Taken together , these data support the conclusion that the wild-type PXR LBD is “tuned” in its heterotetrameric complex with the RXR LBD to produce correlated motions that promote the binding of transcriptional coactivators . Extension of this analysis into other nuclear receptors revealed correlated AF-2 surface motions in “active-capable” forms of ERα and PPARγ ( Figure 6 and Figure S4 ) . Thus , long-range motions may play critical roles in the LBD activation potential of several members of the nuclear receptor superfamily . Our results expand on previous MD investigations of NR LBDs . For example , dynamics studies on ERα [27] showed that the addition of coactivator peptide and ligand to apo ERα lead to increased αAF helix motion in unspecified directions . Similarly , studies on androgen insensitivity syndrome associated androgen receptor Pro-892-Ala and Pro-892-Leu mutations revealed via biochemical assays and MD simulations an increased flexibility and distortion of the αAF helix [28] . We present evidence that the AF-2 domain helices of the Erα , PPARγ , and PXR LBDs move together and in the same direction in each receptor . One may postulate that the uncorrelated motion between the helices in the AF-2 domain observed for inactive receptors ( e . g . , apo PXR-RXR heterodimer , ERα monomer and the PPARγ P467L-RXR mutant ) may represent the initial transition towards an αAF position required for corepressor binding [29] . Alternatively , these anticorrelated motions may simply prevent coactivator binding to LBDs that are not in active-capable oligomeric states . The results presented here are also in agreement with limited proteolysis [15] , fluorescence polarization [20] , and NMR [30] , [31] studies that examined the stabilization of global and local motions of ERα [15] and PPARγ [20] upon ligand binding . Of particular note are time-resolved fluorescence polarization studies by Kallenberger and Schwabe [20] on the human P467L PPARγ mutant that causes insulin resistance and early onset hypertension . This mutation was found to weaken immobilization of αAF against the main body of the receptor . In our molecular dynamics simulations , wild type PPARγ-RXR exhibited a strong degree of correlated AF-2 motion while the PPARγ P467L-RXR mutant showed uncorrelated motion in its AF-2 domain ( Figure 6C , D ) . This is the first model of nuclear receptor dynamics that relates changes in motion to a mutation causing a disease state . While nuclear receptors are well-established targets for small molecule modulators that treat a wide range of conditions , current drugs function as agonists and antagonists via the ligand binding pocket . However , recent data have indicated that nuclear receptor LBDs can be antagonized using small molecules that block coregulator binding to the AF-2 surface . For example , thyroid receptor antagonists discovered by high-throughput screening were found to act at the AF-2 site of that receptor [32] , [33] . In addition , the azole family of antifungal compounds has recently been shown to antagonize the action of human PXR via the AF-2 domain [24] , [25] . The dynamics data presented here further elucidate the nature of motions essential for AF-2 active-capable function , and may facilitate the improved design or development of therapeutics targeted to specific NR AF-2 surfaces .
Molecular dynamics simulations were run on the apo PXR-RXR LBD heterodimer and heterotetramer . MD simulations were also performed for the nuclear receptors ERα ( monomer and homodimer ) and PPARγ ( wild-type heterodimer with RXR and mutant P467L heterodimer with RXR ) . A summary of these simulations containing their oligomeric states , starting structure PDB IDs , and activity is provided in Table 1 . All starting structures were obtained from the protein databank ( www . rcsb . org ) . The PXR-RXR heterodimer and heterotetramer models as proposed in Noble et al . [12] were generated by first generating a PXR-RXR heterodimer model , followed by overlaying two copies of the heterodimer onto each protomer of the PXR homodimer structure . The PXR-RXR heterodimer model was created by superimposing the PXR LBD onto the LBD of PPARγ in the PPARγ-RXRα heterodimer crystal structure ( PDBID: 1FM6 ) . Upon creating this model , the PXR LBD was found to make nearly identical salt bridges , hydrogen bonds and hydrophobic interactions with the RXRα LBD as seen in the PPARγ-RXRα heterodimer crystal structure . All MD simulations were carried out with a 2 fs time step using the AMBER 2003 force field [34] . Molecular graphics figures were generated in Pymol ( http://pymol . sourceforge . net ) . All production runs employed the PMEMD module from Amber 9 . 0 [35] . Frames were recorded every 0 . 4 ps . Topology and parameter files were created using the LEaP program within AMBER [35] . The simulation system consisted of the protein surrounded by a truncated octahedron of water and sodium ions to maintain charge neutrality . An explicit solvent model was used with TIP3P water molecules filling 12 . 5 Å between the surface of each protein and the edge of the box [36] . Electrostatic interactions were calculated using the particle-mesh Ewald algorithm [37] with a cutoff of 10 Å applied to Lennard-Jones interactions . The SANDER package within AMBER was used for 5000 steps of energy minimization . Equilibration included 20 ps of constant volume conditions with heating from 100 to 300 K followed by 100 ps constant temperature conditions . Constant volume heating from 200 to 300 K was applied to the system for 20 ps before beginning the production run with the NPT ensemble . Simulations were analyzed using the PTRAJ package in Amber [35] . All-atom moving average root-mean-square deviations ( RMSD ) were calculated for each trajectory using the initial crystal structure as reference with an interval of 100 data points . Quasiharmonic analysis was employed for each trajectory using PTRAJ [35] . In all PXR simulations , a disordered loop ( PDB ID 1ILG , residues 178–197 ) missing from the apo PXR LBD crystal structure was modeled using the MODELLER module of InsightII with database searching ( www . accelrys . com ) [38] . The N and C termini of the modeled loop segment were reconnected to the missing sections of the crystal structure to avoid the termini from unrealistic flopping during simulations . The loop was examined for its potential impact on the RMSD from starting crystal structure by analyzing the simulations of the PXR LBD with and without the loop . The loop was found to impact the overall magnitude , but not the variability of the RMSD , suggesting that these regions move more than others , but do not effect stable conformations sampled during the simulation . Therefore , we have omitted the loop from subsequent analyses . However , we chose to include this loop in our simulations because it is a more realistic biological representation of the receptor . The pair-wise correlation coefficient as described in Sharma et al . [18] , Cij , was computed between α-carbons of two residues , i and j , with values ranging from −1 to +1 . The more positive the value of Cij , the more correlated ( moving in the same direction with one another ) the two residues , i and j , move . Likewise , the more negative the value of Cij , the more anticorrelated ( moving in the opposite direction to one another ) the two residues , i and j , move . The single-linkage clustering method [39] was applied to identify distinct sets of residues that move correlated with each other or anticorrelated to each other . In this method , a graph is initially built where each entity corresponds to individual residues . The clustering method proceeds by first finding two entities that have the highest similarity ( i . e . , the correlation coefficient ) between them . After clustering those two entities into one , the similarities between this new entity and the rest are updated . This process is repeated until there are no more entities to cluster or the correlation coefficient cutoff is satisfied . In a single-linkage clustering method , the similarity between two clusters is defined as the largest similarity or the highest correlation coefficient between any two members from the two clusters . Residues chosen to describe motion in Tables 2–4 were not chosen at random in the AF-2 domain of PXR LBD . Glu-427 of αAF and Lys-259 of α3 are the “charge clamp” residues of PXR; the charge clamp is a common structural motif in nuclear receptor-coactivator interactions and involves contacts between the LBD and the termini of the coactivator LxxLL helix . Lys-277 of α4 was chosen because it is conserved in many receptors and Leu-424 of αAF directly contacts the coactivator SRC-1 [14] . Angle analysis was performed using Equation 1 to find the average angle between vectors for α-carbon a and b . ( 1 ) The effective modes of vibrational motion can be obtained using quasiharmonic analysis by calculating a force field relative to the average structure based on the fluctuations generated from an MD simulation . Quasiharmonic modes , unlike standard principal component methods , are mass weighted just as normal modes and thus may be compared directly with normal mode analysis . However in quasiharmonic modes , anharmonic effects are implicitly included and thus may be different from normal modes [40] . The percent contribution of each quasiharmonic analysis mode to the overall motion can be evaluated by analyzing the eigenvalues of the first 50 modes . The percent contribution of each mode can be determined by taking the reciprocal of the eigenvalue of one mode and dividing by the sum of the inverse eigenvalues for all 50 modes . The eigenvalue is equivalent to the square of the frequency ( cm−1 ) . The percent contribution of each mode ( Figure S5 ) drops off quickly with only the first few modes showing any significant contribution to the overall motion . Modes 1 and 2 in the PXR-RXR heterodimer and heterotetramer simulations represent proximal percent contributions , while in ERα and PPARγ-RXR simulations mode 2 contributed 50% less to overall motion than mode 1 ( Figure S5 ) . In order to sample the most relevant motions , the first two modes were analyzed for PXR-RXR simulations and only the first mode was analyzed in the ERα and PPARγ-RXR simulations . In all cases , the first mode ( s ) were sufficient to describe between 18–33% of the overall motion ( Figure S5 ) . To simplify analysis of the PXR-RXR simulations , the sums of the x , y and z vector components of each atom in each mode were obtained and weighted against the percent contribution . Normal mode analysis ( NMA ) is based on a harmonic approximation of the potential energy function around a minimum energy conformation [41] , [42] . ELNEMO uses a Hookean potential described by Tirion [41] , [43] , which assumes that the total energy potential function of the reference 3D structure ( in this case the crystal structure ) is at an energy minimum . In NMA , the lowest energy modes ( below 30–100 cm−1 ) have the largest contribution to the amplitude of atomic displacements . However the first six normal or vibrational modes represent rotational and translational motion and are disregarded [44] . Normal mode theory has been shown to accurately describe large conformational transitions in proteins such as hexokinase [45] , lysozyme [46] , [47] and citrate synthase [48] which occur at microsecond or millisecond time scales . Fifty normal modes were generated using the ELNEMO server for each state of the three nuclear receptors [44] . The only change made was the removal of the modeled loop region ( residues 178–197 ) in the PXR-RXR complexes , as these residues resulted in low frequency modes with low collectivity . Collectivity is a measure of the fraction of residues affected by a given mode . Computed normal modes sometimes have localized motion that corresponds to extended parts of the protein and are usually ignored [44] . This was done to confirm that the high degree of correlated motion we observed in simulations involving the active-capable forms of nuclear receptors were relevant at longer time scales . Just as in the quasiharmonic analysis of the all-atom molecular dynamics simulations , we first sought to determine the minimum number of modes required to obtain an accurate description of the overall motion . Figure S6 shows the percent contribution of each mode , up to the first 50 modes . The first six modes of motion are trivial and have been removed from the analysis . Except for the tetramer , the percent contribution of each of the normal modes appears to drop off more slowly than those of the QH analysis ( Figure S5 , S6 ) . We chose to analyze modes 7–20 , which describe from 48–81% of the overall motion of each nuclear receptor ( Figure S6 ) . To simplify the analysis of the modes , we calculated the vector sum of each atom for modes 7–20 , weighted by the percent contribution of each mode .
|
Long-range motions play essential roles in protein function but are difficult to appreciate from static crystal structures . We sought to understand how macromolecular motion affects the formation of transcriptional complexes central to controlling gene expression . Using 20- to 30-ns molecular dynamics simulations , we examined three nuclear receptors that function as ligand-regulated transcription factors: the pregnane X receptor , the peroxisome proliferator-activator receptor-γ , and estrogen receptor-α . We found that each of these receptors exhibits a high degree of correlated motions within the domain responsible for forming functionally essential protein–protein interactions with transcriptional coactivators . We further found that specific long-range ( up to 30 Å ) motions play an important role in these dynamics . Our results show that “active-capable” nuclear receptors are prepared for coactivator contacts by maintaining a mobile but preformed protein–protein interaction surface .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"computational",
"biology/molecular",
"dynamics",
"computational",
"biology/macromolecular",
"structure",
"analysis",
"diabetes",
"and",
"endocrinology/endocrinology"
] |
2008
|
Active Nuclear Receptors Exhibit Highly Correlated AF-2 Domain Motions
|
Thromboembolic events were described in patients with Chagas disease without cardiomyopathy . We aim to confirm if there is a hypercoagulable state in these patients and to determine if there is an early normalization of hemostasis factors after antiparasitic treatment . Ninety-nine individuals from Chagas disease-endemic areas were classified in two groups: G1 , with T . cruzi infection ( n = 56 ) ; G2 , healthy individuals ( n = 43 ) . Twenty-four hemostasis factors were measured at baseline . G1 patients treated with benznidazole were followed for 36 months , recording clinical parameters and performance of conventional serology , chemiluminescent enzyme-linked immunosorbent assay ( trypomastigote-derived glycosylphosphatidylinositol-anchored mucins ) , quantitative polymerase chain reaction , and hemostasis tests every 6-month visits . Prothrombin fragment 1+2 ( F1+2 ) and endogenous thrombin potential ( ETP ) were abnormally expressed in 77% and 50% of infected patients at baseline but returned to and remained at normal levels shortly after treatment in 76% and 96% of cases , respectively . Plasmin-antiplasmin complexes ( PAP ) were altered before treatment in 32% of G1 patients but normalized in 94% of cases several months after treatment . None of the patients with normal F1+2 values during follow-up had a positive qRT-PCR result , but 3/24 patients ( 13% ) with normal ETP values did . In a percentage of chronic T . cruzi infected patients treated with benznidazole , altered coagulation markers returned into normal levels . F1+2 , ETP and PAP could be useful markers for assessing sustained response to benznidazole .
Chagas disease ( CD ) is one of 17 neglected tropical diseases recognized by the World Health Organization . Caused by the protozoan parasite Trypanosoma cruzi , it mainly affects people with poor socioeconomic status and limited health care access in endemic and nonendemic countries . [1 , 2] Thrombosis is considered as a pathological deviation of haemostasis , and it is characterized by intravascular thrombus formation and vessel occlusion . Perturbation of hemostasis is an important factor in the pathogenesis of thromboembolic events , which can be caused by blood flow dysregulation , endothelial injury , and coagulation system alterations . Recently , is has been described that under certain circumstances thrombosis is a physiological process that constitutes an intrinsic effector mechanism of innate immunity , and the process has been defined as “immunothrombosis” . [3] It is activated after the recognition of pathogens and damaged cells , and inhibits pathogen dissemination and survival . Immunothrombosis can therefore be regarded as a newly identified , crucial element of intravascular immunity , which is a part of the immune system that encompasses a wide range of host strategies to detect and protect against pathogens in the vasculature . Dysregulation of immunothrombosis is likely to constitute a key event in the development of thrombotic disorders . [3] Infectious disease can cause a hypercoagulable state through the upregulation of tissue factor in monocytes , the generation of procoagulant microparticles , the activation of the coagulation intrinsic pathway , platelet activation , and NETs ( Neutrophil Extracellular Traps ) release . [3] Different infectious agents may cause different responses but a final degree of hypercoagulability can be a common trait as one of the biological endpoints . Additionally , patients with chronic inflammation may also present platelet adhesion events , which are considered inflammatory processes and can be observed in patients with chronic T . cruzi infection , even in the asymptomatic stages . [4] Infection itself can cause vasculitis , increasing proinflammatory cytokine levels and perpetuating the risk of thrombotic events . [5] In the case of the Chagas’ disease the effect of hemostasis in the bradikinin formation , through the effect of factor XII activation in the Kallikrein-Kinin system , can modify the type 1 immune response and then modulate the antiparasite immunity as suggested in a mice model of subcutaneous infection by T . cruzi . [6] Thromboembolic events and dilated cardiomyopathy , ventricular aneurysms , and intracavitary thrombosis are associated with CD . [7 , 8] Rheological factors can induce intraluminal thrombus formation with the risk of embolism . [9] Alterations of molecular markers of coagulation system activation have been described in T . cruzi infection individuals with or without clinical thrombosis . [9–12] Other factors , such as injury to vessel walls by parasites or changes in blood viscosity due to host immune response , may influence in the development of thromboembolic events in T . cruzi-infected individuals without Chagas cardiomyopathy or other vascular risk factors . [13]Based on studies performed in humans with chronic T . cruzi infection , there are controversial results regarding the existence of a prothrombotic status in T . cruzi-infected patients . [13 , 14] There is an study in which a of higher prothrombotic status in the CD group was not found , but the control group were individuals without T . cruzi infection and heart failure . [14] In previous studies performed in murine models , several abnormalities of the heart microcirculation of individuals with chronic CD were pointed out , but they did not find evidence of thrombi and neither thromboembolism . [15 , 16] Higher levels of the hypercoagulability markers prothrombin fragment 1+2 ( F1+2 ) , thrombin-antithrombin complexes ( TAT ) , fibrinogen/fibrin degradation products , plasminogen activator inhibitor type 1 ( PAI-1 ) , and D-dimer have been reported in T . cruzi–infected patients compared with healthy individuals . [10 , 11] A pilot study performed by our group showed that endogenous thrombin potential ( ETP ) and F1+2 levels were outside normal ranges in 73% and 80% of T . cruzi–infected patients without advanced heart disease , respectively . [12] We demonstrated a 100% and 73% decrease in these levels six months after treatment with benznidazole . Thus , if they prove to remain stable in time , hypercoagulability factors could be used as biomarkers of therapeutic response in CD . Besides , although whether or not chronic Chagas disease is an independent vascular risk factor remains to be confirmed . [17 , 18] While specific treatment is recommended in both acute and chronic stages of infection [19 , 20] , there are only two drugs ( i . e . , benznidazole and nifurtimox ) available for the treatment of CD . The mechanism of action of benznidazole relates to the nitro-reduction of components of the parasite , the binding of metabolites of the nuclear DNA and k-DNA of T . cruzi and the lipids and proteins of the parasite . [21] In adults , benznidazole has a high rate of adverse effects , which can be classified into three groups: ( i ) hypersensitivity , including dermatitis with cutaneous eruptions ( usually appearing between days 7 and 10 ) , myalgias , arthralgias , and lymphadenopathy; ( ii ) polyneuropathy , paresthesias , and polineuritis usually during the 4th week of treatment ) ; and ( iii ) bone marrow disorders , such as thrombopenic purpura and agranulocytosis ( usually after the second week of treatment ) . [22]Furthermore , the effectiveness of these drugs in the chronic stage of infection is still a topic of debate due to inconsistent studies’ results [23–25] and a lack of early biomarkers of response to specific T . cruzi treatment with benznidazole . [26] Following on from our pilot study [12] , here we increased the sample size and extended follow-up to further investigate the value of hypercoagulability factors as biomarkers of treatment response in CD . We also added current treatment response parameters measured by conventional serology , serology for lytic anti-α-galactosyl ( anti-α-Gal ) antibodies against T . cruzi [27–29] , and quantitative reverse transcription polymerase chain reaction ( qRT-PCR ) . [30] The aims of the study were to investigate alterations of hypercoagulability factors in patients chronically infected with T . cruzi and determine whether there is an early and sustainable improvement of the hypercoagulability factors after antiparasitic treatment .
Written informed consent was obtained from participants before being recruited ( all of them were adults ) . Approval for the protocols and for the informed consent was obtained from the Hospital Clínic of Barcelona Ethics Review Committee . This is a descriptive study of 99 individuals ( 56 with T . cruzi infection and 43 healthy individuals ) from Latin American , where CD is endemic . All the individuals were evaluated at the Centre for International Health at Hospital Clínic in Barcelona , Spain . Ninety-nine individuals from CD-endemic areas living in Barcelona were invited to participate . Inclusion criteria were an age of over 18 years and provision of signed informed consent . Exclusion criteria were pregnancy , non-Chagasic cardiopathy , late chronic cardiac or digestive forms of CD , other acute or chronic infections , inflammatory or immunological diseases , and chronic systemic diseases ( high blood pressure and diabetes ) . After signing the informed consent form , participants were asked for clinical and epidemiological data , including area of origin and risk factors for the CD transmission . The information recorded included vascular risk factors , toxic habits , and cardiological and/or vascular events . Conventional serology of T . cruzi infection was established using two ELISA kits: a commercial kit with recombinant antigens ( BioELISA Chagas , Biokit S . A . , Barcelona-Spain ) and an in-house kit with whole T . cruzi epimastigote antigen , as described . [12 , 31] . Diagnosis was confirmed by a positive result on both tests . [19] Following serological tests results , participants were divided into two groups: those with T . cruzi infection ( Group 1 [G1] ) and those without ( Group 2 [G2] ) . All the participants underwent human immunodeficiency virus testing , basic blood and biochemical tests ( including renal and liver function ) , and specific evaluation of hemostasis factors . For the hemostasis studies , blood was collected in citrate-containing tubes ( Becton Dickinson ) , samples were centrifuged , and platelet-poor plasma aliquots were frozen at –80°C until assayed . Prothrombin time , activated partial thromboplastin time , coagulation factor VIII , protein C activity , free and total protein S levels , antithrombin and plasminogen activity , F1+2 , plasmin-antiplasmin complexes ( PAP ) , factor VIIa , PAI-1 , P-selectin , factor V Leiden and prothrombin gene G20210A mutation , lupus anticoagulant and anticardiolipin antibodies were measured as previously described . [12] D-dimer was measured using an automated turbidimetric test ( Siemens Healthcare Diagnostics ) and ETP was assessed using a continuous chromogenic thrombin generation assay and ETP Curves software ( Siemens ) . The ETP coagulation test was initiated by using human recombinant tissue factor , phospholipids , and calcium ions . ADAMTS-13 was measured using a commercial chromogenic method ( American Diagnostica ) . Factor XIIa was determined by a direct quantitative commercially available immunoassay ( Shield Diagnostics ) with a highly specific monoclonal antibody that does not recognize its zymogen factor XII . [32] Plasma tissue factor levels were determined using a commercial kit ( American Diagnostica ) according to the manufacturer’s protocol . Plasma levels of von Willebrand factor antigen were determined by enzyme-linked immunosorbent assay ( ELISA ) ( Corgenix ) . Procoagulant activity of microparticles was measured using a functional assay with the addition of factors Xa , Va , and prothrombin after microparticle capture in the solid phase using annexin V ( Hyphen Biomed ) . Soluble CD40L was measured by ELISA ( R&D Systems ) . qRT-PCR [30] and a chemiluminescent ELISA assay based on a highly purified , trypomastigote-derived glycosylphosphatidylinositol-anchored mucin ( tGPI-mucin ) antigen for the serological detection of lytic anti-α-Gal antibodies against T . cruzi ( AT CL-ELISA ) [27–29 , 33–36] , were performed in G1 at month 0 ( baseline ) , and 6 , 12 , 18 , 24 , 30 , and 36 months post-treatment . For AT CL-ELISA , a serum sample was considered positive when the titer was ≥1 . 0 and negative when it was ≤0 . 9 . Inconclusive or equivocal results were determined by a titer between 0 . 9 and 1 . 0 . [27 , 35]All sera were tested in duplicate and the results were expressed as the mean of two simultaneous determinations . G1 patients were studied using a protocol that included a 12-lead electrocardiogram , chest X-ray , and echocardiogram . They were followed up every 6 months for at least 36 months . At each visit , clinical data were collected and the following tests were performed: ELISA , AT CL-ELISA , qRT-PCR , and hemostasis tests . Other tests were performed according to individual symptoms . Specific treatment with benznidazole ( 5 mg/kg/day for 60 days ) was offered to all T . cruzi–infected patients , and those treated were monitored fortnightly for clinical and analytical assessment . Treatment was considered complete when at least 80% of the total dose was reached . A hypercoagulable state is defined as the presence , in certain individuals , of thrombotic potentialities that activate the endothelium and the formative elements of the blood ( mainly , platelets ) that favors plasma kinetics that lead to the formation of thrombin , which disturbs fibrinolytic activity and produces hemorheological changes with turbulence phenomena that predispose to thrombogenesis . [18] Quantitative variables were presented as medians and interquartile range ( IQR ) and were compared between groups using the Wilcoxon rank sum test . Qualitative variables were reported using absolute frequencies and percentages and between-group comparisons were made using Fisher’s exact test . Hypercoagulability biomarker variation over time was assessed using a mixed-effect linear regression model with a random intercept structure . Hypercoagulability factors were used as dependent variables and follow-up time as the explanatory variable , with one category for each time point: baseline , month 6 ( reference for comparisons ) , and months 12 , 18 , 24 , and 36 . This type of model allows for the inclusion of random effects in addition to the overall error term . Random intercept regression was also used to assess whether antibody levels measured by ELISA and AT CL-ELISA approached the negative threshold during follow-up . The response variable was the distance from this threshold ( i . e . , the difference between each ELISA or AT CL-ELISA value and the negative cutoff ) and the explanatory variable was the follow-up time from month six ( reference ) to month 36 . The regression coefficients express the effect estimate of follow-up on the outcome variable . The pattern of the relationships between hypercoagulability biomarkers was assessed by multiple correspondence analysis ( MCA ) using the Burt matrix approach . [37 , 38] The MCA represents a method for analyzing multi-way contingency table containing measure of correspondence between row ( subjects ) and columns ( levels of variables ) . The interpretation is based upon proximities between levels of variables ( or points ) in a low-dimensional map . The firsts dimensions ( usually one or two ) account for meaningful amounts of variance and are those retained for the map definition and interpretation . The first dimension accounts for a maximal amount of total variance in the observed variables . Under typical conditions , this means that the first component will be correlated with at least some of the observed variables . The second dimension has two important characteristics: it accounts for a maximal amount of variance in the data set that is not accounted for by the first dimension , thus it is correlated with some of the observed variables that not display strong correlations with dimension 1; and it is uncorrelated with dimension 1 . Looking at the map , the proximity between levels of different variables means that these levels tend to appear together in the observations . Since the levels of the same variable cannot occur together , the proximity between levels of the same variable means that the groups of observations associated with these levels are themselves similar . A level far away from the origin ( of the dimensions ) means that is well-represented in the map , thus that level is meaningful for the interpretation of the dimension ( s ) . All levels that are not useful for the solution are near the origin . Supplementary ( passive ) variables are those not used for the solution but mapped in the graph in order to help in the interpretation . The biomarkers were classified into three categories: normalization of values throughout follow-up , non-sustained normalization during follow-up and normal values at baseline . Two additional variables were considered: qRT-PCR results during follow-up ( categories: always negative and sometime positive ) and level of adherence ( categories: 80% and 100% ) . All the tests were 2-tailed and the confidence level was set at 95% . The analyses were performed using Stata 13 ( Stata Corporation , College Station , TX , USA ) .
Ninety-nine individuals ( 76 women ) were studied . Fifty-six of these ( 43 women ) were T . cruzi–positive ( G1 ) and 43 ( 33 women ) were T . cruzi–negative . The mean ages were 34 ( SD , 9 ) years for the overall group ( range 17–56 , median 33 ) , 37 ( SD , 9 ) years for G1 , and 32 ( SD , 7 ) years for G2 . Fifty G1 patients were treated with benznidazole ( six were lost to follow-up before starting treatment due to unexpected work-related changes in the migratory process ) . Forty-five ( 90% ) completed treatment . Eighty-six participants ( 87% ) ( 51 [91%] in G1 and 35 [81%] in G2 ) were from Bolivia . None of the participants traveled to their countries or other CD-endemic areas during follow-up . The clinical and demographic data are summarized in Table 1 . The epidemiological and baseline clinical data were similar in both groups , making them statistically comparable . Comparison of the 24 hypercoagulability biomarkers at baseline between ( untreated ) G1 and G2 individuals showed statistically significant differences for D-dimer ( P = . 0262 ) ; F1+2 ( abnormal values in 43/56 G1 patients [77%] , P < . 0001 ) , PAP ( abnormal values in 17/56 G1 patients [30%] , P = . 0111 ) , P-selectin ( abnormal values in 7/56 G1 patients [13%] P = . 0177 ) , and ETP ( abnormal values in 28/56 G1 patients [50%] , P < . 0013 ) , and circulating microparticles ( P = . 0112 ) ( Table 2 ) . D-dimer levels were normal in all the individuals in G1 and G2 , and microparticles were within the normal range in a high percentage of patients ( 86% in G1 and 93% in G2 , P = . 3402 ) . Our findings showed that a high percentage of patients with chronic T . cruzi infection have a hypercoagulable state regardless the clinical stage of disease , thus confirming the observations of previous studies . [11–13] Thirty-three ( 76% ) of the 43 patients with abnormal baseline F1+2 values achieved normal levels after a median follow-up of 9 month ( IQR , 8 ) . All but one of the 28 patients with abnormal ETP values before treatment showed normal values at 6 months ( IQR , 3 ) . These values were maintained throughout follow-up ( 30 months; IQR , 28 ) in 15 patients ( 60% ) . Fifteen of the 17 patients with abnormal baseline PAP values showed normal values 7 months ( IQR , 7 ) after treatment and nine of these ( 60% ) maintained these values throughout follow-up ( 28 months; IQR , 11 ) . However , PAP values at 12 and 48 months seemed to be higher than those at 6 months , but the confidence interval indicates a lack of precision for both time point effect estimates ( Table 3 ) . Thus , once normalized , F1+2 and ETP levels did not increase again significantly after treatment . Fig 1 shows a graphic representation of these results . F1+2 values are an indirect measure of the amount of thrombin generated in vivo ( mainly due to endothelial injury , even in subclinical states ) [39] , and ETP levels indicate the potential amount of thrombin that can be formed when blood coagulation is activated through the addition of tissue factor . PAP complexes are markers of fibrinolysis . Upon activation , plasmin , which is primarily responsible for a controlled and regulated dissolution of the fibrin polymers into soluble fragments , is immediately inactivated by antiplasmin , forming PAP complexes . [40] Therefore , it is conceivable that the increase formation of PAP complexes stems from excessive formation of fibrin in the blood stream of untreated T . cruzi infected patients . Soluble P-selectin is considered a biomarker of in vivo platelet activation . P-selectin is contained in the α-granules of platelets; following platelet activation , the soluble form is expressed on the platelet surface and then shed by cleavage . P-selectin has been shown to act as a link between thrombosis and inflammation . [41] Additionally , the four biomarkers-F1+2 , ETP , PAP complexes , and P-selectin-reflect are highly stable over time . A hypercoagulable state is a term that pretends to denominate a condition in which there is an increased tendency toward blood clotting . There is not a universally accepted definition for this state based in biomarkers values , but an increase in several of them suggests the possibility of an increase in the person's chances of developing blood clots . The increases in F1+2 , PAP and ETP are congruent with this idea: F1+2 and PAP indicate the actual amount of thrombin and plasmin formed , as markers in procoagulant and fibrinolysis pathways , respectively; and ETP indicates the potential amount of thrombin that can be formed considering globally all the activators , inhibitors and substrates of the hemostasis present in the plasma . The increase observed in these biomarkers is good enough to be an argument to point out a hypercoagulable state in patients with Chagas’ disease . Sixteen ( 33% ) of the 56 G1 patients had a positive qRT-PCR result at baseline , but only four of these had a positive result after treatment ( treatment failure rate of 25% in this subgroup ) . Five of the 34 patients with a negative baseline qRT-PCR result showed a positive result during follow-up . None of the patients with normal F1+2 values during follow-up had a positive qRT-PCR result , but 3 ( 13% ) of the 24 patients with normal ETP values during follow-up did . Of the patients with altered levels of F1+2 , ETP , or PAP complexes at baseline , a positive qRT-PCR result during follow-up was not significantly associated with changes observed in lytic anti-α-Gal antibodies , F1+2 , ETP , and/or PAP levels . A positive qRT-PCR result after treatment in patients who achieved normalization of F1+2 , ETP , and/or PAP could mean that a decrease in parasite load is sufficient to modify the hypercoagulable state or that benznidazole , which acts on the redox system , could modify these biomarkers without eliminating the parasites . This would limit the use of these factors as biomarkers for parasite elimination , although they could be valuable indicators of treatment response and add support to the theory that , by reverting the hypercoagulable state , benznidazole may also prevent clinical thrombotic events . Conventional ELISA results were positive in all the patients in G1 . Although , as expected , antibodies remained positive throughout follow-up , a slight decrease was detected by the commercial and in-house methods during this period . A statistically significant relevant decrease , was only observed with the in-house test from month 18 onwards ( P = . 0006 ) . Lytic anti-α-Gal antibodies were positive in 52 ( 96% ) of the 54 patients tested before treatment , and in all patients AT CL-ELISA remained within positive levels to the end of the follow-up ( Fig 2 ) . Besides , there was no correlation between lytic anti-α-Gal antibody assay and the hemostasis factors evaluated . In relation to previous studies’ results , early decreases in lytic anti-α-Gal antibodies were expected to be observed . On the contrary , a decrease in levels was evident at month 12 and this was significant since month 18 and forward ( P = . 0052 ) . [28 , 34] Adherence to treatment was high , with only five patients not achieving 80% of the total dose . All five patients showed abnormal F1+2 values throughout follow-up and 3 ( 60% ) had abnormal ETP and PAP values . One of the five patients had a positive qRT-PCR result during follow-up , and all five maintained the same positive ELISA and AT CL-ELISA results throughout follow-up . A large cohort of adolescents with T cruzi infection treated with benznidazole showed seronegativity in lytic anti-α-Gal antibodies , as measured by AT CL-ELISA , in 58% and 85% of the patients 36 and 72 months after treatment , respectively . [28 , 34] The differences between those studies and ours may be due to the nature of the cohorts ( adolescents vs . adults ) and the stage of the disease . Nevertheless , both studies showed a similar trend towards a reduction in lytic anti-α-Gal antibodies following treatment with benznidazole . We studied the relationship between normalization of hypercoagulability markers F1+2 , PAP , and ETP and qRT-PCR results by multiple correspondence analyses ( MCA ) . Due to the low rate of positive qRT-PCR results , this variable was used as a supplementary variable jointly with treatment adherence . The MCA results ( Fig 3 ) showed an association between complete normalization of PAP and ETP levels and non-sustained and marginally abnormal values in F1+2 . These factors had the highest contribution and correlation in the positive part of the second dimension , while normal baseline ETP and PAP values had the highest contribution and correlation in the negative part . F1+2 normalization clearly characterized the positive part of the first dimension , while non-sustained normalization of PAP and ETP values clearly characterized the negative part . In other words , the sustained normalization observed post-treatment in PAP and ETP , could , despite the non-sustained normalization of F1+2 values , reflect response to antiparasitic treatment due to the strong correlation between these three variables . The projection of qRT-PCR results and adherence to treatment in the solution space provided little additional information . Consistently negative qRT-PCR results throughout follow-up appear to be related to 100% treatment adherence . In a recent study , the authors found that the serum samples of 37 individuals with chronic Chagas disease showed an upregulation of specific fragments of apolipoprotein A-1 ( Apo A1 ) and one fibronectin fragment , that returned to normal levels in 43% of them three years after a treatment with nifurtimox . [38] Apo A1 and fibronectin fragment were altered in all the 37 patients with T . cruzi infection before treatment , but the number of patients treated with that normalized levels was lower than in our series ( 60% and 96% of patients who normalized F1+2 and ETP values ) . This study has some limitations . Although the sample size was calculated to obtain sufficient statistical power to answer the hypothesis , a larger sample may have detected differences that would be expected to appear earlier ( e . g . , before 12 months ) . The lost to follow-up samples also affected the estimates . Even within Spain , it is difficult to follow individuals with high migratory mobility for long periods . In addition , the fact that only 30% of patients had a positive baseline qRT-PCR result was a constraint for assessing the effect of treatment . In conclusion , patients with chronic T . cruzi infection have a potential hypercoagulable state , regardless of cardiological and/or digestive involvement . The hypercoagulability markers F1+2 and ETP were abnormally expressed in a high percentage of patients with chronic T . cruzi infection before treatment ( 77% and 50% , respectively ) but returned to and remained at normal levels shortly after treatment in 76% and 96% of patients , respectively . Baseline PAP values were altered in just 30% of patients before treatment , but normalized several months after treatment in 88% of these . These three hypercoagulability biomarkers could be useful for assessing short-term response to treatment . However , the fact that normal values were seen in some infected patients , including some with positive post-treatment qRT-PCR results , reduces their usefulness as universal biomarkers . The decrease in hypercoagulability factor levels could be explained by a decrease in parasitemia or by other benznidazole effect .
|
The manuscript describes the results of a study whose aim was to assess the tendency to coagulate in people suffering from a parasitic infection frequent in Latin America named T . cruzi infection or Chagas disease , by the study of several coagulation factors . According to the state of the art in this topic , specific treatment for Chagas disease is recommended in recent ( acute ) and late ( chronic ) stages of the infection . The effectiveness of current available drugs in the chronic stage of infection is still a topic of debate due to inconsistent results across studies and a lack of early measurable parameters of response to specific treatment . Another aim of this study was to determine if the presence of an upregulated procoagulative activity in plasma in people suffering T . cruzi infection could be used as potential marker that indicates therapeutic response in people at chronic stage of the disease . The results of this study suggest that measurements of alterations of procoagulative activity may be useful to indicate specific treatment for T . cruzi chronically infected patients and new data concerning early response to treatment biomarkers .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"and",
"Discussion"
] |
[] |
2016
|
Altered Hypercoagulability Factors in Patients with Chronic Chagas Disease: Potential Biomarkers of Therapeutic Response
|
During the intracellular life of Salmonella enterica , a unique membrane-bound compartment termed Salmonella-containing vacuole , or SCV , is formed . By means of translocated effector proteins , intracellular Salmonella also induce the formation of extensive , highly dynamic membrane tubules termed Salmonella-induced filaments or SIF . Here we report the first detailed ultrastructural analyses of the SCV and SIF by electron microscopy ( EM ) , EM tomography and live cell correlative light and electron microscopy ( CLEM ) . We found that a subset of SIF is composed of double membranes that enclose portions of host cell cytosol and cytoskeletal filaments within its inner lumen . Despite some morphological similarities , we found that the formation of SIF double membranes is independent from autophagy and requires the function of the effector proteins SseF and SseG . The lumen of SIF network is accessible to various types of endocytosed material and our CLEM analysis of double membrane SIF demonstrated that fluid phase markers accumulate only between the inner and outer membrane of these structures , a space continual with endosomal lumen . Our work reveals how manipulation of the endosomal membrane system by an intracellular pathogen results in a unique tubular membrane compartmentalization of the host cell , generating a shielded niche permissive for intracellular proliferation of Salmonella .
Bacterial pathogens have evolved sophisticated mechanisms to modify host cell functions in order to avoid antimicrobial defense or to use host cell-derived material for their own proliferation . Intracellular pathogens evade humoral immune responses of the host by hiding inside cells of the host organism , using these cells to support their own proliferation and to disseminate within the host organism [1] . These activities require the manipulation of the normal host cell processes to avoid killing by the host cell and to establish a replication-permissive intracellular niche . One group of intracellular bacteria resides in specialized membrane compartments that derive from the endosomal system of the host cell . Salmonella enterica is a facultative intracellular pathogen that modifies eukaryotic host cells in order to establish a unique parasitophorous vacuole , the Salmonella-containing vacuole or SCV [2] , [3] . The SCV is a membrane-bound compartment that allows the survival and replication of S . enterica in a variety of mammalian host cells . At later time points after infection , the SCV acquires certain characteristics of late endosomal/lysosomal compartments such as ( i ) the presence of a subset of Rab GTPases , such as Rab7 , ( ii ) lysosomal glycoproteins ( lgp ) , such as LAMP1 , ( iii ) acidification , and ( iv ) juxtanuclear positioning . However , the SCV remains permissive for intracellular replication of Salmonella . Several studies investigated the interaction of the SCV with the host cell endosomal system and indicate that various types of interaction take place during the maturation of this specialized compartment [4]–[7] . One unique phenotype resulting from these interactions is the induction of long tubular membrane compartments extending from the SCV , termed Salmonella-induced filaments or SIF [8] . SIF were first observed in Salmonella-infected epithelial cells , but we extended this observation also to phagocytic cells [9] . SIF are characterized by presence of specific protein markers similar to those on the SCV membrane and exhibit dynamic properties . The formation of SIF in epithelial cells starts from 3–4 h after invasion by fusion events of the SCV with endosomal vesicles . These early SIF are highly dynamic tubules which often exhibit a thinner appearing part at the distal end , termed leading SIF ( LS ) , in comparison to the proximal part , termed trailing SIF ( TS ) . From 8 h post infection ( p . i . ) a complex tubular network is established which appears stable and less dynamic . [9] , [10] . Despite a large body of work on the intracellular activities of Salmonella , the biogenesis and function of SIF are still enigmatic . In addition , recently distinct tubular membrane compartments have been identified induced by intracellular Salmonella: Golgi-derived Salmonella-induced SCAMP3 tubules ( SIST ) [11] and LAMP1-negative tubules ( LNT ) in cells infected with a sifA sopD2 double mutant strain [12] . Throughout this paper we will use the general term Salmonella-induced tubules ( SIT ) which does not discriminate between SIF , SIST , LNT or another ( un ) classified subtype of SIF . Characteristic features of SIT are summarized in Table 1 . Of central importance for the intracellular lifestyle of Salmonella is the function of the type III secretion system ( T3SS ) encoded by Salmonella Pathogenicity Island 2 ( SPI2 ) . Intracellular Salmonella deploy the SPI2-T3SS to translocate a set of effector proteins across the membrane of SCV [2] . Collectively , these effector proteins enable the intracellular survival and proliferation of Salmonella . For the majority of SPI2-T3SS effectors the exact mode of action and molecular targets in the host cell remain unknown . However , some effectors are likely relevant for both , induction of specific intracellular phenotypes and systemic pathogenesis . Formation of SIF requires at least five Salmonella effector proteins , i . e . SifA , SseF , SseG , SopD2 and PipB2 , and the integrity of the microtubule cytoskeleton [13] , [14] . The most severe phenotype is mediated by the SPI2 effector SifA [15] . Mutant strains lacking SifA are highly attenuated in intracellular replication and systemic virulence . Bacteria deficient in sifA fail to induce SIF and escape into the cytoplasm of the host cell due to a loss of SCV membrane integrity [15] . SseF and SseG also contribute to the intracellular lifestyle , although the defects in intracellular replication of the corresponding mutant strains are less pronounced compared to sifA or SPI2-T3SS deficient strains . The strict correlation between intracellular fitness of Salmonella and its ability to form SIF prompted us to investigate the nature of these tubular endosomal compartments in Salmonella-infected cells . We examined the biogenesis of the SCV and Salmonella-induced tubular compartments by a combination of confocal laser scanning microscopy ( CLSM ) of live infected cells , followed by ultrastructural analyses by transmission electron microscopy ( TEM ) and electron tomography ( ET ) . Our observations based on correlative light and electron microscopy ( CLEM ) provide a novel view on the intracellular activities of Salmonella leading to unique compartmentalization of the host endosomal membranes and we propose new models for SIF biogenesis .
We previously reported the ultrastructure of tubular membrane compartments in connection with SCV [9] . In addition to SIF with the characteristic presence of lgp , recent studies revealed additional types of tubular membrane compartments in Salmonella-infected cells including SIST [11] and LNT [12] . Here we use the general term ‘Salmonella-induced tubules’ or SIT if no further distinction is possible . For an overview of possible ultrastructural differences of SIT we first performed conventional TEM of Salmonella WT-infected HeLa cells ( Figure 1 ) . Ultrathin horizontal sections through flat-embedded cells revealed both , longitudinal and cross-sections through SIT . At the time points of sampling , i . e . 8–12 h p . i . , the majority of intracellular WT Salmonella ( 82 . 7% , N = 208 bacteria ) were contained within membrane compartments . These SCV enclosed one or more bacteria into a complex membrane organelle with a lumen enriched in electron-dense granules and multi-lamellar vesicles ( Figure 1 ) . Membrane tracing on cross-sections through the SCV revealed a continual ‘outer’ membrane of the SCV which encircled bacteria but also numerous membrane-enclosed bubble-like compartments within the lumen of the SCV ( Figure S1 ) . Although these compartments were reminiscent of cytosolic pockets enclosed within the complex 3D structure of the SCV , they may as well result from invaginations of the SCV ‘outer’ membrane . Extended analyses of ultrathin sections by TEM revealed additionally a variety of SIT morphotypes and an unexpected membrane organization of the tubular compartments ( Figure 1 ) . Apart from very thin single membrane SIT of rather uniform diameter of 46±8 nm and moderately electron-dense content ( Figure 1Ai ) , we also observed thicker single membrane SIT that appeared more electron-dense ( Figure 1Bii , iii ) . The latter tubules were characterized by a lumen containing electron-dense granules and vesicles , reminiscent of the luminal content of late endosomes or endolysosomes [16] , and will be referred to as type 1 SIT . Most frequently , we observed extended tubules with luminal content of low electron density ( Figure 1Cvi , vii ) . These SIT were delimited by two adjacent membranes which we discriminate as outer and inner membrane . These structures will be referred to as type 2 SIT . Both type 1 and type 2 SIT were also found in the same cells , and in some sections , interconnection between both types of tubules was detected . From observations by light microscopy , Salmonella is known to induce tubular networks [8] , [11] , [12] . We identified and traced the observed tubular structures by TEM throughout serial sections of Salmonella WT-infected cells ( about 100 sections per cell ) , indicating also a complex 3D network of SIT . Thus , quantification of the frequency of SIT phenotypes in ultrathin sections by TEM was not feasible given the complexity , size and 3D organization of SIT . However , inspection of TEM ultrathin sections allows an approximation of the frequency of SIT phenotypes . Sections of HeLa cells ( N = 124 cells ) infected with Salmonella WT for 8–12 h revealed that 13 . 7% of cells displayed type 1 SIT , 74 . 2% displayed type 2 SIT , and 6 . 5% of cells showed both SIT types . Measurements of diameters revealed for type 1 SIT average diameters for single membrane tubules of 120±46 nm ( N = 7 cells with 5 measurements per SIT ) and for type 2 SIT , the average diameter of double membrane tubules was 221±65 nm ( N = 30 cells with 5 measurements per SIT , Figure S2A ) . The mean distance between inner and outer membrane of the tubule was 31±12 nm . We observed that the diameter of type 2 SIT was rather constant over the segment present in one ultrathin section . Examples of longitudinal sections and cross sections through type 2 SIT are shown in Figure S2 . The closer examination of type 2 SIT by conventional TEM indicated that the inner lumen of the double membrane tubules contains numerous ribosomes and occasionally also filamentous structures . Filaments of smaller and larger diameter were detected with a distinct morphological similarity to actin filaments and microtubules , respectively ( Figure 2 ) . We determined the diameter of the larger filaments as 22 . 05±2 . 29 nm ( N = 15 ) which is consistent with the diameter of microtubules ( 22–24 nm ) ( Figure 2A–D ) [17] . Microtubules were detected only in a subpopulation of sections through type 2 SIT , suggesting that their presence or absence may be related to a certain developmental stage of the tubules . The thinner filaments had diameters of less than 5 nm , but a precise measurement was not possible due to limits of resolution in 40–60 nm sections ( Figure 2E ) . Based on the filament thickness and morphology in accordance with previous observations [17] , we propose that the thinner filaments are F-actin microfilaments . Immuno-gold labeling of cytoskeleton in TEM sections was not successful , probably due to insufficient preservation . Thus , our conclusion of F-actin and microtubules within SIT so far is based on the typical ultrastructural morphology . For an unambiguous determination of the composition of the SIF lumen , our future work will make use of genetically encoded tags [18] . Co-localization of SIF and LNT with microtubules [9] , [12] and F-actin ( DC , KG , MH , unpublished data ) was already observed , but light microscopy cannot reveal if cytoskeletal filaments are on the outside of SIT , or within the lumen . Examination of various ultrathin sections of HeLa cells ( N = 124 cells ) 8–12 h after infection with Salmonella WT revealed that 27 . 2% or 4 . 3% of type 2 SIT contain microtubules or F-actin filaments , respectively . The density of cytoskeletal filaments inside type 2 SIT was variable , but some compartments appeared filled with a large amount of filaments ( Figure 2Ai ) . The ultrastructure-based localization of cytoskeletal filaments both outside [9] and inside type 2 SIT ( this paper ) clearly emphasizes their prominent role in the biogenesis of SIT . These findings suggest that the interior space of type 2 SIT is either continuous with the host cell cytosol or represents a partial volume of cytosol enwrapped by double membranes during formation of the tubule . In order to investigate the continuity between the lumen of SIT and endosomes , we probed the intracellular environment of Salmonella by pulse-chase experiments in combination with live cell imaging . Our previous analyses demonstrated that BSA-Rhodamine-conjugated gold nanoparticles are a useful experimental tool that efficiently accumulate and label endosomal compartments [19] . This tracer is also suitable for CLEM applications by virtue of electron-dense gold particles of defined size , as well as the possibility to use Rhodamine for photo-conversion of diaminobenzidine ( DAB ) [20] . Additionally , in order to obtain a moderate and consistent expression level of an endosomal membrane marker , we used lentiviral transfection to generate a HeLa cell line constitutively expressing LAMP1-GFP which also localizes to SIF . This cell line was indistinguishable from the parental cell line with respect to the intracellular phenotypes of Salmonella ( Figure S3 , Figure S4 , Figure S5 ) and was therefore used in all microscopic approaches . For fluid tracer experiments , HeLa LAMP1-GFP cells were pulse-chased with BSA-Rhodamine-gold at various times prior or after the infection with Salmonella WT and analyzed by live cell imaging ( Figure 3 ) . Endocytosed BSA-Rhodamine-gold distributed rapidly within the network of SIF and all LAMP1-positive compartments were also positive for the fluid tracer . Compared to the positive control ( co-localization rate of 78 . 5±1 . 3% and Pearson's correlation coefficient of 0 . 82±0 . 01 ) , LAMP1-GFP and Gold-BSA-Rhodamine showed comparable high co-localization in either non-infected cells ( co-localization rate of 80 . 22±6 . 7% and Pearson's correlation coefficient of 0 . 85±0 . 02 ) or cells post infection ( e . g . , at 8 h p . i . , co-localization rate of 79 . 7±3 . 5% and Pearson's correlation coefficient of 0 . 74±0 . 02 ) . We also observed the fluid tracer within the SCV , decorating the bacterial cell body . The intensity of Rhodamine fluorescence increased with the duration of the pulse and consequently with the amount of the marker accumulated in the endosomal system of the cell . In conclusion , these data show that SIF lumen is accessible to endosomal content and support our observation in TEM of the endosomal-like content of type 1 SIT ( Figure 1B ) . The precise interpretation of the SIF integrity requires the analysis of Salmonella infection in living cells [9] , [21] . Therefore , we used the HeLa LAMP1-GFP cell line for live cell imaging of infected cells followed by processing for CLEM . This approach allowed us first to follow the integrity and dynamics of LAMP1-GFP-positive tubular structures in living infected cells , and then to investigate the ultrastructural features of the same tubular structures by TEM . We observed that at 8 h after infection with Salmonella WT all double membrane type 2 SIT were positive for LAMP1-GFP ( Figure 4 ) . Thus , we conclude that type 2 SIT are of late endosomal/lysosomal origin and are identical to the previously described SIF . Any similar tubular compartments were neither observed in non-infected HeLa cells ( Figure S6 ) , nor in cells infected with an ssaV-deficient strain unable to translocate SPI2-T3SS effector proteins ( data not shown ) . We next analyzed infected cells shortly after onset of SIF formation , i . e . 4–5 h p . i . ( Figure 5 , Figure S7 ) . Previous work showed that SIF are highly dynamic at this early stage and undergo extension , branching and contraction [9] , [10] . Furthermore , in living cells , we frequently observed connections between LAMP1-positive tubules of two kinds: ( i ) thinner tubules with weak fluorescence intensity often localized at the periphery of the cell and ( ii ) , centrally localized thicker tubules with strong fluorescence intensity connected to SCV ( Figure S5 ) . This phenotype was reminiscent of the previously described leading SIF ( LS ) and trailing SIF ( TS ) [9] , [10] . When observed by light microscopy LS extend further from the distal end of TS . Importantly , frequently detected gradual transitions of LS into TS suggest that they both represent one tubular structure undergoing morphological transition ( s ) . We therefore set out to analyze these morphological features of WT-infected HeLa LAMP1-GFP cells by CLEM , additionally pulse-chased with BSA-Rhodamine . Our results indicated two types of SIF morphologies . SIF representing LS in light microscopy correlated with single membrane SIT of thinner diameter and electron-dense content , similar to type 1 SIT . These were often found in continuum with double membrane SIT of extended diameter and cytoplasmic content , similar to type 2 SIT , correlating with TS in light microscopy ( Figure 5 ) . Based on these observations we propose that the dynamic conversion of leading into trailing SIF observed by light microscopy , in fact corresponds to the conversion of a single membrane tubule ( type 1 SIT ) to a double membrane tubule ( type 2 SIT ) . Consequently , the observed type 1 SIT or LS are of late endosomal/lysosomal origin and serve as precursor of type 2 SIT or TS , while both morphological structures represent LAMP1-positive SIF in Salmonella WT-infected cells . To test if the formation of double membrane SIF ( type 2 SIT ) is restricted to HeLa cells or a more general phenomenon , we analyzed Salmonella-infected RAW264 . 7 macrophages by live cell CLEM . First , a RAW264 . 7 cell line with stable expression of LAMP1-GFP was generated by lentiviral transfection and the transfection had no influence on the intracellular phenotypes of Salmonella ( data not shown ) . To achieve a flat , adherent cell morphology for microscopic studies , the macrophages were stimulated with Interferon-γ ( IFNγ ) . We first investigated non-infected RAW264 . 7 macrophages activated by IFNγ , since long tubular endosomal compartments were reported for phagocytic cells [22] . CLEM of non-infected activated RAW264 . 7 cells showed thin LAMP1-GFP-positive single membrane tubules that contain many small vesicles ( Figure S8A–E ) . These tubular structures in RAW264 . 7 cells were sometimes extending throughout the entire cell . In infected RAW264 . 7 cells the LAMP1-GFP-positive SIF were identified by virtue of their connections to the SCV . These tubules were substantially thicker and with stronger LAMP1-GFP signal compared to LAMP1-GFP-positive tubules in non-infected cells . Macrophages were infected with stationary phase Salmonella to avoid invasion-induced pyroptosis , thus the onset of SIF formation in infected RAW264 . 7 macrophages is delayed compared to infected HeLa cells [9] . Accordingly , we adjusted time points for microscopic analyses . CLEM of activated and Salmonella WT-infected RAW264 . 7 LAMP1-GFP cells revealed single membrane tubules at 8 h p . i . ( Figure S8F–J ) , comparable to type 1 SIT in HeLa cells at 4 h p . i . , as well as double membrane tubules at 12 h p . i . ( Figure 6 ) , comparable to type 2 SIT in HeLa cells at 8 h p . i . Pulse-chase experiments with BSA-Rhodamine and subsequent photo-conversion of DAB in infected RAW264 . 7 macrophages showed similar results as for HeLa cells ( data not shown ) . In summary , these data demonstrate the presence of type 2 SIT in RAW264 . 7 macrophages comparable to double membrane SIF in HeLa cells , and suggest that induction of double-membrane tubules by intracellular Salmonella is a common phenomenon . Double ( and multiple- ) membrane compartments are typically observed during the formation of autophagosomes [23] , and autophagy has been reported as a host cell factor involved in the control of intracellular Salmonella [24]–[27] . We thus tested whether SIT are related to autophagosomes . Microtubule-associated protein 1 light chain 3 ( LC3 ) , a mammalian homolog of yeast ATG8 , is a well-established marker of autophagy and its proteolytically processed form , LC3-II , associates with membranes of the phagophore , autophagosome and autolysosome [28] , [29] . HeLa cells were transfected with GFP-LC3b , infected with Salmonella WT and the endosomal system was then labeled with fluorescent fluid-phase marker ( Figure S9A ) . Epifluorescence microscopy of living cells showed that a small fraction of intracellular bacteria ( app . 5–10% ) was associated with LC3b , a finding in line with previous observations on the cytosolic presence of a subset of intracellular bacteria and the targeting of some of these bacteria by autophagy [30] . Starting at 3–4 h p . i . , GFP-LC3b-transfected HeLa cells also showed the presence of SIT , but these were not associated with LC3b at all and the distribution of the GFP-LC3b signal was clearly distinct from the tubular arrangement of the endocytosed fluid tracer ( Figure S9A ) . To further examine the possible involvement of autophagy , we inhibited ATG5 , a key regulator of all known human ATG8 homologous [reviewed in 31] . Therefore HeLa LAMP1-GFP cells were transiently transfected with a plasmid encoding mCherry-ATG5-K130R , a dominant-negative form of ATG5 [32] , [33] . Transfected cells were infected with Salmonella WT , analyzed by light microscopy and prepared for CLEM . Interestingly , CLSM revealed a reduction in SIF formation compared to non-transfected HeLa cells . Non-transfected HeLa cells infected by Salmonella WT showed 72 . 3±3 . 5% of infected cells with SIF formation . Salmonella WT-infected cells previously transfected with dominant-negative ATG5 revealed only 56 . 6±6 . 1% SIF formation ( Figure S9B ) . This observation is consistent with work of Birmingham et al . [30] , who investigated Salmonella-infected atg5−/− MEF . However , we observed no obvious changes in SIF morphology . Micrographs also clearly showed that LAMP1-GFP-positive SIF in HeLa cells transfected with mCherry-ATG5-K130R display a double membrane structure similar to type 2 SIT ( Figure S9C–F ) . Based on these observations , we rule out that biogenesis of type 2 SIT is related to formation of autophagosomes and anticipate a distinct mechanism of double membrane formation . The observation of double membrane SIF raised the question of the origin of the SIF inner lumen and of the lumen between inner and outer membrane , termed SIF outer lumen . To analyze the lumen of SIF with ultrastructural resolution , we applied pulse-chase experiments with fluid-phase markers suitable for detection by both , CLSM and EM . Although Rhodamine-gold provided intense and uniform labeling of the SIF lumen ( Figure 3 ) , the frequency of gold NP in thin sections analyzed by TEM was rather low ( <10 particles per SIF in a ultrathin section ) . As an alternative , we used Rhodamine for the UV light-induced photo-conversion of diaminobenzidine ( DAB ) [20] . The conversion ( oxidation ) of DAB leads to a precipitation of an alcohol-insoluble , granular polymer which , after further processing , becomes an electron-dense product detectable by TEM . Alternatively , we used the non-fluorescent horseradish peroxidase ( HRP ) as a fluid-phase marker catalyzing the conversion of DAB in presence of peroxide [34] . For this analysis , infected cells were first pulse-chased with either BSA-Rhodamine ( without gold NP ) ( Figure 7 ) or HRP ( Figure S10 ) to allow accumulation of the reagent within endocytic compartments including SCV and SIF , and then subjected to live cell imaging before fixation . The reaction with DAB was performed post fixation prior to sample processing for TEM . If endocytosed BSA-Rhodamine or HRP interact with SIF , we anticipated that DAB reaction product is located either in the inner lumen , the outer lumen , or both lumen of the double membrane SIT . For LAMP1-GFP-expressing HeLa cells 8 h p . i . and pulse-chase with BSA-Rhodamine , we detected the Rhodamine signal in smaller LAMP1-positive spherical vesicles , as well as inside SIF and within the SCV ( Figure 7J ) , thus showing the same distribution as fluorescent nanoparticles described in Figure 3 . However , the correlation with TEM images revealed double membrane tubules with cytosolic content inside the inner lumen ( i . e . type 2 SIT ) and electron-dense DAB product inside the whole outer lumen . Higher magnification TEM clearly showed that electron-dense DAB product localizes inside the SCV with direct contact to Salmonella and inside the outer lumen of double membrane SIF , while the inner lumen was virtually devoid of the marker ( Figure 7F ) . Comparable results were obtained using HRP . Since no fluorescence signal is present for HRP as a marker , the DAB product was aligned using bright field microscopy . CLEM analysis of infected HeLa LAMP1-GFP cells pulse-chased with HRP also revealed DAB product inside the outer lumen of double membrane SIF ( type 2 SIT ) and within the SCV at 8 h p . i . ( Figure S10 ) . Next , we analyzed the early stage of Salmonella infection to test if DAB product localizes within leading SIF ( LS ) . CLEM of HeLa cells 4 h p . i . revealed LAMP1-GFP-positive single membrane SIF ( type 1 SIT or LS ) , completely filled with the DAB product ( Figure S11 ) . In conclusion , these results indicate that after the conversion of single membrane tubules ( type 1 SIT , LS ) to double membrane tubules ( type 2 SIT , TS ) , the endocytosed content of LS eventually accumulates inside the outer lumen of TS . We have previously reported that mutant strains deficient in SPI2-T3SS effector proteins SseF or SseG induce the formation of SIF with altered morphology [35] , i . e . less intense labeling by fluid phase markers and LAMP1-GFP fluorescence , thinner appearance and higher susceptibility to fragmentation by chemical fixation with para-formaldehyde ( PFA ) . PFA-fragmented SIF in sseF or sseG-infected HeLa cells are also known as pseudo-SIF [21] . We found that tubular membrane compartments induced by the sseF or sseG strain were maintained after fixation with glutaraldehyde ( GA ) used conventionally for TEM . Compared to PFA-fixation of sseF-infected HeLa cells with 67 . 6±5 . 1% pseudo-SIF formation , fixation by GA resulted in 4 . 3±2 . 1% pseudo-SIF and 95 . 7±2 . 1% continuous SIF , as judged by light microscopy ( N = 100 cells , 3 replicates ) . However , SIF in sseF-infected cells appeared more heterogeneous in morphology in ultrastructure and were reminiscent of LS observed at early time points of infection ( Figure 5 ) . CLEM revealed that SIF in sseF-infected HeLa LAMP1-GFP cells at 8 h p . i . were delimited by a single membrane with electron-dense content ( Figure 8 ) , like type 1 SIT . The mean diameter of SIF in sseF-infected cells was 107±18 nm ( N = 6 cells with 5 measurements per SIF ) . This diameter is less than a half of the mean diameter of the double membrane SIF ( 221±65 nm ) but in the range of the diameter of single membrane SIF ( 120±46 nm , see above ) . We next performed CLEM of sseF-infected HeLa LAMP1-GFP cells after pulse-chase labeling with BSA-Rhodamine followed by DAB photo-conversion . TEM analysis of SCV-derived membrane tubules in these cells revealed that electron-dense DAB product localizes inside the entire lumen of the single membrane SIF at 8 h p . i . ( Figure S12 ) . CLEM experiments with HRP showed a similar phenotype ( data not shown ) . In cells infected with the sseF-deficient strain complemented with WT sseF , the SIF double membrane phenotype was restored ( Figure S13 ) . An sseF strain expressing episomal sseFΔ200–205 was previously reported to be unable to induce normal SIF [36] and we found that cells infected with sseF strain expressing sseFΔ200–205 contained only thinner single membrane SIF ( Figure S14 ) . Our previous work showed that a defect in sseG phenocopies the sseF mutation . CLEM analyses of cells infected with the sseG-deficient strain showed results similar to the sseF strain , since sseG-infected cells displayed single membrane SIF with electron-dense content ( Figure S15 ) . The SPI2-T3SS effector protein SifA has a key role in virulence since a sifA-deficient strain is highly attenuated in systemic virulence and intracellular replication . The sifA strain fails to induce SIF and loses the SCV membrane during intracellular replication , thereby escaping into the cytoplasm [15] . Interestingly , Boucrot et al . [37] showed that a transient ectopic expression of SifA in epithelial cells leads to LAMP1-positive aggregations and also thin SIF-like tubular structures . We investigated the ultrastructure of SifA-induced SIF-like tubules in non-infected HeLa cells co-transfected with LAMP1-mCherry and GFP-SifA by CLSM and CLEM . CLSM indicated that 9 . 3±2 . 5% of co-transfected HeLa cells showed a SIF-like phenotype ( Figure S16A ) , compared to 15±5% reported before [37] . Micrographs of the observed thin LAMP1-mCherry and GFP-SifA-positive SIF-like structures clearly show single membrane tubules ( Figure S16B–E ) , comparable to type 1 SIT , but much thinner ( 41±10 nm ) . Thus , SifA appears to be sufficient for the formation of LAMP1-positive single membrane tubules in uninfected HeLa cells , but is not sufficient to induce double membrane tubules . In summary , our results show that the effector proteins SseF and SseG are not required for formation of single membrane SIF ( type 1 SIT ) , but are essential for biogenesis of double membrane SIF ( type 2 SIT ) , while ectopic expression of SifA is sufficient to induce single membrane tubules . Finally , we addressed the complexity of SCV and SIT membrane organization by ET to reveal the ultrastructure of these compartments in 3D . HeLa cells were infected with WT Salmonella , fixed 10 h p . i . and 300 nm sections were used to generate tilt series from −60° to +60° . This approach identified Salmonella inside SCV with connecting SIT ( Figure 9A , B , D , E ) . The 3D reconstruction of partial cell volumes revealed both types of SIT previously observed by TEM on ultrathin sections: ( i ) single-membrane delimited SIT with a lumen containing electron-dense granules and multi-lamellar vesicular structures ( type 1 SIT ) ( Figure 9A , Movie S2 , Figure 9B , Movie S3 , Figure 9C , Figure 9D , Movie S5 , Movie S6 ) , and ( ii ) double-membrane delimited SIT with a hollow lumen containing cytoplasmic components ( type 2 SIT ) ( Figure 9C , Movie S4 , Figure 9E , Movie S7 , Movie S8 , Figure 9F , Movie S9 , Movie S10 ) . In addition to the complexity of membrane folding within lumen of SIF , tomograms also revealed structural continuity between distant portions of SIF which otherwise appeared as separate structures on ultrathin sections . Interestingly , we also found branched SCV-derived tubular membrane structures that appeared as intertwined type 1 and type 2 SIT ( Figure 9C , Movie S4 ) . This once more suggests that the two structurally distinct membrane tubules may represent different developmental stages , a process we described earlier as leading-to-trailing SIF phenotype during dynamic membrane remodeling in living cells [9] , and in this study as a possible conversion of single membrane tubules ( type 1 SIT , LS ) to double membrane tubules ( type 2 SIT , TS ) . Closer inspection of membrane arrangement within type 2 SIT emerging from the SCV revealed that both , inner and outer membrane , are in continuum and that Salmonella , together with other endosomal content resides between these two membranes ( Figure 10A–C , Movie S11 , Movie S12 ) . Volume rendering of the distal end of the type 2 SIT disclosed that the inner membrane can enwrap portions of cytoplasm including electron-dense ribosomes . The tip of this SIT was captured within the data set of the tilt series and the 3D-rendering indicates that the tip of the captured SIT is closed ( Figure 10D–F , Movie S13 , Movie S14 ) . These findings further support our hypothesis that the inner lumen of type 2 SIT , at a certain stage of SIT development , is in continuum with the cytoplasm before it is compartmentalized , while the space between the inner and outer membrane of these structures is the bona fide luminal space of SCV/SIT containing endocytosed cargo and bacteria .
To our knowledge , this is the first systematic ultrastructural analysis of the intracellular environment of Salmonella in host cells and the fine structure of host cell compartments modified by activities of intracellular Salmonella . Our approach used combinations of live cell imaging , TEM of ultrathin sections , EM tomography and cytochemistry to reveal various novel aspects of Salmonella-driven manipulation of the host cell endosomal system . We anticipate that current models for the intracellular lifestyle of Salmonella have to be revised due to the findings reported in this study . The limit of spatial resolution in light microscopy does not allow to obtain sufficient details in the organization of SCV and SIF , as for instance the distinction of single , double or multiple membranes forming these compartments . Consequently , the interpretation of presence or absence of canonical markers of the endosomal maturation has to be made with care . For example , given that SIF are composed of two membranes , different forms of biogenesis could explain the presence of lysosomal glycoproteins in SIF membranes . The marker may be located i ) in the inner and outer membrane ( both membranes are derived from late endosomes or lysosomes ) , ii ) only in the outer membrane ( outer membrane derived from late endosomes or lysosomes , origin of inner membrane is different ) , or iii ) only in the inner membrane ( situation inverse to ii ) . The data shown in this communication clearly highlight the importance of ultrastructural analyses in addition to analyses by light microscopy . Besides new insight into the complexity of SCV membrane arrangement , our study revealed for the first time two morphologically distinct types of Salmonella-induced tubules ( SIT ) in Salmonella-infected epithelial cells and macrophages: ( i ) type 1 SIT , delimited by a single membrane with a lumen containing electron-dense granules and vesicles , reminiscent of the luminal content of late endosomes , endolysosomes or in part lysosomes [16] , and ( ii ) type 2 SIT , composed of two membranes with the inner lumen containing host cell cytosol and the outer lumen containing the bacteria and endocytosed material . CLEM showed that both type 1 and 2 SIT are dynamic , lgp-positive membranes , thus identical with the previously described Salmonella-induced filaments or SIF ( Figure 4 , Figure 5 , Figure S11 ) . Moreover , type 1 SIT were identified as the earlier described leading SIF ( LS ) and type 2 SIT as the appropriate trailing SIF ( TS ) ( Figure 5 ) . Thus , we suggest that dynamic conversion of leading to trailing SIF corresponds to the conversion of single membrane tubules ( type 1 SIT ) to double membrane tubules ( type 2 SIT ) and that type 1 SIT represent developmental precursors of type 2 SIT . Double membrane compartments are typical for autophagosomes . Double membrane formation has been reported during the maturation of Coronaviruses [38] , [39] , or Hepatitis C virus [40] whereby an LC3-associated , but autophagy-independent recruitment of host membrane was considered . It is well known that autophagy targets Salmonella that has been released into host cytosol due to loss of SCV integrity [25] , but we showed that double membranes of type 2 SIT are formed independently from autophagy . How can we explain the biogenesis of double membrane SIF with internal cytoskeletal elements ? Figure 11 shows our working model with key events of SIF biogenesis . After invasion of host cells ( Figure 11A ) , Salmonella within an early SCV start to translocate SPI2-T3SS effector proteins and to manipulate the host cell endocytic system ( Figure 11B ) . We propose that single membrane SIF ( type 1 SIT ) emerge by tubulation of SCV membranes along cytoskeletal filaments ( Figure 11C ) . This tubulation occurs longitudinal by fusion processes of endosomes and we and others previously reported the dynamic features of SIF in this initial phase of intracellular life of Salmonella [9] , [10] . Our model implies that membrane extension also occurs laterally , resulting in the formation of a double membrane sheath enclosing microtubules , F-actin and portions of cytosol ( Figure 11D , M ) . Since SIF cross sections and tomograms demonstrate continuity of SIF inner and outer membrane , we postulate that such a process ultimately results in membrane fusion and formation of double membrane SIF ( type 2 SIT ) ( Figure 11E , M ) . How does the transition from early , single membrane SIF ( type 1 SIT , LS ) to double membrane SIF ( type 2 SIT , TS ) occur ? In light microscopy , the dynamic formation of thin LS with low LAMP1-GFP intensity , and bolder TS was observed [9] , [10] , see Figure 11F for an example . We think this observation is in line with a zipper-like mechanism of membrane fusion depicted in Figure 11O . The conversion into double membrane SIF is incomplete in the early stage of SIF formation , thus explaining the reversion of LS to TS and back . The limited resolution of light microscopy does not allow to trace the proposed zipper-like fusion events in living cells and the molecular mechanism underlying the membrane fusion has to be revealed by future work . What is the origin of SIF membranes and luminal material ? Previous studies indicated the interaction of the SCV with early and late endosomes [6] , as well as with exocytic vesicles [7] . We demonstrated that SIF networks are in continuous interchange with endocytosed material , as observed for fluid tracers and fluorescent nanoparticles ( [19] , this study ) . The multiple interactions may result in a continuous network of SIF lumen with a very large volume connected to SCV ( Figure 11G , H ) . Simple TEM micrographs in this study revealed electron-dense granules and vesicles for single membrane SIF , indicating a luminal content comparable to late endosomes , endolysosomes or , in part , lysosomes . CLEM experiments with HRP or BSA-Rhodamine demonstrated that the single membrane SIF lumen incorporates these fluid tracers and our model implies that , upon conversion into double membrane tubules , this luminal content represents the content of the outer lumen of double membrane SIF that is also in contact to Salmonella within the SCV . We propose that Salmonella within the SCV have a continuous interchange with endocytosed material such as medium components , but remain segregated from cytosolic components of the host cells that are entrapped in the inner SIF lumen of double membrane SIF ( Figure 11H ) . This model clearly supports a role of SIF for the nutritional supply of Salmonella within the SCV . The high abundance of lgp such as LAMP1 suggests a late endosomal/lysosomal origin of SIF membranes , however , the exposure of Salmonella within the SCV to antimicrobial effectors appears to be limited [4] . Recent work by the Holden group [41] demonstrated that by action of the effector protein SifA , intracellular Salmonella actively interfere with the proper activation and delivery of lysosomal hydrolases such as Cathepsin D to the SCV . Furthermore , there is evidence for distinct routes of delivery of lysosomal membrane material including lgp ( ‘LAMP carriers’ ) and the lysosomal hydrolases to late endosomes ( ‘MPR carriers’ ) [42] . By selectively recruiting ‘LAMP carriers’ and/or avoidance of ‘MPR carriers’ , intracellular Salmonella would be able to generate extensive lgp-positive membrane compartments without exposure to lysosomal hydrolases . A further explanation could be the dilution of antimicrobial effectors in a large volume of SCV continual with volume of extensive network of SIF ( Figure 3 , [19] ) . _ENREF_16Luminal content of vesicles fusing with the SCV or SIF could be rapidly mixed with the luminal content of SIT , resulting in decreased concentrations of antimicrobial activities . Indeed , we also observed a decreased acidification of SCV if connections to SIF were present ( YZ , MH , unpublished observations ) . This model may also explain the previous controversial observation of low degree of delivery of lysosomal enzymes to the SCV [4] . We clearly detect delivery of fluid tracers ( NP , BSA-Rhodamine , HRP , etc . ) to SCV and SIF . However , rapid dilution of incoming tracers within the luminal space of the network may explain why prolonged exposures ( longer pulse times ) are required to obtain detectable signals . Which molecular mechanisms lead to fusion events resulting in double membrane SIF ? Induction of SIF is dependent on function of the SPI2-T3SS and a subset of effector proteins , predominantly SifA , SseF , SseG , SopD2 and PipB2 . Most likely , host cell-intrinsic mechanisms are manipulated by activities of SPI2-T3SS effectors . The interaction of SifA with SKIP is an excellent example [41] , [43] , [44] . SifA is essential for the induction of SIT and sifA mutant strains are defective in maintenance of the SCV ( Figure 11I , J ) . Expression of GFP-SifA in uninfected epithelial cells leads to LAMP1-positive aggregations and also thin SIF-like tubular structures [37] . Here , we showed by CLEM that these SIF-like tubular structures represent single membrane tubules ( Figure S16 ) . Therefore , we propose that the SPI2-T3SS effector SifA could be sufficient for the formation of single membrane SIF . Yet action of SifA is insufficient to induce double membrane SIF . The exact mechanism of membrane fusion and longitudinal tubulation caused by SifA has to be examined in future work . Interestingly , mutant strains lacking SseF and SseG are attenuated in intracellular replication [35] . In sseF-and sseG-infected HeLa cells only thin SIF are formed ( Figure 11K , L ) and our CLEM analysis revealed single membrane SIF for both mutants ( Figure 8 , Figure S12 , Figure S15 ) . Consequently , we propose that SseF and SseG are the key SPI2-T3SS effectors involved in the lateral tubulation of single membrane SIF and the final membrane fusion leading to double membrane SIF . For SseF and SseG , the host cell targets appear less clear , the target candidates for SseF and SseG [45] , [46] do not indicate involvement in membrane fusion . The molecular mechanism behind SseF-and SseG-mediated membrane remodeling still remains open , but our previous study [47] suggested that SseF is involved in Dynein recruitment to the SCV membrane , what is likely essential for extension of SIF from SCV along microtubules . The absence of double membrane SIF after infection with sseF- or sseG-deficient Salmonella may also be explained by a missing direct interaction of these effectors after membrane insertion . SseF is an integral protein of host endosomes [36] and a direct interaction with SseG has been demonstrated [48] . This model would imply that Salmonella , by means of the SPI2-T3SS , transfers an autonomous membrane fusion machinery into the host cell in order to induce extensive membrane aggregations resulting in double membrane compartments . Such novel function of T3SS effectors will require detailed experimental investigation . Alternative models for biogenesis of SIF are shown in Figure 11N , P , Q . An inner tubule may also be formed by fusion of invaginated membrane vesicles located within the single membrane SIF ( Figure 11N , Q ) . We observed some single membrane SIF with a multivesicular content ( Figure 1B , Figure 8A–D ) . However , we neither observed invagination processes nor intermediate phases of fusion processes of the intraluminal vesicles inside single membrane SIF . Furthermore , the model proposed in Figure 11N , Q fails to explain mechanistically the enclosure of microtubules and F-actin , and we consider this model as less likely . Alternatively , vesicles moving along microtubules together with single membrane SIF could be entrapped the same way as cytoskeleton during the double membrane formation and afterwards the inner tubule could grow due to fusion events of the vesicular content with the SIF inner membrane from the inside ( Figure 11M , P ) . Such double membrane SIF with vesicles inside the inner lumen were also observed in micrographs for early time points of Salmonella infection ( Figure S7 ) . At later time points after infection , vesicles were seldom observed inside the SIF inner lumen , indicating their disappearance , probably due to fusion with the inner membrane . It is also possible that events depicted in Figure 11O and P occur simultaneously , since both originate from lateral tubulation ( Figure 11M ) . The data reported here give rise to a number of further questions and experimental challenges regarding the intracellular lifestyle of Salmonella . The most important question is: How could intracellular Salmonella benefit from the formation of a double membrane SIF network ? We consider three consecutive scenarios: i ) By means of the SPI2-T3SS effector SifA intracellular Salmonella induce the formation of single membrane tubules extending from the SCV . This extensive tubular SIF network may generate a compartment in which incoming endocytic cargo accumulates and bactericidal lysosomal content is diluted . Such mechanism would provide nutrients for Salmonella within the SCV , but also decrease the local concentration of antimicrobial activities . For the dilution effect vesicles with various types of luminal content have to fuse to the SIF network . ii ) Transport of both , endocytic and exocytic vesicles towards SCV and SIF and eventual fusion is critical for the successful intracellular lifestyle of Salmonella . A large number of fusion events may result in excessive amounts of membrane material at the SCV/SIF and dramatically increase their surface area . Here we speculate that a shortage of solid content within the volume of SCV/SIF causes insufficient physical support to maintain their circular profile and , together with a surplus of membrane , leads to a structural collapse of these membrane compartments . This hypothesis is supported by our findings that SCV membranes often fold into double-membrane curved sheets and invaginate or extend into SIF tubules ( examples in Figure 1C , Figure 10C ) . Our model suggests how single membrane SIF ( type 1 SIT ) can extend both , longitudinally and laterally along microtubules and eventually wrap up and close to form double membrane SIF ( type 2 SIT ) . It remains to be clarified to which extent and how this membrane rearrangement is controlled by Salmonella , e . g . through the action of SPI2-T3SS effectors SseF and SseG . iii ) The formation of double membrane SIF through the entrapment of cytosol and especially the entrapment of cytoskeletal filaments into double membrane SIF will result in a stabilized tubular network . Within double membrane SIF , F-actin and microtubules are segregated from exchange with cytosolic components . The previously observed phenotype of microtubule bundling in Salmonella-infected cells may be explained by accumulation of microtubules inside double membrane SIF [14] . Furthermore , the accumulation of cytoskeleton within double membrane SIF could explain the previous observation of decreasing dynamics of SIF at later time points after infection [9] , [10] . The stabilized SIF network would allow the SCV to maintain the juxtanuclear , Golgi-associated position that is essential for efficient intracellular proliferation [5] , [47] , [ reviewed in 49] . A stabilized SIF network would ensure the supply of Salmonella within the SCV with endocytosed nutrients and probably nutrients delivered by fusion events of distinct vesicles . Finally , the morphology and function of other SIT like SIST and LNT remains an open question . In our study performed on HeLa or RAW264 . 7 cells at early ( 4–5 h ) or late ( 8–12 h ) time points p . i . , all tubular structures connected to SCV observed were LAMP1-GFP-positive . For SIST , formation at very late time points after infection of HeLa cells was described [11] , thus our analyses most likely did not cover SIST formation . The highest number of LNT was reported for the infection of HeLa cells with a sifA sopD2 double mutant strain [12] , which was not analyzed here . Future work has to reveal the details of the complex tubular membrane compartments induced by intracellular Salmonella .
Salmonella enterica serovar Typhimurium strain NCTC12023 was used as wild-type strain and isogenic mutant strains used in this study are listed in Table 2 [15] , [35] , [50] . For complementation , the sseF mutant strain was used harboring PsseA sscB sseF::HA or PsseA sscB sseFΔ200–205::HA [36] . If required for detection in live cell imaging , the strains harbored pFPV25 . 1 [51] , pFPV-mCherry/2 [10] or pETcoco1-PrpsM-mCherry2 [52] for constitutive expression of GFP or mCherry , respectively . Bacteria were grown in LB broth at 37°C with aeration . For all experiments the non-polarized epithelial cell line HeLa ( American Type Culture Collection , ATCC no . CCL-2 ) was used . HeLa cells were cultured in Dulbecco's modified Eagle's medium ( DMEM ) containing 4 . 5 g×l−1 glucose , 4 mM stable glutamine and sodium pyruvate ( Biochrom ) and supplemented with 10% inactivated fetal calf serum ( iFCS ) ( Sigma-Aldrich ) at 37°C in an atmosphere containing 5% CO2 and 90% humidity . The Lentivirus-transfected stable HeLa cell line expressing LAMP1-GFP was cultured under same conditions . The murine macrophage-like cell line RAW264 . 7 ( ATCC no . TIB-71 ) stably transfected with LAMP1-GFP via Lentivirus transfection was cultured in DMEM containing 4 . 5 g×l−1 glucose and 4 mM stable glutamine supplemented with 6% iFCS at 37°C in an atmosphere containing 5% CO2 and 90% humidity . For activation of RAW264 . 7 cells 7 . 5 ng×ml−1 IFNγ ( BD Heidelberg ) was added to the cell culture medium 24 h before infection . Prior to infection the cells were provided with fresh medium without IFNγ . HeLa or HeLa LAMP1-GFP cells were cultured for one day in various culture vessels depending on the experimental setup and transfected with FUGENE HD reagent ( Promega ) according to manufacturer's instruction . In brief , 0 . 5–2 µg of plasmid DNA were solved in 25–100 µl DMEM without iFCS and mixed with 1–4 µl FUGENE reagent ( ratio of 1∶2 for DNA to FUGENE ) . After 10 min incubation at room temperature ( RT ) the transfection mix was added to the cells in DMEM with 10% iFCS for at least 18 h . Before infection the cells were provided with fresh medium without transfection mix . For infection of HeLa or HeLa LAMP1-GFP cells , Salmonella strains were grown in LB broth overnight ( ON ) , diluted 1∶31 in fresh LB and subcultured for 3 . 5 h in order to induce maximal SPI1-dependent invasion . The infection of HeLa cells was performed at different multiplicities of infection ( MOI ) for 25 min . RAW264 . 7 LAMP1-GFP cells were infected with ON cultures of Salmonella strains for 25 min . Subsequently , cells were washed thrice with PBS and incubated for 1 h with medium containing 100 µg×ml−1 gentamicin ( Applichem ) to kill non-invaded bacteria . Finally the medium was replaced by medium containing 10 µg×ml−1 gentamicin for the rest of the experiment . Fluid phase markers Gold-BSA-Rhodamine and BSA-Rhodamine were synthesized as described before [19] and applied to the cells after Salmonella infection at various time points prior to imaging . Co-localization analyses were performed using Leica LAS AF software ( Leica , Wetzlar , Germany ) . As positive control , HeLa cells simultaneously pulse-chased with Dextran-Alexa Fluor 488 ( Invitrogen ) and Dextran-Alexa Fluor 568 ( Invitrogen ) were used . The same threshold was applied to each dataset for LAMP1-GFP and Gold-BSA-Rhodamine to calculate co-localization rates and Pearson's correlation coefficient ( Imaris , Bitplane ) . For 3 , 3′-diaminobenzidine tetrahydrochloride ( DAB , Sigma ) as TEM marker , cells were pulse-chased for indicated periods of time with 400 µg×ml−1 BSA-Rhodamine or 10 mg×ml−1 horseradish peroxidase ( HRP , Type IV , Sigma ) in complete DMEM medium . After live cell imaging of a region of interest ( ROI ) by CLSM , the cells were fixed and blocked as described in ‘Sample preparation for CLEM’ . For the DAB conversion fixed cells were covered with freshly-prepared ice-cold 1 mg×ml−1 DAB in 0 . 2 M HEPES buffer . For BSA-Rhodamine , the ROI was viewed again by CLSM and DAB photo-conversion was started by irritating the ROI with blue light ( Xenon lamp , full power ) until a brown DAB polymer was visible by eye . For HRP the DAB conversion was initialized by adding H2O2 to a final concentration of 0 . 01% for 2 min in the dark . After DAB conversion in both cases , the DAB solution was removed and the cells were washed several times in HEPES buffer . The HRP-fed cells were checked by light microscopy for DAB conversion . Subsequently , the samples were further processed for TEM as described in ‘Sample preparation for CLEM’ . For live cell imaging DMEM was replaced by imaging-medium consisting of Minimal Essential Medium ( MEM ) with Earle's salts , without NaHCO3 , without L-glutamine and without phenol red ( Biochrom ) supplemented with 30 mM HEPES ( 4- ( 2-hydroxyethyl ) -1-piperazineethanesulfonic acid ) ( Sigma-Aldrich ) , pH 7 . 4 . The imaging studies were performed using the confocal laser-scanning microscope ( CLSM ) Leica SP5 equipped with an incubation chamber maintaining 37°C and humidity during live cell imaging . The software LAS AF ( Leica , Wetzlar , Germany ) was used for setting adjustment , image acquisition and image processing . At various time points post infection images were acquired with the 100× objective ( HCX PL APO CS 100× ) ( Leica , Wetzlar , Germany ) and the polychroic mirror TD 488/543/633 for the three channels GFP and mCherry/Rhodamine and DIC . A permanent LAMP1-GFP expressing HeLa cell line was generated by lentiviral transfection as described in Text S1 . We confirmed that intracellular phenotypes of Salmonella were identical to those observed in transiently LAMP1-GFP infected HeLa cells . HeLa cells ( 2 . 5×104 ) were seeded in eight well chamber slides ( Ibidi ) , allowed to adhere overnight , and subsequently transfected with plasmid GFP-LC3b ( Addgene , pBABEpuro GFP-LC3 ) . The next day , cells were infected with WT Salmonella at an MOI of 100 . Gold- Rhodamine nanoparticles were pulsed at 3 h p . i . for 1 h in order to indicate the Salmonella-induced tubular structures . Living cells were imaged at various time points post infection to track the co-localization of GFP-LC3b with Salmonella or Salmonella-induced tubular structures . As control for induction of autophagy by starvation , transfected HeLa cells were washed thrice with PBS and incubated in PBS at 37°C for 1 h . HeLa LAMP1-GFP cells ( 1×105 ) or RAW264 . 7 LAMP1-GFP cells ( 1 . 5×105 ) were seeded in a petri dish with a gridded coverslip ( MatTek , Ashland , MA ) two days prior to the infection with Salmonella . One day before infection HeLa cells were transfected , if necessary , and RAW264 . 7 cells were activated by IFNγ . After the infection with an MOI of 75 , if required , cells were pulse-chased with fluid-phase markers for indicated periods of time post infection . At indicated time points after infection , a ROI was observed by live cell imaging and cells were fixed as fast as possible directly on stage with pre-warmed 2 . 5% glutaraldehyde ( Electron Microscopy Sciences ) in buffer ( 0 . 2 M HEPES , pH 7 . 4 , 5 mM CaCl2 ) for 1 h at 37°C . After rinsing the cells several times in buffer , unreacted glutaraldehyde was blocked by 50 mM glycine in buffer for 15 min , followed by rinses in buffer . Post-fixation was performed with 2% osmium tetroxide ( Electron Microscopy Sciences ) in buffer containing 1 . 5% potassium ferricyanide ( Sigma ) and 0 . 1% ruthenium red ( Applichem ) for 1 h at 4°C in the dark . After several washing steps the cells were dehydrated in a cold graded ethanol series and finally one rinse in anhydrous ethanol and two rinses in anhydrous acetone at room temperature . The gridded coverslip was removed from the Petri dish and cells were infiltrated and flat-embedded in mixes of acetone and EPON812 ( Serva ) . During the removal of the gridded coverslip from the polymerized EPON block the engraved coordinates were transferred to the EPON surface and allowed trimming around the ROI . Serial 70 nm sections were cut with an ultramicrotome ( Leica EM UC6 ) and collected on formvar-coated EM copper grids . After staining with uranyl acetate and lead citrate , cells were observed with TEM ( Zeiss EFTEM 902 A ) , operated at 80 kV and equipped with a 2K wide-angle slow-scan CCD camera ( TRS , Moorenwies , Germany ) . Images were taken with the software ImageSP ( TRS image SysProg , Moorenwies , Germany ) . For image analysis , software packages LAS AF ( Leica . Wetzlar ) , ImageJ ( http://rsbweb . nih . gov/ij/ ) and Imaris ( Bitplane , Zürich ) were used . Stitching and overlay of CLSM and TEM images was done using Photoshop 5 . 5 ( Adobe ) . For superior membrane preservation , samples were processed by high pressure freezing-freeze substitution ( HPF-FS ) . Cells were grown in glass-bottom dishes ( MaTek Corp . ) on top of sapphire discs with a carbon finder-grid mask and coated with poly-L-lysine . At appropriate time post infection with MOI of 100 , the dishes were processed directly for HPF as follows: Discs were removed from media , placed between hexadecane-treated aluminum specimen carriers with a 0 . 1 mm cavity and immediately transferred to LEICA HP010 holder and processed by HPF . Discs were then transferred from liquid nitrogen to the cryovials containing freeze-substitution medium ( 1% OsO4 and 0 . 2% uranyl acetate in Acetone ) and then to FS device ( LEICA AFS1 ) for FS processing with final embedding in EPON . Thin-sections were post-stained with 2% lead citrate in water and examined using a Morgagni electron microscope ( FEI ) . Thick sections ( 300 nm ) of HPF-FS samples were placed on a slot grid covered with a formvar film and decorated with 10 nm protein-A gold particles on both sides for image alignment . Grids were placed in a high-tilt holder ( Fischione Model 2020 ) and dual-axis ET were carried out using a Tecnai F30 ( FEI ) electron microscope ( operated at 300 kV ) equipped with a field emission gun and a 4084×4084 pixels CCD camera ( Eagle , FEI ) . Tomographic tilt ranges were typically from +60° to −60° with an angular increment of 1° pixel size ranging from 0 . 5 to 1 nm . Alignments , 3D reconstructions , and merging of serial tomograms were done with IMOD software suite [53] . The volume segmentations were performed with the Amira 4 . 1 visualization package ( Visage Imaging , Berlin , Germany ) .
|
Salmonella enterica is an invasive , facultative intracellular bacterial pathogen . Within mammalian host cells , Salmonella inhabits a specialized membrane-bound compartment , the Salmonella-containing vacuole ( SCV ) , redirects host cell vesicular transport and massively remodels the endosomal system . These activities depend on the function of a type III secretion system and its translocated effector proteins . Intracellular Salmonella induces several types of tubular compartments termed Salmonella-induced tubules ( SIT ) , but the biogenesis and biological function of SIT is only partially understood . Our work combines live cell imaging with correlative light and electron microscopy to provide ultrastructural insight into SIT . We report that SIT emerge as single membrane tubules that convert into double membrane tubules entrapping cytosol and cytoskeletal filaments . Labeling of the endosomal compartment and cytochemistry demonstrate that the space between inner and outer SIT membrane is composed of internalized material and connected to Salmonella within the SCV . The effector proteins SseF and SseG translocated by intracellular Salmonella are essential for the conversion of single to double membrane SIT . These findings challenge current models for the intracellular lifestyle of Salmonella and the composition of its intracellular habitat .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"bacteriology",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"host-pathogen",
"interactions",
"medical",
"microbiology",
"microbial",
"pathogens",
"biology",
"and",
"life",
"sciences",
"microbiology",
"pathogenesis"
] |
2014
|
Reorganization of the Endosomal System in Salmonella-Infected Cells: The Ultrastructure of Salmonella-Induced Tubular Compartments
|
The circadian clock is a set of regulatory steps that oscillate with a period of approximately 24 hours influencing many biological processes . These oscillations are robust to external stresses , and in the case of genotoxic stress ( i . e . DNA damage ) , the circadian clock responds through phase shifting with primarily phase advancements . The effect of DNA damage on the circadian clock and the mechanism through which this effect operates remains to be thoroughly investigated . Here we build an in silico model to examine damage-induced circadian phase shifts by investigating a possible mechanism linking circadian rhythms to metabolism . The proposed model involves two DNA damage response proteins , SIRT1 and PARP1 , that are each consumers of nicotinamide adenine dinucleotide ( NAD ) , a metabolite involved in oxidation-reduction reactions and in ATP synthesis . This model builds on two key findings: 1 ) that SIRT1 ( a protein deacetylase ) is involved in both the positive ( i . e . transcriptional activation ) and negative ( i . e . transcriptional repression ) arms of the circadian regulation and 2 ) that PARP1 is a major consumer of NAD during the DNA damage response . In our simulations , we observe that increased PARP1 activity may be able to trigger SIRT1-induced circadian phase advancements by decreasing SIRT1 activity through competition for NAD supplies . We show how this competitive inhibition may operate through protein acetylation in conjunction with phosphorylation , consistent with reported observations . These findings suggest a possible mechanism through which multiple perturbations , each dominant during different points of the circadian cycle , may result in the phase advancement of the circadian clock seen during DNA damage .
Circadian rhythms are biological oscillations occurring with an approximately 24-hour period affecting many processes . In mammals , these oscillations are centrally controlled in the brain by the suprachiasmatic nuclei ( SCN ) . The SCN synchronizes the peripheral circadian clocks that exist in nearly every cell . Disruption of the circadian clock can lead to higher incidence of certain forms of cancer , and circadian timing can affect both the tolerability and efficacy of cancer therapeutics though the underlying mechanisms for these effects are still not well-understood [1 , 2] . Mutations of core circadian components in tumors can affect several properties of circadian oscillations , including: changes in amplitude , phase shifts , and period [3] . Investigation into the molecular components of the circadian clock has revealed much about how these biological rhythms function . In mammals , the core of the circadian clock is coordinated by four components that operate in a transcription-translation feedback loop . The positive ( i . e . transcriptional activation ) arm of the circadian clock involves a transactivating heterodimer complex composed of Brain and Muscle Arnt-Like protein-1 ( BMAL1 ) and Circadian Locomotor Output Cycles Kaput ( CLOCK ) that induces the transcription of many genes; the current model and its simplifications are described in the Model section . Gene expression microarray analyses have shown that as much as 10% of an organism's transcriptome could be under circadian influence with expression exhibiting circadian oscillations; this value depends on experimental conditions and the tissue of origin [4] . The BMAL1/CLOCK transactivating complex operates on E-box regions of gene promoters . Additionally , CLOCK is an acetyltransferase involved in chromatin remodeling that is required for the proper operation of the circadian clock [5] . The negative ( i . e . transcriptional repression ) arm of the circadian clock involves the Cryptochrome ( CRY1 and CRY2 ) and Period ( PER1 , PER2 , and PER3 ) genes that act as inhibitors of the BMAL1/CLOCK transcription factor complex . CRY/PER heterodimers in the nucleus suppress CLOCK/BMAL1-mediated transcription completing the feedback loop , which then repeats to result in increased transcriptional activity as the levels of CRY/PER complex diminish [6] . The degradation of CRY/PER levels is partially triggered by CKI-epsilon ( Casein Kinase I-epsilon ) mediated phosphorylation , which marks the PER proteins for proteasomal degradation [7] . Period ( PER ) proteins have been shown to interact with ATM and CHK2 , two key proteins involved in DNA damage response; the Neurospora ortholog for CHK2 , PRD-4 , has been shown to promote the phosphorylation of the PER protein analogue in Neurospora , FRQ [8 , 9] . Several studies show the existence of interplay between the pathways regulating circadian rhythms and those regulating DNA damage response . For example , disruptions to the core components can lead to alterations in DNA damage response pathways through altered expression patterns [10] . The reverse has also been observed , in that circadian oscillations can be reset by genotoxic stress [11 , 12] . Rat-1 fibroblasts were subjected to pulses of ionizing radiation resulting primarily in phase advancements of circadian oscillations [11] . In contrast , other forms of perturbation produce phase advancements and delays , such as in the case of pharmacological perturbation with dexamethasone [13] . Dexamethasone is a glucocorticoid agonist capable of resetting the circadian phase of asynchronous cells by triggering the expression of PER1 [14] . The molecular basis for the regulation of the circadian clock in the presence of genotoxic stress continues to be explored [11 , 12] . As our understanding of circadian regulation expands , so do the interconnections with other biological processes . Several recent studies have shown the circadian clock to be regulated by proteins , such as SIRT1 , involved with DNA damage response and cellular metabolic state through their consumption of nicotinamide adenine dinucleotide ( NAD ) [15 , 16] . NAD participates in many oxidation-reduction reactions and functions , including ATP production [17] . Supplies of NAD are under circadian regulation due to circadian oscillation of nicotinamide phosphoribosyltransferase ( NAMPT ) that controls a rate-limiting step in the salvage of NAD [16 , 18] . In its DNA damage response role , NAD is involved in cell fate decisions through its utilization by PARP1 and SIRT1 , as recently reviewed [19] . PARP1 is an ADP-ribosyltransferase where the ADP-ribosyl moieties are obtained from the cleavage of NAD . PARP1 is activated in the presence of DNA strand breaks ( its activity can increase 10–500 fold ) and helps to recruit DNA repair proteins [20 , 21] . At severe levels of DNA damage , energy depletion due to loss of NAD and ATP may trigger necrosis [20 , 22] . SIRT1 is an NAD-dependent protein deacetylase that can help inhibit transcription through histone deacetylation . The acetylation of histones leads to the activation of gene expression by inducing a relaxed chromatin confirmation at gene promoters , which permits the access of DNA transcription proteins [15] . Histone acetylation is counter-balanced through deacetylation causing a condensed chromatin state and transcriptional silencing . SIRT1 is involved in DNA damage responses through interaction with several key proteins , such as p53 , where the deacetylation of p53 inhibits p53 and promotes cell survival [23] . More recently , SIRT1 has been implicated in the regulation of the circadian clock in several ways . First , SIRT1 destabilizes the interaction between CRY and BMAL1 through the deacetylation of BMAL1; the deacetylation of BMAL1 is counter-balanced at the same position through the acetyltransferase activity of CLOCK [15 , 24] . Second , SIRT1 has been shown to deacetylate PER destabilizing the protein and promoting its degradation , which may promote transcription during circadian oscillations [25] . Finally , SIRT1 is recruited to promoters of PER2 and NAMPT and is involved in the chromatin remodeling of the vicinity of each of the two promoters [16] . The circadian clock has been the subject of several mathematical models that have helped in our understanding of the molecular mechanisms underlying regulation of the circadian clock [26 , 27] . Our understanding of the NAD circadian regulation dynamics and the molecular mechanism regulating the phase resetting response of the circadian clock upon exposure to genotoxic stress remains incomplete; given the interactions mentioned above , it is possible that NAD utilization may be involved . We have developed an ordinary differential equation ( ODE ) model that includes the role of NAD in the regulation of SIRT1 . The current study explores the potential role of NAD depletion in phase resetting of the circadian clock through the activities of the NAD consumers , SIRT1 and PARP1 . Also , we examine the effect of multiple perturbations on the circadian cycle and how these perturbations may account for this observed behavior of the primarily phase advancement resetting of the circadian clock seen during DNA damage .
We have developed a simple model ( referred to here as the current model ) representing the circadian clock of mammals , which extends a previous model developed by Hong et al . ( referred to here as the Hong 2009 model ) [28] . As in the Hong 2009 model , we only consider the activity of the PER protein and have subsumed the paralogs of the CRY ( Cryptochrome ) and PER ( Period ) genes into a single species CP in order to simplify the model . Within the model , PER can exist as a monomer , dimer , or in complex with BMAL1/CLOCK . BMAL1/CLOCK is inactivated when it exists in a complex with the PER dimer . Each form of PER contains a phosphorylation term that simulates the phosphorylation that triggers proteasomal degradation [7] . Fig . 1 shows a wiring diagram for the current model using the Molecular Interaction Map ( MIM ) notation for bioregulatory networks and drawn using PathVisio-MIM [29 , 30] . Each interaction is labelled and described in Table 1; these descriptions are used to label the reactions in the SBML model file . The original form of CRY/PER mRNA transcription in the Hong 2009 model used a Hill function , but this is zeroed out in the current model using kms ( kms = 0 ) in Equation 1 ( below ) . We extend the Hong 2009 model to account for the effects of acetylation on transcription for both PER and NAMPT by using Equation 1 through Equation 15 from Smolen et al . re-worked for the system described in the current model; these equations become the method for describing transcription rather than usage of a Hill function [31] . Deacetylation of histones results in chromatin compaction and decreased transcription as a result of lowered accessibility of DNA polymerase to these regions of condensed chromatin . In the case of PER , the first term of Equation 8 accounts for the fractional levels of histone acetylation . The rate of promoter acetylation is a function of acetylation regulated by the BMAL1/CLOCK ( TF ) complex through CLOCK acetyltransferase activity and inhibited by the effects PER dimer , Equation 13 . Further , it is known that CLOCK is able to acetylate histones at positions deacetylated by SIRT1 [15] . The rate of histone acetylation is regulated by the basal rate of histone deacetylation and the SIRT1 deacetylation activity simulated as a two substrate Michaelis-Menten reaction that utilizes NAD in the process; the activity of SIRT1 is discussed further below . Therefore , unlike Smolen et al . , we do not use a single , fixed deacetylation rate [31] . This is consistent with the work of Nakahata et al . , which showed that peak SIRT1 deacetylation activity coincided with the lowest acetylation levels of histone H3 [15] . This level of single histone acetylation is then used to generate an overall promoter accessibility value , Equation 9 . Lastly , this promoter accessibility value is multiplied by a maximal rate of transcription to denote the expression of PER , Equation 1 . The same mechanism is used to denote the expression of NAMPT . Neither SIRT1 expression nor protein levels are under circadian control , yet its deacetylation activity is regulated in a circadian manner [15] . Therefore , we do not consider changes to SIRT1 levels and only consider the ability of SIRT1 to utilize NAD to deacetylate three species ( PER , BMAL1/CLK , and acetylated histone ) within the model , thereby affecting circadian rhythms via separate mechanisms . First , SIRT1 deacetylates PER2 destabilizing the protein and promoting its degradation [25] . Second , acetylation of BMAL1 promotes the binding of CRY1 to BMAL1 and BMAL1 is a target of SIRT1 deacetylation [32] . Third , as a histone deacetylase SIRT1 is able to deacetylate lysine residues of histones helping to produce transcriptionally silenced chromatin that exists with a closed chromatin structure [33] . Two parameters specify the activity of SIRT1 in the model . The first parameter VSIRT1c regulates the deacetylation of PER ( either monomer , dimer , or in complex with BMAL1/CLOCK ) and the second parameter , VSIRT1d , regulates the histone deacetylation . The levels of NAD are regulated using a first-order reaction dependent on the availability of NAMPT . The model includes perturbation inputs from the Hong 2009 model , dexamethosone ( Dex ) and the CHK2 phosphorylation ( kchk2 affecting PER monomer and dimer and kchk2c affecting PER in complex with BMAL1/CLOCK ) . All simulations were conducted using MATLAB ( http://www . mathworks . com ) . Copies of our model as a Systems Biology Markup Language ( SBML ) generated using COPASI ( http://www . copasi . org ) are published as supporting information on the PLOS website ( S1 File ) . The model is a system of 11 equations described above and shown below . Equation 12 and Equation 13 denote the rate promoter acetylation for the NAMPT and PER promoters , respectively . Equation 14 denotes the level of inactive complex , while Equation 15 is the total amount of PER that exists in the system . Equation 1: CRY/PER mRNA Equation 2: BMAL1/CLOCK complex Equation 3: CRY/PER protein monomer Equation 4: CRY/PER protein dimer Equation 5: NAMPT mRNA Equation 6: NAMPT protein Equation 7: Single histone acetylation ( NAMPT promoter ) Equation 8: Single histone acetylation ( CRY/PER promoter ) Equation 9: DNA accessibility value ( CRY/PER promoter ) Equation 10: DNA accessibility value ( NAMPT promoter ) Equation 11: NAD Equation 12: Rate of NP promoter acetylation Equation 13: Rate of CP promoter acetylation Equation 14: Inactive complex ( BMAL1/CLOCK and PER dimer ) Equation 15: Total amount of PER Kinetic parameters used for the current model are described in Table 2; the table also lists the parameter values necessary to reconstitute the Hong 2009 model . Rate constants were based on previously published circadian models [28 , 31] . Kinetic parameters unique to the current model were then optimized to generate oscillations in the current work . Rate constants are in units of h-1 . The resulting amplitudes have similar orders of magnitude to the original Hong 2009 model . Initial values used in the current model are described in Table 3; initial values to reconstitute the Hong 2009 model are also listed in Table 3 . The concentrations of proteins and metabolites are in arbitrary units ( AU ) because these are currently not known for many circadian clock proteins . Damage was simulated by altering levels of kparp and kchk2 as described in the Results section using the parameters in Table 4 . The period was calculated by finding the mean of the simulated results and then finding the time points where a selected time point was greater than the mean and the subsequent time point was less than the mean . For each of the selected time points , the previous time point was subtracted to produce the period value . The resulting values were then averaged for the final period value; a requirement was imposed that at least seven oscillations were necessary to produce this value otherwise an error value , negative one , was produced . The period was calculated using the time series for the CRY/PER ( CP ) protein . Differences in phase were calculated after 19 days ( 19 circadian oscillations ) between the unperturbed and perturbed systems . The phase shift ( advancement or delay ) was calculated using the difference between oscillation peaks for the two systems . Treatments were induced at each circadian hour , and the phase response curve was calculated using the time series data for the CRY/PER ( CP ) protein .
Fig . 2A illustrates the oscillatory behavior simulated by the model for the core circadian components using the current parameter set outlined in Table 2 . The system oscillates with an autonomous period of 23 . 8 hours , which is well within the range seen in circadian oscillations of mice [34] . The current model simulates a free-running circadian clock without external stimuli or cues ( zeitgebers ) periodically synchronizing the clock and this is the state in which current model results are described . The model can account for entrainment by varying the Dex as a square-wave increasing the value of Dex to 0 . 125 for 12 hours and decreasing it to 0 for another 12 hours . Circadian models , such as the one by Leloup and Goldbeter in 2003 , make use of varying PER transcription to simulate the effect of light entrainment . Dexamethasone with its ability to trigger PER transcription therefore is a suitable substitute for entrainment by light [14 , 35] . Fig . 2B illustrates the oscillations in the histone acetylation levels for both PER and NAMPT mRNA . Histone acetylation levels peak at approximately hour 22 in Fig . 2B , helping the relaxation of DNA to permit transcription to be initiated . The peak levels of PER and NAMPT mRNA are then reached after a lag of ~6 hours . Experimentally , peaks in the acetylation levels of histones H3 and H4 have been observed 4 and 8 hours in advance of the PER1 and PER2 mRNA peaks [36] . Acetylated histone and NAD levels oscillate in antiphase , as seen when comparing Fig . 2B and Fig . 2C . In the context of the model , this is due to a feedback mechanism involving NAD production and SIRT1 activity where NAD levels ( NAD ) rise to their peak measured levels ~5 hours after the peak levels of NAMPT mRNA ( N ) . This is the time when SIRT1 activity is at its maximum and acetylated histone levels decline to their minimum ~5 hours later . NAD levels oscillate by approximately 40% during each circadian cycle , as shown in Fig . 2C , in response to oscillations in NAMPT protein levels; NAD levels oscillate in phase with NAMPT levels . Similar changes in oscillations levels have been seen experimentally [16 , 18] . This decline in the NAD levels is a product of several SIRT1 deacetylation processes captured by the current model , as well as the basal degradation of NAD levels via processes external to the model . Fig . 3 shows that the current model retains the phase dynamics present in the Hong 2009 model that are critical in the modeling of circadian systems . There is a lag of ~3 hours between the peak of PER mRNA and the peak in PER monomer levels; this is similar to experimental results seen for mammalian circadian rhythms [6] . Peaks in the PER monomer levels then proceed prior to the peak in the PER dimer levels several hours later , and peak levels in the PER dimer are then antiphase to the levels of the transcription factor BMAL1/CLOCK . The Hong 2009 model possesses an autocatalytic positive feedback loop involving PER that is necessary to sustain oscillations [28] . This feedback loop requires that differential stabilities exist between PER monomer and PER in complexes , either the dimeric form alone or in the dimeric form complexed with BMAL1/CLOCK . This mechanism arises from experimental evidence in the Drosophila circadian clock by Kloss et al . wherein PER complexes were shown to be less susceptible to degradation [37] . The current model exhibits the same autocatalytic requirement with a smaller value for the degradation of the PER dimer ( kcp2d ) than for the degradation of the monomeric PER form ( kcpd ) by two magnitudes of order . In contrast to the Hong 2009 model which possesses values for the two parameters ( kcpd and kcp2d ) with a smaller difference , in the current model we assume the activity of SIRT1 ( VSIRT1c ) in the destabilization of PER in either monomeric or in complexes to be equivalent , which means that kcpd2d accounts for a smaller portion of the degradation of the PER dimer . Due to the importance of circadian rhythms in the synchronization of biological processes , circadian oscillations must be robust to minor perturbations and must stably oscillate in the presence of varied parameters resulting from individual variation . The results of a study of the circadian rhythms of 72 mice from 12 inbred mouse strains showed this robustness of circadian oscillations [34] . Across the combined strains , the period mean was 23 . 53 ( range 22 . 94 to 23 . 93 ) hours . We expected a similar robustness in the current model and tested the sensitivity of the model to perturbations of each parameter individually using a method that has been used in computational studies previously [31 , 38] . Model robustness was tested by increasing and decreasing parameter values individually by 20% and plotting the resulting amplitude changes in PER mRNA ( often used as an experimental proxy in PER luciferase experiments ) against the oscillation periods . The results of this testing are shown in Fig . 4 , and this testing suggests that the model is robust to parameter perturbations . Out of the perturbations tested , none of the parameter sets resulted in periods that deviated from 24 hours by more than 3 hours . A majority of the parameter perturbations clustered near the current model parameter values from Table 2 ( this is shown in red in Fig . 4 ) with only slight increases or decreases of the period and amplitude . Stress input variables: Dex , kchk2 , kchk2c , and kPARP are set to 0 in the current model parameter set , and therefore , they are not expected to , nor did they , have any effect during the sensitivity testing . Three parameters resulted in periods less than 23 hours and PER mRNA amplitudes less than 0 . 4 AU . All three of these parameters affected PER , either mRNA or protein , levels . Decreases of 20% to PER protein synthesis rate ( kcps ) and PER mRNA synthesis rate ( VM ) , resulted in this behavior , while an increase of 20% to the PER mRNA degradation ( kmd ) also resulted in a similar behavior with a decreased amplitude and period . A 20% decrease in PER mRNA degradation resulted in the opposite behavior with both an increase in amplitude and a period; as shown in Fig . 4 , this is the only parameter that resulted in periods greater than 26 hours . Next , phase response curves ( PRCs ) were generated using pulses of dexamethasone ( Dex ) which trigger the transcription of PER to show that the current model is able to produce both Type 1 and Type 0 PRCs as with the Hong 2009 model . Phase response curves illustrate the relationship between the timing of a perturbation and the effect of the perturbation on a circadian oscillation in the form of a phase shift [39] . There are two types of PRCs , Type 1 and Type 0 . The resulting PRC is often dependent on the strength of the perturbation with Type 1 PRCs occurring at lower perturbations than Type 0 . As shown in Fig . 5B , low values of Dex ( Dex = 0 . 15 ) result in a Type 1 PRC ( shown in Fig . 5A ) whereby there is a continuous transition between phase advancements ( positive values on the PRC ) and delays ( negative values ) in response to the dexamethasone stimulus . At high values of Dex ( Dex = 20 ) , a Type 0 PRC is produced with a discontinuity between the phase advancements and delays of the system . We next examined the roles of NAD biosynthesis and SIRT1 activity in the current model given the multiple deacetylation interactions in the model utilizing NAD via SIRT1 activity . Current literature contains a contradiction as to the effect of SIRT1 inhibition on PER2 mRNA levels . Nakahata et al . have shown that the inhibition of SIRT1 activity leads to an increased maximal level of PER2 mRNA [15 , 16] . Asher et al . have shown the reverse—that an inhibition SIRT1 activity results in a decrease in PER2 mRNA levels [25] . Both increases and decreases may be theoretically possible via SIRT1 activity , since SIRT1 can affect the positive ( i . e . transcriptional activation ) and negative ( i . e . transcriptional repression ) regulation arms of circadian rhythms . We began to address this apparent contradiction in our simulations by decreasing the rate of NAD biosynthesis . As shown in Fig . 6 , this result agreed with the Asher et al . experimental results by qualitatively producing a decrease of approximately 12% in CRY/PER mRNA ( M ) levels following a decrease of 75% from the original VNADc parameter value [25] . We then further investigated this behavior by decreasing SIRT1 activity by reducing VSIRT1c ( non-histone deacetylation activity ) and VSIRT1d ( histone deacetylation activity ) to determine if either of these parameters would result an increase of CRY/PER mRNA levels . Similar to Asher et al . , a decrease in VSIRT1c results in CRY/PER mRNA level decreases , as shown in Fig . 7 . Similar to Nakahata et al . , a decrease in VSIRT1d results in an increase of CRY/PER mRNA levels , as shown in Fig . 8 , due to a smaller repressive effect by SIRT1 on transcription [15 , 16] . Fig . 9 shows the percentage change in maximal levels of CRY/PER mRNA levels over the parameter values that exhibit stable oscillations for the SIRT1-related parameter values . We find these results to be robust by reducing each of these three parameters to 30% of the original value ( this is near the lower limit where parameter decreases for VSIRT1c , VSIRT1d , and VNADc continue to result in oscillations ) and conducting a sensitivity analysis as described above . Sensitivity analysis for each of these parameters shows increases in the maximal levels of CRY/PER mRNA consistently for VSIRT1d and decreases for both VSIRT1c and VNADc . While both VSIRT1c and VSIRT1d parameters contribute to the overall state of the system , the parameters VSIRT1c and VSIRT1d have opposing effects and parameter VSIRT1c has a stronger overall effect within the model . Next , we examined the effect of DNA damage on circadian rhythms , which has been experimentally studied by Oklejewicz , et al . using Rat-1 fibroblasts [11] . In the current model we have examined this effect via the two possible mechanisms . First , the current model allows the examination of DNA damage as simulated by the activation of CHK2 ( kchk2 ) to phosphorylate PER monomer and dimer that triggers their degradation , and the second being sharp decreases in NAD levels on the circadian clock using changes in kPARP to simulate PARP1 activity . As a major participant in DNA damage response , PARP1 activity becomes greatly increased in response to DNA strand breaks and is recruited to the sites of DNA damage in a matter of minutes [20] . Since ionizing radiation results primarily in phase advancement , we asked whether perturbations in PARP1 , singly or in combination with CHK2 , could produce similar phase responses , and if so by what mechanism these phase advancements arise . To compare the phase responses between simulations , we use the ratio of the maximum phase advancement in a PRC to the maximum phase delay in the PRC [28] . The PRC for the Hong 2009 is described in Fig . 2 of Hong et al . [28] . For comparison , Table 4 shows these PRC ratio results for both the Hong 2009 model using the current model and re-parameterized ( using the parameters from Table 2 ) and for the current model under various parameter conditions . Additionally , in Table 4 we provide the fraction of the area under the PRC that is positive; these values are largely consistent with the ratio metric . With the re-parameterized model , we first perturb the model using the same kchk2 ( kchk2 = 0 . 2 ) from Hong et al . There is a discrepancy in values for the ratio ( 3 . 54 as originally published versus 3 . 0193 here ) , but we believe this may be a by-product of numerical analysis and we use our value as the point of comparison . Perturbing the current model using the same kchk2 ( kchk2 = 0 . 2 ) value results in a larger positive fraction of the area under the phase response curve . We next calculated the positive area fraction using only kPARP ( kPARP = 20 ) for a treatment duration of two hours . This yielded a PRC where the majority of the area was positive , similar to the one observed for the re-implemented Hong 2009 model; 0 . 6235 versus 0 . 8513 , respectively . We next wondered whether a combination of perturbations would yield a larger positive area fraction . Using the values kchk2 = 0 . 1 and kparp = 10 , we calculated a positive area fraction slightly greater than the fraction value for the CHK2 perturbation alone in the current model . This is with a CHK2 value of half the value used for the Hong 2009 re-parameterization . At kchk2 = 0 . 2 and kparp = 20 , we produce a positive area fraction that is almost completely positive . These results are robust when we conduct a sensitivity analysis at these high levels of perturbation with parameter changes of 20% . The model is less robust to changes in parameters that further decrease NAD production; this is expected given the strain on NAD levels due to PARP activity . The resulting PRC has a near bimodal appearance . Within the context of the model this effect has a direct relation on the activities of SIRT1 in the model both as an inhibitor of transcription and as a mechanism for the destabilization of PER protein . This effect of this CHK2 perturbation occurs at a circadian time of 10 hours , shown in Fig . 10 , which is during peak of PER dimer levels ( the dominant form of the repressor in the system ) , shown in Fig . 2A . This degradation allows mRNA levels of PER and NAMPT to rise in advance of the unperturbed model thereby resulting in a strong phase advancement . The delays for this CHK2-dependent PRC occur at troughs of PER dimer levels . This degradation of the PER dimer repressor at this point causes a slight increase in the maximum PER mRNA level relative to the unperturbed model in the subsequent circadian cycle resulting in the delay observed in the CHK2-dependent PRC . The CHK2-dependent PRC is in contrast to the PARP-dependent PRC , shown in Fig . 10 , at the highest value tested ( kparp = 20 ) . At this value , a Type 1 PRC is also produced , but whereas the CHK2 perturbation degrades PER dimer levels , the simulated consumption of NAD by PARP removes an inhibitory effect ( the deacetylation of PER leading to its degradation by the activity of SIRT1 ) on this repressor causing an opposite effect; the peak of the PARP-dependent PRC occurs at roughly circadian time 20 hours and its trough at circadian time 10 hours . This increase in PER dimer levels causes an inactivation of transcription by repressing the activity of BMAL1/CLOCK , which acts as both the transcription factor complex and as the histone acetylatransferase in the model . Therefore , these two perturbations , NAD depletion and PER degradation , may have different effects depending on the circadian time . The disparate effects of these two perturbations are seen in Fig . 10; advance-delay ratio and positive area fraction results are listed in Table 4 . In combinations of the two perturbations , a bimodality in the PRC emerges at larger values of the two perturbations , which is not directly seen experimentally in the observations by Oklejewicz et al . suggesting that if this is a mechanism that exists biologically , then the balance between these two forms of perturbation may be under additional regulation [11] . Yet , the phase response curves seen experimentally in response to DNA damage are undoubtedly the products of several forms of perturbation each that may have a dominant effect depending on the phase of the system during perturbation .
Here we have developed a simple model that expands on the work of both Hong et al . and Smolen et al . to produce a mathematical model that connects circadian rhythms to DNA damage response and metabolism via the regulation of chromatin remodeling [28 , 31] . The current model predicts a molecular mechanism through which multiple forms of perturbation , as a result of DNA damage , and multiple post-translational modifications can reproduce the experimentally observed phase response curve as shown in Oklejewicz et al . in Fig . 1 of that publication [11] . We began with the hypothesis that the activities of SIRT1 and PARP1 in regulating the circadian rhythm could impact on the primarily phase advancement seen in circadian oscillations during the response to genotoxic stress given their known interactions with core circadian clock components . To investigate this question , we expanded a previous model to account for the activity of SIRT1 in the regulation of transcription and circadian clock components and the activity of PARP1 during DNA damage response . The model reveals that the regulation of the circadian clock may be wired in a way that integrates multiple forms of post-translational modifications as a mechanism to respond to environmental stress; in the case of acetylation , this post-transcriptional modification is controlled using a circadian feedback mechanism through regulation of NAMPT . We examined phase response curves resulting from various conditions by using the simulated effects of CHK2 and PARP1 activity . The results of our in silico study help to confirm the potential for CHK2 involvement in producing the experimentally observed PRC in the presence of an autocatalytic positive loop regulating PER . Our model suggests that additional regulatory mechanisms may factor into the observed PRC . The expanded model shows that NAD depletion via PARP1 activity may produce a similar PRC result as experimentally observed through the removal of the SIRT1 inhibitory effect . Models with both NAD depletion and CHK2 activity reproduce the observed PRC best . This raises the possibility , that multiple perturbations may work in concert to produce the observed PRC . The current model also addresses an apparent contradiction in the literature as to the effect of SIRT1 inhibition on the levels of PER2 mRNA [15 , 16 , 25] . We showed that differential SIRT1 activity targeting specifically either histone or non-histone component deacetylation may account for this contradiction . Alternatively , this contradiction may suggest an additional mechanism that was not controlled between the two sets of experimental observations and may also be missing from the current model . A recent publication by Xydous et al . has suggested that the byproduct of the SIRT1 activity , nicotinamide , ( though not specific to its reaction and not accounted for in the current model ) may be able to affect histone methylation levels leading to alterations in gene expression in a manner that is independent of SIRT1 activity [40] . Though the inclusion of histone methylation is outside the scope of the current model , this is a potential avenue for future examination and may further add to our understanding of how multiple post-translational modifications are co-regulated to affect circadian activity . One part of the SIRT1-PARP1 system that obviously remains to be explored through a more comprehensive model would include a more complete description of the salvaging of NAD , including the activity of NMNAT1 that yields an intermediate step in this process . Although , NAMPT is the rate-limiting step in the salvage process , it catalyzes the first step in the conversion of nicotinamide ( the by-product of SIRT1 and PARP1 catalysis ) into nicotinamide mononucleotide; a substrate that is subsequently converted into NAD by NMNAT1 [17] . In the current model , only NAMPT has been included , because it is under circadian control , and because it is known to be rate limiting in the production of NAD . Yet several publications have shown that SIRT1 can bind to nicotinamide mononucleotide adenylyltransferase 1 ( NMNAT1 ) , and it has been hypothesized that this activity may help to stimulate SIRT1 activity [41] . This would be an interesting next step to pursue , as well as the more detailed PARP1 dynamics that account for the negative feedback cycle in these dynamics due to its auto-modification capability [20] . As the underlying mechanisms regulating the circadian clock become better understood with respect to the effects of post-translational modifications , such as acetylation , methylation , sumolyation , and ubiquitination , the addition of these factors can be used to refine current models of circadian rhythms .
|
Many physiological processes are regulated by the circadian clock , and we are continuing to learn about the role of the circadian clock in disease . Research in recent years has begun to shed light on the feedback mechanisms that exist between circadian regulation and other processes , including metabolism and the response to DNA damage . A challenge has been to understand the dynamic nature of the protein interactions of these processes , which often involve protein modification as a means of communicating cellular states , such as damaged DNA . Here we have devised a model that simulates an alteration of the circadian clock that is observed during DNA damage response . A novel aspect of this model is the inclusion of SIRT1 , a protein that regulates core circadian proteins through modification and helps to repress gene expression . SIRT1 is dependent on a metabolite regulated by the circadian clock and is depleted during DNA damage . In conjunction with a second form of protein modification , our results suggest that multiple forms of protein modification may contribute to the experimentally observed alterations to circadian function .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Predicted Role of NAD Utilization in the Control of Circadian Rhythms during DNA Damage Response
|
Chagas disease is a parasitic disorder caused by the infection with the flagellated protozoan Trypanosoma cruzi . According to the World Health Organization , more than six million people are currently infected in endemic regions . Genetic factors have been proposed to influence predisposition to infection and development of severe clinical phenotypes like chronic Chagas cardiomyopathy ( CCC ) . Interleukin 18 ( IL18 ) encodes a proinflammatory cytokine that has been proposed to be involved in controlling T . cruzi infection . In this study , we analyzed the possible role of six IL18 gene variants ( rs5744258 , rs360722 , rs2043055 , rs187238 , rs1946518 and rs360719 ) , which cover most of the variation within the locus , in the susceptibility to infection by T . cruzi and/or CCC . In total , 1 , 171 individuals from a Colombian region endemic for Chagas disease , classified as seronegative ( n = 595 ) , seropositive asymptomatic ( n = 175 ) and CCC ( n = 401 ) , were genotyped using TaqMan probes . Significant associations with T . cruzi infection were observed when comparing seronegative and seropositive individuals for rs187238 ( P = 2 . 18E-03 , OR = 0 . 77 ) , rs360719 ( P = 1 . 49E-03 , OR = 0 . 76 ) , rs2043055 ( P = 2 . 52E-03 , OR = 1 . 29 ) , and rs1946518 ( P = 0 . 0162 , OR = 1 . 22 ) . However , dependence analyses suggested that the association was mainly driven by the polymorphism rs360719 . This variant is located within the promoter region of the IL18 gene , and it has been described that it creates a binding site for the transcription factor OCT-1 affecting IL-18 expression levels . In addition , no evidence of association was observed between any of the analyzed IL18 gene polymorphisms and the development of CCC . In summary , our data suggest that genetic variation within the promoter region of IL18 is directly involved in the susceptibility to infection by T . cruzi , which provides novel insight into disease pathophysiology and adds new perspectives to achieve a more effective disease control .
Host genetic factors have been suggested to play an important role in the susceptibility to human infectious diseases [1] . An example of such conditions is Chagas disease , which is caused by infection of the protozoan Trypanosoma cruzi . Recent estimations indicate that more than 70 million people live in endemic areas for this parasite , with around 6 million people being currently infected and a reported incidence of the disease of almost 30 , 000 cases [2 , 3] . Two phases , acute and chronic , are clearly defined in Chagas disease . The early stages are characterized by acute symptoms like fever , headache or swollen lymph nodes . After 8–12 weeks from the bite , infected individuals enter the chronic phase of the disease , in which most of them will remain asymptomatic for the rest of their lives . However , around 30% of patients will develop further symptoms , including chronic cardiomyopathy and/or digestive complications [4] . During the last decade , several studies have investigated the possible role of gene polymorphisms in the predisposition to T . cruzi infection and/or chronic Chagas cardiomyopathy in patients from endemic countries , reporting promising results [5–17] . Interleukin 18 ( IL18 ) is one of the genes that have been proposed to influence the development of Chagas disease . It encodes a proinflammatory cytokine that was originally described as an interferon-gamma ( IFN-γ ) inducing factor . Because of this , IL-18 was classified among the Th1-inducing family of cytokines , along with IL-2 , IL-12 and IL-15 [18] . Due to its crucial role in the induction of IFN-γ production by T cells and NK cells , thus promoting the Th1 response , IL-18 is considered a relevant molecule for controlling intracellular pathogens [19 , 20] . In Chagas disease , IFN-γ is essential for parasite control during the early stages of the infection . It has been described that knockout mice for IFNG are highly susceptible to infection due to defective macrophage activation and nitric oxide production [21] . Interestingly , mice inoculated with T . cruzi displayed elevated IL-18 levels 6 days after infection followed by an increase of IL-12 and IFNγ [22] . Indeed , IL-18 can mediate IFN-γ induction in T cells in an IL-12 independent manner [23] . Consistent with the above , previous studies have suggested a genetic influence of both IFNG and IL18 gene variants in the susceptibility to infection by T . cruzi and Chagas cardiomyopathy , respectively [9 , 15] , adding additional evidences to the high relevance that this pathway may have in Chagas disease development . Taking into consideration all this knowledge , we decided to perform a comprehensive analysis of the IL18 variation , in a well-powered cohort from an endemic region of T . cruzi , in order to dissect the possible genetic association of the region with predisposition to infection by this parasite and/or the development of cardiomyopathy in Chagas patients .
A total of 1 , 171 Colombian individuals from an endemic region for Chagas disease ( Guanentina and Comunera provinces , at the department of Santander localized between 5°26’ and 8°08’ north and 72°26’ and 74°32’ west ) were included in this study ( S1 Fig ) . The population in this region of Colombia is a homogeneous mixture , with no specific concentration of any ethnicity . All participants underwent a serological diagnosis for T . cruzi infection by means of the enzyme-linked immunosorbent assay ( ELISA ) and a commercial indirect hemagglutination test . According to the results of these tests , 576 individuals were classified as seropositive for T . cruzi antigens and 595 were classified as seronegative , with this latter group being used as controls . Subsequently , and based on the results of the clinical evaluation , an electrocardiogram and echocardiogram were recorded to detect any conduction alteration and/or structural cardiomyopathy . As a result , 175 seropositive individuals were classified as asymptomatic and 401 individuals were classified as having chronic Chagas cardiomyopathy . From this last group , Chagas patients were further subclassified accordingly to the severity of cardiomyopathy as follows: CII ( n = 166 , radiology indicative of light heart hypertrophy or minor ECG alterations ) , CIII ( n = 200 , moderate heart hypertrophy and considerable ECG alterations , mainly conduction abnormalities ) and CIV ( n = 35 , severe cardiomegaly and marked ECG alterations , predominantly frequent and/or complex forms of ventricular arrhythmia ) . The mean age of participants was 45 . 86 years for seronegative individuals , 58 . 00 for asymptomatic individuals and 63 . 14 for chronic Chagas cardiomyopathy patients . The sex distribution for the entire group was 55% female and 45% male . None of the patients included in this study received any treatment ( i . e . Benznidazole ) for the infection . The Ethics Committees from the ‘Universidad Industrial de Santander and Fundación Cardiovascular de Colombia’ approved this study ( entitled “Identificación de factores de riesgo genético para Cardiopatia Chagásica crónica” [Identification of genetic risk factors for chronic Chagas cardiomyopathy] and approved on June 27th 2005 in the Act No . 15 of 2005 ) in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki . Written informed consent was obtained from all subjects prior to participation . In order to comprehensively analyze the possible role of IL18 on the genetic susceptibility to Chagas disease , a total of six single-nucleotide polymorphisms ( SNP ) within the locus were selected for genotyping following a combined candidate gene/tagging strategy . These include: 1 ) two promoter variants ( rs187238 and rs1946518 ) that have been reported to affect gene expression [24–27]; 2 ) one intronic variant ( rs2043055 ) previously implicated in Chagas disease outcome in a Brazilian population [15]; 3 ) an additional promoter variant ( rs360719 ) that has been described to interact with the transcription factor OCT-1 [28]; and 4 ) two intronic tag SNPs ( rs5744258 and rs360722 ) covering the remaining variability of the IL18 gene . These latter variants were selected with the software Haploview V4 . 2 [29] on the basis of both pairwise tagging ( r2>0 . 80 ) and minor allele frequencies >0 . 1 , using Colombian-Medellín ( CLM ) genotype data from the 1000 genomes phase III project ( http://www . 1000genomes . org ) [30] , encompassing the coding and promoter regions of IL18 and considering the four previously selected candidate variants . Genomic DNA was isolated following standard procedures and the genotyping was performed using TaqMan assays ( Applied Biosystems , Foster City , California , USA ) on a LightCycler 480 real-time PCR system ( Roche Diagnostics , Basel , Switzerland ) . All statistical analyses were performed with the statistical software package Plink V1 . 07 ( http://pngu . mgh . harvard . edu/purcell/plink ) [31] . Deviance from Hardy-Weinberg equilibrium was determined at the 1% significance level in all groups of individuals . To test for possible allelic and genotypic associations , we analyzed the allelic , genotypic and haplotypic frequencies by comparing seronegative vs . seropositive individuals and asymptomatic vs . chronic Chagas cardiomyopathy individuals using the χ2 test and logistic regression analyses , when necessary . The Benjamini & Hochberg step-up false discovery rate ( FDR ) correction was used in all analyses to control for multiple testing . Permutation tests ( 10 , 000 permutations ) were also performed in the haplotype analysis to estimate empirical P-values as implemented in Plink . Odds ratios ( OR ) and 95% confidence intervals ( CI ) were calculated according to the Woolf’s method . P-values lower than 0 . 05 were considered as statistically significant . Pairwise linkage disequilibrium ( LD ) ( D’ and r2 ) and haplotypes were estimated using an expectation–maximization algorithm implemented in Haploview . In addition , to determine whether the haplotype model better explained the observed effects than the model considering the individual SNPs , we compared the goodness of fit of both models by a likelihood ratio test as described elsewhere [32] . The statistical power of our study ( Table A in S1 Text ) was estimated with the Power Calculator for Genetic Studies 2006 ( CaTS ) software ( http://www . sph . umich . edu/csg/abecasis/CaTS/ ) [33] .
The six IL18 SNPs were in Hardy-Weinberg equilibrium in all the analyzed subgroups ( P>0 . 01 ) , suggesting that a possible inbreed in Guanentina and Comunera provinces is not likely . The genotyping success rate was over 95% and the allele frequencies in all cases were similar to those described for the Colombian population ( CLM ) of the 1000 genomes phase III project ( http://www . 1000genomes . org ) [30] . A relatively high LD was observed throughout the gene in the analyzed population ( Fig 1 ) . Particularly , rs187238 and rs360719 showed an r2 value = 0 . 98 , indicating that these two variants are almost completely linked and , consequently , they may be considered as the same marker for this study . In order to evaluate the possible association between IL18 gene variants and susceptibility to T . cruzi infection , the allelic and genotypic frequencies of seronegative and seropositive individuals were compared ( Table 1 ) . The allelic frequencies of four out of the six IL18 genetic variants were significantly different between these two groups of individuals . In this regard , rs187238*C and rs360719*C were significantly increased in seronegative individuals compared to the seropositive subset ( P = 2 . 18E-03 , PFDR = 5 . 04E-03 , OR = 0 . 77 , CI = 0 . 65–0 . 91; and P = 1 . 49E-03 , PFDR = 5 . 04E-03 , OR = 0 . 76 , CI = 0 . 64–0 . 90; respectively ) , thus suggesting that these variants are associated to protection against infection by T . cruzi . On the contrary , the frequencies of rs2043055*C and rs1946518*C were reduced in the seronegative sample set in comparison with the seropositive one ( P = 2 . 52E-03 , PFDR = 5 . 04E-03 , OR = 1 . 29 , CI = 1 . 10–1 . 53; and P = 0 . 0162 , PFDR = 0 . 0243 , OR1 . 22 , CI = 1 . 04–1 . 44; respectively ) , indicating that they are associated with a higher infection risk . No statistical significance was observed when the allelic and genotypic frequencies of both rs5744258 and rs360722 were compared between seropositive and seronegative individuals . The marked difference of the average age between the different subgroups of patients ( i . e . , seronegative and seropositive individuals ) could represent a limitation in this study , as Chagas disease is a parasitic disorder in which patients could develop symptoms many years after the infection [3 , 4] . To control for this possible confounding factor , we decided to perform a logistic regression analysis accordingly with the serological status using age as covariate . The statistical significance was maintained in this analysis thus supporting the consistency of our results ( Table B in S1 Text ) . Due to the high LD among the four associated SNPs , which are located either within or nearby the promoter region , dependency of the associations could be masking a possible unique causal variant . To check this , conditional logistic regression analyses were conducted by conditioning each associated SNP to the remaining variants ( except for rs187238 that was almost completely dependent to rs360719 ) . The results of these analyses pointed to rs360719 as the most likely causative variant of the IL18 association with Chagas disease , as the statistical significance of both rs2043055 and rs1946518 were clearly lost after conditioning to it , and a trend was maintained when conditioning rs360719 on these two latter variants ( Table 2 ) . Subsequently , in order to investigate the possible association between IL18 and chronic Chagas cardiomyopathy , we compared the allelic and genotypic frequencies of the IL18 SNPs between seropositive asymptomatic individuals and chronic Chagas cardiomyopathy individuals ( Table 3 ) . No statistically significant differences were observed between asymptomatic individuals and chronic Chagasic cardiomyopathy patients for any of the analyzed polymorphisms . In addition , to further evaluate the possible association between IL18 and progression of cardiomyopathy , we compared IL18 allelic and genotypic frequencies by grouping asymptomatic + CII individuals and CIII + CIV individuals; however , similar to that observed in the previous analysis , no statistically significant differences were yielded ( Table C in S1 Text ) . Finally , we also investigated a possible haplotype effect between the associated SNPs and susceptibility to T . cruzi infection ( Table 4 ) . Five possible haplotypes were observed ( rs2043055|rs187238|rs1946518|rs360719: TCAC , CGCT , TGCT , TGAT and CGAT ) , with the haplotypes TCAC and CGCT showing the higher frequencies ( 37 . 30% and 37 . 40% , respectively ) . In relation to the seronegative vs seropositive analysis , the frequency of TCAC was increased in the former individuals , being this difference statistically significant ( P = 1 . 50E-03 , PFDR = 7 . 50E-03 , OR = 0 . 76 , CI = 0 . 65–0 . 90 ) . On the other hand , the frequency of CGCT was significantly lower in seronegative individuals compared with seropositive individuals ( P = 0 . 0116 , PFDR = 0 . 0290 , OR = 1 . 25 , CI = 1 . 05–1 . 47 ) , whereas the frequencies of TGCT , TGAT , and CGAT did not differed significantly between seropositive and seronegative individuals . Similar results were observed when the haplotype analysis was performed using permutation test with 10 , 000 permutations instead of Chi-square ( Table 4 ) . However , the haplotype model did not better explain the IL18 association to risk of infection than the model considering the SNPs independently ( likelihood P-value = 0 . 1454 ) , indicating no additive effects ( that is , the associated haplotypes were a consequence of the independent associations of the considered variants ) . In relation to the haplotype analysis according to the presence/absence of chronic Chagas cardiomyopathy and/or to the progression of cardiomyopathy , no statistically significant differences among different subgroups of individuals were observed ( Tables D and E in S1 Text ) .
This study evidenced that four genetic variants , namely rs2043055 , rs187238 , rs1946518 and rs360719 , are statistically associated to differential risk of infection by T . cruzi in a Colombian population . However , our data suggested that the association is mainly driven by a single SNP , likely rs360719 . Evidences supporting this fact include: 1 ) this IL18 variant showed the most significant P-value and the higher effect size; 2 ) the statistical significance of both rs2043055 and rs1946518 was lost after conditioning on rs360719 , whereas a trend towards significance was clearly observed for rs360719 after conditioning on rs2043055 or rs1946518; 3 ) no improvement in the goodness of fit for the model considering the association with rs360719 was observed for any of the haplotypic models; and 4 ) this IL18 variant has a demonstrated functional implication in the gene expression [28] . On the other hand , IL18 does not seem to be involved in later parasitic burden in the tissues of chronic infected patients . It should be noted that the analysis between symptomatic and asymptomatic patients was performed with lower statistical power than that between seronegative and seropositive patients ( S1 Table ) . Hence , a possible type II error may not be rule out . Another possibility could be that additional genetic/environmental factors other than this gene may have a higher relevance for the disease progression [5–7 , 11–14] . In addition , a lack of association with infection by T . cruzi was observed for rs5744258 and rs360722 . These two polymorphisms had the lower minor allele frequency and , therefore , their analysis could be limited in terms of statistical power . However , the power considering our study cohort was not reduced ( 92% to detect associations with OR = 1 . 5 at the 5% significance level ) and the allele frequencies of the tested groups were very similar ( rs5744258: 11 . 57% vs 11 . 58% , OR = 1 . 00; rs360722: 11 . 59% vs 12 . 33% , OR = 1 . 07 ) . Analysis of larger cohorts would be required to definitively discard these IL18 variants as susceptibility markers for Chagas disease . A possible limitation in the inclusion methodology of our study could be that the seronegative group comprised individuals that underwent a seroconversion . However , in our opinion , it is more likely that seronegative individuals with putative spontaneous cure avoided antibody production due to a quick innate immune response by killer cells and macrophages instead . No consistent seroconversion rates have been reported in Chagas patients so far , and seroconverted individuals were reported only after treatment when there is not persistence in the infection [34–36] . None of the seronegative individuals included in our study were either reported to have Chagas disease or to have a previous therapy . In addition , it should be noted that there is a considerable high prevalence of cardiac patients in our study cohort , which could suggest that the seropositive population is biased to the patients with Chagas cardiomyopathy . However , we would like to state that the participants were recruited after a medical visit to the endemic area . In this regard , individuals coming to the citation underwent serological analyses , and those showing seropositivity were subsequently subjected to electrocardiograms and further medical analyses in which they were classified as asymptomatic seropositive or CCC patients . In any case , this sample set has been used in previously published studies by our group [13] and we are confident about its homogeneity . IL-18 is a cytokine which induces IFN-γ production activating several immune cells in response to intracellular pathogens , including T . cruzi [20] . IL-18 was shown to play an important role in early immunity to Chagas disease [22 , 23] . Moreover , the susceptibility to T . cruzi depends on the capability of releasing IFN-γ during early stages of infection and this is directly related to release of IL-18 during this phase [37] . Regarding this , it has been reported that rs360719 may be located within a repressor site of the gene , and individuals carrying the C allele showed a higher IL18 expression due to the creation of a binding site for the transcription factor OCT-1 [28] . This is consistent with the protective role that we observed for this allele in Chagas disease development , and support the hypothesis that major IL-18 levels could increase parasite clearance in early stages of infection . Besides , our results are also in concordance with a previous study performed by our group reporting an association between IFNG and susceptibility to infection by T . cruzi [9] . Altogether , these findings clearly point to IL-18 along with IFN-γ as crucial players in the immune response against infection by this parasite . In any case , additional functional analyses are needed to confirm this assumption , and to have an accurate estimation of the putative IL-18 and IFN-γ levels that may discriminate the different subgroups of Chagas patients from each other and from the healthy population . On the other hand , a previous study showed that IL18 rs2043055 may modulate Chagas disease severity in a Brazilian population [15] . In our study we were not able to find an association of any of the analyzed IL18 SNPs ( including this one ) with the severity of Chagas disease . We speculate that the discrepancy could be due to a different genetic background between the analyzed cohorts from Brazil and Colombia . Despite being both populations a mixture from Amerindian , west-European and African populations , the proportion of these ancestries could differ between them , which would affect the LD and haplotypic block architecture across the genome [38–40] . It would be interesting , therefore , to examine whether the association described for rs2043055 is dependent upon rs360719 in the Brazilian population , as our data suggest based on the LD structure observed in our Colombian cohort . In any case , both studies open a new window to understand differences in Chagas disease outcome and susceptibility . Another explanation for the observed differences between both studies could be the existence of different T . cruzi strains in the studied regions from Colombia and Brazil , as the two strains present in such areas ( I and II , respectively ) have been described to be implicated in CCC development and severity [41] . The analysis of the specific strains affecting both our population and the Brazilian one was out of the scope of this study , but it could represent an interesting future complementary analysis to this reported here . The influence of IL18 gene variants on the susceptibility to infection or severity of other protozoan infectious diseases has been also evaluated . In this context , a weak association between the IL18 SNP rs1946519 , and a higher risk to develop Leishmaniasis was described in Iranians [42] . Nevertheless , the authors did not find evidence of association between Leishmania infection and rs187238 , which was associated with T . cruzi infection in our study . As stated before , the discrepancy could be due to population-specific genetic architectures within the gene , but also to the fact that our study had a considerably higher statistical power . Additionally , the possible role of the IL18 gene variants rs187238 and rs1946518 in severe malaria anemia and mortality were also investigated in a Kenyan children population [43] . The authors of that study observed that homozygosity for the rs1946518*A allele conferred protection against severe malaria , and that the allelic combination of rs187238*G and rs19463518*C had a higher frequency in the severe malaria group compared to the non-severe group [43] . In our study , the frequency of the AA genotype for rs1946518 was increased in asymptomatic individuals compared to chronic Chagas cardiomyopathy patients , but this difference was not statistically significant . Similarly , the frequency of the rs187238*G|rs19463518*C haplotype was increased in chronic cardiomyopathy individuals compared with asymptomatic patients , although the difference did not reach statistical significance either . Additional studies encompassing larger cohorts of seropositive patients with different degrees of disease severity may shed light into these putative associations . IL18 gene variants have also been evaluated in other infectious conditions such as hepatitis B or C viruses . Cumulating data indicate that IL-18 may influence the clearing of the viral load [44–46] , as well as the severity of the infection in some cases of hepatic carcinomas or cirrhosis [26 , 47 , 48] . In this context , it has been proposed that differential expression levels of IL18 could be directly involved in the predisposition to infection by the above mentioned viruses and in the severity of hepatitis [26 , 44–48] , consistent with what we observed in Chagasic patients . In conclusion , our results suggest that IL18 variation plays an important role in the susceptibility to infection by T . cruzi , probably by influencing IL-18 production during the immune response in the early stages of the infection . The promoter polymorphism rs360719 is likely the causal variant of this association , at least in the Colombian population . In any case , further studies on this gene on different ancestries and larger samples sizes , as well functional analyses , would be desirable to validate our findings .
|
Chagas disease is a parasitic disorder caused by the infection with the protozoan Trypanosoma cruzi . In Latin America , this disease represents a major public health concern , as almost 6 million people are currently infected . During the last years , great efforts have been made in health policy to control the disease; however , there is still a long way ahead to achieve this challenging goal . Most affected people remains asymptomatic after infection for the rest of their lives , but around one third of infected people may develop cardiomyopathy , a condition that reduces dramatically the quality of life and life expectancy in Chagas patients . The causes of the marked differential disease outcomes are currently unknown , but it is believed that a genetic predisposition could play a relevant role in the host . We investigated in an endemic region of Colombia whether the IL18 gene , which is involved in the immune response to intracellular pathogens like T . cruzi , is related to a higher susceptibility to infection or disease severity . Our results suggest that IL18 is a relevant gene in Chagas disease , and could represent a valuable insight that may help to better understand the disease pathogenesis and the development of more efficient therapeutic strategies .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"cardiomyopathies",
"medicine",
"and",
"health",
"sciences",
"population",
"genetics",
"tropical",
"diseases",
"variant",
"genotypes",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"genetic",
"mapping",
"genetic",
"predisposition",
"protozoans",
"neglected",
"tropical",
"diseases",
"population",
"biology",
"cardiology",
"protozoan",
"infections",
"trypanosoma",
"cruzi",
"haplotypes",
"trypanosoma",
"chagas",
"disease",
"heredity",
"genetics",
"biology",
"and",
"life",
"sciences",
"evolutionary",
"biology",
"genetics",
"of",
"disease",
"organisms"
] |
2016
|
IL18 Gene Variants Influence the Susceptibility to Chagas Disease
|
There is a current interest in quantifying time-varying connectivity ( TVC ) based on neuroimaging data such as fMRI . Many methods have been proposed , and are being applied , revealing new insight into the brain’s dynamics . However , given that the ground truth for TVC in the brain is unknown , many concerns remain regarding the accuracy of proposed estimates . Since there exist many TVC methods it is difficult to assess differences in time-varying connectivity between studies . In this paper , we present tvc_benchmarker , which is a Python package containing four simulations to test TVC methods . Here , we evaluate five different methods that together represent a wide spectrum of current approaches to estimating TVC ( sliding window , tapered sliding window , multiplication of temporal derivatives , spatial distance and jackknife correlation ) . These simulations were designed to test each method’s ability to track changes in covariance over time , which is a key property in TVC analysis . We found that all tested methods correlated positively with each other , but there were large differences in the strength of the correlations between methods . To facilitate comparisons with future TVC methods , we propose that the described simulations can act as benchmark tests for evaluation of methods . Using tvc_benchmarker researchers can easily add , compare and submit their own TVC methods to evaluate its performance .
Time-varying connectivity ( TVC ) is being applied to an increasing number of topics studying the brain’s networks . Topics that have been explored with TVC include development [1] , various pathologies [2 , 3] , affect [4] , attention [5] , levels of consciousness [6] , and temporal properties of the brain’s networks [7–9] . There are many concerns raised regarding methodological issues . These issues span biased variance [10 , 11] , movement artefacts [12] , and appropriate statistics [13 , 14] . Methods used to derive TVC estimates are as diverse as its range of applications . Examples of different methods include: the sliding window method , sometimes tapered [15] , multiplication of temporal derivatives [16] , methods using Euclidean distance between spatial configurations [8] , k-means clustering methods [7 , 17] , eigenconnectivities [18] , point process methods [19 , 20] , Kalman filters [21 , 22] , flexible least squares [23] , temporal ICA [24] , sliding window ICA [25] , dynamic conditional correlation [26] , phase differences [27] wavelet coherence [4] , hidden Markov models [28] , and variational Bayes hidden Markov models [29] . This list of TVC methods is not exhaustive , and even more methods can be found in the literature . While these methods and their applications may offer new insights into the functions of the brain and cognition , it becomes difficult to compare results when different studies use different methods to estimate brain dynamics . Each method is often introduced and evaluated by the authors’ own simulations , empirical demonstrations , and/or theoretical arguments . However , apparent differences in time-varying connectivity in different studies may have been influenced , or even caused , by differences in the underlying methodology used to derive connectivity estimates . In order to maximize reproducibility of reported findings , it is important that comparisons of proposed TVC methods can be made with a common set of simulations . To this end , we have developed four simulations that aim to show how well results from different TVC methods correlate with each other and evaluate their performance of tracking time varying covariance . The proposed methods and simulations are included in the Python package tvc_benchmarker , ( available at www . github . com/wiheto/tvc_benchmarker ) . Researchers can evaluate their own TVC methods in tvc_benchmarker . The software also allows for new methods to be submitted to us for inclusion in future reports . Here we demonstrate the functionality and results obtained by tvc_benchmarker by evaluating the performance of the following five methods: sliding window ( SW ) , tapered sliding window ( TSW ) , spatial distance ( SD ) , jackknife correlation ( JC ) , and multiplication of temporal derivatives ( MTD ) .
All methods for TVC derivation were implemented in Teneto v0 . 2 . 7b [8] . Bayesian statistics for evaluating performance of TVC methods were calculated in PyMC3 V3 . 1 [30] , simulations and analysis were done using Numpy V1 . 13 . 1 [31] , Scipy V0 . 19 . 1 [32] , and Pandas V0 . 19 . 2 . Matplotlib V2 . 0 . 2 [33] and Seaborn V0 . 7 . 1 [34] were used for figure creation . As discussed in the introduction , the list of published TVC methods that are designed to be applied to fMRI imaging data is long . In an ideal world all methods will be contrasted under the same conditions such that an evaluation of those methods that give appropriate results can be performed . However , it was not our intention to provide a complete comparison of all published methods . Instead we have made all simulation tools freely available so that researchers can evaluate their own TVC methods . Before describing the simulations and the results , we provide a brief overview of the five methods that are evaluated in this article . This section provides an overview of the simulations that are conducted and the general methodology used . See each simulation’s subsection in the results section for full details of each simulation . To compare accuracy and performance for the five TVC methods , we performed four different simulations . The first simulation investigated the similarity of the different TVC methods by correlating their respective connectivity estimates . The second simulation targeted how well the different methods were able to track a fluctuating covariance parameter . The third simulation tested how robust the estimated fluctuating covariance is when the mean of the time series fluctuates , mimicking the haemodynamic response function . The forth simulation considered whether TVC methods can accurately track abrupt changes in covariance . All simulations considered two time series each consisting of 10 , 000 samples generated from multivariate Gaussian distributions . At each time point , the covariance between the time series could vary ( see below ) . A full account of all model assumptions made as well as a justification for our model parameter settings for the four simulations models used in the present study are given in S2 Appendix . Simulations 2 , 3 , and 4 all consisted of a fluctuating covariance parameter ( rt ) that was used to generate the covariance between the time series . TVC methods were evaluated based on their ability to track the rt parameter . How rt was generated could vary for different simulations . In simulation 2 , rt varied throughout the time course based on a normal distribution . The simulation was run multiple times allowing for different autocorrelation of rt through time . In simulation 3 , rt varied in the same way as simulation 2 but it was applied to time series that had a non-stationary mean that mimicked a HRF . This simulation was also run multiple times with different autocorrelations . In simulation 4 , rt varied based on two different “states” that lasted for varying amounts of time . This method was run two times when states could be short ( 2-6 time points long ) or long ( 20-60 time points long ) . By evaluating the correlation of different TVC methods with each simulation’s rt , we can evaluate which time varying properties a method is sensitive to . Simulation 1-3 have all their parameters justified on empirical data in S2 Appendix . Simulation 4 has its state lengths based on what has been identified by different TVC studies . It is important to stress that these different state lengths may have been identified due to the methods which were used and may not reflect real dynamic properties . In principle , it is possible to simply correlate the results from the different TVC methods with the rt values of each simulation to statistically evaluate their performance . However given the inherent , but known , uncertainty in rt , we deemed it was appropriate to create a statistical model which accounts for this uncertainty . Thus , for each TVC method , a Bayesian statistical model was created to evaluate the relationship between the TVC estimate and the signal covariance . The Bayesian model aims to predict y , which is the vector of the known sampled covariances ( i . e . rt ) with x , which is the connectivity estimate for each TVC model . All TVC estimates and the values of rt were standardized prior to calculating the models with a mean of zero and standard deviation of one . This was done to facilitate the interpretation of the posterior distribution parameter β . The different TVC methods vary in the number of time points estimated ( e . g . the beginning and end of the time series cannot be estimated with the sliding window method ) . In order to facilitate model comparison between methods , we restrained the simulations to include only the time points that had estimates from all TVC methods ( i . e the limit was set by the SW and TSW methods which can estimate the covariance for 9 , 972 out of 10 , 000 time points ) . The statistical models were estimated through 5 , 500 draws from a Markov Chain Monte Carlo ( MCMC ) with a No-U-Turn Sampler [38] sampler implemented in pymc3 . The first 500 samples were burned . The statistical models for the different TVC methods can be contrasted in two ways: ( 1 ) model comparison by examining the model fit; ( 2 ) by comparing the posterior distribution of β for the different TVC methods . To evaluate the model fit , the Watanabe-Akaike information criterion ( WAIC , [39] ) was used . The posterior distribution of β illustrates the size and uncertainty of the relationship between x and y . To aid the interpretation of these results for readers unfamiliar with Bayesian statistics , the mode of the distribution corresponds approximately to a maximum-likelihood estimated β value in a linear regression ( if uniform priors are used for the parameters the posterior mode and the maximum-likelihood estimator would have been exactly the same ) . In simulation 1 , the different TVC estimates are compared with each other to evaluate how similar these estimates are . To do this , a Spearman correlation is used to evaluate the relationship .
The first simulation aimed to quantify the similarity of the different TVC time series estimates . If two TVC methods are strongly correlated , this is a positive sign that they are estimating similar aspects of the evolving relationship between time series . A negative correlation between two methods would suggest that they do not capture the same dynamics of the signal . In this simulation we created two time series ( X ) , each consisting of 10 , 000 time points in length . The time series were constructed by: X t = α X t - 1 + ϵ ( 7 ) The autocorrelation with lag of 1 is determined by αXt−1 and the covariance at t is determined by ϵ . ϵ was sampled from a multivariate Gaussian distribution ( N ) : ϵ ∼ N ( μ , Σ ) ( 8 ) where μ is the mean and Σ being the covariance matrix of the multivariate Gaussian distribution . Both time series were set to have a mean of 0 , variance of 1 and a covariance of 0 . 5 . In summary: μ = 0 , 0 Σ = ( 1 0 . 5 0 . 5 1 ) ( 9 ) The autoregressive parameter α controls the size of the autocorrelation in relation to the preceding time point ( i . e . the proportion of the previous time point that is kept ) . Here , it was set to 0 . 8 which was deemed to be an appropriate degree of autocorrelation for BOLD time series ( see S2 Appendix ) . A portion of the two simulated time series is found in Fig 2A together with the plots of their respective autocorrelation ( Fig 2B and 2C ) and a plot of the correlation between the two time series ( Fig 2D ) . The resulting connectivity time series for the different TVC methods when applied to the simulated data is shown in Fig 3 . From Fig 3 , several qualitative observations can be made about the methods . Firstly , there was a very strong similarity between the SD and JC methods , despite the fact that they consist of quite different assumptions . Further , the SD , JC , and MTD methods were all able to capture considerably quicker transitions than the SW and TSW methods . The long window lengths ( SW-29 and TSW-29 ) were smoother than the SW-15 and TSW-15 methods . Finally , the variance of the JC method was considerably smaller than all other methods , illustrating the variance compression as previously discussed . To assess the degree of similarity of the estimates of functional connectivity time series obtained from all TVC methods , a Spearman correlation was computed for each TVC method pairing ( Fig 4 ) . The connectivity time series estimates from all methods correlated positively with each other ( Fig 4 ) . Some methods showed strikingly strong correlations ( SD & JC: 0 . 976; SW-15 & TSW-15: 0 . 999; SW-29 & TSW-29: 0 . 978 ) . Between the different window lengths the correlation was slightly smaller ( SW: 0 . 644; TSW: 0 . 755 ) . The lowest correlation was found between the JC and MTD methods ( ρ = 0 . 138 ) . The results from Simulation 1 showed that the connectivity estimates provided by the tested methods are , to a varying extent , correlated positively with each other . It also illustrated how the different methods differ in their resulting smoothness of the connectivity time series . The results from this simulation cannot validate whether any TVC method is superior to any other , it merely highlights which methods produce similar connectivity time series . In Simulation 1 , it was not possible to evaluate how well the different TVC methods perform . To evaluate the performance , the simulated data must change its covariance over time and how this changes must be known beforehand . The aim of this simulation was to see how well the derived TVC estimates can infer the covariance that the data was sampled from when the covariance is fluctuating . Two time series were generated ( X ) . Each time point t is sampled from a multivariate Gaussian distribution: X t ∼ N ( μ , Σ t ) ( 10 ) where the covariance matrix was defined as: Σ t = ( σ r t r t σ ) ( 11 ) and where the variance , σ = 1 , was set to 1 . At each time point , rt was sampled from another Gaussian distribution: r t ∼ N ( μ r , σ r ) ( 12 ) The mean of the time series ( μ ) was set to 0 , the mean of the covariance ( μr ) was set to 0 . 2 . The simulation was run three times where the parameter for the variance of the fluctuating covariance ( σr ) was set to three different values {0 . 08 , 0 . 1 , 0 . 12} . This ensured that the different TVC methods are robust to different variances of connectivity changes . The covariance at time ( rt ) was sampled from a Gaussian distribution . Each time point received a new value of rt . This allowed us to compare each TVC method’s connectivity estimate in relation to the time varying covariance parameter rt . Note , that at each time point the relationship between the two time series is dictated by a single realization from a Gaussian distribution where rt is the covariance . Thus , we should not expect the connectivity estimate from any method to correlate perfectly with rt . However , it is possible to compare which method correlate better or worse with rt to evaluate the overall performance . The above model will have a temporally fluctuating covariance . It fails to include any autocorrelation in the time series . Not accounting for this may bias the results for some of the tested methods that utilize nearby temporal points to assist estimating the covariance . Merely adding an autocorrelation , like in Simulation 1 , will also increase the covariance between the two time series and this will not be tracked by rt . To account for this , we placed a 1-lag autoregressive model for the fluctuating covariance at rt: r t = α r t - 1 + ϵ ( 13 ) ϵ ∼ N ( μ r , σ r ) ( 14 ) Where α is the autocorrelation parameter . The values for μr and σr were the same as above . When t = 1 , ϵ was set to 0 . This revised formulation of our simulation model allowed for the covariance to fluctuate , but with an added autocorrelation on the covariance parameter . In simulation 2 , three different settings of the parameter α were used ( α = 0 , 0 . 25 , 0 . 5 ) . When α = 0 it is equivalent to the original model outlined above with no autocorrelation . With an increased α it entails a greater influence of the covariance from t − 1 in sampling the covariance at t . α = 0 . 5 is reasonable given highly correlated BOLD time series . An α = 0 is more to be expected when time series are less correlated . 10 , 000 time points were sampled for each of the three different settings of the autocorrelation parameter . See also S2 Appendix for a justification of the parameter settings chosen here based on empirical fMRI data . Simulation 2 was run with 9 different simulation parameter combinations: three different values of α and three different values of σr . A sample of time series generated with the model using different settings for the autocorrelation parameter α is shown in Fig 5A , 5D and 5G . Due to the varying degree of autocorrelation , the mean covariance for time series changes as a function of α , but rt still depicts a Gaussian distribution ( Fig 5B , 5E and 5H ) . The degree of crosscorrelation between the two time series followed the specified α parameter for the autocorrelation of the covariances ( Fig 5C , 5F and 5I ) . The results from Simulation 2 are shown in Tables 1–3 ( for σr = 0 . 1 ) and Tables A-F in S3 Appendix ( for σr = 0 . 08 and 0 . 12 ) . The JC method had the lowest WAIC score for all settings of α , followed by the SD method . The MTD method came in third place for all but one parameter configurations . All WAIC values , their standard error and Δ WAIC scores are shown in Tables 1–3 . The posterior distribution of the β parameter for each of the TVC methods for all parameter choices are shown in Fig 6 when σr = 0 . 1 ( for other values of σr see Figs A-B in S3 Appendix ) . Larger values in the β distribution for a method ( i . e . correlating more with rt ) conforms with the best fitting models ( i . e . lower WAIC score ) . The SW-15 , SW-29 , TSW-15 , TSW-29 and MTD methods performed equally poor when α = 0 , and all improved as α increased . The MTD method improved the most as the α value increased , followed by the TSW-15 and SW-15 methods . SD and JC showed the best performance , with similar posterior distributions of β , although the JC was always slightly higher . There was little difference between the methods when changing the variance of the fluctuating covariance ( σr ) ( See S3 Appendix ) . The β values do however scale when σr changes . When σr is smaller , β values decrease due to there being more uncertainty when sampling each realization from similar distributions . At times parts of the posterior distributions of the SW , TSW and MTD methods were below 0 to the extent that they would be not classed as “significant” . For example , these methods performed worst when σr = 0 . 08 and α = 0 . Here the percentage of the posterior distribution above 0 was: SW-15: 80% , SW-29: 47% , TSW-15: 84% , TSW-29: 54% , MTD: 89% . The JC and SD methods always had the entire posterior distributions above 0 . In sum , the JC method , followed closely by the SD method , showed the best performance in terms of tracking a fluctuating covariance between two time series as performed in Simulation 2 . The MTD method ranked in third place when there is a higher crosscorrelation between the time series present . The SW and TSW methods showed the worst performance , both in the WAIC score and posterior distributions of β . The aim of Simulation 3 was to examine the behaviour of different TVC methods when there were non-stationarities present in the data . A typical scenario when this will occur is in a TVC analysis in task fMRI . Simulation 3 is identical in structure to Simulation 2 apart from the following two changes: ( 1 ) A non-stationarity , aimed to mimic the occurrence of an event related haemodynamic response function ( HRF ) . Specifically μ , which was set to 0 for both time series in Simulation 2 , received a different value at each t ( see next paragraph ) . ( 2 ) σr was set to 0 . 1 instead of varying across multiple values . This is because Simulation 2 showed no large differences when varying σr . μt was set , for both time series , according to the value of a simulated HRF , that was twenty time points in length and repeated throughout the simulation . The HRF was simulated , with a TR of 2 , using the canonical HRF function as implemented in SPM12 using the default parameters [40] . This HRF , which has a length of 17 time points , was padded with an additional 3 zeros . The amplitude of the normalized HRF was multiplied by 10 to have a high amplitude fluctuations compared to the rest of the data . μt is thus the padded HRF repeated throughout the entire simulated time series . This represents a time series that includes 250 “trials” that each lasts 40 seconds . This simulation helps illustrate how well TVC methods could be implemented in task based fMRI . Examples of the time series generated using different autocorrelation are shown in Fig 7 . The results from Simulation 3 are shown in Fig 8 ( posterior distributions of β ) and Tables 4–6 ( model fit ) which evaluated each TVC’s method performance at tracking the fluctuating covariance ( rt ) . Results were similar with Simulation 2 . In the case when the autocorrelation of the covariance was 0 , the SW , TSW and MTD methods performed quite poorly , but again all improved to varying degrees as this increased . The longer windows ( SW-29 and TSW-29 ) methods were generally the worst method , followed by shorter sliding window methods ( SW-15 and TSW-15 ) . The MTD method came in third place . The JC method has the best performance , followed closely by the SD method , in all parameter conditions . When α = 0 , some methods had only portions of their posterior distribution above 0 ( SW-15: 73% , SW:-29: 30% , TSW-15: 78% , TSW-29: 65% , MTD: 84% ) . The JC and SD methods had 100% of their distributions above 0 for all parameter conditions . In sum , the results from Simulations 2 and 3 suggests that the JC method has the best performance in terms of detecting fluctuations in covariance compared to the other four TVC methods . This result also holds when a non-stationary event related haemodynamic response was added to the mean of the time series . Simulation 4 aimed to test how sensitive different TVC methods are to large and sudden changes in covariance ( i . e . changes in “brain state” ) that previously have been postulated to exist in fMRI data ( e . g . [11 , 15 , 17] ) . We here start in a similar fashion as we did in Simulation 2 where samples for the two time series are drawn from a multivariate Gaussian distribution X t ∼ N ( μ t , Σ t ) ( 15 ) Σ t = ( σ r t r t σ ) ( 16 ) Similar to simulation 2 , we set μt = 0 and σ = 1 . The covariance parameter rt was sampled from a Gaussian distribution where the mean was shifted r t ∼ N ( μ state t , σ r ) ( 17 ) and where σr = 1 . At each state transition , μ state t was randomly chosen from a set M ( M = {0 . 2 , 0 . 6} ) . The duration of each state was randomly sampled from L . Two different scenarios for state transitions were simulated . In the fast transition condition L = {2 , 3 , 4 , 5 , 6} and in the slow transition condition L = {20 , 30 , 40 , 50 , 60} . These values correspond to the number of time points a “state” lasts . Beginning at t = 1 , μ state t to μ state t + l was randomly sampled from M where l was sampled from L . This procedure was continued until Xt was 10 , 000 samples long . These choices for brain state changes provide time scales of state transitions between 40-120 seconds ( slow condition ) or 4-12 seconds ( fast condition ) in simulated fMRI data with a TR of 2 ( Fig 9A and 9D ) . The statistical model for evaluating the different TVC methods performance was the same as Simulation 2 and 3 . A summary of data generated in Simulation 4 is shown in Fig 9 . The results from Simulation 4 are shown in Fig 10 and Tables 7 and 8 . In the quick transition condition , the JC and the SD showed the best performance for both the WAIC scores and the posterior distribution of β ( Fig 10A; Table 7 ) . This was followed by the SW-15 and TSW-15 methods . In the slow transition condition the two sliding window methods outperformed the other methods ( Fig 10B; Table 8 ) , with the longer windows ( TSW-29 and SW-29 ) being outperforming the shorter windows . The JC and SD methods perform similarly for both conditions . Thus , when there are shifts in covariance that occur relatively slowly , the sliding window methods are sensitive at tracking these changes . All methods had 100% of their posterior distributions above 0 .
In this study we have developed four simulations to test the performance of different proposed time-varying connectivity methods . The first simulation showed which methods yield similar connectivity time series . Notably , all methods correlated positively with each other , but to a varying degree . The second simulation generated data in which the autocorrelated covariance between simulated time series varied in time . In this case , the JC method , followed closely by the SD method , showed the best performance . In the third simulation , the generated time series contained a non-stationary mean related to haemodynamic responses . Again , our simulations suggested that the JC method performed best . The fourth simulation included nonlinear shifts in covariance ( in an attempt to simulate brain state shifts ) . When the states changes were quick , the JC method performed best . When the state changes were slow , the TSW ( followed by the SW ) performed best . In a previous simulation that evaluated the sliding window method , the sensitivity of the SW and TSW methods was found to be good at detecting state shifts [41] . Here , at least when the transitions are slow , we found similar results . The sliding window methods is optimal if there are slow state changes . However it is unclear if “state changes” are the best yardstick for time-varying connectivity . In particular , non-stationarities in time-varying connectivity have been attributed to spurious sources such as movement [12] . Given the unknowns of the “true” connectivity , methods which are robust over conditions are more likely the safer options—in this case the JC or SD method performed similarly in both conditions . However , as mentioned in the methods section , the SD method tested here is the bivariate version of the method and not the multivariate version previously proposed in [8] ( see also S1 Appendix for more the relationship between these methods ) . Overall the jackknife correlation method performed the best across all simulations . We have shown it to be robust to numerous changes in parameters . However , the JC method is not without some considerations . First , it introduces variance compression that reduces the absolute variance , while preserving the relative variance within the time series . This variance compression also scales with the length of the time series . The consequence of this is that direct comparisons of the TVC variance between cohorts/conditions become hard to interpret as time-varying fluctuations , especially when the length of the data varies . However , this is the case for most methods and it should be remembered that the variance is proportional to the static functional connectivity [7 , 9 , 10] . Simply put , the JC method ( like all other methods ) should not be used for a direct contrast of the variance of TVC time series . Second , the JC method sensitivity means that noise will be carried over per time point instead of being smeared out over multiple time points . This is actually beneficial as it allows for further processing steps to be applied that aim to remove any remaining noise ( e . g . motion ) which cannot be done when the noise has been smeared across the connectivity time series ( e . g . in windowed methods ) . The simulations and results presented in this study should not be taken as an exhaustive and complete assessment of all aspects of a given method to conduct TVC . Rather , the four simulations described here represents a subset of possible scenarios in terms of different methodological characteristics that might be of interest . The current four simulations are marked tvc_benchmarker simulation routine V1 . 0 . If modifications or additional scenarios are considered to be improvements to the current simulations , these will get an updated version number . Many additional simulations could be conceived on top of this original routine . For example , one could include multiple time series , adding movement type artifacts , adding frequency relevant characteristics , a stationary global signal etc . These have not been included here , as the focus in these simulations was to primarily assess tracking of a fluctuating covariance . Input from researchers about appropriate additions to the simulations is welcome . We encourage researchers designing TVC methods to benchmark their own results with tvc_benchmarker ( www . github . com/wiheto/tvc_benchmarker ) . Researchers need only to write a Python function for their method and use it as an input for tvc_benchmarker . run_simulations ( ) and their method will be compared to the TVC methods presented in this paper ( see online documentation ) . Functions can then be submitted through the function tvc_benchmarker . send_method ( ) . All valid methods submitted will be released in summaries of the submitted benchmarked results so that researchers can contrast the performance of different methodologies .
|
Time-varying connectivity attempts to quantify the fluctuating covariance relationship between two or more regions through time . In recent years , it has become popular to do this with fMRI neuroimaging data . There have been many methods proposed to quantify time-varying connectivity , but very few attempts to systematically compare them . In this paper , we present tvc_benchmarker , which is a python package that consists of four simulations . The parameters of the data are justified on fMRI signal properties . Five different methods are evaluated in this paper , but other researchers can use tvc_benchmarker to evaluate their methodologies and their results can be submitted to be included in future reports . Methods are evaluated on their ability to track a fluctuating covariance parameter between time series . Of the evaluated methods , the jackknife correlation method performed the best at tracking a fluctuating covariance parameter in these four simulations .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"diagnostic",
"radiology",
"functional",
"magnetic",
"resonance",
"imaging",
"markov",
"models",
"neural",
"networks",
"random",
"variables",
"neuroscience",
"covariance",
"simulation",
"and",
"modeling",
"magnetic",
"resonance",
"imaging",
"regression",
"analysis",
"mathematics",
"statistics",
"(mathematics)",
"brain",
"mapping",
"neuroimaging",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"imaging",
"techniques",
"hidden",
"markov",
"models",
"mathematical",
"and",
"statistical",
"techniques",
"probability",
"theory",
"statistical",
"models",
"radiology",
"and",
"imaging",
"diagnostic",
"medicine",
"linear",
"regression",
"analysis",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"statistical",
"methods"
] |
2018
|
Simulations to benchmark time-varying connectivity methods for fMRI
|
Meta-analysis of genetic association studies increases sample size and the power for mapping complex traits . Existing methods are mostly developed for datasets without missing values , i . e . the summary association statistics are measured for all variants in contributing studies . In practice , genotype imputation is not always effective . This may be the case when targeted genotyping/sequencing assays are used or when the un-typed genetic variant is rare . Therefore , contributed summary statistics often contain missing values . Existing methods for imputing missing summary association statistics and using imputed values in meta-analysis , approximate conditional analysis , or simple strategies such as complete case analysis all have theoretical limitations . Applying these approaches can bias genetic effect estimates and lead to seriously inflated type-I or type-II errors in conditional analysis , which is a critical tool for identifying independently associated variants . To address this challenge and complement imputation methods , we developed a method to combine summary statistics across participating studies and consistently estimate joint effects , even when the contributed summary statistics contain large amounts of missing values . Based on this estimator , we proposed a score statistic called PCBS ( partial correlation based score statistic ) for conditional analysis of single-variant and gene-level associations . Through extensive analysis of simulated and real data , we showed that the new method produces well-calibrated type-I errors and is substantially more powerful than existing approaches . We applied the proposed approach to one of the largest meta-analyses to date for the cigarettes-per-day phenotype . Using the new method , we identified multiple novel independently associated variants at known loci for tobacco use , which were otherwise missed by alternative methods . Together , the phenotypic variance explained by these variants was 1 . 1% , improving that of previously reported associations by 71% . These findings illustrate the extent of locus allelic heterogeneity and can help pinpoint causal variants .
Meta-analysis has become a critical tool for genetic association studies in human genetics . Meta-analysis increases sample sizes , empowers association studies , and has led to many exciting discoveries in the past decade [1–5] . Many of these genetic discoveries have informed new biology , provided novel clinical insights [6 , 7] , and led to novel therapeutic drug targets [8 , 9] . Conditional meta-analysis has been a key component for these studies , which is useful to distinguish novel association signals from shadows of known association signals and to pinpoint causal variants . Existing methods for conditional meta-analysis were proposed based upon the assumptions that summary association statistics from all variant sites are measured and shared . Yet , in practice , the score statistics from contributing studies often contain missing values , possibly due to the use of different genotyping arrays , sequencing capture assays , or quality control filters by each participating cohort . While genotype imputation is an effective approach to fill in missing genotype data for participating cohorts , many scenarios may preclude accurate genotype imputation . For example , a targeted genotyping array/sequencing assay ( e . g . exome array ) may not provide sufficient genome-wide coverage for imputation . In addition , it is challenging to impute low frequency variants even with the highest quality reference panels . Imputed genotypes of low quality are often filtered out based upon the recommendations from the best practices [10] , since these variants are more prone to artefacts and can lead to inflated type I errors . Therefore , missing data in meta-analysis of genetic association studies are unavoidable . Some existing meta-analysis strategies can be highly biased in the presence of missing data . First , a commonly used method for conditional analysis , COJO , can lead to biased results when contributed summary association statistics from participating studies contain missing values [11] . The COJO method approximates the variance-covariance matrix between association statistics with the linkage disequilibrium ( LD ) information from a reference panel . When the association statistics from contributed studies are missing at some variant sites , the correlation matrix of the meta-analysis statistics can differ greatly from the LD matrix . Consider the simple example of a meta-analysis of two independent studies , where variant 1 is only measured in study 1 and variant 2 is only measured in study 2 . The meta-analysis association statistics for the two variants are independent , which cannot be approximated by the LD . COJO only uses meta-analysis results as input . Therefore , it cannot distinguish the scenario where only study 1 measures both variants ( and study 2 measures none ) , and the scenario where study 1 only measures variant 1 and study 2 only measures variant 2 . In the presence of missing data , COJO can be highly biased and lead to inflated type I errors . Second , the strategy of imputing missing data from contributed association statistics and using imputed association statistics in meta-analysis can also lead to inflated type I errors in conditional analysis . A simple imputation strategy for marginal ( or unconditional ) analysis is to replace missing summary statistics with zeros ( REPLACE0 ) , which are their expected value under the null hypothesis [2 , 3] . This method yields valid type I errors for marginal association analysis . Taking this simple approach for conditional analysis , however , is problematic . The genetic variants at conditioned sites are likely to have non-zero effects . Replacing missing summary data with zeros will bias the genetic effect estimates at conditioned variant sites , and can lead to highly inflated type I errors for conditional analysis ( see RESULTS ) . Similarly , the methods that seek to impute missing summary statistics based upon LD ( e . g . impG [12] ) may introduce substantial biases to the effects of missing variants . Plugging in the imputed Z-score statistics into conditional analysis ( impG+meta ) can lead to inflated type I errors . Finally , discarding studies with missing summary statistics ( DISCARD , or complete case analysis ) will give valid type I errors , but at the cost of reduced power . In the statistics literature , synthesis methods have previously been developed to meta-analyze joint effects from different studies , where the participating studies measure different predictors [13 , 14] . The scenario is similar to the meta-analysis of genetic association studies with missing data . Yet , in genetic association analysis , usually only marginal effects are reported and joint effects have to be approximated from marginal effects . The synthesis methods also lack an implementation for genetic association studies , which greatly limits their impact . To explore the usefulness of synthesis methods , we proposed and implemented an extension of the synthesis methods termed SYN+ , which can be applied in genetic association meta-analysis . To overcome these limitations of existing GWAS meta-analysis methods and improve power , we developed an improved conditional meta-analysis method called partial correlation based score statistic ( PCBS ) that borrows strength across multiple participating studies and consistently estimates the partial variance-covariance matrices between genotypes and phenotypes . We conducted extensive simulations , and showed that our PCBS method has valid type I error and the highest power among all the methods . On the other hand , COJO , impG+meta and REPLACE0 can lead to highly inflated type I errors in the presence of missing data . SYN+ , while having valid type I errors , is consistently less powerful than PCBS , especially when the missingness is high or the conditioned variants have larger effects . We also demonstrated the clear advantage of PCBS in the meta-analysis of cigarettes per day phenotype . PCBS identified many more independently associated variants from known loci , compared to alternative approaches . We implemented the proposed methods in the open-source software tools RAREMETAL [15] and R package rareMETALS and made them publically available ( https://genome . sph . umich . edu/wiki/Rare_Variant_Analysis_and_Meta-Analysis ) . RAREMETAL and rareMETALS use marginal score statistics and exact variance-covariance matrix as input , which is suitable for rare variant association analysis . We also implemented the same method in rareGWAMA ( https://github . com/dajiangliu/rareGWAMA ) , which conducts meta-analysis using approximate covariance matrix from a reference panel . These methods and tools have been applied and tested in a few large scale meta-analyses . We expect these methods to play an important role in sequence-based genetic studies and lead to important genetic discoveries .
We denote the genotype for individual i at variant site j in study k as Gijk , which can take values of 0 , 1 or 2 , representing the number of the minor ( or alternative ) alleles in the locus . When the genotypes are imputed or generated from low pass sequencing studies , genotype dosage can be used in association analysis . In this case , Gijk will be the expected number of minor ( or alternative ) allele counts . We denote the non-genotype covariates as Zik , which includes a vector of 1’s to incorporate the intercept in the model . Single variant association can be analyzed in a regression model: Yk = Gjkβj + Zkγk + ek . The score statistic for single variant association takes the form: Ujk=1σ^02∑iGijk ( Yik−y^ik ) ( 1 ) where y^ik=Zikγ^k , γ^k is the covariate effect , and σ^0 is the standard deviation of the phenotype residuals estimated under the null model M0 Yk=Zkγk+ek , ek∼MVN ( 0 , σ^02I ) ( M0 ) Without the loss of generality , we assume that the phenotype residuals are standardized in each study as in commonly done in practice . So σ^0 is often equal 1 in practice . We denote the vector of score statistics in a genetic region as Uk = ( U1k , … , UJk ) . The variance-covariance matrix between scores statistics is equal to Vk=1/σ^02[Gk′Gk−GkTZk ( ZkTZk ) −1ZkTGk] ( 2 ) For our illustration of the method , we focus on the analysis of continuous outcomes . Yet , the meta-analysis and conditional meta-analysis methods work for both continuous outcomes and binary outcomes . The meta-analysis score statistics and their covariance matrices are calculated using the Mantel-Haenszel method , i . e . U = ∑k Uk and V = ∑k Vk . The meta-analysis statistics can be used to estimate the joint effects for variants 1 , … , J , i . e . β^=V−1U . We denote the score statistics at candidate and conditioned variant sites as U= ( UG , UG* ) , where G and G*represent the genotypes from the candidate and conditioned variants respectively . The variance covariance matrix for U equals to V= ( VGVGG*VG*GVG* ) The conditional score statistic can be calculated by UG|G*= ( UG−VGG*VG*−1UG* ) σ^02/σ^c2 ( 3 ) where σ^c2 is the residual variance estimated from the conditional analysis model Yk=Gk*βG*+Zkγk+ek , ek∼MVN ( 0 , σ^c2I ) ( Mc ) After conditioning on the genotypes G* , the residual variance equals to σ^c2=σ^02 ( 1−1NUG*′VG*−1UG* ) . It is easy to verify that the variance of the conditional score statistics under Mc is equal to VG|G*= ( VG−VGG*VG*−1VG*G ) σ^02/σ^c2 ( 4 ) The single variant and gene-level tests in conditional analysis can be calculated based upon the conditional score statistics UG|G* and the covariance matrix VG|G* . Details are provided in S1 Text . Reviewing formulae ( 3 ) and ( 4 ) , we note that the conditional score statistics and their variances only depend on the partial variance-covariance matrix between the phenotypes and the genotypes after the adjustment of covariates . The key idea underlying our approach is to derive a consistent estimator for the partial covariances in the presence of missing summary statistics and to use it for unbiased conditional analysis . In statistics , to calculate the partial covariance between random variables Gjk and Yk adjusting for variable Zk , we first regress out covariate Zk from both Gjk and Yk , and then calculate the covariance between the residuals . Specifically , ρ^GjkYk|Zk=1Njkσ^02Gjk′ ( Yk−Zkγ^ ) ( 5 ) For a given study , it is easy to check that the partial covariances are in fact scaled score statistics , i . e . Therefore , in meta-analysis , we propose to estimate the partial covariance between genotype Gij , phenotype Yi after adjusting the covariate effect Zi using all available summary statistics: ρ^GY|Z , j=∑k∈{k:Mjk=1}Ujk∑k∈{k:Mjk=1}Njk ( 8 ) ρ^GG|Z , j1j2=∑k∈{k:Mj1k=Mj2k=1}Vj1j2k∑k∈{k:Mj1k=Mj2k=1}Njk ( 9 ) Here Mjk is an indicator variable that takes the value of 1 when the summary statistic at variant site j is measured in study k . For notational convenience , we define the matrices of partial covariance as ρ^GY|Z= ( ρ^GY , j ) j=1 , … , J and ρ^GG|Z= ( ρ^GG|Z , j1j2 ) j1 , j2=1 , … , J . Under the fixed effect model , we have E ( Vk−1Uk ) =β for all k . We showed in S1 Text that E ( ρ^GG|Z−1ρ^GY|Z ) =β . Therefore , the partial covariance matrices can be consistently estimated even in the presence of missing summary statistics . We define partial correlation based score statistics as U˜G|G*=ρ^GY|Z−ρ^GG*|Zρ^G*G*|Z−1ρ^G*Y|Z ( 10 ) The covariances for U˜G|G* are equal to V˜G|G*=cov ( ρ^GY|Z ) +ρ^GG*|Zρ^G*G*|Z−1cov ( ρ^G*Y|Z ) ρ^G*G*|Z−1ρ^G*G|Z−ρ^GG*|Zρ^G*G*|Z−1cov ( ρ^G*Y|Z , ρ^GY|Z ) −cov ( ρ^GY|Z , ρ^G*Y|Z ) ρ^G*G*|Z−1ρ^G*G|Z ( 11 ) It is easy to verify that the conditional analysis using the estimator U˜G|G* is equivalent to the standard score statistics when no missing data are present . In the presence of missing data , the partial correlation based statistic U˜G|G* remains consistent . The conditional association analysis can be performed by replacing the standard score statistic with a partial correlation based score statistic . Details for calculating single variant and gene-level conditional association statistics can be found in S1 Text . When the contributed summary association statistics from participating studies contain missing values , a natural strategy is to replace the missing values using imputation . Several imputation methods were previously developed . One method is REPLACE0 , which is to replace the missing values by 0 . We denote the resulting statistics as U0 and V0 . To mathematically describe this method , we define an indicator variable Mjk , which takes value 1 if the summary statistics at site j in study k is measured and 0 if missing . The meta-analysis score statistic is calculated by Uj0=∑k∈{k:Mjk=1}UjkandVj1j20=∑k∈{k:Mj1k=Mj2k=1}Vj1j2k We proved in S1 Text that replacing missing summary association statistics with zero will bias the genetic effect estimate , i . e . E ( UG*0 ) ≠VG*0βG* . As a consequence , under the null hypothesis that the candidate variant is not associated with the phenotype , the expectation of the conditional score statistics is not equal to 0 , i . e . E ( UG|G* ) =VGG*βG*−VGG*0 ( VG*0 ) −1E ( UG*0 ) ≠0 . The type I error for conditional analysis can be highly inflated . A more sophisticated set of methods is to impute missing summary statistics based upon LD information . Yet , the genetic effect estimates based upon the imputed Z-score statistics are often biased , unless the following condition holds E[Zimp]=Σimp , tagΣtag−1E[Ztag] where Zimp and Ztag are Z-score statistics at the missing and tagSNP sites , Σimp , tag and Σtag are genotype correlation matrices . A special case for this condition is that both the tagSNP and missing variants have null effects . Similar to REPLACE0 , applying impG+meta method can lead to inflated type I errors . We conducted extensive simulations to evaluate the performance of PCBS as well as 5 alternative approaches , including 1 ) impG+meta; 2 ) COJO; 3 ) REPLACE0; 4 ) DISCARD and 5 ) SYN+ using simulated data . We simulated genetic data following a coalescent model that we previously used for evaluating rare variant association analysis methods [2] . The model captures an ancient population bottleneck and recent explosive population growth . Model parameters were tuned such that the site frequency spectrum and the fraction of the singletons of the simulated data match that of large scale sequence datasets . For quantitative traits , phenotype data from each cohort were simulated according to the linear model: Yi=β0+∑j=1JGijβj+∑j=1JGij*γj+ϵi where Gij and Gij* denote the candidate and conditioned variant genotypes , and βj and γj are their effects respectively . The model assumes that the genetic variants have additive effects on the phenotype . The genetic effects for candidate variants follow a mixture normal distribution , which accommodates the possibility that a genetic variant can be causal ( with probability c ) or non-causal ( with probability 1 − c ) : βj∼ ( 1−c ) ×I ( 0 ) +c×N ( 0 , τβ2 ) . The genetic effects for the conditioned variants follow: γj∼N ( 0 , τγ2 ) . To evaluate the influence of missing data , we randomly chose a certain fraction ( 10% 30% or 50% ) of the sites from each study and masked them as missing . We then applied the new method PCBS , along with impG+meta , COJO , DISCARD , REPLACE0 and SYN+ to the data . In our evaluations , we used the exact LD with COJO and impG+meta , in order to remove the influence of approximate LD and focus on the impact of missing summary statistics on the power and type I error . We evaluated the type I errors and power for each approach under a variety of scenarios with different genetic effect sizes , fractions of causal variants in the gene region , and the fractions of missing data . To evaluate the effectiveness of methods in real datasets , we applied our methods to a meta-analysis of seven cohorts with a cigarettes-per-day ( CPD ) phenotype , a key measurement for studying nicotine dependence . Participating studies were the Minnesota Center for Twin and Family Research ( MCTFR ) [17–19] , SardiNIA[20] , METabolic Syndrome In Men ( METSIM ) [21] , Genes for Good [22] , COPDGene with samples of European ancestry[23] , Center for Antisocial Drug Dependence ( CADD ) [24] , and full UK Biobank . Genotypes were imputed using the Haplotype Reference Consortium panel [25] and the Michigan Imputation Server [26] ( with the exception of UK Biobank dataset , which was imputed centrally by the UK Biobank team ) . Summary association statistics from the seven cohorts were generated using RVTESTS [27] , and meta-analysis performed using rareMETALS with the PCBS statistics and other alternative approaches . Detailed descriptions of the cohorts are available in S1 Text section 4 , including the methods for association analyses and the adjusted covariates . To ensure the validity of our association analysis results , we conducted extensive quality control for the imputed genotype data . We filtered out variant sites with the imputation quality metric R2 < . 7 , and sites that showed large differences in allele frequencies from the imputation reference panel . Imputation dosages were used in the association analysis . For each sentinel SNP with genome-wide significance ( α = 5×10−8 ) , we defined the locus as the 1 MB window surrounding it . We applied iterative single variant conditional analysis to identify independently associated variants in each locus . We started by conditioning on the most significant variant from marginal association analysis . After each round of the association analysis , if the top variant remained statistically significant , we added the top variant to the set of conditioned variants , and performed an additional round of association testing . We applied the six methods to analyze the data , including the PCBS statistic , SYN+ , impG+meta , REPLACE0 , DISCARD and COJO . In order to examine if the low frequency variants in aggregate can be explained by the identified independently associated variants , we also performed gene-level association analysis for rare variants with MAF<1% , conditional on the identified independently associated variants .
We evaluated the type I errors for the six conditional analysis methods PCBS , SYN+ , COJO ( with exact LD ) , impG+meta , REPLACE0 , and DISCARD . Scenarios were considered for different combinations of the fractions of missing data , the genetic effects of the variants in the candidate gene , and the genetic effects of the conditioned variants . First , we noted that PCBS , SYN+ and DISCARD are the only three methods that have controlled type I errors across all scenarios , consistent with our theoretical expectation ( Table 1 ) . The type I error rate for the other three methods , i . e . impG+meta , REPLACE0 and COJO are inflated in a number of scenarios . The inflation tends to increase with the effect of the conditioned variant ( s ) and the rate of missingness . In many scenarios , the type I error can be >100X inflated over the significance threshold ( α = 5×10−8 ) . For example , when the conditioned variant effect is . 04 , and the association statistics from 30% of the variant sites are missing , type I errors for impG+meta , COJO and REPLACE0 are . 015 , . 57 and . 74 under the significance threshold of α = 0 . 005 . When the missing rate is 50% , and the conditioned variant effects is . 08 , the type I errors for the three methods become . 25 , . 65 , and . 60 . Second , among the methods with the controlled type I error rates ( i . e . SYN+ , PCBS and DISCARD ) , PCBS is consistently the most powerful method ( Table 1 ) . The power advantage of PCBS over the other two approaches increases when 1 ) the conditioned variant ( s ) have larger effects or 2 ) the fraction of missing summary association statistics is larger . For example , when candidate variant effect is . 04 , the conditioned variant effect is . 08 , and the missing rate of score statistics is 30% , the power for PCBS is . 21 , which is 75% higher than the power for SYN+ ( . 12 ) . When the candidate variant effect is . 08 , the conditioned variant effect is . 08 , and score statistics from 50% of the variant sites in each participating study are missing , the power for PCBS and SYN+ are respectively . 83 and . 74 . Due to the obvious limitations of complete case analysis , the DISCARD method of discarding the studies with missing data can lead to considerable loss of power ( Table 1 ) . The power for DISCARD is substantially lower than PCBS and SYN+ . In some scenarios where the missingness is high , the power is barely larger than the significance threshold . Interestingly , gene-level association tests are affected by two types of missing data with opposite consequences: Missing values at causal variant sites reduce power but missing values at non-causal variant sites tend to reduce noise and thus improve power ( Table 2 ) . When missingness is higher , the power of gene-level tests is lower , but the power loss is small . For instance , when a causal variant in the candidate gene has effects sampled from N ( 0 , 0 . 22 ) , the conditioned variant has effect . 1 , and 30% of the contributed summary statistics in each study have missing values , the power for burden/SKAT/VT tests are 58%/58%/56% , which are only slightly reduced compared to the power of analyzing the complete datasets ( 60%/61%/60% ) . On the other hand , the method that discards studies with missing data has much reduced power ( 0 . 011/0 . 011/8 . 8×10−3 ) . Our method was developed for the fixed effect meta-analysis , where the genetic effects are assumed to be constant across different studies . But since PCBS first aggregates association statistics from across studies and then performs conditional analysis , the impact of genetic effects heterogeneities does not invalidate the test and the type I error remains well controlled . The power is slightly reduced , but the advantages over other methods remain . To confirm this , we performed simulation analysis assuming that the genetic effects across studies are heterogeneous ( S1 Table , S2 Table ) . In our simulations , the genetic effects for a given variant in different studies were simulated from a normal distribution N ( μβG* , ( μβG*/2 ) 2 ) , allowing for substantial between-study heterogeneities . The power comparison for different methods remains similar to the scenarios where the genetic effects are the same across studies . We performed a meta-analysis of CPD phenotype in 7 cohorts . The locus CHRNA5-CHRNB4-CHRNA3 was previously identified as associated with CPD [28] . After careful quality control , 42 , 669 , 770 variants were meta-analyzed . A majority ( 32 , 796 , 258 ) of these variants had minor allele frequencies <1% . It is important to note that even with high quality imputation panels , such as the haplotype reference consortium panel [25] , there was still considerable missing data in the imputed datasets . A fraction of 76 . 1% of the variants were missing from at least one participating study post imputation , due to filtering on the imputation quality ( R2> . 7 ) . Compared to common variants , rare variants were considerably more likely to be missing: 95 . 3% of the variants with MAF<1% were missing from at least one cohort , compared to the fraction of 20 . 1% for the common variants with MAF>1% . The Quantile-Quantile plot for–log10 ( p-value ) is well calibrated ( S1 Fig ) . The genomic control value is 1 . 14 for common variants with MAF>0 . 01 , and 1 . 00 for rare variants with MAF<0 . 01 . The genomic control value is consistent with that of large scale GWAS for highly polygenic traits [29 , 30] . The intercept for LD score regression [31] was 1 . 01 , which shows little influence from potential population structure . The meta-analysis of 7 cohorts identifies 9 loci ( S2 Fig ) , including the well-known CPD associated loci , the nicotine receptor genes CHRNB2 , CHRNB3-CHRNA6 , CHRNA5-CHRNB4-CHRNA3 , the gene CYP2A6 that encodes cytochrome P450 protein , the gene PDE1C that encodes Phosphodiesterase 1C , FAM163B-DBH , YTHDF3 and GRM4 . Among these loci , CHRNB2 and FAM163B-DBH are associated with CPD at the genome-wide significance threshold for the first time . While smoking behaviors are known to be heritable , only the CHRNA5-CHRNB4-CHRNA3 and CYP2A6 loci have been consistently implicated in human GWAS to date . The other nicotine receptor gene CHRNB3-CHRNA6 was first identified with genome-wide significance in an isolated population for associations with nicotine dependence and nicotine use [32] . CHRNB2 was implicated in the nicotine dependence trait , but not at genome-wide significance . To our knowledge , there is no report that this gene is associated with CPD at genome-wide significance [33] . In order to understand the allelic architecture of the CPD phenotype and compare different methods on real data , we performed sequential forward selection with the new PCBS method , and identified 5 independently associated variants for the CHRNA5-CHRNB4-CHRNA3 locus and 4 independently associated variants for the CYP2A6 locus at genome-wide significance threshold ( with p-values < 5 × 10−8 ) ( Table 3 ) . The other loci do not have additional independently associated variants besides the sentinel variant . As a comparison , we also performed sequential forward selection using the five alternative approaches ( S3 Table ) . Using the SYN+ method , fewer independently associated variants are identified . At the CHRNA5-CHRNB4-CHRNA3 locus , 3 independently associated variants are identified , and also at the CYP2A6 locus , only 3 independently associated variants are identified . DISCARD also identifies fewer number of independently associated SNPs . The results from real data analysis is consistent with our simulation study that PCBS has higher power than alternative approaches . Among the approaches that have inflated type I errors in simulations , impG+meta identifies a lot of SNPs with very significant p-values . Many of these identified SNPs have substantial missingness among the participating cohorts ( e . g . N<50 , 000 ) . Given the inflated type I errors that we observed in simulations , as well as the small available sample sizes for the top variants , the validity of the results using impG+meta is of concern . Most of the top variants identified by COJO and REPLACE0 have low missingness , so there are not many false positive results . Yet , COJO and REPLACE0 identified fewer independently associated SNPs compared to PCBS and SYN+ ( Table 3 and S3 Table ) . Together , the analysis of real data confirmed our simulation experiments . We examined if our independently associated variants explained previously known association signals . To do this , we looked up GWAS catalog [34] using key words “CPD” or “cigarettes per day” and found 11 associated variants in the loci that we identified ( S4 Table ) . We first analyzed these 11 variants conditional on our independently associated variants . All of these variants became insignificant , which indicated that our newly identified independently associated variants can explain previously known association signals . We also performed conditional analysis in the opposite direction to examine if our identified association signal may be explained by the known variants . We found that variants within the CPY2A6 locus remained highly significant and variants within the CHRNA5-CHRNB4-CHRNA3 locus remained marginally significant . Together , our independently associated variants explained 1 . 1% of the phenotypic variance , which substantially improves the phenotypic variance ( . 64% ) explained by the 11 known signals . Finally , in addition to single variant association , we investigated if rare variants within each of the 9 loci were independently associated with the CPD phenotype ( S5 Table ) . 27 genes were analyzed using simple burden , SKAT and VT tests under a MAF threshold of 0 . 01 . Only one gene ( CHRNA5 ) has gene-level p-values less than 0 . 05/27 , which is the Bonferroni threshold . None of the genes have exome-wide significant gene-level association p-values .
We proposed a simple yet effective meta-analysis method to estimate joint and conditional effects of rare variants in the presence of missing summary statistics from contributing studies . The method leads to the optimal use of shared summary association statistics . It has well controlled type I error and much higher power than alternative approaches even when a large number of contributing studies contain missing summary statistics . Several approaches were previously developed to combine genetic effects across studies when different studies may measure different genetic variants e . g . Verzilli et al [35] and Newcombe et al [36] . These methods have some noticeable limitations . The method by Verzilli et al requires the individual level genotype and phenotype data as input . Also the method focuses on random effects meta-analysis , while our approach focuses on fixed effect meta-analysis . The method by Newcombe et al models the haplotype counts in cases and controls . The method does not allow for the adjustment of covariates , which is a serious limitation . Both methods use MCMC for fitting the model , which may not scale well for contemporary meta-analysis with tens of millions of variants and dozens of studies . It is important to note that our method , PCBS is developed for proper conditional and joint analysis when imputation fails to work . As we showed in our meta-analysis of smoking phenotypes , even with the state-of-the-art imputation methods and high quality reference panels , there are still considerable amount of association statistics filtered out from participating studies . The rate of missingness is much higher for rare variant association statistics than for common variant association statistics . PCBS will be particularly useful for the meta-analysis of sequence data , where the measured variants are predominantly low frequency or rare [37] . Our method is not developed to replace genotype imputation . Genotype imputation fills in missing genotypes with imputed values , and increases effective sample sizes and power . Our method does not increase the effective sample size for tested variants . In practice , imputation method should first be applied in each participating cohort . Our method should be applied at the meta-analysis stage for valid and powerful conditional meta-analysis , especially when contributed summary statistics from participating cohorts contain missing values . Missing data will continue to be a persistent issue in the next generation of large-scale genetic studies . Major biobanks have started to develop their own genotyping arrays and imputation reference panels to incorporate customized content . Combining these newly genotyped studies with existing datasets will result in missing summary statistics . Our method will continue to be useful when analyzing these newly generated datasets . Another major application of the proposed method is in the meta-analysis of sequence data . Given the use of targeted sequencing assays and variability in batch processing and quality control across studies , it would be difficult to impute missing genotype data or missing summary statistics . One of the challenges in sequence-based meta-analysis is to properly represent monomorphic sites , as the polymorphic variant sites are not known a priori . Neither un-called variant sites ( e . g . due to insufficient coverage or failed quality control ) nor monomorphic sites contribute to the single variant meta-analysis statistic . Yet they should be treated differently in joint and conditional meta-analysis . Summary statistics from monomorphic variants should be replaced by zeros . On the other hand , summary statistics from un-called variants should be treated as missing data , and the conditional association analysis can be performed using our partial correlation based score statistics . While not the focus of this article , the proposed method is also helpful for downstream analyses that make use of the joint effects of multiple variants , e . g . estimating the phenotypic variance explained by variants in LD or fine mapping causal variants ( e . g . using methods such as RIVERA [38] , FINEMAP [39] , CAVIARBF [40] ) The validity of these analyses relies critically on the proper estimates of the joint effects , which are usually obtained from single variant association statistics and the LD information from a reference panel . When summary statistics from contributing studies contain missing data , the correlations between resulting marginal meta-analysis association statistics may not be properly approximated by the LD estimated from a reference panel . In this case , PCBS can be used to obtain valid joint effect estimates , which can potentially lead to better calibrated estimates phenotypic variance explained and more accurate fine mapping analysis . Taken together , our partial correlation based score statistic is a simple yet effective method for estimating joint and conditional effects from a meta-analysis . With its efficient implementations in RVTESTS , RAREMETAL and rareGWAMA , this method will have broad application in current array-based meta-analysis , as well as the upcoming imputation-based meta-analysis ( e . g . based upon the haplotype reference consortium panel ) and sequence-based meta-analysis . Correct inference on the joint and conditional effects using these methods will pave the way for a more accurate characterization and a more complete understanding of the genetic architecture of complex traits .
|
It is of great interest to estimate the joint effects of multiple variants from large scale meta-analyses , in order to fine-map causal variants and understand the genetic architecture for complex traits . The summary association statistics from participating studies in a meta-analysis often contain missing values at some variant sites , as the imputation methods may not work well and the variants with low imputation quality will be filtered out . Missingness is especially likely when the underlying genetic variant is rare or the participating studies use targeted genotyping array that is not suitable for imputation . Existing methods for conditional meta-analysis do not properly handle missing data , and can incorrectly estimate correlations between score statistics . As a result , they can produce highly inflated type-I errors for conditional analysis , which will result in overestimated phenotypic variance explained and incorrect identification of causal variants . We systematically evaluated this bias and proposed a novel partial correlation based score statistic . The new statistic has valid type-I errors for conditional analysis and much higher power than the existing methods , even when the contributed summary statistics contain a large fraction of missing values . We expect this method to be highly useful in the sequencing age for complex trait genetics .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] |
[
"genome-wide",
"association",
"studies",
"variant",
"genotypes",
"random",
"variables",
"covariance",
"genetic",
"mapping",
"mathematics",
"statistics",
"(mathematics)",
"genome",
"analysis",
"research",
"and",
"analysis",
"methods",
"genome",
"complexity",
"genomics",
"mathematical",
"and",
"statistical",
"techniques",
"genetic",
"loci",
"research",
"assessment",
"probability",
"theory",
"phenotypes",
"heredity",
"meta-analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"computational",
"biology",
"research",
"errors",
"statistical",
"methods",
"introns",
"human",
"genetics"
] |
2018
|
Proper conditional analysis in the presence of missing data: Application to large scale meta-analysis of tobacco use phenotypes
|
Although they have become a widely used experimental technique for identifying differentially expressed ( DE ) genes , DNA microarrays are notorious for generating noisy data . A common strategy for mitigating the effects of noise is to perform many experimental replicates . This approach is often costly and sometimes impossible given limited resources; thus , analytical methods are needed which increase accuracy at no additional cost . One inexpensive source of microarray replicates comes from prior work: to date , data from hundreds of thousands of microarray experiments are in the public domain . Although these data assay a wide range of conditions , they cannot be used directly to inform any particular experiment and are thus ignored by most DE gene methods . We present the SVD Augmented Gene expression Analysis Tool ( SAGAT ) , a mathematically principled , data-driven approach for identifying DE genes . SAGAT increases the power of a microarray experiment by using observed coexpression relationships from publicly available microarray datasets to reduce uncertainty in individual genes' expression measurements . We tested the method on three well-replicated human microarray datasets and demonstrate that use of SAGAT increased effective sample sizes by as many as 2 . 72 arrays . We applied SAGAT to unpublished data from a microarray study investigating transcriptional responses to insulin resistance , resulting in a 50% increase in the number of significant genes detected . We evaluated 11 ( 58% ) of these genes experimentally using qPCR , confirming the directions of expression change for all 11 and statistical significance for three . Use of SAGAT revealed coherent biological changes in three pathways: inflammation , differentiation , and fatty acid synthesis , furthering our molecular understanding of a type 2 diabetes risk factor . We envision SAGAT as a means to maximize the potential for biological discovery from subtle transcriptional responses , and we provide it as a freely available software package that is immediately applicable to any human microarray study .
Since their inception over 13 years ago [1] , DNA microarrays have become a staple experimental tool used primarily for exploring the effects of biological interventions on gene expression . Microarrays have enabled a range of experimental queries , including a survey of gene expression across dozens of mammalian tissues [2] , a comparison of human cancers in over 2000 tumor samples [3] , and the identification of differentially expressed ( DE ) genes between pairs of conditions . Identifying DE genes is especially common , as it is often the first means of characterizing differences between two poorly understood conditions . As of 2009 , there are publicly available microarray data for human conditions ( at the Gene Expression Omnibus [4] ) . These data make possible a huge number of pairwise comparisons for DE gene analysis . Given this sizable opportunity for biological discovery , we focus our attention on the task of DE gene identification . Microarrays are notorious for generating noisy or irreproducible data [5]–[8] . This is partially due to the inherent technical noise of the experiment , which can be modeled and often removed from the resulting data . However , biological noise also plays a significant role , and effects of this noise source are not as easily corrected [9] . A common solution to biological noise involves replicating the experiment many times in order to “average out” noise effects . In the context of DE gene prediction , we define a replicate as a biologically independent comparison of RNA levels between the experimental conditions of interest . Unfortunately , assay cost and a limited supply of biological material often limit the efficacy of a replication-based strategy . To circumvent these difficulties , we need analytical methods which increase DE gene prediction accuracy at no additional cost . One inexpensive source of microarray replicates comes from prior experiments . In the last decade , researchers have generated data from hundreds of thousands of microarrays , and many of these are publicly available at repositories like the Gene Expression Omnibus ( GEO ) . It is unlikely that any of these arrays ( hereafter referred to as “knowledge” ) represent exact replicates of data from a novel study ( referred to as “data” ) , but a subset of these experiments may describe similar underlying biology and could be considered “partial replicates” . Because it is not clear a priori which of the prior experiments ( if any ) would qualify as partial replicates , pre-existing microarray knowledge cannot be used directly to identify DE genes in a novel dataset . It is therefore worth considering indirect methods for using this knowledge . Two previously existing methods use microarray knowledge to compute more accurate variance estimates for each gene [10] , [11] . Both methods replace sample variance estimates for each gene by gene-specific variances calculated across a compendium of microarrays from GEO . This approach was shown to be most useful with small data sample sizes , and no further benefits were seen when the microarray knowledge exceeded arrays . A different approach might involve identifying transcriptional modules: groups of genes that exhibit coordinated or correlated expression changes across a range of conditions . A complete and accurate understanding of module structure would reveal expression dependencies between genes , such that on average , genes in the same module would be coexpressed more often than genes chosen at random . Thus , knowledge of one gene's expression would confer information about the expression of the other genes in the module . Several studies [12]–[18] have used microarray knowledge to identify transcriptional modules . Of these , five have been tested on yeast datasets of 1000 arrays or fewer [12]–[14] , [16] , [17] and one has been applied to human cancer datasets [15] . Only one [18] was applied to a diverse human microarray knowledge set , in this case containing arrays . Given that tens of thousands of arrays are publicly available for some individual microarray platforms , a larger-scale identification of transcriptional modules is certainly possible . Knowledge of transcriptional modules and their constituent genes is not directly applicable to DE gene identification , and most existing methods ignore these relationships . Of the few that provide a means to incorporate expression modules [19]–[21] , none provide a mechanism for extracting these modules from large-scale microarray knowledge sets . Consequently , there is a need for a method that can identify relevant transcriptional modules from huge compendia of microarray knowledge and use this information to better predict DE genes . In this work , we present the SVD Augmented Gene expression Analysis Tool ( SAGAT ) , a mathematical approach that identifies expression modules from microarray knowledge and combines these with novel data to identify DE genes . To accomplish these tasks , SAGAT employs Singular Value Decomposition ( SVD ) in concert with pseudoinverse projection . SVD has been used previously to decompose microarray knowledge into mathematically independent transcriptional modules ( eigengenes ) and the corresponding independent cellular states where these modules are active ( eigenarrays ) [22] . Most non-SVD module-finding methods identify discrete modules where module membership for each gene is a binary feature . In contrast , SVD assigns a continuously-valued weight for each gene , which allows varying strengths of coexpression to be present in the same module and genes to be part of multiple modules . SVD models the expression of each gene as a linear combination of the eigengenes' expressions , and a number of studies have used this technique to define modules on smaller scales . Raychaudhuri et al . [23] and Alter et al . [22] each initially applied SVD ( the former in the form of PCA ) to yeast time course data to identify fundamental modes of expression response that vary over time . The latter study also demonstrated the ability of SVD to remove noise or experimental artifacts present in the data . Shortly thereafter , Troyanskaya et al . [24] used SVD to identify eigengenes in gene expression data for the purposes of missing value estimation . Alter and colleagues subsequently employed generalized [25] and higher order [26] versions of SVD for the integration and decomposition of heterogeneous microarray datasets . Horvath and Dong [27] used SVD of microarray data in combination with coexpression analysis to generate eigengene coexpression networks . Finally , in a large scale study , SVD was shown to reduce noise when used in the integration of disparate microarray datasets [28] . The technique of pseudoinverse projection has also previously been applied to genome-scale data . Alter and Golub demonstrated the utility of SVD coupled with pseudoinverse projection by reconstructing one genomic dataset in terms of the eigenarrays of another [29] . This enabled the observation of a set of cellular states in one dataset that were also manifested in the other . Subsequent work used pseudoinverse projection in concert with an alternative matrix decomposition technique ( non-negative matrix factorization ) to classify gene expression states of one organism in terms of another [30] . In the current work , using SAGAT , we combine SVD-derived modules , pseudoinverse projection , and a rigorous statistical model to adjust gene expression error estimates in a dataset of interest . This yields a knowledge-informed differential expression score for each gene . We demonstrate SAGAT in several ways . First , we investigate whether transcriptional modules are readily detectable in a large compendium of microarray knowledge . Second , we test SAGAT on a range of simulated datasets to assay its performance with respect to a known gold standard . Third , we evaluate SAGAT's ability to increase DE gene predictive power in three highly replicated real world datasets . Finally , we apply SAGAT to a new human dataset investigating transcriptional profiles in the setting of insulin resistance ( IR ) , a risk factor for type 2 diabetes . Though a known relationship exists between obesity and insulin resistance [31] , [32] , it is not always consistent [33] , [34]; in addition , many studies characterizing IR do not deconvolve the effects of obesity [35] . This novel microarray dataset builds upon previous work [35]–[37] to investigate obesity-independent transcriptional effects of insulin resistance . We illustrate the improved sensitivity of SAGAT over existing methods by identifying IR candidate DE genes , and we validate a subset of these using quantitative PCR assays . Results of this analysis contribute to a more comprehensive molecular understanding of human insulin resistance .
To demonstrate that transcriptional modules are detectable in a multi-condition microarray knowledge compendium , we characterized the degree of modularity in a collection of 4440 arrays from the HGU95Av2 platform . We consider an expression module a group of genes exhibiting coordinated expression across some subset of the entire compendium . Genes in such a group will have relatively large positive or negative pairwise covariances; thus , degree of modularity refers to the number of genes in the compendium that belong to one or more groups of significantly covarying genes . Figure 1A displays a binarized representation of the sample covariance matrix for the entire HGU95Av2 compendium , whereby each covariance value whose magnitude is is colored black ( white otherwise ) . This matrix was then subjected to hierarchical biclustering ( Figure 1B ) , which resulted in many blocks of nonzero binary covariance , ranging in size from a few genes to nearly 1000 . Furthermore , this covariance pattern does not appear to be due to chance , as the biclustering results from 100 randomized knowledge matrices ( see Materials and Methods ) showed no covariance blocks exceeding a 15 gene cutoff . To parameterize a simulation study ( details below ) , we used a 1000-gene compendium to characterizee the mean number of genes per module , the mean percentage of DE genes found in modules , and the mean percentage of non-DE genes found in modules . This was achieved by subsetting the HGU95Av2 compendium and coupling it with a human prostate cancer dataset [38] . The mean number of genes per module was 15 , the mean percentage of DE genes found in modules was 60% ( 673/1122 ) , and the mean percentage of non-DE genes found in modules was 47% ( 3752/7983 ) . These values were employed in the simulation study that follows . SVD identifies eigengenes whose expression is mutually orthogonal across all arrays in the compendium . To demonstrate that mathematical orthogonality correlates with biological orthogonality ( as manifested by biologically independent eigengenes ) , we performed a Gene Ontology ( GO ) term enrichment analysis of a subset of the eigengenes from the HGU95Av2 compendium ( using the gene weights of each eigengene as scores ) . Table 1 displays the top three significant Biological Process terms with fewer than 500 annotated genes for eigengenes 1–5 , 10 , 20 , 50 , and 200 . The terms within each eigengene are largely consistent , and each eigengene describes a relatively distinct biological process . We note that there is not an absolute correspondence between the modules displayed in Figure 1B and the eigengenes identified by SVD , as the methods used to identify these structures are algorithmically different . However , we detected substantial overlap in the enriched Biological Process terms associated with the largest covariance modules and highest ranking eigengenes ( e . g . the largest module and first eigengene were both strongly enriched for translation and biosynthesis terms ) . We first tested the validity of the SAGAT model using simulated data . We simulated knowledge compendia with structures ranging from that shown in Figure 2A , where 60 of the 100 DE genes are in 15-gene modules and none of the 900 non-DE genes are , to that shown in Figure 2C , where the same number of DE genes are in modules and all 900 non-DE genes are also . Figure 2B depicts a modularity structure that is approximately equivalent to that of the prostate cancer dataset , where 60% of prostate cancer DE genes are found in modules and 47% of non-DE prostate cancer genes are found in modules . After running SVD on each simulated compendium to calculate the appropriate W matrix , we tested SAGAT on all combinations of data and knowledge . As SAGAT relies on a single parameter specifying the number of eigengenes ( M ) , we first estimated the optimal value for this parameter by trying all possible values on several configurations of data and knowledge ( results not shown ) . The best performance was achieved with ; we used this value for all subsequent simulation runs . Figure 3 displays results from running SAGAT on two compendia with modularity structures identical to Figures 2A ( 3A , 3B ) and 2B ( 3C , 3D ) coupled with datasets having either one or 15 replicates . The mean AUC improvement over the fold change metric ( Mean ) , ranging from . 0042 to . 0708 , is shown . Within the range of structures bounded by these two compendia and for both sample sizes , SAGAT consistently improves the AUC of DE gene prediction . The two trends observed are: ( 1 ) increasing performance improvement with decreasing numbers of array replicates , and ( 2 ) increasing performance improvement with decreasing numbers of non-DE gene modules . Performance begins to degrade below that of fold change if the simulated compendia adopt modularity structures between those of Figures 2B and C ( results not shown ) , but we have evidence suggesting that the modularity of real world datasets resemble configurations falling between Figures 2A and B ( see Discussion ) . To demonstrate that use of SAGAT could yield improved statistical power without concurrently increasing the false positive rate of prediction , we repeated the above experiments using true positive rate ( TPR ) evaluated at a fixed false positive rate ( FPR ) of . 05 ( in place of AUC ) . These results are shown in Figure S3 , and the performance improvements with respect to fold change closely resemble those displayed in Figure 3 . To evaluate SAGAT performance on real data , we tested it on subsets of three highly replicated human microarray datasets ( see Materials and Methods for details ) . As a gold standard , we used either the fold change or limma t [39] metrics to identify significant DE genes from each dataset in its entirety; this resulted in 1122 ( 12 . 3% ) , 588 ( 4 . 4% ) , and 6002 ( 29 . 9% ) DE genes for the prostate cancer , letrozole treatment ( GEO ID: GSE5462 ) , and colorectal cancer ( GSE8671 ) datasets , respectively . After downloading the three corresponding knowledge compendia ( minus the highly replicated datasets ) and running SVD on each , we determined the optimal number of eigengenes by training on the prostate cancer dataset . Figure 4 shows results of SAGAT run on two non-overlapping subsets of this dataset and the HGU95Av2 compendium while varying the number of eigengenes ( parameter M ) . The AUCs of the fold change metric are displayed as red horizontal lines . SAGAT outperforms fold change for many values of M , and for both subsets there is a distinct maximum in the AUC curve for a particular value of the parameter . For these two subsets and several others tested ( not shown ) , the optimal value for M is approximately half the number of arrays in the compendium . We used this value for subsequent analyses on all datasets and compendia , which translates to 2220 , 7238 , and 6108 eigengenes for the HGU95Av2 , HGU133A , and HGU133plus2 . 0 platforms , respectively . To show that SAGAT's performance as a function of M was not due to chance , we randomized the expression values of the compendium and re-ran the same test in Figure 4B . These results are shown in gray . In this case , SAGAT never outperforms fold change , suggesting that the performance improvement from the original compendium is not spurious . Next we applied SAGAT to multiple subsets of each of the three datasets . Figure 5 displays the performance of SAGAT coupled with the appropriate W matrices . For comparison , we feature AUC differences with respect to fold change of both SAGAT and the limma t-statistic . Figures 5A and B display performance on the prostate cancer dataset using a fold change and limma t-derived gold standard , yielding mean AUC improvements of . 023 and . 018 , respectively . Given that the relative performance trends are similar , Figures 5C and D show performance on the letrozole treatment and colorectal cancer datasets using only the fold change-derived gold standard , yielding AUC improvements of . 009 and . 019 , respectively . In all three datasets , irrespective of sample size , SAGAT nearly always improves the AUC over fold change; in cases where this does not occur , AUC is left essentially unchanged . In contrast , the t-statistic consistently lowers the AUC of DE gene prediction and is not applicable when the number of replicates is 1 . Though the limma t performance improves when using a limma t gold standard , it is still unable to outperform the other two metrics . AUC improvement for SAGAT generally decreases with increasing sample size , and the improvement is largest for the prostate cancer and colorectal cancer datasets . To express the performance of SAGAT in a more tangible form , we estimated the effective number of arrays added by using the method . Table 2 shows results for each of the three highly replicated datasets at four initial sample sizes . On average , with one exception in 12 tests , use of SAGAT always increased the effective number of arrays . In some cases , this improvement was quite significant: a two-array prostate cancer subset coupled with SAGAT effectively performed as well as a 4 . 72 array dataset . As before , the number of arrays added generally decreases with increasing sample size . As with the simulated data , we also repeated the highly replicated dataset experiments using TPR calculated at an FPR of . 05 as an evaluation metric . These results are displayed in Figure S4 , and the performance improvements very closely resemble those shown in Figure 5 . We evaluated the GEO method ( both standard and “voting” methods ) on the prostate cancer dataset and HGU95Av2 compendium and compared its performance to SAGAT . Figure S1 shows the results , which demonstrate that SAGAT ( and fold change ) outperform the GEO method in much the same way as when compared to the limma t-statistic above . We also measured the sensitivity of SAGAT performance to compendium size . As Figure S2 shows , SAGAT continues to improve performance as the compendium increases to its full size . The performance begins to level off near 4400 arrays , but further improvement would still be expected with an even larger compendium . Given encouraging performance of SAGAT on simulated and real human datasets , we applied it to an unpublished experimental dataset investigating expression differences between human insulin resistant and insulin sensitive adipose tissue . The obesity-independent relationship between insulin resistance and adipose gene expression has previously been characterized on a small scale [40] , but no large-scale studies have attempted to decouple the effects of obesity from insulin resistance [35] . In this experimental design , patients were otherwise healthy and matched for levels of obesity; thus , we expected to identify more subtle expression changes associated with insulin sensitivity status . As detailed in Materials and Methods , the same 12 pairs of RNA samples were applied to three different microarray platforms: Affymetrix , Agilent , and Illumina . We initially attempted to identify DE genes using the limma t metric on data from each platform individually . After correcting the results for multiple tests , we did not detect any significant genes at a . 05 FDR cutoff . Next , we integrated results from all three platforms to try to capture subtle but consistent signals . We applied the method of Rank Products ( RP ) [41] to lists of genes ranked by either fold change or SAGAT . Table 3 shows results from this procedure . As we wanted to evaluate only the most confident predictions , we corrected for multiple testing by controlling the PFER ( per family error rate ) . This is a strict multiple hypothesis test correction method that is generally more conservative than the FDR ( false discovery rate ) or FWER ( family wise error rate ) [42] . A total of 19 genes were found to be significantly DE at a PFER of . 05 . When ranking genes by fold change before applying RP , 12 genes were found to be significantly DE—five upregulated and seven downregulated . When using SAGAT to rank the genes instead , 18 genes were significantly DE—seven upregulated and 11 downregulated . SAGAT with RP detected all but one of the genes found using fold change with RP , and seven genes were identified only through use of SAGAT . We refer to the 11 genes detected by both fold change and SAGAT rankings as Group I; Group II genes are those that were detected exclusively using SAGAT . We searched the literature for evidence implicating the genes of Table 3 in insulin resistance , diabetes , or fatty acid metabolism ( an important function of adipose tissue ) . Genes for which evidence was found are marked with an asterisk . Four of the Group I genes [FOSB ( Entrez Gene ID: 2354 ) , FADS1 ( 3992 ) , SELE ( 6401 ) , PPBP ( 5473 ) ] had some literature describing their involvement; five of the Group II genes [ATP1A2 ( Entrez Gene ID: 477 ) , FASN ( 2194 ) , FOS ( 2353 ) , CXCR4 ( 7852 ) , ELOVL6 ( 79071 ) ] were also implicated . To experimentally validate these candidates , we performed quantitative RT-PCR ( qPCR ) using 23 of the original 24 RNA samples subjected to an amplification reaction . We tested 11 of the 19 genes from Table 3: five from Group I and six from Group II . We also tested four genes that were not significant by Rank Products; these genes serve as negative controls . For each gene , we calculated the mean fold change over the -actin ( Entrez Gene ID: 60 ) housekeeping gene for the insulin resistant and insulin sensitive samples . Results are displayed in Figure 6 . Of the Group I and II genes tested , all had qPCR expression differences that matched the direction of those identified using Rank Products . We then tested the significance of each gene's expression difference using a Wilcoxon rank-sum test . Three of the genes had p-values smaller than a . 05 threshold: CSN1S1 ( Entrez Gene ID: 1446 ) , FOSB , and CXCR4 ( marked by asterisks in Figure 6 ) . The first two genes are from Group I; the third is from Group II . Of the four negative controls tested , none were found significantly different in expression between the two groups .
In this work , we present SAGAT , a principled method for integrating pre-existing microarray knowledge with a dataset of interest to identify DE genes . From prior knowledge , SAGAT extracts “eigengenes” , or mathematically independent transcriptional modules , which collectively describe observed expression dependencies between genes . These dependencies are combined with the expression changes of each gene in the data to form the SAGAT score , which enables expression information to be shared between genes that are coexpressed in the knowledge . To validate SAGAT , we first demonstrated that a compendium of microarray knowledge showed significant modularity . This result , which was not sensitive to varying compendium sizes ( not shown ) , was not surprising , as it has been shown before on knowledge sets of a smaller scale . Nevertheless , it was not clear whether such modules would be detectable on a much larger and more heterogeneous collection of microarrays . Next , we demonstrated favorable SAGAT performance in identifying DE genes on a series of simulated datasets . We note that our model for simulating data represents an oversimplification of realistic coexpression relationships between genes ( see Materials and Methods ) , but with it we can create distinct numbers of modules in DE and non-DE genes to test the limits of SAGAT performance . As detailed in the Results , SAGAT most improves performance with respect to the fold change metric when transcriptional modules are only composed of DE genes . As the number of non-DE gene modules increases , the performance improvement decreases , but at a realistic ratio of DE gene modules to non-DE gene modules ( Figure 2B , which closely matches the configuration of the prostate cancer dataset ) , SAGAT still outperforms fold change for all numbers of replicates tested . We evaluated SAGAT on three highly replicated microarray datasets . We chose datasets with many replicates so we could approximate a gold standard DE gene list for each one . Ideally , results from an independent and more accurate experiment like quantitative RT-PCR would provide the DE gene truth for a given dataset , but quantifying expression differences of every gene on a microarray would be prohibitively expensive . Instead , we assume that for each of the three datasets , the number of replicates is large enough that DE genes calculated using fold change on all arrays is approximately correct . Then the task becomes using small ( often noisy ) subsets of each dataset to predict the true DE genes . We applied the fold change , limma t , and SAGAT metrics to multiple non-overlapping subsets of varying numbers of replicates . SAGAT always outperforms the t-statistic , often by a large margin . With sample sizes of only 1 replicate , the limma t is not applicable as it requires a fold change variance estimate . Compared to fold change , SAGAT nearly always better identifies DE genes; in the worst case it leaves performance unchanged . These results suggest that SAGAT would be consistently beneficial for predicting DE genes from a dataset of interest . Importantly , the results displayed in Figure S4 demonstrate that use of SAGAT leads to improved statistical power at a small fixed false positive rate , which is a necessity for the effective analysis of high-throughput biological experiments . We expressed SAGAT's performance improvement over fold change in terms of the effective number of arrays added . This shows that , except in a small number of cases , use of SAGAT always increases the effective sample size of an experiment . In some cases this increase is substantial: for one two-array subset of the prostate cancer dataset , the effective sample size became 4 . 72 arrays , or more than double the initial sample size of the experiment . As expected , the number of arrays added decreases as the initial number of arrays increases , due in part to the lower capacity for prediction improvement when starting with a larger sample size . We also demonstrated that SAGAT outperforms the related GEO method when evaluated on the prostate cancer dataset . As even the fold change method consistently outperforms the GEO method , it appears that more accurate estimation of gene variances is not the most effective way to improve performance for this dataset . In contrast , use of gene module information from an SVD of microarray knowledge gives consistent improvement over fold change . We determined the sensitivity of SAGAT performance to the number of arrays in the knowledge compendium . It was shown in [11] that the GEO method does not give further performance improvement when knowledge exceeds arrays . To compare , we evaluated the effect of compendium size on SAGAT performance using the prostate cancer dataset . Unlike the GEO method , SAGAT continues to improve performance as the compendium increases in size . The improvement starts leveling off near the compendium's full size ( 4400 arrays ) , but an even larger knowledge compendium should still give better performance . Thus , SAGAT is able to extract useful information from much larger microarray compendia than the GEO method . Given SAGAT's potential to improve DE gene identification , we applied the method to a novel insulin resistance dataset obtained from three different microarray platforms . An initial attempt to identify DE genes on each platform separately yielded no candidates , suggesting that the transcriptional response in question was noisy and/or subtle . A Gene Ontology term enrichment analysis on data from each platform consistently identified terms related to immune response ( results not shown ) , implying that a reproducible biological signal was present in the data . To improve the signal to noise ratio at the gene level , we used the method of Rank Products ( RP ) across all three platforms to identify subtly but consistently changing DE genes . An application of RP to genes ranked by fold change yielded 12 DE gene candidates with a per family error rate of . 05 or smaller . A similar analysis on genes ranked by SAGAT yielded 18 genes , 11 of which overlapped with the fold change list . This suggests that the incorporation of transcriptional module information resulted in an increased sensitivity to detecting DE genes . We intentionally used a very strict significance threshold to select a small number of DE genes that were most consistently changed ( and which hopefully represent true biological differences ) , but relaxation of this threshold would lead to additional candidates . We next performed a literature search on each significant gene for information implicating it in insulin resistance , diabetes , or fatty acid metabolism . This uncovered evidence for multiple genes from three biological processes: inflammation [SELE , IL6 ( Entrez Gene ID: 3569 ) , PPBP , CXCR4] , cell differentiation [FOSB , FOS] , and fatty acid synthesis [FADS1 , FASN , ELOVL6] [43]–[45] . A role for inflammation in IR has previously been suggested by a similar study [35] , but of the four pro-inflammatory genes listed above only IL6 was also detected in that work . In this study , SELE , IL6 , and CXCR4 were upregulated in insulin resistant patients , reinforcing the positive role of inflammation in IR . Cell differentiation has also been implicated in insulin resistance in the sense that insulin resistant adipose tissue displayed lower expression of differentiation markers than their insulin sensitive counterparts [37] . In this work FOSB and FOS were upregulated in IR , which is compatible with the above since both gene products have been shown to trigger de-differentiation [46] , [47] . Fatty acid synthesis has long been known to be relevant to insulin resistance [48] . The details of this relationship are not always consistent: FADS1 is known to be downregulated in IR [49] , while ELOVL6 has shown the opposite effect [50] and FASN has shown conflicting results [51] . To our knowledge , no single study has analyzed the effects of all three of these fatty acid synthesis genes with respect to insulin resistance in adipose tissue . Our results show a coherent decrease in the gene expression of all three genes , suggesting that obesity-independent insulin resistance is associated with altered fatty acid synthesis and storage in adipose tissue . We speculate that such an occurrence may lead to inappropriate fatty acid accumulation elsewhere ( i . e . circulating in serum ) , which has been known to lead to IR [51] . One explanation for the inconsistent results in previous studies is the potentially confounding effects of obesity ( a condition where fatty acid synthesis increases ) and insulin resistance . The current study explicitly attempts to remove the former effect . Taken together , the above results emphasize the importance of increased inflammation , differentiation , and decreased fatty acid synthesis to adipose tissue-based insulin resistance . We note that our confidence in this assertion was greatly helped by SAGAT , as four of the nine genes involved in these processes were only identified using this method . This is particularly true for genes like CXCR4 , whose PFER received a substantial boost upon application of SAGAT ( 0 . 6058 to 0 . 0311 ) . We expect that further experimentation will reveal the precise relationships between these processes and IR . The remaining significant genes detected only by SAGAT exhibited varying levels of insulin resistance-related literature evidence . ATP1A2 , which codes for an ATPase , was previously found to be differentially expressed between insulin resistant and insulin sensitive muscle tissue , though in the opposite direction than was found in this study [52] . PMP2 ( Entrez Gene ID: 5375 ) and SRGN ( 5552 ) , coding for a myelin protein and hematopoietic proteoglycan , respectively , lack any literature evidence for a relationship to IR; illumination of their specific roles would require further study . To confirm the validity of some of the above DE gene candidates , we performed qPCR using RNA samples from 23 of the original 24 patients ( one IR sample did not have sufficient RNA for the procedure ) . We tested five genes found to be significant using both fold change and SAGAT , six genes found only with SAGAT , and four negative controls . All of the qPCR expression differences of the non-control genes matched the direction of those from the microarray data , suggesting that these changes are reproducible . We then tested the significance of these changes using a Wilcoxon rank-sum test ( RST ) . We note that the RST is one of the more conservative two-sample tests available [53] , and we anticipated noisy data due to the amplification reactions needed prior to qPCR ( see Materials and Methods section ) . Nevertheless , three genes—two identified by fold change and SAGAT , one by only SAGAT—were found to be significant . In contrast , none of the negative control genes showed significant expression differences . Combining the qPCR results together with the literature evidence implicating four of the eight genes not confirmed by qPCR suggests a false positive rate of 0 . 4 ( 2/5 ) for fold change and 0 . 36 ( 4/11 ) for SAGAT . Though the difference between these values may not be statistically significant , this result suggests that SAGAT was able to improve the sensitivity of DE gene detection in this experiment without increasing the false positive rate . We did not explicitly test IL6 using qPCR , although we note that previous work has shown this gene to be overexpressed in insulin resistant adipose tissue [35] . This is the only gene detected using fold change that was not also detected using SAGAT , which may reflect discordant expression patterns of IL-6 between previously existing datasets and this one . We now explore the means by which SAGAT improves prediction of DE genes . Results from the simulation study demonstrate that the method improves performance to the extent that DE genes are more likely to be in transcriptional modules than non-DE genes . This is realized through the standard error term ( denominator ) of the SAGAT score ( see Materials and Methods ) . For a given gene in a module ( eigengene ) , the standard error for that gene's mean expression difference receives contributions from measurements of the other genes in that module , leading to a smaller error ( more precise estimate of expression ) . Thus , genes in modules will on average have slightly boosted SAGAT scores compared to genes acting in isolation . In the process of characterizing modularity of the HGU95Av2 knowledge set to parameterize our simulation , we have discovered that DE genes are more likely to be in modules than non-DE genes . Given that the performance improvements in the letrozole treatment and colorectal cancer datasets were similar to the prostate cancer case , we expect this feature of DE genes ( and the corresponding performance improvement by SAGAT ) to be generalizable to a wide variety of biological datasets . To support this hypothesis , we note that genes which are frequently differentially expressed are more likely to be associated with a disease [54] , and genes implicated in the same disease show higher levels of coexpression ( modularity ) than randomly selected genes [55] . A closer look at the functional form of the SAGAT score shows its similarity to versions of the t-statistic , including the limma t and SAM [39] , [56] . The difference between these metrics lies in their method for calculating the standard error of each gene's mean expression difference . Though the limma t-statistic borrows information for calculating this term from other genes , SAGAT is the only approach that identifies and uses expression dependencies between genes in the computation of gene-wise variances . Fortunately , this addition is not computationally expensive , as SAGAT utilizes efficient algorithms . Eigengenes are identified using SVD , which must only be run once per knowledge compendium . Computation of the SAGAT score requires projection of a small ( with respect to the size of the knowledge ) dataset into eigengene space followed by a simple dot product for each gene . Practically , the running time of SAGAT is approximately the same as that of related methods like the limma t-statistic . We note , however , that the distribution of the SAGAT score is complex , and unlike the t-statistic , it does not provide for a straightforward estimation of statistical significance . Thus , we advocate data permutation-based methods ( similar to those used by SAM ) to calculate SAGAT p-values . Use of SAGAT does require some explicit assumptions about microarray knowledge . First , we assume that ( detectable ) multi-gene transcriptional modules give rise to the expression values in a compendium of microarray knowledge . Previous work [12]–[18] detecting reproducible , biologically plausible transcriptional modules ( along with results from our characterization of the HGU95Av2 compendium ) suggest that this is a valid assumption . Second , representing the transcriptional levels of each gene as a weighted combination of eigengene levels assumes that each gene's expression can be modeled in a linear fashion . While some evidence exists to support this assumption [57] , it is more realistic that expression is a non-linear phenomenon . Nevertheless , linear approximations have proven useful and even quite accurate in the modeling of non-linearity [58] . We find empirical support for this accuracy in the coherence of the GO terms significantly enriched in eigengenes of the HGU95Av2 compendium . Third , though SVD does not make any distributional assumptions about the knowledge , the analytical derivation of the SAGAT score requires the eigengene expressions to be statistically independent . When the underlying eigengenes are distributed as multivariate normal ( MVN ) random variables , they will exhibit independence , but otherwise this may not be the case . Given that we did not explicitly enforce this assumption in either the simulated data ( here , genes were MVN , not eigengenes ) or the highly replicated real datasets , this assumption does not appear to be detrimental to SAGAT performance . An implicit assumption in the use of prior microarray knowledge to inform a novel dataset is that the expression dependencies from the knowledge are conserved in the novel dataset . In a worst-case scenario , a novel dataset would exhibit a transcriptional response completely unlike anything assayed previously . Given the modular nature of transcription , we expect this to be unlikely , and the favorable performance of SAGAT on three independent biological datasets supports this assertion . Additionally , as even more microarray experiments are performed and their data become available , the likelihood of such a scenario occurring will tend to zero . As SAGAT requires a large compendium of microarray knowledge , it is worth examining potential biases in currently available compendia . Due to their popularity among researchers , the vast majority of publicly available human microarray datasets are from Affymetrix platforms . Thus , the three compendia and highly replicated datasets used in this study represent the three most popular human Affymetrix GeneChips . One concern would be that a non-biological bias ( perhaps due to cross-hybridization between specific probesets ) exists in Affymetrix data which cannot be detected and removed without considering data derived from other platforms . This might lead to artifactual coexpression relationships . Another concern would be that the dependency information inferred from Affymetrix microarray knowledge is not extensible to non-Affymetrix datasets , due to differences in probesets or the artifactual coexpression phenomenon discussed above . While these concerns may have some merit , we note that in applying SAGAT to a novel insulin resistance dataset we incorporated microarray knowledge from an Affymetrix platform with data from Affymetrix , Illumina , and Agilent platforms . Given the ability of SAGAT to correctly identify novel DE genes in this case , we do not believe such a large Affymetrix-specific bias is present . Finally , as the value for parameter M ( specifically , the fraction M/P—see Materials and Methods ) was set for all three Affymetrix compendia based on performance observed using the HGU95Av2 compendium , there is an implicit assumption that the optimal parameter value is identical between platforms . We evaluated this by testing SAGAT on several data subsets from the HGU133A and HGU133plus2 . 0 platforms across a range of M values . Results suggest that a value of M that is approximately half the number of arrays in the compendium is nearly optimal for all three compendia ( not shown ) . Nevertheless , more principled approaches of effectively choosing platform-specific values for M likely exist , and future work will include identifying these approaches . We provide SAGAT as an R package ( sagat ) , which is available at https://simtk . org/home/sagat . The package includes all necessary functions to run the method along with preprocessed versions of the W matrix for the three Affymetrix platforms analyzed in this work . Given its abilities to improve the prediction of DE genes , we expect that SAGAT will be useful to microarray researchers studying a wide range of biological phenomena .
The insulin resistance study was approved by the Stanford University Human Subjects Committee and the National Institute of Digestive Diseases and Kidney Disease ( NIDDK ) Institutional Review Board , and all subjects gave written informed consent . We downloaded all available expression data for the Affymetrix HGU95Av2 microarray ( GPL91 ) from the Gene Expression Omnibus ( GEO: http://www . ncbi . nlm . nih . gov/geo/ ) in August 2007 . These data are hereafter referred to as “knowledge” , or a knowledge compendium . The Robust Multi-array Average ( RMA ) algorithm was first used to compute averages between probes in a probeset . Probesets were then mapped to a non-redundant list of Entrez Gene IDs ( provided by the Bioconductor R package hgu95av2 version 1 . 16 . 0 ) , and expression values for multiple probesets of the same gene were averaged using an arithmetic mean . This resulted in a matrix of 9105 genes by 4440 arrays , which is available for download at https://simtk . org/home/sagat . We log transformed and quantile normalized the arrays to ensure that they were on the same scale , and we computed the gene-gene covariance matrix across all 4440 arrays , ignoring missing values . In order to simplify characterization of the covariance structure , we discretized the covariance matrix such that diagonal entries and entries whose absolute value was greater than the mean covariance value ( . 25 ) were set to one , and all others were set to zero . We then hierarchically biclustered the rows and columns of the binarized covariance matrix ( using a distance metric of and complete linkage ) to enable visualization of gene groups with significant covariances . Here , we define an expression module as a group of genes of size , identified upon hierarchical biclustering of the covariance matrix , whose pairwise binarized covariance values are all nonzero . To test whether the observed modularity was due to chance , we generated 100 permuted versions of the knowledge matrix , whereby the columns of each row were permuted independently of the other rows . We followed the subsequent steps of calculating covariance , discretizing , and clustering as above , and we counted the number of diagonal covariance clusters containing genes ( i . e . expression modules ) . To characterize expression modularity with respect to differentially expressed ( DE ) or non-DE genes , we coupled the HGU95Av2 compendium with a human prostate cancer microarray dataset [38] . Beginning with the clustered , binarized covariance matrix of Figure 1B , we generated five 1000-gene covariance matrices by randomly subsetting the full matrix . In each one , we zeroed all covariance values in off-diagonal clusters and those in diagonal clusters with fewer than five genes ( in the 1000-gene matrix , we relax the cutoff for expression modules to five genes ) . We calculated the mean number of genes per module in the remaining covariance modules across the five matrices and used this for simulating new compendia ( details below ) . Using the prostate cancer dataset , we identified DE genes as those having a limma t-statistic with FDR ( calculated with the limma R package version 2 . 8 . 1 ) . We split each of the five covariance matrices above into DE or non-DE subsets , and we calculated the mean percentages of genes in covariance modules for each . These values were also used for simulating compendia ( below ) . An overview of the SVD procedure is illustrated in Figure 7A . In equation form , SVD transforms an ( genes×arrays ) knowledge matrix X into the product of three matrices U , S , and V: ( 1 ) where and T represent matrix multiplication and transposition , respectively . As detailed in [22] , the dimensions of U , S , and are genes×eigenarrays , eigenarrays×eigengenes , and eigengenes×arrays , respectively . We follow the notation used in [59] and treat the dimensions of the product as “scaled eigengenes”×arrays . As SVD requires complete data , we either exclude arrays of the knowledge matrix with missing values ( if fewer than 10% of the total number of arrays are incomplete ) or impute missing values using the K-nearest neighbor algorithm implemented in the impute R package ( version 1 . 6 . 0 ) [24] . We center and scale the rows of the complete data matrix and run the svd R function . To confirm the validity of an eigengene theory of gene expression , we first ran SVD on the HGU95Av2 knowledge matrix with missing values imputed . We then identified enriched Biological Process Gene Ontology terms for each eigengene by applying the Kolmogorov-Smirnov statistic ( implemented in the topGO R package version 1 . 2 . 1 ) . Specifically , within each eigengene , all 9105 genes were ranked ( in descending order ) by the magnitudes of their weights ( determined from the appropriate column of U ) . GO terms significantly enriched at the top of each ordered list were then identified using the getSigGroups R function . SVD constructs a linear relationship between genes and eigengenes such that each gene's expression can be formulated as a linear combination of the eigengene expressions ( Figure 7B ) . We can explicitly represent this in equation form by approximating ( 1 ) as follows: ( 2 ) where W is simply a matrix containing the first M columns of U ( M most significant eigenarrays ) , and E is the product of the first M rows of matrix S with . Intuitively , E represents the knowledge matrix X transformed from array space into eigenarray space , and W provides the map between genes and scaled eigengenes . Given a novel dataset D with m replicates ( referred to as “data” ) , we obtain data-specific eigengene expressions by solving the following approximation for : ( 3 ) where we use W from ( 2 ) , and represents dataset D transformed into eigengene space . We obtain a mathematically rigorous solution to ( 3 ) by premultiplying both sides by the transpose of W . This is possible due to the orthogonality properties of SVD and is equivalent to a projection using the pseudoinverse of W . Such a projection gives the optimal ( in the least squares sense ) approximation of dataset D in terms of the knowledge set X . We note that pseudoinverse projection has previously been successfully used in other areas of microarray analysis , particularly with respect to noise reduction in data [29] , [30] , [60] . Knowledge of ( and D ) allows us to calculate a mean log expression ratio for each gene ( ) and a log expression ratio sample variance for each eigengene ( ) ( Figure 7C ) . To perform hypothesis tests for differential expression , we created a probabilistic model for each gene's mean log expression ratio . The properties of SVD allow us to approximate this quantity in the following manner: ( 4 ) where implies “approximately equal to” , represents scalar multiplication , represents the unknown true mean log expression ratio of gene , M is the number of eigengenes used to reconstitute the gene expressions , the weights come from W , and the are mean log ratios for mean-centered eigengenes ( assumed to be normally distributed ) : ( 5 ) where implies “distributed as” , specifies a normally distributed random variable , and represents the population expression variance for eigengene . Thus , the acquire the following distribution: ( 6 ) By using the empirical Bayes variance estimators ( calculated using the limma R package ( version 2 . 8 . 1 ) [39] ) in place of the unknown , we arrive at the test-statistic for gene , analogous to the one sample t-statistic: ( 7 ) This “SAGAT score” borrows information regarding expression variability for each gene from covarying genes via their shared eigengenes . Though the statistical model used to derive this metric assumes normally distributed eigengene log expression ratios , it will still provide quantitatively useful scores when this assumption is not met . In the case when , the are undefined and a slight modification is required . We discovered that the following form of the SAGAT score gave performance consistent with that achieved on datasets with m greater than 1: ( 8 ) where is the single log ratio for eigengene i calculated by transforming the data into eigengene space and implies absolute value . In ( 2 ) , ( 4 ) , ( 6 ) , ( 7 ) , and ( 8 ) above , the correct value for M is unknown , so we treat it as a parameter to be learned from data . Details of the learning procedure for simulated and highly replicated real data are found below in the corresponding sections . We simulated 1000-gene compendia of microarray knowledge by generating 1000 multivariate normal random variables ( using the mvrnorm function in the R MASS package version 7 . 2–48 ) . The mean vector used for the simulation was derived from sample means of 1000 random genes from the HGU95Av2 compendium; the covariance matrix contained all zeros except in positions needed to create the desired modularity structures ( Figure 2 ) . In these positions , we used a covariance value of 4 , which was chosen to be large enough to generate knowledge compendia that led to noticeable differences in SAGAT performance . We simulated 1000-gene microarray data with numbers of replicates ranging from 1–15 using the procedure listed in [19] , parameterized with values derived from the prostate cancer dataset . Each dataset was engineered to contain 100 DE genes . We ran SVD on each simulated compendium and used the resultant W matrix to test SAGAT on all combinations of data and knowledge . To estimate M , we evaluated SAGAT performance as a function of varying M across a range of simulated data ( 1–15 replicates ) and knowledge compendia ( all configurations between Figures 2A and B ) . We chose a value of M that gave optimal performance across all tested configurations; this value was used for all subsequent tests on simulated data . We compared the results of these tests ( in the form of ROC AUC and TPR at a fixed FPR of . 05 ) to that achieved by fold change to determine the range of data/knowledge configurations in which SAGAT outperformed fold change . We evaluated SAGAT's potential to improve DE gene prediction on real data by testing the method on three highly replicated datasets . This approach is similar to that used by [11] , except that we choose area under the ROC curve and true positive rate as our evaluation metrics . The first dataset , listed above , measures differences in expression between prostate cancer tissue and matched non-cancer prostate [38] . This dataset measures expression of 9105 genes ( identified by mapping probe names to Entrez Gene IDs as above ) across 47 pairs of samples ( “replicates”: as Affymetrix arrays measure one RNA sample at a time , one experimental replicate is equivalent to two arrays ) . The second dataset compares breast cancer tissue before and after letrozole treatment [61] . These data were collected across 58 pairs of samples on the HGU133A Affymetrix platform , which measures expression of 13410 Entrez Genes . The final dataset measures expression differences between colorectal cancer tissue and matched non-cancer tissue [62] . This dataset was generated for 32 pairs of samples on the HGU133plus2 . 0 Affymetrix platform , which encompasses 20099 Entrez Genes . For each dataset we determined truly DE genes by calculating either mean fold changes or limma t statistics across all replicates and counting genes with the largest scores ( irrespective of sign ) as DE . The number of DE genes in each case was set to the number of genes whose t-statistic was significant at a . 05 FDR cutoff . We performed all analyses using the limma R package ( version 2 . 8 . 1 ) . To obtain knowledge for each dataset , we downloaded all publicly available microarray datasets from GEO ( minus the highly replicated datasets listed above ) for each of the corresponding Affymetrix platforms . As mentioned above , the HGU95Av2 compendium contained 4440 arrays , while the HGU133A ( GPL96 ) and HGU133plus2 . 0 ( GPL570 ) compendia consisted of 14476 and 12217 arrays , respectively ( as of March 2008 ) . For each knowledge source , we either imputed missing data ( HGU95Av2 ) or excluded incomplete arrays ( HGU133A , HGU133plus2 . 0 ) to arrive at the number of arrays listed above . As with the above datasets , we mapped probe names of each knowledge compendium to the corresponding Entrez Genes . We ran SVD as detailed above on each knowledge matrix , generating the matrices , , and , each containing the maximal number of eigengenes . We evaluated SAGAT on its ability to identify DE genes from subsets of each dataset that best match the truly DE genes discovered using all replicates . For each dataset , we generated the maximal number of non-overlapping subsets of size 1 , 2 , 5 , and 15 ( 14 for Letrozole treatment ) replicates . We ran SAGAT on each data subset with the appropriate W matrix ( defined below ) , calculated fold changes and limma t-statistics for comparison , and computed the ROC AUCs and TPRs evaluated at FPR = . 05 for all three metrics with respect to the truly DE genes . We used the R package ROCR ( version 1 . 0–2 ) [63] for AUC and TPR calculations . To determine the optimal number of eigengenes ( M parameter ) to use in the W matrices for each dataset , we tested all possible numbers of eigengenes from 5 to 4400 ( in multiples of 5 ) on several subsets of the Prostate cancer dataset . The number of eigengenes that gave the best performance overall was used as the value for , and the values for and were set such that they yielded an identical fraction of M/P , where P is the total number of arrays . From these values of M we subset the matrices , , and by only including the first M columns of each to form , , and , respectively . We used these modified matrices in the SAGAT analysis described above . We also characterized SAGAT performance in terms of the effective number of arrays added . For each of the highly replicated datasets , we calculated ROC AUCs of the fold change metric applied to all non-overlapping replicate subsets ranging in size from 1 to the total number of replicates . These AUCs enabled us to fit a “standard curve” for each dataset , from which we could interpolate the mean number of arrays gained by using SAGAT given initial numbers of 2 , 4 , 10 , and 30 ( 28 for Letrozole dataset ) arrays [equivalent to 1 , 2 , 5 , and 15 ( 14 ) replicates , respectively] . We compared SAGAT performance to that of the GEO method , which was implemented as described in [11] using both the standard method and “voting” scheme . The comparison was made as above on subsets of the prostate cancer dataset , using the HGU95Av2 compendium as knowledge . We also evaluated the effect of smaller compendium sizes on SAGAT performance by taking random subsets of 100 to 4000 arrays ( 10 subsets per size ) of the HGU95Av2 compendium and calculating the mean AUC improvement over fold change across all subsets of the prostate cancer dataset . We applied SAGAT to an unpublished biological dataset investigating human insulin resistance . Briefly , 33 moderately obese but otherwise healthy female patients were tested for insulin resistance using a modified insulin suppression test [64] . RNA was isolated from the adipose tissue of the 12 most and 12 least insulin resistant patients and hybridized to three different microarray platforms: Affymetrix HGU133plus2 . 0 , Agilent G4112A , and Illumina HumanRef-8 v2 . The data from the Affymetrix platform were normalized using a bias correction algorithm [65]; data from the other two platforms were normalized using default algorithms accompanying the respective feature extraction programs . Raw data for each of the three platforms are available for download as Datasets S1 , S2 , S3 . We first used the limma t-statistic to identify DE genes using the data from each platform individually . To utilize data from all three platforms simultaneously , we applied the method of Rank Products to lists of genes from each platform ranked either by fold change or SAGAT score ( in both cases separating up and downregulated genes ) . Predicted DE genes were validated by quantitative RT-PCR experiments . 200ng of total adipose tissue RNA was amplified using the Ambion MessageAmp II aRNA Amplification Kit ( cat #AM1751 ) according to manufacturer's instructions . 1ug of amplified product was then used for quantitative PCR analysis using Taqman primer/probe sets for ACTG2 ( Entrez Gene ID: 72 ) , CSN1S1 , FOSB , SELE , FAM150B ( 285016 ) , PMP2 , ATP1A2 , CXCR4 , ELOVL6 , FASN , SRGN , EPHX2 ( 2053 ) , F2 ( 2147 ) , CEBPD ( 1052 ) , and LIPG ( 9388 ) as well as Human -actin endogenous control . Primer/probe sets were purchased from Applied Biosystems ( Foster City , CA ) . Amplification was carried out in triplicate on an ABI Prism 7900HT at for 2 min and for 10 min followed by 40 cycles of for 15 s and for 1 min . A threshold cycle ( CT value ) was obtained from each amplification curve and a value was first calculated by subtracting the CT value for -actin from the CT value for each sample . A value was then calculated by subtracting the value of a single insulin-sensitive subject ( control ) . Fold-changes compared with the control were then determined by raising 2 to the power . We tested the significance of each gene's qPCR-derived expression differences using a one-sided Wilcoxon rank-sum test ( two-sided test was used for negative controls ) . Genes with p-values smaller than a . 05 threshold were considered significant .
|
Though the use of microarrays to identify differentially expressed ( DE ) genes has become commonplace , it is still not a trivial task . Microarray data are notorious for being noisy , and current DE gene methods do not fully utilize pre-existing biological knowledge to help control this noise . One such source of knowledge is the vast number of publicly available microarray datasets . To leverage this information , we have developed the SVD Augmented Gene expression Analysis Tool ( SAGAT ) for identifying DE genes . SAGAT extracts transcriptional modules from publicly available microarray data and integrates this information with a dataset of interest . We explore SAGAT's ability to improve DE gene identification on simulated data , and we validate the method on three highly replicated biological datasets . Finally , we demonstrate SAGAT's effectiveness on a novel human dataset investigating the transcriptional response to insulin resistance . Use of SAGAT leads to an increased number of insulin resistant candidate genes , and we validate a subset of these with qPCR . We provide SAGAT as an open source R package that is applicable to any human microarray study .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"computational",
"biology/systems",
"biology",
"genetics",
"and",
"genomics/gene",
"expression",
"computational",
"biology/genomics",
"genetics",
"and",
"genomics/bioinformatics",
"diabetes",
"and",
"endocrinology/type",
"2",
"diabetes"
] |
2010
|
Using Pre-existing Microarray Datasets to Increase Experimental Power: Application to Insulin Resistance
|
A fundamental question in understanding neuronal computations is how dendritic events influence the output of the neuron . Different forms of integration of neighbouring and distributed synaptic inputs , isolated dendritic spikes and local regulation of synaptic efficacy suggest that individual dendritic branches may function as independent computational subunits . In the present paper , we study how these local computations influence the output of the neuron . Using a simple cascade model , we demonstrate that triggering somatic firing by a relatively small dendritic branch requires the amplification of local events by dendritic spiking and synaptic plasticity . The moderately branching dendritic tree of granule cells seems optimal for this computation since larger dendritic trees favor local plasticity by isolating dendritic compartments , while reliable detection of individual dendritic spikes in the soma requires a low branch number . Finally , we demonstrate that these parallel dendritic computations could contribute to the generation of multiple independent place fields of hippocampal granule cells .
Neurons possess highly branched , complex dendritic trees , but the relationship between the structure of the dendritic arbor and underlying neural function is poorly understood [1] . Recent studies suggest that dendritic branches form independent computational subunits: Individual branches function as single integrative compartments [2] , [3] , generate isolated dendritic spikes [4] , [5] linking together neighbouring groups of synapses by local plasticity rules [6]–[8] . Coupling between dendritic branches and the soma is regulated in a branch-specific manner through local mechanisms [9] , and the homeostatic scaling of the neurotransmitter release probability is also regulated by the local dendritic activation [10] . The computational power of active dendrites had already been demonstrated by several computational studies [11]–[16] , but how local events influence the output of the neuron remained an open question . Using the cable equation [17] or compartmental modelling tools one can calculate the current or voltage attenuation between arbitrary points in a dendritic tree [14] , which is in good agreement with in vitro recordings . However , cortical networks in vivo are believed to operate in a balanced state [18] , [19] , where the inhibitory drive is continuously adjusted such that the mean activity of the population is nearly constant [20] , [21] . In this case , the firing of an individual neuron is determined , beyond its own input , by the activity distribution of the population . A simple cascade model [22] incorporating numerous dendritic compartments allowed us the statistical estimation of the activity distribution of neurons within the population . We used this model to study how localized dendritic computations influence the output of the neuron . The present study focuses on hippocampal granule cells . Compared to pyramidal neurons granule cells have relatively simpler dendritic arborization: They lack the apical trunk and the basal dendrites , but are characterized by several , equivalent dendritic branches , extended into the molecular layer [23] ( Figure 1A ) . Recordings from freely moving rats revealed that like pyramidal neurons , granule cells exhibit clear spatially selective discharge [24] , [25] . However , granule cells had smaller place fields than pyramidal cells , and had multiple distinct subfields [24] , [26] . It has also been recently shown that these subfields are independent , i . e . , their distribution was irregular and the transformation of the environment resulted in incoherent rate change in the subfields [26] . The dendritic morphology of granule cells suggest that parallel dendritic computations could contribute to the generation of multiple , distinct subfields of these neurons . In the present study we analyzed how synaptic input arriving to dendritic subunits influence the neuronal output . First , we introduce the model used in this study and we define statistical criteria to measure if a dendritic branch alone is able to trigger somatic spiking . We show , that generally neurons perform input strength encoding i . e . , input to the whole dendritic tree but not activation of a single branch is encoded in the somatic firing . Next we demonstrate that if the local response is enhanced by active mechanisms ( dendritic spiking and synaptic plasticity ) then neurons switch to feature detection mode during which the firing of the neuron is usually triggered by the activation of a single dendritic branch . Furthermore we show that moderately branched dendritic tree of granule cells is optimal for this computation as large number of branches favor local plasticity by isolating dendritic compartments , while reliable detection of individual dendritic spikes in the soma requires low branch number . Dendritic branches of dentate granule cells could therefore learn different inputs; and the cell , activated through different dendritic branches , could selectively respond to distinct features ( locations ) , participating in different memories . Finally using spatially organized input we illustrate that our model explains the multiple independent place fields of granule cells and these dendritic computations increase the pattern separation capacity of the dentate gyrus .
Supposing that firing rates of presynaptic neurons ( uj ) are independent and identically distributed we assume that the total input of the dendritic branches Ui = Σjwijuj is drawn randomly from a Gaussian distribution with mean μ and variance σ2: ( 5 ) where p[U] indicates a probability distribution over U ( Figure 1C; see Eq . 17 in Methods for parameters specific to hippocampal granule cells ) . More specifically , indicates the distribution of the magnitude of possible total inputs to a single dendrite over many different instances . Based on the distribution of the total input , we can compute the distribution of the somatic activation and determine the firing threshold ( β ) according to the proportion of simultaneously active cells ( the sparseness of the representation , spDG ) in the DG [24] . First , we rearrange Eq . 3 using the input distribution to express the distribution of : ( 6 ) where indicates that the inputs of the dendritic branches are randomly sampled from a Gaussian distribution . We substitute Eq . 6 into Eq . 4 , and we get ( 7 ) We can assume again , that the inputs ( Ui ) of the dendritic branches are independent and identically distributed variables . ( Note , that while the activations are not independent because of the back-propagation of currents from the soma , the inputs are . ) If N is high enough , we can approximate the sum in Eq . 7 with a Gaussian distribution , and rewrite the equation: ( 8 ) where indicates a probability distribution over , while μF and are the expected value and the variance of the dendritic integration function F ( U ) given the input distribution : ( 9 ) ( 10 ) We calculated the integrals 9–10 with two different forms of dendritic integration of synaptic inputs: a linear and a quadratic function ( Figure 1C ) . The details of these calculations are in the Supporting Information ( Text S4 ) . In this paper we do not model inhibitory neurons in the dentate gyrus , however , we assume , that they play a substantial role in continuously adjusting the firing threshold of principal neurons and regulating the activity of the network [20] , [21] . As a result of this regulation always the most depolarized neurons are able to fire , and the proportion of simultaneously active neurons is characteristic for different hippocampal areas [24] , [29] . Given that all neurons share a common input statistics and have similar internal dynamics , equation 8 also describes the distribution of across the granule cell population at a given time . If only the most depolarized 1–5% of the population are able to fire [29] , this also means that only those neurons exceed their firing threshold whose activation is within the uppermost 1–5% of the distribution described by Eq . 8 . Therefore , the proportion of simultaneously active neurons within the dentate gyrus spDG [24] , [29] also determine the firing threshold β for granule cells . We approach the dendritic independence by focusing on the statistical distributions of the input to dendritic branches , as these branches form the basic computational subunits in our model . We ask whether the input of a single branch could be sufficiently large to significantly depolarize not only the given branch but also the soma of the neuron . We defined two conditions to study whether the spiking of the neuron is caused by the activation of a single dendritic branch or by the simultaneous depolarization of multiple branches . First , the conditional probability is the probability of firing given that any branch k has total input Uk = Σjwkjuj , while inputs to all other branches are random and independent samples from the distribution of ( Figure 2A ) . At those Uk values where this probability is close to 1 the cell tends to fire when any of the dendritic branches gets that input . Second , the conditional distribution is the distribution of the synaptic input of the most active branch at the time the depolarization of the soma exceeds the firing threshold ( β ) , where U* is the total synaptic input arriving to the most active branch ( Figure 2A ) . K ( U* ) can be regarded as the marginal distribution of above the firing threshold ( Figure 2B ) . The probability mass of this function shows the typical maximal input ( U* ) values when the neuron fires . These two conditions together determine whether a single branch can be sufficiently depolarized to trigger somatic spike or not . If the probability of firing is high ( H ( U ) ≈1 ) at typical input values ( K ( U* ) ) then the firing of the cell is caused by a single branch . With the definition of Gasparini and Magee [30] we call this form of information processing as independent feature detection . On the other hand , if the firing probability is low ( H ( U ) ≪1 ) even if one of the branches receive extremely large input ( U* is high ) then the cell mostly fires when the overall dendritic activation is high , and even the most depolarized branch usually fails to make the neuron fire . We use the expression input strength encoding [30] to denote this second type of computation . The calculation of the two functions H ( U ) and K ( U* ) is described in the Methods section .
First we chose unstructured synaptic input , i . e . , the firing of entorhinal neurons were independent and the strength of all synapses were equal . In this case we approximated the total synaptic input U to a branch with a Gaussian distribution ( Eq . 5 , Figure 1C ) . Given the input distribution we asked whether the excitation of single branches can be sufficiently large to cause significant depolarization in the soma . The typical largest input values , indicated by the probability mass of K ( U* ) ( Figure 2C–D ) are unable to sufficiently depolarize the soma and determine the neuronal output ( indicated by the low H ( U ) values ) in the case of both the linear ( Figure 2C ) and the quadratic ( Figure 2D ) integration functions . Wherever K ( U* ) has high values , H ( U ) is low in both cases , which indicate , that these branches are not able to independently influence the output of the neuron . Only coactivation of several branches could make the neuron fire in this case , and the output of the neuron encodes the strength of all dendritic inputs . As H ( U ) converges to 1 for high input values extremely high inputs to a single dendrite could reliably trigger somatic firing . In the next sections , however , we study how synaptic plasticity selectively modifies individual synapses and contributes to the sparse occurrence of extraordinarily high input values . During Hebbian learning synapses contributing to postsynaptic activation are potentiated while other synapses may experience compensatory depression [31] , [32] . We simulated the learning process by showing a finite number of uncorrelated samples from the input distribution ( see Methods ) to the model neuron initiated with uniform synaptic weights . The synaptic weights of those dendritic branches where the activation exceeded a threshold , βd were modified according to the following Hebbian plasticity rule [33] that incorporates heterosynaptic depression [31]: ( 11 ) where is the local dendritic activation , uj is the presynaptic firing rate and wij is the synaptic strength . is the Heaviside function and γ<1 is a constant learning parameter . Note , that the learning rule is local to the dendritic branches: the synaptic change depends on the local activation but not on the somatic firing . Next , we calculated the total input to the branches Ui = Σjwijuj after modification of synapses ( Figure 3A ) , and recalculated the two functions H ( U ) and K ( U* ) defined previously with the new input distribution ( Eq . 18 ) . As shown on Figure 3A the total synaptic input in response to a learned pattern increases significantly after learning ( compare blue and grey curves on Figure 3A ) , while untrained patterns generate smaller synaptic inputs ( compare grey and black curves on Figure 3A ) . The main consequence of synaptic plasticity is that the trained patterns generate much larger local response than untrained patterns , which raise the possibility of their detection in the soma . Note , that an unspecific increase of synaptic weights would result in an upward shift of both the input distribution ( Eq . 5 ) and the firing threshold , but would not affect the somatic detection of individual dendritic events . The neuron is able to selectively respond to the dendritically learned patterns if a single branch , when facing with its preferred input , is able to induce significantly more depolarization at the site of the action potential initiation compared with the case when all of the branches get random , not learned input . Figure 3B–E shows the dendritic input and the activation of the soma after learning . If the maximal input U* is small ( left bumps on Figure 3B , D ) and none of the branches got its preferred input then the somatic activation is usually small . If U* is high ( Figure 3B , D; right bumps ) , which means that one of the branches receives its preferred input pattern , then the somatic activation is increased . The increase of the somatic activation with learned input is only moderate in the linear case ( Figure 3B , C ) resulting in an incomplete separation of learned and not learned inputs by the somatic firing threshold . However , if synaptic inputs are supra-linearly ( quadratically ) integrated within the dendritic branches , efficient separation is possible: the probability that the presentation of a learned pattern elicits subthreshold somatic response , called dendritic spike detection probability was over 95% ( Figure 3D , E ) . In this case the output of the neuron encodes whether or not one of the stored features was present in the neuron's input and not simply the strength of the total input arriving to the whole dendritic tree . In other words , if dendritic nonlinearity enhance the response of a given branch to its preferred input , then this branch alone is able to trigger somatic spiking . In the following sections we use the term dendritic spiking to refer to these supra-linear dendritic events . Although there is no data available on the synaptic induction of local dendritic spiking in hippocampal granule cells , voltage dependent Ca2+ currents are present in the membrane of granule cells [34] , [35] and whole-cell recordings from these neurons suggest that T-type Ca2+ channels can generate dendritic action potentials at least in young neurons [36] or under hyper-excitable conditions [34] , [37] . Next , we explored how the independent feature detection ability of the model depends on the resistance between the somatic and dendritic compartments with nonlinear dendritic integration . In the passive cable model of dendritic trees the space constant of the membrane λm≈ ( Rm/Ri ) 1/2 plays a substantial role in determining the voltage attenuation among two sites . Consequently , an increase in the intracellular resistivity Ri or a similar decrease in the membrane resistance Rm will contribute to the separation of dendritic subunits by decreasing the membrane's space constant λm . In the present study we used the inverse of the space constant R≈Ri/Rm to characterize the degree of electrical resistivity between the somatic and dendritic compartments . Indeed , an increased resistivity ( R ) between the compartments ( smaller space constant ) induced larger degree of electrical isolation as the somatic response to the same amount of dendritically applied current decreased ( compare Figure 4A left and right panels ) . However , this isolation did not modify the dendritic spike detection probability in the soma: Large dendritic spikes localized to a single compartment could be reliably separated from subthreshold events with a somatic firing threshold at a large range of resistances R ( Figure 4A–B ) . This was also true for the selective alternation of the somatic or the dendritic membrane resistance ( Figure 4B ) . On the other hand , the resistance parameter had a substantial impact on the isolation of different dendritic compartments which might be necessary for the independence of synaptic plasticity . To measure the isolation of the dendritic subunits we calculated the influence of other compartments on the activation of a given branch ( external influence ) quantified by the standard deviation of . Figure 4C shows the activation of a dendritic branch in the function of its input at different R values . If the resistance is small ( , Figure 4C , left ) , then the local activation depends only slightly on the local input and the external influence is high ( Figure 4D ) . In this case the local input spread out to the entire dendritic tree and activates similarly all branches . On the other hand , if the resistance is high ( R = 1 , Figure 4C , right ) then the external influence is small , and the depolarization of a dendritic branch depends mostly on the local input . Interestingly , decreasing the resistance of the perisomatic membrane ( ) alone was more efficient in separating the dendritic subunits than decreasing the resistance of the dendritic membrane or both ( Figure 4D ) . The extensive GABAergic [38] , [39] and glutamatergic [40] innervation of the proximal dendritic and perisomatic region of granule cells may therefore contribute significantly to the isolation of the dendritic compartments . The impact of a single branch on the somatic activation , and also the coupling between dendritic branches may depend highly on the structure of the dendritic tree . Therefore we varied the number of dendritic subunits , N , and calculated the probability of detecting dendritic spikes in the soma and the external influence on the dendritic subunits ( Figure 5 ) . The probability of detecting a dendritic spike in the soma decreased gradually after a few ( N≈30 ) number of branches from 1 to 0 . 3 ( N≈1000 , Figure 5A–B ) . If the number of branches was low , then the effect of a single branch on the soma was relatively high , and the somatic detection of single dendritic events was reliable . Conversely , one out of hundreds of branches had relatively low impact on the neuron's output even if the local depolarization was significant . The electrical coupling between the dendritic subunits characterized by the external influence on the local activation also decreased with the number of branches , ( Figure 5C–D ) . In the model the branches are connected through the somatic compartment , and because the variance of the somatic activation decreases if N increases ( Eq . 8 ) , the external influence will also decrease . However , in a complex dendritic tree containing higher number of subunits the branches are electronically more isolated which is required for local plasticity . To keep the probability of dendritic spike detection high and the dendritic coupling low at the same time , the number of branches should therefore be as high as possible , but not higher than N≈60 . As we showed on Figure 4 , the dendritic coupling depends on the resistance R , as high resistance separates better the subunits . Therefore we conclude , that a medium number of branches with relatively high resistance is ideal for parallel dendritic computations . The optimal number of dendritic subunits , however , depends on the size of the dendritic event determined by the local integration of the synaptic inputs ( Figure 5B ) . Appropriate detection of dendritic responses to learned patterns with linear integration is possible only in very small dendritic trees , whereas supra-linear integration allows the detection of individual dendritic events also in a larger dendritic arbor . Nonlinear integration by dendritic spiking therefore permits the neuron to selectively respond to a larger number of distinct input pattern . During the calculation above we assumed , that the activity of the presynaptic neurons are independent and that the samples from the distribution are uncorrelated . It is known , however , that the firing of entorhinal neurons are not independent: At least half of layer II cells in the medial entorhinal cortex ( EC ) are grid cells , whose firing depend mostly on the position of the animal [27] . Moreover , in reality animals do not face with discrete uncorrelated samples , but they experience the continuous change of their environment which is mirrored by the activity of the entorhinal neurons . In order to test our model under more realistic conditions , we simulated the activity of the rodent's EC during exploratory behavior as input to our modeled granule cell . The EC consisted of two neuron population: A population of grid cells ( 1000 neurons , 5 spacing , 5 orientations ) representing a path integrator system [41] and a population of visual cells ( 1200 units ) , representing highly processed sensory information available in the EC [42] . In these simulations we used the Webots mobile robot simulator [43] . The firing statistics of the entorhinal neurons was the same as used in the analytical calculation except that the activity of the neurons was location dependent . Moreover , as we simulated the trajectory of the rat during continuous foraging for randomly tossed food pellets [26] the subsequent input patterns were highly correlated . We simulated a single granule cell with N = 20 dendritic branches each of them receiving a total number of M = 100 synaptic contacts from entorhinal neurons . The resistance was R = 1 , we used the quadratic integration function and the neuron was tested in 5 different environments . During the 5 min . learning period ( while 2000 spatial locations was sampled with an average running speed of 0 . 22 m/s ) 0–8 branches learned usually at different spatial locations in each of the 5 environments . In most of the time synaptic plasticity in different branches occurred at different places , therefore the subunits were able to learn independently . Moreover , learning occurred only in naive branches , i . e . , each branch learned only in one environment at a specific location and synapses of trained branches did not engage in learning at a different location . After the training period the synaptic weights of those branches that were subthreshold for synaptic plasticity ( βd = 1 . 11 ) in all environments were scaled down manually . Next we studied the spatial activity pattern of the somatic and dendritic compartments while the robot was moving on a different track in the same environments . The dendritic branches responded with high activation ( “dendritic spikes” ) to subsequent visit of places close to their preferred locations leading to the formation of dendritic place fields ( Figure 6 ) . Moreover , since the activation of the soma was substantially increased in each of these dendritic place fields , the neuron had a multi-peaked activity map in several environments ( Figure 6 ) . Finally we explored the effect of the size of the dendritic tree on the spatial firing pattern of the neuron ( Figure 7 ) . If there were only a few functional dendritic subunit than the neuron obviously had a small number of dendritic place fields ( Figure 7A ) , but the individual branches had strong influence on the somatic activity . Therefore the correlation between the somatic activation as and the maximal dendritic input U* was high ( Figure 7B , C ) , as predicted by the analytical calculations . On the other hand , in neurons with large number of dendritic subunits there were more dendritic place fields ( Figure 7A ) , but a single branch had only a little impact on the activity of the neuron ( Figure 7D ) . Accordingly , the correlation between the maximal dendritic input and somatic activation was reduced ( Figure 7B ) . In these cases the cell fired when the overall excitation was high or when more than one branch were simultaneously excited . Therefore , the moderately branching dendritic tree of granule cells seems optimal for parallel dendritic computations since extensive branching inhibits the detection of individual dendritic events . We conclude , that clustered plasticity together with dendritic spiking may be an adequate cellular mechanism to explain the generation of multiple place fields in the DG [24] , [26] .
Dendritically generated spikes mediated by voltage-gated Na+ [3] and/or Ca2+ channels [44] as well as glutamate-activated N-methyl-D-aspartate ( NMDA ) channels [45] have been described in a variety of neurons ( for a review see [46] or [47] ) including hippocampal granule cells [34]–[37] . We used a quadratic integration function in order to analytically model supra-linear dendritic integration [15] which differs from the sigmoid form of nonlinearity realized by dendritic spiking ( Text S1 , [3] , [4] , [45] ) . We believe , however , that at this level of abstraction the exact form of nonlinearity does not affect our results: As that is the difference between the dendritic responses to learned and not learned patterns that influence the somatic detection of dendritic events , a sigmoid integration function give qualitatively similar results ( Text S2 ) . Moreover , we studied only passive interactions between individual dendritic events as the effect of voltage and calcium dependent currents ( including A-type and Ca2+-dependent potassium [48] and the H-current [49] ) regulating the propagation of dendritic spikes were not included in the model . Future studies using a compartmental model equipped with dendritic spiking could support our results and clarify further details . Our analysis has revealed that a moderately branched dendritic tree is optimal for the independent branches model , and we have shown that this mechanism could contribute to the spatial firing properties of granule cells in the DG . The dendritic tree of cerebellar Purkinje cells as well as the apical dendrites of hippocampal and neocortical pyramidal cells is typically larger , and more ramifying [50] . Their morphology is suitable for local plasticity within single branches [6] , [8] , and although it seems that individual branches may function as single integrative compartments [3] , [4] , [51] , [52] , dendritic spikes localized to these compartments fail to propagate to the soma and directly influence the neuron's output [53] . Larger dendritic events , active spread of dendritic spikes towards the soma or interactions among dendritic subunits could contribute to the generation of somatic action potentials in this case . The dendritic tree of pyramidal neurons is , however , far more complex than that of granule cells: it has several morphological and functional subregions with different afferent inputs and membrane excitability [50] . Understanding how their spatial firing characteristics arise from their cellular properties would require at least a different model structure and is beyond the scope of this paper . Whether individual dendritic events influence the output of the neuron depends - beyond the structure of the dendritic tree - on the size and the frequency of the large dendritic events and the output sparsity . The size of the events depends on the exact form of the dendritic integration function and the plasticity rule while the input statistics determine the frequency of such events . We have shown that given the sparseness of the output , sufficiently large , localized dendritic events arriving with appropriate frequency are able to separately determine the output of the neuron . Whether a local event is sufficiently large depends on the geometry of the dendritic tree: A smaller event may be sufficient if there are only a few subunits , or if the events actively propagate to a large part of the entire dendritic tree ( e . g , the apical tuft in pyramidal neurons , [54] ) . Conversely , in neurons such as cerebellar Purkinje cells with large , ramifying dendritic tree , where individual events are localized to small branches , very large dendritic spikes would be required to influence the output . Indeed , detailed compartmental modelling of dendritic morphology revealed that the forward propagation of the action potential initiated in the apical trunk of pyramidal neurons was very effective , while in Purkinje cells dendritic action potentials were rapidly attenuated [53] . Clustered plasticity allows the neuron to simultaneously learn several different patterns but requires the electrical and/or biochemical isolation of the dendritic compartments [47] , [55] . However , the intracellular resistance ( ) in dentate granule cells is relatively low and granule cells are usually regarded as electrically compact neurons [28] . Indeed , signal propagation from somata into dendrites in vitro is more efficient in granule cells compared with CA1 pyramidal cells and distal synaptic inputs from entorhinal fibers can efficiently depolarize the somatic membrane of granule cells [28] . However , in vitro studies do not take into account that neurons are embedded in a network of spontaneously active cells . As thousands of synapses bombard the dendritic tree in vivo , the dendritic membrane becomes “leakier” and , consequently , the membrane's space constant decreases significantly [56] . Moreover perisomatic inhibition [57] and feed-back excitation ( via hilar mossy cells [40] ) further decrease the resistance of the proximal membrane contributing to the separation of the somatic and dendritic compartments [54] , [58] . More specifically , we predict , that the membrane resistance of granule cells is considerably smaller at the perisomatic region than in the distal dendrites . Indeed , computational studies predict a 7–30 fold increase in the somatic leak conductance due to the synaptic background activity [59] . On the other hand , large space constant at long terminal branches facilitate interactions among synapses distributed on the same branch . Therefore the long dendritic branches of dentate granule cells may act as single integrative computational subunits , separated from each other by the perisomatic region of the cell . Furthermore , in the present paper we used steady-state approximations and we neglected temporal characteristics of the input and the integration . For rapidly varying inputs the coupling between dendritic sites and the soma is much smaller than for slowly varying currents since the distributed capacitance throughout the tree will absorb the charge before it reaches the soma [14] . Therefore dendritic compartments in a passive tree are more isolated for transient events such as dendritic spikes than for steady-state current . Finally , biochemical compartmentalization is likely to play a substantial role in the cooperative induction of LTP in both hippocampal [60] and neocortical neurons [7] . If , on the other hand , dendritic branches are not isolated during the learning process and synapses across the whole dendritic tree are modified simultaneously then different dendritic branches will be sensitive for different component ( modalities ) of the same episode . A new episode with partial overlap with the previously learned one may trigger dendritic spiking in the corresponding dendritic branch . As the somatic detection probability of dendritic spikes does not depend on the degree of electrical isolation ( Figure 4 ) , individual branches trigger somatic spiking , and , in this way the dentate gyrus contributes to the associative recall of the previously encoded episode in the hippocampus . Since the first description of LTP at perforant path - granule cell synapses [61] synaptic plasticity has become widely accepted as the physiological basis of memory [62] . As Hebbian plasticity is intrinsically unstable , simply because it is a positive feed-back mechanism multiple stability-promoting mechanisms have been proposed , including heterosynaptic depression [31] , [63] . Indeed , in the present model synaptic plasticity results in an average decrease of synaptic strengths ( Figure 3A ) , which have several functional consequences: First , as the dendritic response to untrained patterns and likewise the baseline activation of the cell decreases during training , the somatic detection of individual , large dendritic events becomes easier . Consequently , feature detection is less efficient in semi-trained neurons where synaptic weights at only a part of the dendritic tree has already been modified due to the learning precess . Therefore , in this model , appropriate training of each dendritic branch is required for proper functioning . Second , increased excitability stimulates learning in naive branches , while decreased responsiveness of previously trained branches prevents overlearning . Indeed , newly generated granule cells are more excitable than the neighboring old neurons [36] , and they are preferentially incorporated into functional networks in the dentate gyrus during acquisition of new memories [64] . One of the most interesting prediction of the present model is how the number of presynaptic spikes required for the postsynaptic induction of dendritic spiking changes during the course of learning . We can calculate this by dividing the total input U needed for dendritic spiking with the mean synaptic weight parameter ( μw ) before and after learning . Our model predicts , that while in young neurons the simultaneous occurrence of ≈70–80 presynaptic spikes ( randomly distributed across the presynaptic neurons ) would trigger a postsynaptic dendritic spike , after learning ( i . e . , in matured neurons ) ≈130–160 would be required . A recent study showed that the homeostatic regulation of the neurotransmitter release probability at neighbouring synapses depends on the local dendritic activity [10]: Increased dendritic depolarization elicits a local homeostatic decrease in the release probability and vice versa . This mechanism may also prevent overlearning in trained branches where dendritic spikes has sufficiently high rate by reducing the excitability of that branch . On the other hand the same mechanism may stimulate learning new patterns in naive or disused branches where dendritic spikes are not present . One of the key elements of our model was the local nature of the synaptic plasticity , i . e . , the change of the synaptic weights was controlled by the local dendritic but not the somatic activity [6]–[8] . Specifically , in hippocampal granule cells the induction of LTP was shown to be independent of the discharge of the neurons during the high-frequency stimulation [65] . Our model predicts that , if the postsynaptic signal for synaptic plasticity is localized to individual dendritic branches than , due to the associative nature of the LTP , the synapses from entorhinal cells with overlapping firing become potentiated . If LTP is accompanied by structural remodeling , than the entorhinal neurons with overlapping place fields project to the same dendritic branches of granule cells as also proposed by Hayman and Jeffery ( 2008 ) [66] . The variation in the strength of perforant path-granule cell synapses was found to be critical in the generation of multiple place fields in a recent modelling study [67] . This heterogeneity caused a greater average synaptic excitation in a fraction of granule cells . This extra excitation therefore selects the subpopulation of neurons active within a given environment similar to the proposed role of contextual inputs in the model of Si and Treves [68] . One possible source of synaptic heterogeneity is synaptic plasticity [69] which was also crucial in the present model to amplify the local responses to learned patterns . Hippocampal granule cells receive afferent fibers from the medial and the lateral portion of the entorhinal cortex , and these two pathways differ both in their pattern of termination [70] , [71] and information content [72] . Fibers originating in the lateral EC display weak spatial selectivity and terminate on the most distal branches of granule cells , while medial entorhinal neurons innervate the middle third of their dendritic tree and show strong spatial selectivity [72] , [73] . It has been recently suggested by modelling studies [66] , [68] that inputs originating from the lateral EC conveys contextual information to granule cells . In these models the contextual input select a subpopulation of neurons ( or dendritic branches in [66] ) that can be activated within the given context ( environment ) while medial entorhinal fibers determine the exact location of the place fields . The selection of a subpopulation by contextual inputs can also contribute to the multiple firing fields of granule cells by reducing the number of available neurons within the given environment [67] , [68] . However , the spatial distribution of the individual place fields become regular ( grid-like ) if the multiple firing peaks are the consequence of an incomplete competition between neurons , especially if the input grid cells are organized into a finite number of ensembles [74] , [75] . In the present paper we have shown that synapses , irrespective of their origin , arriving at different branches of hippocampal granule cells can be modified at different spatial locations . We have also shown , that in granule cells each dendritic branch is able to activate the neuron , therefore each subfield on the cell's multi-peaked activity map corresponds to a dendritic place field . The segregation of contextual and positional information could explain the sensitivity of the subfields to contextual manipulations [26] , [66] and is consistent with the role of DG in context discrimination [76] . Along with the laminar organization of excitatory input , different interneurons innervate different dendritic domains of granule cells [39] , [77] . It appears , that distinct types of interneurons have evolved to selectively and locally modulate the computations performed by the postsynaptic membrane [57] , [78] . According to our model , basket and axo-axonic cells may continually adjust the inhibitory drive such that the mean activity of the population remains nearly constant; HICAP cells , targeting the proximal dendritic domain of granule cells together with the excitatory mossy cells [40] may increase electrical isolation of distal dendritic regions by raising the conductance of the proximal membrane; whereas MOPP and HIPP cells associated with the entorhinal afferents may contribute to the de-inactivation of calcium channels required to dendritic spiking by providing rhythmic hyperpolarization to distal dendritic branches . Hippocampal interneurons have also a substantial role in shaping the temporal dynamics of the network [78] . The firing of neurons in the hippocampal formation is strongly modulated by the theta rhythm [25] , [79] , [80] which is a prominent , large amplitude field potential oscillation in the rodent hippocampus during exploratory behavior [81] . The relative synchronization of presynaptic spikes by the theta rhythm allows the temporal integration of their postsynaptic potentials despite the relatively small time constant of granule cells' membrane [28] . Moreover , the synchronization of synaptic inputs can also influence the form of dendritic integration by switching from linear to nonlinear integration [30] . Extending the present model with temporal dynamics could be an exciting direction for future research . What is the additional computational power gained from the present model ? We argue , that smaller and uncorrelated place fields may help pattern separation in the dentate gyrus . Theoretical considerations suggest that the DG helps the hippocampal storage of new episodes by producing sparse representations via competitive learning [82] , [83] . It was demonstrated by modelling studies that competitive learning on spatially organized input results in the formation of place fields [68] , [74] , [84] , [85] that is a sparse and orthogonal representation of the input space . In the present paper we proposed that parallel dendritic computations explain the formation of multiple , independent place fields of hippocampal granule cells even within a relatively small environment [24] , [26] . Pattern separation by the DG can be more efficient if granule cells have multiple , irregularly placed fields and the individual fields are smaller . The neural representation of neighbouring locations is more similar if neurons have one , larger field than if they have several but smaller fields ( Text S3 ) . In our model the place fields of a dendritic branches are analogous to the to the single place field of an electrically compact neuron . The multi-peaked somatic firing of the granule cells mirrors the several dendritic fields of the same neuron . We argue , that if the size of the somatic firing fields is limited by competition between simultaneously active neurons [86] , then the place fields of granule cells could be smaller than the corresponding dendritic fields . If the individual place fields of granule cells become smaller , than the neural representation of adjacent places becomes less correlated which further increase the pattern separation ability of the DG . Therefore independent dendritic subunits increase the computational power of the DG while keeping the number of cells and their sparsity constant . Moreover , clustering of different inputs into different dendritic domains could explain the remapping of hippocampal place cells under several experimental conditions [26] , [66] . The impact of both dendritic nonlinearity and clustered plasticity on the computational power of neurons was rarely addressed by modeling studies . Poirazi and Mel [16] predicted , that nonlinear dendritic integration with local ( structural ) plasticity rule increase the representational capacity of neural tissue . They showed on binary input , that the number of attainable input-output functions ( representational capacity ) is maximal if the neuron has many , relatively short branches , and the performance of the model in a linear classification task correlates remarkably well with the logarithm of representational capacity . However , in order to approach the combinatorial bound of the representational capacity in a neural tissue and to amplify slight differences in the input extremely large subunit nonlinearity was required ( they used F ( U ) = U10 ) . In the present study we showed that a moderate increase in the memory-capacity can be achieved with local , Hebbian learning rule and slightly supra-linear dendritic integration . We emphasized that under certain conditions a single branch is able to evoke somatic output . However , if the amplitude of the individual events is smaller , a larger spatial extent involving the depolarization of additional branches will be required to trigger output spiking . This mechanism could induce a combinatorial increase in the representational capacity as shown by [16] . According to our model hippocampal granule cells can be regarded as a two layer neural network of abstract integrate and fire elements: In the first layer corresponding to the terminal branches the units integrate separately their inputs and they innervate a common output unit ( second layer , the somatic compartment ) that implements a logical OR computation . The idea that a dendritic tree may perform logical computations was originally proposed by [87] to explain directional selectivity of retinal ganglion cells . Shepherd and Brayton [88] further elaborated this approach but instead of branches they used dendritic spines as basic computational subunits . Our approach is more similar to how Poirazi et al . [89] describe hippocampal pyramidal cells , however , in that model the output unit performs ( nonlinear ) summation prior to final thresholding . Another similar model was proposed by Gasparini and Magee [30] , in a paper where they showed that the apical trunk of hippocampal pyramidal neurons integrate spatially clustered and synchronously arriving synaptic inputs nonlinearly , whereas distributed or asynchronous inputs are linearly integrated . They suggest that processing in the nonlinear mode could functionally separate the dendritic arbor into a large number of independent nonlinear computational units , each sending its own output to the soma . In the present paper , we showed that a single computational units is powerful enough to determine the output of the neuron only if there are not too much similar units ( N<100 ) and if the local integration is sufficiently nonlinear . A similar picture emerged form a recent series of in vitro experiments performed on the basal dendrites of neocortical pyramidal neurons: These branches behave as independent computational subunits as nearby inputs on the same branch summed sigmoidally due to the presence of local NMDA spikes [2] , [45] , [90] and synaptic plasticity required the pairing of local NMDA spikes with biochemical signals [7] . Moreover , an NMDA-spike localized to a single basal dendrite could efficiently induce somatic UP-state like depolarization accompanied by bursts of action potentials [5] . These results suggest that our model describes remarkably well the neuronal computations performed by the basal dendritic tree of pyramidal neurons . Although we tried to fit our model to the available experimental data we had to make some assumptions regarding the integration of neighbouring inputs in dentate granule cells . Moreover , based on the model described in the present paper we make some explicit predictions . Both the assumptions and the predictions of our model should be tested experimentally .
We used the data from [28] to estimate the passive membrane parameters of the granule cells the DG ( Table 1 ) . First we computed the membrane area of a single branch ( Adend ) falling into the perforant path termination zone ( the outer two third of the dendritic tree ) : ( 12 ) where ld is the total length of the dendritic tree , db = 1 . 1 µm is the average diameter of a single branch , N is the number of branches and α = 1 . 9 is a correction factor for the membrane area of dendritic spines . Similarly , the area of the somatic compartment ( Asoma ) , assuming a sphere with diameter ds: ( 13 ) The area of the cross section of a single branch is , and the length of the proximal third of the branches , that do not receive input from the entorhinal cortex is lds = 50 µm . Finally , we estimate the parameters in Eqs . 1–2: ( 14 ) ( 15 ) ( 16 ) where Rm and Ri are the membrane resistance and the intracellular resistivity , respectively . As the somatic and the dendritic membrane area ( and hence the resistance ) were similar , we used that . The parameter R used in our calculations was R = Ra/Rm0 . 01 for a passive granule cell in the DG . Note that due to the synaptic conductances activated in vivo the membrane resistances of functioning granule cells are certainly lower than its in vitro estimates [59] . A single dentate granule cell receive synaptic input from nEC–DG≈2500–4000 entorhinal layer II cells distributed on N≈25–40 branches , whereas a single branch receives M≈100 synapses in the rat's hippocampus [28] . According to Amaral and Lavenex [71] , there are nDG≈1 . 2 · 106 granule cells in the rat's DG , and nEC≈0 . 11 · 106 projection cells in the layer II of the entorhinal cortex . It is known , that a given location in the hippocampus may receive inputs from more than 25% of the dorsomedial-to-ventrolateral axis of the medial entorhinal cortex [94] , [95] . Therefore , while a single dendritic branch get its M≈100 synaptic inputs randomly from nearly 25000 entorhinal cortical neuron , we assume that each synapse on a dendritic branch comes from different entorhinal neurons . By electrical recordings from different hippocampal regions one can estimate the proportion of simultaneously active cells within a reasonable time window . We call this number the sparseness of the representation in the given area . Specifically , 1–5% of the granule cells are active simultaneously in the DG [24] , [29] , therefore we used spDG = 0 . 05 . The sparseness of the entorhinal input is somewhat larger , spEC = 0 . 2 [27] , [80] , [96] . Experimental data provide a good estimate for the mean firing rate of these neurons , however , they give the variance of the mean across neurons , but not the variance in the firing rate of individual cells . To estimate the variance in the firing rate of an individual cell , we generated random spike trains based on the ISI histogram on Figure 5 of [80] . The expected value and the variance of the number of spikes in a 100 ms time bin ( corresponding to one period of the hippocampal theta rhythm ) was μEC′ = 0 . 32 and and there was at most 4 spikes during 100 ms in the case of an entorhinal excitatory cell . We scaled these values relative to the maximal firing rate , so we had μEC = μEC′/4 = 0 . 08 and characterizing the distribution of the presynaptic firing uj . Possible differences in firing statistics across different ( medial-lateral or dorsal-ventral ) regions in the EC and across individual neurons are neglected here . Next , we start with originally equal synaptic weights , wij = w = 3 · μEC . In this case , if we assume that the firing of entorhinal neurons are independent and identically distributed , we can approximate the total input to a branch with a Gaussian distribution: ( 17 ) where and . The distribution of the total input U is shown on Figure 1C . Learning alters the distribution of the total input Ui = Σjwijuj of dendritic branches ( Eq . 17 ) by modifying synaptic weights . From Eq . 11 used to describe synaptic plasticity , we can see that synaptic weights converge to a fixed point wij = uj whenever the activity of the postsynaptic branch i is above threshold βd . In the stationary state , the weight vector reflects a presynaptic firing pattern . In other words , the learned presynaptic firing pattern is stored in the corresponding synaptic weights . In order to stimulate initial plasticity in naive branches and prevent learning in those branches that have already learned a pattern , we initialized the synaptic weights to wij = w = 3μEC , which is higher than their expected value at the fixed point ( μEC ) . This initialization ensured that the response ( U ) to unlearned inputs decrease during the process of learning , and prevented interference in branches that already have learned a specific pattern . Indeed , synaptic plasticity is enhanced in newly generated granule cells of the hippocampus compared with mature neurons already integrated into functional circuits [36] , [64] , [97] . After learning we can approximate the distribution of the total synaptic input U to a branch by the sum of two Gaussians representing the total input in the case of learned ( ) and not learned patterns ( ) , respectively: ( 18 ) where pl ( pn ) is the probability that one of the branches receive a learned ( not learned ) input , and μl and ( μn and ) are the mean and the variance of the response to learned ( not learned ) inputs . If we have a finite number ( NS ) of different inputs , and each branch learns one of them , then ( 19 ) Distribution of the total input to the dendritic branches before and after learning is shown on Figure 3A . Parameters μn = 1 , σn = 0 . 39 , μl = 5 . 6 and σl = 0 . 58 were estimated numerically based on the reconstructed firing characteristics of entorhinal neurons . We assumed that each branches learned one of the samples and the probability that one of the branches receive its learned input ( N/NS ) was the sparseness in the DG ( spDG≈0 . 05 . Note , that the distribution of is the theoretical distribution of the responses to learned inputs , from which each branch draw only a few ( perhaps one ) sample because learning is very sparse . We recalculated the two functions H ( U ) and K ( U* ) with the new input distributions by replacing μ and σ with μn and σn in Equations 9–10 and 27–28 , and by changing the distribution of U in Eq . 29 from Eq . 17 to Eq . 18 . In these calculations , we neglected the possibility that two ( or more ) branches may both get their learned input at the same time . Finally , we determined the firing threshold by solving the following integral to β ( see Eq . 24–25 ) : ( 20 ) In the case of continuous variables we can write that H ( Ui ) = H ( U ) . The function H ( U ) has the form: ( 21 ) The conditional probability has a form similar to Eq . 8 , except that we have only N−1 random variables from the Gaussian distribution of U ( Eq . 17 ) with parameters μF and , therefore we can write that: ( 22 ) We can compute the second function K ( U* ) as follows: ( 23 ) ( 24 ) ( 25 ) where is the conditional distribution of the somatic activation as and the maximal dendritic input U* . The distribution is similar to the distribution of in Eq . 8 with two important differences: First , we have only N−1 random variables . Second , we know that U<U* , therefore the distribution of the inputs to other branches is different from the Gaussian in Eq . 17 . Hence we can write , that ( 26 ) where and are the conditional expectation and variance of the distribution p[F ( U ) |U<U*] . We calculate and by integrating Equations 9–10 from −∞ to U*: ( 27 ) ( 28 ) where is a normalization factor . Finally we calculate the last term of Eq . 24 , the distribution of U* as follows: ( 29 ) where P ( U ) is the cumulative distribution function ( CDF ) of U and [X]′ marks derivation . The intuition behind Equation 29 is that: First , P ( U ) is the probability that a given input is smaller than U . Second , P ( U ) N is the probability that all inputs are smaller than U , also the ( CDF ) of U* . Third , its derivative [P ( U ) N]′ gives us the probability density function ( PDF ) of U* . The PDF of U is a Gaussian function , its CDF can be expressed with the Gauss error function ( erf{} ) . To calculate the dependence of the dendritic activation on the inputs , we first repeat Eq . 6: ( 30 ) Next , we substitute in Equation 30 with Eq . 7: ( 31 ) The two terms of the sum in Eq . 31 are independent , because Ui is independent from Ujs , therefore we can calculate the distribution of by the convolution of two distributions ( corresponding to the two terms in the sum ) . The second term in Eq . 31 is the sum of independent random variables and we approximate it with a Gaussian ( similarly as we did it for previously , Eq . 8 ) . The distribution of Ui is a Gaussian ( Eq . 17 ) , that we can transform into the first term of Eq . 31 by a Jacobian factor [98]: ( 32 ) where V = F ( U ) . We get the distribution by substituting the first term of Eq . 31 by a Dirac delta distribution . Similarly , we can calculate by first computing a conditional sum in the second term ( Uj+Σk≠{i , j} Uk ) as described by Eq . 22 and then performing the convolution . The R software environment [99] was used to analyze the data and to prepare the figures .
|
Neurons were originally divided into three morphologically distinct compartments: the dendrites receive the synaptic input , the soma integrates it and communicates the output of the cell to other neurons via the axon . Although several lines of evidence challenged this oversimplified view , neurons are still considered to be the basic information processing units of the nervous system as their output reflects the computations performed by the entire dendritic tree . In the present study , the authors build a simplified computational model and calculate that , in certain neurons , relatively small dendritic branches are able to independently trigger somatic firing . Therefore , in these cells , an action potential mirrors the activity of a small dendritic subunit rather than the input arriving to the whole dendritic tree . These neurons can be regarded as a network of a few independent integrator units connected to a common output unit . The authors demonstrate that a moderately branched dendritic tree of hippocampal granule cells may be optimized for these parallel computations . Finally the authors show that these parallel dendritic computations could explain some aspects of the location dependent activity of hippocampal granule cells .
|
[
"Abstract",
"Introduction",
"Model",
"Results",
"Discussion",
"Methods"
] |
[
"neuroscience/theoretical",
"neuroscience",
"computational",
"biology/computational",
"neuroscience"
] |
2009
|
Parallel Computational Subunits in Dentate Granule Cells Generate Multiple Place Fields
|
Even with identification of multiple causal genetic variants for common human diseases , understanding the molecular processes mediating the causal variants’ effect on the disease remains a challenge . This understanding is crucial for the development of therapeutic strategies to prevent and treat disease . While static profiling of gene expression is primarily used to get insights into the biological bases of diseases , it makes differentiating the causative from the correlative effects difficult , as the dynamics of the underlying biological processes are not monitored . Using yeast as a model , we studied genome-wide gene expression dynamics in the presence of a causal variant as the sole genetic determinant , and performed allele-specific functional validation to delineate the causal effects of the genetic variant on the phenotype . Here , we characterized the precise genetic effects of a functional MKT1 allelic variant in sporulation efficiency variation . A mathematical model describing meiotic landmark events and conditional activation of MKT1 expression during sporulation specified an early meiotic role of this variant . By analyzing the early meiotic genome-wide transcriptional response , we demonstrate an MKT1-dependent role of novel modulators , namely , RTG1/3 , regulators of mitochondrial retrograde signaling , and DAL82 , regulator of nitrogen starvation , in additively effecting sporulation efficiency . In the presence of functional MKT1 allele , better respiration during early sporulation was observed , which was dependent on the mitochondrial retrograde regulator , RTG3 . Furthermore , our approach showed that MKT1 contributes to sporulation independent of Puf3 , an RNA-binding protein that steady-state transcription profiling studies have suggested to mediate MKT1-pleiotropic effects during mitotic growth . These results uncover interesting regulatory links between meiosis and mitochondrial retrograde signaling . In this study , we highlight the advantage of analyzing allele-specific transcriptional dynamics of mediating genes . Applications in higher eukaryotes can be valuable for inferring causal molecular pathways underlying complex dynamic processes , such as development , physiology and disease progression .
Identifying the causative genetic variants associated with complex human diseases is only the first step [1] . The major challenge is to understand how these genetic variants cause the disease . The mediating molecular pathways connecting these variants to phenotypes have been more systematically understood in model organisms than in humans [2] . However , even in model organisms there are several examples where a causal genetic variant is not a component of the annotated pathways associated to a trait , making it difficult to fully understand its molecular basis [3] . Having this complete knowledge for complex diseases has a huge potential for development and evaluation of available therapeutic and preventive strategies to counter these diseases [4] . Studying gene expression variation is a standard approach for identification of the causal path from a genetic variant to disease [5 , 6] . Many of these causal genetic variants have been resolved to single nucleotide polymorphisms ( SNPs ) . Several studies in multiple organisms have been performed to study the effects of these variants called as expression quantitative trait loci ( eQTLs ) [7 , 8] . However , for making predictions for the molecular mechanisms underlying a disease , trans-acting SNPs are more challenging than cis-acting . This is due to the difficulty in distinguishing causative effects of these SNPs from the correlative effects since a SNP can: i ) either affect gene expression and the phenotype independently , or ii ) modulate gene expression of downstream molecular players , which in turn causes phenotypic variation ( causal mediators ) , or iii ) modulate the phenotype which then affects the gene-expression [5] . A few pragmatic approaches have been recently tested in model organisms to identify the causal mediators by studying gene expression changes . One approach , for instance , involved utilizing expression information for the causal genetic variants from multiple environments , which was a better predictor to identify the causal molecular intermediates by the fact that they interact persistently with the variant [9] . For developmental and physiological processes , gene expression follows complex dynamic patterns [10] and so the effect of eQTLs on gene expression can be highly context-sensitive , depending on the developmental stage , physiological phase or tissue type [11–13] . Therefore , when the causative molecular effects of a genetic variant are being studied by measuring gene expression , knowledge of the particular temporal phase when the causal variant transduces its molecular effects is crucial . Allele replacement strains have been used extensively for fine-mapping the effects of causal genetic variants associated with a trait [14] . Studying allele-specific gene expression could be yet another useful approach which could be exploited in model organisms such as yeast , to study the precise molecular effects of the causal variant on the trait . This can be done by performing genome-wide expression profiling in a pair of allele replacement strains having the same genetic background except for the allele . Using allele replacement strains , MKT1 ( 89G ) was identified as a causal genetic variant for an efficient completion of sporulation in yeast , called its sporulation efficiency [15] . MKT1 is a putative endonuclease and its molecular role is beginning to be , but not completely understood [9 , 16] . MKT1 has been mapped as a causative gene for several stress-related complex phenotypes , highlighting its extensive pleiotropy [9 , 17–22] , but its functional role in sporulation remains unclear . The developmental process of sporulation in yeast encompasses two meiotic divisions followed by spore formation [23 , 24] . A study performed parallel phenotyping analysis for the yeast deletion collection and identified around 200 genes required for optimal sporulation efficiency [25] . These genes are both sporulation-specific ( i . e . , required only during meiotic processes ) and majorly sporulation-associated ( i . e . , required for general cellular functions during sporulation such as nutrient metabolism and respiration ) . However , the study did not identify MKT1 as one of these genes . It is also not known if MKT1 ( 89G ) affects any of these 200 genes or any other gene to increase sporulation efficiency . The first association of MKT1 and sporulation process was reported in the linkage mapping study between segregants of SK1 and S288c strains [15] . Moreover , MKT1 ( 89G ) was mapped for sporulation efficiency , the end-point of sporulation process . We do not know at which temporal phase during the course of sporulation ( early entry into meiosis , middle progression through meiotic phases , or late spore wall formation ) , MKT1 affects meiosis . In this study , we hypothesized that the use of allele replacement strains for studying genome-wide gene-expression during the temporal phase when the causal variant contributes to the phenotype could provide useful insights for identifying the causal molecular mediators underlying complex trait variation . In a pair of allele replacement strains differing solely for MKT1 causal allele , we characterized the molecular role of MKT1 ( 89G ) in yeast sporulation efficiency variation . Using genetic assays and mathematical modeling for the meiotic events , we identified the role of MKT1 ( 89G ) in the early phases of sporulation . In the specific context of MKT1 ( 89G ) , we studied the genome-wide transcriptional response particularly in the early phase of sporulation and then genetically tested the candidate mediators . Using such an approach , we identified and confirmed novel pathways mediating the effects of MKT1 ( 89G ) in sporulation efficiency variation . The molecular findings resulting from our study demonstrate the advantage of studying allele-specific temporal gene expression dynamics to identify the causal pathways linking genetic variant to complex traits .
Allele replacement of MKT1 in the S288c strain from the endogenous adenine ( 89A ) to guanine ( 89G ) , of SK1 strain , resulted in increased sporulation efficiency [15] . Whole-genome re-sequencing of the MKT1 allele replacement strain followed by a series of backcrosses ( Methods ) , was done to confirm that MKT1 ( A89G ) was the only sequence difference between the S288c parent ( MKT1 ( 89A ) indicated as “S strain” ) and the allele replacement strain ( MKT1 ( 89G ) indicated as “M strain” ) , the two strains used in this study . After 48h , the high sporulating SK1 strain and the M strain showed increased sporulation efficiency compared to the S strain , which was consistent with the previous report [15] ( Fig 1 , Table 1 , Methods ) . Compared to the S strain , the SK1 and M strains showed a 17- and a 9-fold increase , respectively ( P = 1 . 9x10-28 , P = 1 . 0x10-25 , respectively , pair test in Methods ) . Deletion of MKT1 in the S strain resulted in sporulation efficiency similar to the S strain , showing that MKT1 ( 89A ) is a loss-of-function allele for its function in sporulation ( Fig 1 , Table 1 ) . However , it is possible that the MKT1 ( 89A ) gene product may have an activity for other phenotypes . To define the temporal phase during sporulation when MKT1 ( 89G ) contributes to sporulation efficiency , firstly the proportion of yeast cells completing Meiosis I and II ( MI and MII ) in the S , M and SK1 strains were quantified ( Fig 2A , Methods ) . M strain started entering MI/II within 10h in sporulation medium , while S strain did not enter MI/II even after 48h . Using these data , multi-stage modeling for the M strain and the parent strains S and SK1 was done to study the distribution of the cell population in different stages of meiosis ( Methods , S1 File ) . As expected , the model predicted that the difference between the M and the S strains occurred during entry into meiosis ( initial lag phase of sporulation , S1 Fig ) . Hence , our observations and the model suggested an early role of the causal variant of MKT1 in sporulation , which was in agreement with a recent study that showed the contribution of causal variants in critical decision-making steps in the early stages of a phenotypic process [26] . In order to confirm this early role of MKT1 ( 89G ) in sporulation efficiency variation , a tetracycline-repressible dual-system was used to conditionally express MKT1 ( 89G ) ( Methods ) . MKT1 ( 89G ) expression was switched off until 3h after initiation of sporulation , which led to a reduction in the sporulation efficiency of the M strain ( PTet-MKT1 ) equivalent to the S strain ( Fig 2B , S2 Fig ) . This result showed that activity of MKT1 ( 89G ) allele was essential within the first 3h of sporulation . Meiotic initiation is regulated by multiple nutrient signaling pathways [27] . The functional allele of MKT1 has a fitness advantage during growth in glucose-rich conditions [9] . Therefore , we tested if increased sporulation efficiency of the M strain is influenced by expression of MKT1 ( 89G ) during the rich growth medium stage preceding sporulation ( Methods ) . We observed that switching off MKT1 ( 89G ) during growth in glucose had no effect on sporulation efficiency of the M strain ( Fig 2B ) . Altogether , these results indicated that the role of MKT1 ( 89G ) during sporulation was independent of its role during growth in glucose and that the allele played a role in the early response to sporulation . To identify the pathways through which the MKT1 ( 89G ) allele affects early sporulation , we studied the entire range of transcriptional response in the S and M strains during the first 10h of sporulation , with denser sampling in the early phase of sporulation ( Methods ) . An extensive remodeling of gene expression was observed in both strains , which increased as time progressed through sporulation ( S5 Fig ) . As expected , the genes involved in sporulation showed a higher expression in the M strain than in the S strain ( P = 2 . 0 x 10–37 , permutation P = 0 . 16 , Methods , S6 Fig ) . Amongst all genes , we identified 862 gene transcripts showing a statistically significant ( 10% FDR , Methods ) differential expression as a function of time between the M and S strains . No enrichment of any functional category within these differentially expressed genes was observed , indicating the pleiotropic role of MKT1 ( 89G ) and that it might be affecting various aspects of the sporulation process . Comparison of expression profiles of the few known meiotic regulators in the M and S strains showed that IME1 , the master regulator of meiosis [28] , was not differentially expressed . However , NDT80 , the other crucial regulator of meiosis , involved in meiotic commitment [29] , was differentially expressed ( S7 Fig , S4 Table ) . These results suggested that MKT1 ( 89G ) could affect sporulation at the post-transcriptional level of IME1 or at the transcriptional level of NDT80 , both of which could have early regulatory consequences during meiosis [30] . This observation also suggested that the role of MKT1 ( 89G ) during sporulation might be early and upstream to the regulators of meiosis , in agreement with our earlier results ( see Fig 2A and 2B ) . To capture the early role of MKT1 ( 89G ) during sporulation , genes upregulated early in the M strain and either downregulated or expressed later in the S strain , were considered . Thus , differentially expressed genes were clustered based on their expression profiles , separately for the M and S strains ( Methods ) . Clustering gave six and seven clusters in the M and S strains , respectively , from which four major clusters were identified in each strain ( Fig 3A , S5 Table ) . Clusters I and II consisted of genes mostly expressed in the early stages of meiosis with an enrichment for the target genes of IME1 and NDT80 , respectively . In particular Cluster I contained some of the earliest expression changes in the M strain . Comparison of this early cluster between the M and the S strains showed that while 46% ( 71/143 ) of its genes overlapped ( Fig 3B , S5 Table ) , the remaining 72 early expressing genes were uniquely differentially expressed in the M strain ( S6 Table ) . We posited that transcription factor ( s ) whose target genes were significantly enriched within these unique 72 early expressing genes of the M strain might be involved in regulating entry into meiosis . Forty one such transcription factors ( P ≤ 0 . 05 , odds ratio ≥ 1 . 5 ) were identified , which consisted of the regulators of metabolic and mitochondrial signaling ( Methods , S7 Table ) , including sporulation-specific genes , such as IME1 , SIN3 and WTM2 ( a UME1 paralog ) . To evaluate if the approach we used indeed identified the causal mediating genes contributing to sporulation efficiency variation in the context of MKT1 ( 89G ) , we selected a few candidate genes from this list of regulators for further investigation . One of the major concerns while studying gene expression is that transcriptional changes can be buffered at the level of phenotype and so do not always manifest themselves in phenotypic variation [31] . Hence , to avoid this buffering while identifying causal regulators of sporulation downstream MKT1 ( 89G ) , a comprehensive literature survey was done for the selected 41 transcription factors to identify the prime candidate regulators . We did not consider those genes , which have been previously shown to have a causal relationship with sporulation efficiency variation [25] . While prioritizing candidate genes , specifically those regulators were chosen whose functional annotations were related to the processes associated with early regulation of sporulation , such as mitochondrial function and nutrient starvation , but a causal role in sporulation was not known [24 , 27 , 32–35] . From this list , RTG1 , a regulator of mitochondrial retrograde signaling [36] and DAL82 , a regulator of nitrogen metabolism [37] ( Fig 4 , S8–10 Figs , S8 Table ) were selected for further investigation . To test the role of RTG1 and DAL82 in sporulation efficiency variation , their deletions in both M and S strains were phenotyped . Another regulator of retrograde signaling RTG3 [38] , a physical interactor and target gene of RTG1 , showing differential expression in our data , was also deleted in the two strains . Deleting RTG1 , RTG3 or DAL82 reduced the mean sporulation efficiency in the M strain significantly , by about two-fold ( P = 6 . 2x10-10 , P = 2 . 8x10-10 , P = 1 . 6x10-7 respectively , Fig 5A , Table 1 , pair test in Methods ) . This effect was specific to the M strain , because deletion of these genes in the S strain did not affect their mean sporulation efficiency ( Fig 5A , Table 1 , pair test in Methods ) ; and for RTG1 and RTG3 , significant interaction terms were found between the backgrounds ( S and M strains ) and the deletion for these genes ( P = 5 . 8x10-5 , P = 4 . 7x10-3 respectively , interaction test in Methods ) . RTG1 , RTG3 and DAL82 have not been previously identified as involved in sporulation efficiency as determined from a genome-wide deletion screen [25] . Since this deletion collection was made in the S288c background , carrying the non-functional allele MKT1 ( 89A ) , this could be a possible reason for the lack of functional implication . A deletion study in the SK1 strain that contains the functional MKT1 ( 89G ) allele , did not investigate the association of these early sporulation regulators with the process [39] . However , interestingly , an up-regulation of RTG1 in the early phase of sporulation has been observed in SK1 [40] . These results , thus , support our approach of studying the early effects of the causative allele and implicate novel roles for RTG1 , RTG3 and DAL82 in the early phase of sporulation efficiency downstream to MKT1 ( 89G ) . To further investigate if RTG1/3 and DAL82 belonged to the same pathway ( epistatic effect ) or were in separate pathways ( additive effect ) , double deletions for RTG3 and DAL82 were phenotyped in the M strain . Deletion of RTG3 and DAL82 together reduced the mean sporulation efficiency of the M strain by approximately 3-fold ( Fig 5A , Table 1 ) . A non-significant interaction term was obtained between RTG3 and DAL82 ( interaction test in Methods ) , indicating that they regulated sporulation efficiency additively , downstream to MKT1 ( 89G ) . Furthermore , because deletion of RTG3 and DAL82 in the M background only partially reduced the sporulation efficiency to that of the S strain ( P [M ( rtg3∆ dal82∆ ) vs . S] = 2 . 5x10-7 , Fig 5A , pair test in Methods ) , these results indicated that these genes explained a partial role of MKT1 ( 89G ) , and additional complementary pathways were at play . The mitochondrial retrograde signaling pathway gets upregulated in response to altered mitochondrial function and nutrient starvation . This pathway fine-tunes the metabolic and stress response pathways of the cell by affecting glutamate synthesis and mitochondrial DNA maintenance [33 , 41] . Since mitochondrial function with regard to respiration is implicated as a critical regulator of sporulation [42] , we speculated if differential mitochondrial activity was involved in sporulation efficiency variation in the presence of MKT1 ( 89G ) . We evaluated the mitochondrial function in the M and S strains by assaying oxygen consumption flux during early sporulation ( Methods ) . The M strain showed a better mitochondrial function than the S strain ( Fig 5B ) at 1h in sporulation . Deletion of RTG3 in the M strain decreased this oxygen consumption flux , though dal82∆ had no effect on the flux ( Fig 5B ) . These results suggested a role of differential mitochondrial function in sporulation efficiency variation . However , a better understanding of the role of mitochondrial retrograde pathway in sporulation efficiency would require further investigation . Differential mitochondrial activity in the presence of MKT1 ( 89G ) suggests a role for the Mkt1 interactor , Puf3 , a Pumilio-family protein , which has been suggested to explain the extensive MKT1 ( 89G ) pleiotropy during mitotic growth in rich media as well as in stress environments [16 , 22 , 43] . Puf3 is an mRNA binding protein that regulates the fate of nearly 200 nuclear-encoded mitochondrial transcripts [44] . Even though we found a few PUF3 target genes ( 13/214 genes ) differentially expressed during sporulation , none were in the set of unique early expressed transcripts in the M strain ( S10 Fig ) . To further evaluate if PUF3 had a role in sporulation efficiency variation in the presence of MKT1 ( 89G ) , we deleted PUF3 in the S and M strains and M strain with single deletions of rtg3∆ and dal82∆ . If PUF3 has an independent role in sporulation , reduction in sporulation efficiency by puf3∆ deletion would be independent of the background ( MKT1 , RTG3 or DAL82 ) , and we would observe an additive effect on sporulation efficiency . Any observed significant deviation from this expectation would imply dependence . One extreme case of dependence would be epistasis . In that case , deleting PUF3 in these backgrounds would not lead to decreased sporulation efficiency . We observed that PUF3 deletions in all the four backgrounds: M , S , M ( rtg3∆ ) and M ( dal82∆ ) reduced their sporulation efficiency ( regression line y = 0 . 65x showing around 35% less sporulation efficiency for all strains , Fig 6A and 6B , Table 1 , pair test in Methods ) . Furthermore , interaction terms ( Methods ) were non-significant for deletion of PUF3 between the M and the S strains ( P = 0 . 49 ) , the M and M ( rtg3∆ ) strains ( P = 0 . 53 ) , and only mildly significant between the M and M ( dal82∆ ) strains ( P = 0 . 02 ) . These results indicated that the effect of PUF3 on sporulation efficiency was independent of MKT1 ( 89G ) and its downstream genes RTG1/3 and DAL82 .
Over the past decade a detailed genotype-phenotype map for complex traits including diseases has been determined [45] , however , a functional map defining how causal genetic variants ( alleles ) modulate the underlying pathways resulting in phenotypic variation , is missing . Filling this functional gap will help to identify molecular candidates for therapeutic intervention in human diseases and to make useful predictions regarding response to a particular therapy and survival of a patient [1] . The first step to characterize this functional genotype-phenotype map requires identification of the causal mediating genes in a biological network regulating the phenotype . Investigation of the intermediate phenotypes viz . transcripts , proteins and metabolites , is routinely used to identify these causal mediators [46] . In this study we demonstrate a couple of steps essential for accurate identification of these causal molecular mediators: i ) studying allele-specific temporal dynamics of the biological processes underlying complex traits , and ii ) allele-specific functional validation of the predicted mediators . We report the characterization of molecular pathways modulated by a causal genetic variant in a dynamic biological process using the above approach . In particular , we studied the molecular effects of the essential MKT1 ( 89G ) allele on the yeast transcriptome during sporulation . We not only identified novel pathways regulating the phenotype , but also confirmed the independent role of a known interactor ( Puf3 ) of MKT1 ( 89G ) in the phenotype ( Fig 7 ) . MKT1 ( 89A ) is not a naturally occurring allele , observed only in the S288c strain [20] . However , such rare polymorphisms are receiving increasing attention for their contribution to common human diseases [47] . In this sense , our approach has a general applicability since it can be applied to study the molecular basis of both common and rare variants . Using our approach of studying early gene expression dynamics in response to the MKT1 ( 89G ) allele , we identified that regulators of mitochondrial retrograde signaling and of nitrogen starvation act additively to regulate sporulation efficiency ( Fig 5A ) . Mitochondria responds to a wide array of stresses by inducing various complex cellular responses and promoting cellular adaptation to reduce the impact of further stressors [48] . Mitochondrial retrograde signaling is one of the stress signaling responses of the cell during mitochondrial functional alteration and glutamate starvation [33] . It affects mitochondrial DNA maintenance [49] and hence the respiratory competency of a cell . During meiosis in yeast cells , energy production occurs through the Krebs cycle [32 , 35 , 42] , and hence respiration is a critical regulator of meiosis in yeast [42] and in humans . In humans , low mitochondrial DNA has been associated with ovarian insufficiency [50] . We observed an improved mitochondrial activity during early sporulation in the M strain compared to the S strain ( Fig 5B ) . A reduction in this high mitochondrial activity in the absence of mitochondrial retrograde signaling regulator RTG3 indicated that MKT1 ( 89G ) might confer a better stress response through RTG3 , with increased sporulation efficiency being one of the consequences . This role of retrograde signaling in regulation of developmental processes responding to nutritional stresses has been shown for pseudohyphal growth in yeast [51] . Further investigating this association of differential mitochondrial signaling , particularly retrograde signaling with meiosis and development in general can help provide insights into the factors regulating infertility . In this study , we characterized the essential role of MKT1 ( 89G ) allele in sporulation efficiency . This allele was particularly interesting to study as this coding polymorphism of MKT1 is present in all laboratory strains ( except strains isogenic to S288c ) , as well as clinical and natural isolates of yeast including the SGRP strain collection [15 , 18 , 20 , 52] . Since the previous genetic screens [39 , 25 , 53] or genome-wide expression studies [40 , 54] for sporulation and sporulation efficiency , were done in the S288c background carrying the MKT1 allele which is non-functional in sporulation , this could be a possible reason for not identifying MKT1 to be involved in the process . The founder strain of S288c , EM93 carries the MKT1 ( 89G ) allele suggesting that during domestication of S288c this functional allele was lost [20 , 55] . During evolution of S288c in low-glucose conditions , the native MKT1 ( 89A ) mutated to MKT1 ( 89G ) within 500 generations [56] , also indicating the crucial role of MKT1 ( 89G ) in stress-related conditions . Altogether , these observations demonstrate the limitations of studying genotype-phenotype relationships in a single genetic background , especially in laboratory strains , which might have degenerated their stress response machinery partially or completely , as a result of domestication [57] . Using our approach , we further showed an MKT1 ( 89G ) -independent role of PUF3 in meiosis ( Fig 6A and 6B ) . This was surprising since eQTL mapping studies have suggested MKT1 as a global regulator of gene expression [22 , 58] and have identified its most upstream interactors , such as PUF3 , during mitotic growth in multiple environments [9 , 16] . Puf3 regulates translation and degradation of nuclear-encoded mitochondrial mRNAs by localizing them near mitochondria or P-bodies , which are cytoplasmic sites for mRNA decay and stalling [16 , 44 , 59 , 60] . Since MKT1 has a post-transcriptional regulatory role both in yeast [61] and in trypanosomes [62] , its interaction with PUF3 suggested a probable mechanism for understanding the role of MKT1 . However , for sporulation efficiency , we observed that Puf3 showed an MKT1 ( 89G ) -independent role . We , therefore , speculate that Puf3 might be a mitotic growth-specific interactor of MKT1 ( 89G ) . Its role in sporulation efficiency , though , could involve post-transcriptional regulation of mitochondrial mRNAs through P-bodies during sporulation . In Drosophila , C . elegans , mice and mammals [63 , 64] , P-bodies related RNA granules are known to be involved in translational control of germ cell transcripts . However , in yeast , P-bodies have been observed only during glucose starvation and stress conditions such as ethanol tolerance [22 , 65] . Therefore , our results indicate an interesting interaction between Puf3 and sporulation efficiency variation and this could be a future line of investigation to determine if P-body formation has a regulatory role in yeast meiosis . Through our analysis , we attempted to understand the molecular basis of a complex trait . Using an allele-specific approach , we determined and functionally validated the molecular consequences of a single causative variant in phenotypic variation . This approach helped to identify novel associations between mitochondrial and metabolic pathways with meiosis . Further analyses of these expression data can identify additional regulators and pathways involved in sporulation efficiency variation in the presence of MKT1 ( 89G ) ( Fig 7 , S7 Table ) . This approach demonstrated in yeast can be applied to higher eukaryotes to study transcriptional dynamics of developmental processes or progression of diseases . This will assist in understanding the precise genetic effects of a causal variant , improving the existing genotype-phenotype functional relationship map .
Whole-genome resequencing of the MKT1 allele replacement strain ( S9 Table ) was performed to confirm the presence of the causative SNP ( details in S1 Table , S1 Text ( Section 1 ) ) . Backcrossing the haploid allele replacement strain to the S288c parent strain three consecutive times ( details in S1 Text ( Section 2 ) ) confirmed that homozygous MKT1 ( A89G ) was the only sequence difference between the diploid S288c parent ( S strain ) and the allele replacement strain ( M strain ) . All the S ( MKT1 ( 89A ) ) and M ( MKT1 ( 89G ) ) strains used in this study were derivatives of S288c strain except SK1 strain ( S9 Table ) . The strains were grown at 30°C in YPD ( 1% yeast extract , 2% bacto peptone , 2% dextrose ) and YPA ( 1% yeast extract , 2% bacto peptone , 1% potassium acetate ) . Deletions were performed in the haploids by replacing the specific ORF with one of the dominant drug-resistance cassettes ( hphMX4 , kanMX4 or natMX4 ) which were PCR-amplified from their respective plasmids as described previously [66] . The strains were transformed using the standard lithium acetate-based method [67] and homologous integration of the deletion cassette was confirmed by performing a colony PCR for both the ends . Three confirmed independent transformants were selected to minimize random mutations during the transformation step , diplodized using pHS2 plasmid ( containing a functional HO ) and phenotyped . All further experiments were performed using the diplodized parent strains and their diploid derivatives . The primers for deletions and their confirmations are listed in S10 Table . Sporulation conditions and the calculation of sporulation efficiency was done as previously described [68] in liquid sporulation medium ( 1% potassium acetate supplemented with 20mg/ml uracil , 20mg/ml histidine , 30mg/ml leucine , 20mg/ml methionine and 30mg/ml lysine ) . For each strain , minimum three biological replicates were used and approximately 1 , 000 cells were counted per replicate . Fold difference was calculated as the ratio of mean sporulation efficiencies of the two strains A and B when the sporulation efficiency of A is greater than of B . Two statistical tests were used: the pair test and the interaction test . The pair test tests the null hypothesis that two given strains have the same sporulation efficiency . To this end , the number yi , k of sporulated cells ( 4-nuclei count ) among the total number of cells ni , k of strain i in replicate experiment k was modeled with a quasi-binomial generalized linear model using the logit link function and subject to a common log-odd ratio βi between replicates , i . e . : log ( μi , kni , k−μi , k ) =βifor allk , where μi , k = E ( yi , k ) . The pair test tests the null hypothesis of equality of log odd-ratios for two strains i and j , i . e . H0: βi = βj . The interaction test tests the null hypothesis that the effect of mutation A is independent of the effect of mutation B , taking the M strain as reference background . This test thus compares four strains: mutation A only , mutation B only , both A and B and neither A nor B ( M strain ) . Here , the strain S was considered as a M strain mutated for MKT1 ( 89 ) . For every interaction test , we considered the dataset of the four strains of interest and fitted a quasi-binomial generalized linear model using the logit link function and subject to: log ( μi , kni , k−μi , k ) =β0+βAAi+βBBi+βA , BAiBifor allk , where , Ai and Bi are indicator variables of the mutations A and B in strain i respectively . The interaction test tested the null hypothesis that the odd ratio of sporulation in the double mutant equals the product of the odd ratios of each mutation , i . e . H0: βA , B = 0 . Both the pair test and the interaction test were implemented in the statistical language R with the function glm ( ) assuming a constant variance function fitted by maximizing the quasi-likelihood and using the t-test on tested parameters ( see S2 File for raw data and R script ) . Aliquots of sporulating cells of M strain culture were fixed with ethanol at regular intervals ( as indicated in Fig 2A ) from 0 to 48h in the sporulation medium . These time-points were chosen to capture the progression through meiotic stages in the strain . Samples were stained with DAPI ( 4’-6’ diamidino-2-phenylindole ) using the standard methods [69] for calculating the proportion of cells with 1-nucleus ( Non-sporulating/G1 ) , 2-nuclei ( MI ) and 4-nuclei ( MII ) using Carl Zeiss Axiovert 200 fluorescence microscope . For each strain , proportion of cells were counted till saturation was reached for two consecutive time points . Grey scale images were captured using a CCD camera and pseudo-coloured using the image acquisition software ( Axiovision ) supplied with the microscope . To estimate the sporulation efficiency and DAPI staining , 1 , 000 cells from the three biological replicates for each strain were counted . A multi-stage modeling was performed ( details and raw data in S1 File ) . Cells in G1/S phase of cell cycle are said to be in 1-nucleus state . Cells that have completed MI or MII are said to be in 2-nuclei or 4-nuclei state , respectively . Cells that did not progress from one cell cycle state to another are mentioned as inactive cells . The existence of inactive states is supported by the fact that at steady state , some cells still have one nucleus or 2-nuclei indicating they are trapped at these stages , which could be possibly due to nuclear destruction mechanism resulting in dyads [70] . Hence , cells could be either in a 1-nucleus active , 1-nucleus inactive , 2-nuclei active , 2-nuclei inactive or 4-nuclei state . Moreover the cells were assumed to only progress in one direction ( no back transitions ) from the 1-nucleus active to either the 1-nucleus inactive or the 2-nuclei active stage , and from the 2-nuclei active to either the 2-nuclei inactive or to the 4-nuclei state . The samples contain a large number of cells and thus we used Ordinary Differential Equations to describe the dynamics of the system . The dynamics was modeled with an initial lag phase ( measured as τ ) followed by first order kinetics between the stages ( measured as α , β , γ and δ , as shown below ) . ( X1→αX2→γX4X1→βY1X2→δY2 ) where , X1 is proportion of cells in 1-nucleus active stage , X2 in 2-nuclei active stage , X4 in 4-nuclei active stage , Y1 is proportion of cells in 1-nucleus inactive stage , Y2 in 2-nucleus inactive stage . The model was fitted by minimizing least square errors to the measured proportions of the cells with 1 , 2 , and 4-nuclei , measured along the time . Confidence intervals were obtained by bootstrap of the data . tetO7-based promoter substitution cassette containing kanMX4 , amplified from the plasmid pCM225 [71] , was inserted to replace the endogenous MKT1 promoter ( -300 to -1bp upstream start site ) in the M strain ( PTet-MKT1 ) . M strains with the endogenous promoter ( Pwt-MKT1 ) and the tetO7 promoter ( PTet-MKT1 ) were grown in a glucose-rich medium ( YPD ) and synchronized in pre-sporulation medium ( YPA ) prior to initiating sporulation . To determine the concentration of doxycycline at which the effect of MKT1 ( 89G ) on sporulation efficiency is similar to MKT1 ( 89A ) ( implying MKT1 ( 89G ) is not functional or OFF ) , the PTet-MKT1 strain was grown and sporulated in 2 , 3 and 5μg/ml of doxycycline and phenotyped by estimating the sporulation efficiency after 48h . At 5μg/ml doxycycline , the sporulation efficiency of the PTet-MKT1 strain was similar to the S strain ( S2 Fig ) and this concentration was used for further experiments . To switch off MKT1 ( 89G ) expression only during the growth in glucose , the PTet-MKT1 strain was grown in YPD with doxycycline , washed and added to YPA and the sporulation medium in the absence of doxycycline . For switching off MKT1 ( 89G ) throughout the sporulation process , doxycycline was added to all the three media ( YPD , YPA and sporulation ) . To switch off MKT1 ( 89G ) till 3h in sporulation medium , doxycycline was added in YPD , YPA and sporulation medium . Cells were washed after 3h in sporulation and resuspended in the sporulation medium without doxycycline till 48h , and were phenotyped . A complementary experiment where MKT1 ( 89G ) was switched ON till 3h in sporulation medium and switched OFF from 3h to 48h in sporulation was done by adding doxycycline in sporulation medium post 3h in sporulation medium ( S2 Fig ) . For each strain in each condition , minimum three biological replicates were used and approximately 1 , 000 cells were counted per replicate per condition for estimation of sporulation efficiency . The means and variances were tested for significance using one-way ANOVA followed by Tukey’s multiple comparisons test ( Prism , Graphpad Software Inc . ) . Statistical significance was determined at P < 0 . 05 . Temporal transcriptome profiling was performed for the sporulating yeast cells at 0h , 30m , 45m , 1h10m , 1h40m , 2h30m , 3h50m , 5h40m and 8h30m ( logarithmic time-series ) in the sporulation medium . For this , 100ml aliquots of the culture were pelleted and stored at -80°C . Transcriptome profiling was performed using the S . cerevisiae yeast tiling array ( Affymetrix , Cat# 520055 ) as described previously [72] . Time-series arrays of M and S strains in sporulation were normalized by vsn ( S1 Text ( Section 3 ) , S3 Fig ) [73] . Using log2 transformed expression values , after normalization ( S2 Table ) , the expression profiles of all transcripts of S and M strains were made continuous over time using locfit [74] with the bandwidth parameter ‘h’ optimized at 1 . 21 ( S1 Text ( Section 4 ) , S4 Fig , S3 Table ) . A baseline transformation for each transcript , after smoothing , was done by subtracting each time point value from t = 0h ( t0 ) . y'S ( tn ) =yS ( tn ) −yS ( t0 ) y'M ( tn ) =yM ( tn ) −yM ( t0 ) where , y is the expression value of a transcript for a strain ( S or M ) at a specific time point and y’ is the transformed expression value . To compare the sporulation genes ( obtained from Deutschbauer et al . [25] ) between the M and S strains , their expression in the two strains were tested using 1 , 000 permutations of Wilcoxon test on an equal number of randomly selected genes ( S6 Fig ) . R scripts used for the analyses are given in the S3 File . To identify differentially expressed genes ( after removing tRNAs , snRNAs and transcripts from terminal repeats ) between the two strains , the temporal expression profiles of each transcript was compared using the method implemented in the EDGE ( Extraction of Differential Gene Expression ) software [75] . One thousand permutations were done to calculate the null distribution with a random number seed . EDGE analysis identified transcripts of 862 significant differentially expressed genes across time ( 10% FDR , S4 Table ) . Within these 862 genes , a subset of differentially expressed transcription factors and differentially expressed targets of all the transcription factors ( obtained from the YEASTRACT database , [76] were selected . This subset of 727 genes was used for further analysis . The 727 differentially expressed genes were clustered according to their temporal expression patterns using time abstraction method implemented in the TimeClust software [77] . The smoothened and baseline transformed expression data of the 8 sporulation time-points was analysed with window span parameter set at 3 . An absolute expression change of 0 . 1 was considered as a change . This clustering method was applied on the expression data separately for the two strains resulting in six and seven clusters in the M and S strains , respectively ( S5 Table ) . The gene lists of the M and S strains for the Cluster I , consisting of early expressing genes , were compared . For the genes unique to the M strain in this cluster ( S6 Table ) , the transcription factors regulating them were extracted using the YEASTRACT database ( S7 Table ) [76] . After 1h in sporulation , 5 x 106 cells from each of the three biological replicates were used for the assay . Oxygen consumption flux was determined , in total volume of 2 . 1ml sporulation medium at 30°C with 500 rpm , using OROBOROS O2k high-resolution respirometer ( OROBOROS Instruments Corp . , Innsbruck , Austria ) . Data acquisition and calculation of oxygen flux was done according to the manufacturer’s instruction in DatLab software . Unpaired Student’s t-test ( Prism , Graphpad Software Inc . ) was performed for comparing differences between the means of the two strains . Statistical significance was determined at P < 0 . 05 .
The Supporting information is also available at: http://www . tifr . res . in/~dbs/faculty/hsinha/MKT1Spo
|
The causal path from a genetic variant to a complex phenotype such as disease progression is often not known . Studying gene expression variation is one approach to identify the mediating genes , however , it is difficult to distinguish causative from correlative genes . This becomes a challenge especially when studying developmental and physiological traits , since they involve dynamic processes contributing to the variation and only single static expression profiling is performed . As a proof of concept , we addressed this challenge here in yeast , by studying genome-wide gene expression in the presence of the causative polymorphism of MKT1 as the sole genetic variant , during the time phase when it contributes to sporulation efficiency variation . Our analysis during early sporulation identified mitochondrial retrograde signaling and nitrogen starvation as novel regulators , acting additively to regulate sporulation efficiency . Furthermore , we showed that PUF3 , a known interactor of MKT1 had an independent role in sporulation . Our results highlight the role of differential mitochondrial signaling for efficient meiosis , providing insights into the factors regulating infertility . In addition , our study has implications for characterizing the molecular effects of causal genetic variants on dynamic biological processes during development and disease progression .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] |
[] |
2015
|
Temporal Expression Profiling Identifies Pathways Mediating Effect of Causal Variant on Phenotype
|
The variation in the expression patterns of the gap genes in the blastoderm of the fruit fly Drosophila melanogaster reduces over time as a result of cross regulation between these genes , a fact that we have demonstrated in an accompanying article in PLoS Biology ( see Manu et al . , doi:10 . 1371/journal . pbio . 1000049 ) . This biologically essential process is an example of the phenomenon known as canalization . It has been suggested that the developmental trajectory of a wild-type organism is inherently stable , and that canalization is a manifestation of this property . Although the role of gap genes in the canalization process was established by correctly predicting the response of the system to particular perturbations , the stability of the developmental trajectory remains to be investigated . For many years , it has been speculated that stability against perturbations during development can be described by dynamical systems having attracting sets that drive reductions of volume in phase space . In this paper , we show that both the reduction in variability of gap gene expression as well as shifts in the position of posterior gap gene domains are the result of the actions of attractors in the gap gene dynamical system . Two biologically distinct dynamical regions exist in the early embryo , separated by a bifurcation at 53% egg length . In the anterior region , reduction in variation occurs because of stability induced by point attractors , while in the posterior , the stability of the developmental trajectory arises from a one-dimensional attracting manifold . This manifold also controls a previously characterized anterior shift of posterior region gap domains . Our analysis shows that the complex phenomena of canalization and pattern formation in the Drosophila blastoderm can be understood in terms of the qualitative features of the dynamical system . The result confirms the idea that attractors are important for developmental stability and shows a richer variety of dynamical attractors in developmental systems than has been previously recognized .
Canalization refers to the constancy of the wild type phenotype under varying developmental conditions [1]–[4] . In order to explain canalization , C . H . Waddington hypothesized that there must only be a finite number of distinct developmental trajectories possible , since cells make discrete fate decisions , and that each such trajectory , called a chreod , must be stable against small perturbations [5] . One aspect of canalization , the buffering of phenotypic variability against genotypic variability in wild type , has received considerable experimental [2] , [6]–[10] and theoretical [11]–[13] attention . The phenomenon of canalization of genotypic and environmental variation was seen by Waddington as a consequence of the underlying stability of developmental trajectories , an idea supported by theoretical analysis [13] . But this central idea of Waddington's has heretofore received little attention in real developmental systems because of a lack of relevant quantitative molecular data . The further investigation of Waddington's hypothesis is of great importance because it provides a scientific connection between the reliability and invariance of the formation of cell types and tissues in the face of underlying molecular variability , as we now explain . Quantitative molecular data permitting the study of developmental canalization are now available for the segment determination process in Drosophila [14] . The segmented body plan of the fruit fly Drosophila melanogaster is determined when the embryo is a blastoderm [15] by the segmentation genes [16] . Quantitative spatiotemporal gene expression data show that the maternal protein gradients and the early expression patterns of the zygotic gap and pair-rule genes vary a great deal from embryo to embryo [14] , [17] . The variation of the expression patterns of the gap and pair-rule genes decreases over time so that it is significantly lowered by the onset of gastrulation at the end of cellularization ( [14] , Fig . 1 ) . The observed reduction of variability over time in the segmentation gene system suggests that the developmental trajectory of the Drosophila embryo is stable against perturbation . The characterization of the stability properties of the developmental trajectory is central to our understanding of the mechanisms that underlie canalization [3] . In the case of the gap genes , we have shown elsewhere [18] that variation reduction relative to the maternal gradient Bicoid ( Bcd ) occurs because of gap gene cross regulation . Using a gene circuit model of the gap gene network [18]–[22] we identified specific regulatory interactions responsible for variation reduction in silico and verified their role in canalization experimentally . Importantly , the model reproduces the observed low variation of gap gene expression patterns [18] , which provides an opportunity to analyze the properties of the system that give rise to developmental stability . These results raise two generic problems that occur in the analysis of complex numerical models . First , even if the model describes a natural phenomenon faithfully , understanding the natural phenomenon is only achieved when the model's behavior can be understood as well . The complexity of the model , unsurprising in terms of the underlying complexity of the biological system itself , poses a significant challenge to understanding model function . Second , any model is an approximation to the actual mechanisms operating in an organism . The model's behavior must be robust to perturbation , since organisms develop and function reliably even though the underlying mechanisms are subject to a wide variety of perturbations and stresses . There is extensive molecular variability among cells and embryos ( [14] , [17] , [23]–[26]; reviewed in [27] ) and yet there is functional identity between equivalent cell types or conspecific individuals . René Thom tried to resolve this apparent contradiction between the constancy of biological function and the variability in biological substructure by proposing a qualitative topological view of the trajectories of dynamical models [28] . The term “topology” is used here to refer to properties of developmental trajectories that are invariant under continuous deformation . The preservation of these properties ensures the robustness of model behavior , while a qualitative view often leads to an intuitive understanding of complex mechanisms . One such robust property is an attractor state , or a stable steady state of a dynamical system , that attracts all trajectories in some neighborhood of itself . Attractor states are locally stable under small perturbations of the dynamical model [29] , and for this reason it has been proposed that cell fates are attractors [11] , [13] , [30]–[33] . The presence of an attractor state in the phase space of a system implies that there exists a region of phase space , called the basin of attraction , in which all trajectories approach the attractor asymptotically [34] , [35] . This suggests that an attractor is the kind of qualitative robust property that could explain the stability of trajectories , and hence canalization . There are , however , three important considerations to keep in mind when using attractors to describe the Drosophila blastoderm . First , the reduction of variation due to attractors is only guaranteed at late times , but the reduction in the variation of the gap gene expression patterns takes place over about 100 minutes prior to gastrulation . The reduction of variation before gastrulation is biologically essential as the expression patterns of engrailed and wingless , which form the segmentation prepattern , have a resolution of one nucleus and are created by the precise overlap of pair-rule and gap domains [14] , [36] . Furthermore , at about the time of gastrulation the embryo undergoes the midblastula transition [37] , [38] at which time a qualitative change occurs in the genetic control of the embryo . Second , in general there can be more than one attractor in the phase space [39]–[43] . Thus , the basins of attraction need to correspond to biological initial conditions and be large enough to ensure robustness . Finally , the set of attractors found must succeed not only in explaining canalization but also the morphogenetic properties of the system . One such property is the anterior shift of gap gene domains located in the posterior region [14] , [21] , [44] . These shifts are biologically significant and are difficult to reconcile with stable point attractors . In this paper we show that the variation reduction of gap gene expression patterns is a consequence of the action of robust attracting states . We further show that the complex patterning of the gap gene system reduces to the three qualitative dynamical mechanisms of ( 1 ) movement of attractors , ( 2 ) selection of attractors , and ( 3 ) selection of states on a one dimensional manifold . The last of the three mechanisms also causes the domain shifts of the gap genes , providing a simple geometric explanation of a transient phenomenon . In the Gap Gene Circuits section we briefly describe the gene circuit model; see [18] for a full description . For each nucleus in the modeled anteroposterior ( A–P ) region , we identified the attractors in the gap gene phase space , calculated the trajectories , the basins of attraction and other invariant sets such as one dimensional attracting manifolds ( Stability Analysis of the Trajectories of the Gap Gene System section ) . The stability of the trajectories was tested by varying the initial conditions within a biological range , based on gene expression data , that represents the variability of early gap gene expression . We plotted the attractors and several trajectories corresponding to different initial conditions to make phase portraits that show the global qualitative behavior of the system . Finally , we studied how the phase portraits changed as A–P position was varied to infer qualitative pattern formation mechanisms . The biological conclusions about canalization and pattern formation arising from the dynamical characterization are presented in the Mechanisms of Canalization and Pattern Formation section .
The gene circuit used in this study models the spatiotemporal dynamics of the protein expression of the gap genes hunchback ( hb ) , Krüppel ( Kr ) , giant ( gt ) , and knirps ( kni ) during the last two cleavage cycles ( 13 and 14A ) before gastrulation [37] in the Drosophila blastoderm . The protein products of these genes localize to nuclei [45]–[48] so that the state variables are the concentrations of the proteins in a one dimensional row of nuclei along the A–P axis of the blastoderm . The concentration of the protein in the nucleus at time is denoted by . In the model we considered a region , from 35% to 92% egg lenth ( EL ) along the A–P axis , which corresponds approximately to the region of the blastoderm fated to form the segmented part of the adult body [49] , [50] . The gap genes are expressed in broad domains ( Fig . 2A , B; [14] ) under the control of maternal cues . The anterior maternal system acts primarily through the protein gradient Bcd [51]–[53] which is essentially stationary and has an exponential profile ( Fig . 2C; [14] , [51] , [54] ) during the modeled time period . The posterior maternal system is represented by the maternal Hb gradient ( Fig . 2C; [55]–[57] ) . The terminal system regulates gap gene expression by activating tailless ( tll ) and huckebein ( hkb ) [58]–[61] . The terminal system is represented in the model by the Tll gradient , which is expressed posterior to 80% EL in the modeled region during cycles 13 and 14 ( [14] and Fig . S1B ) . tll is considered upstream of the gap genes since its expression pattern is unchanged in gap gene mutants [62] . The concentration of Bcd in nucleus is denoted by and was determined using Bcd data from a representative cycle 13 embryo by an exponential fit , so that ( see [18] for details ) . The concentrations of Tll and another upstream regulator , Caudal ( Cad ) [63] , [64] , were determined by interpolating average data in time [18] . The concentrations of Tll and Cad are denoted by respectively , with an explicit dependence on time , since these gradients are not stationary ( Fig . S1 ) . The dynamical equations governing are given by ( 1 ) where in a gene circuit with genes and nuclei . The first term on the right hand side of Eq . ( 1 ) represents protein synthesis , the second one represents protein transport through Fickian diffusion and the last term represents first-order protein degradation . The diffusion constant , varies inversely with squared internuclear distance , and is the degradation rate . The synthesis term is set to zero during the mitosis preceding the thirteenth nuclear division as synthesis shuts down [65] . Following this mitosis , the nuclei are divided and daughter nuclei are given the same state as the mother nucleus . is the maximum synthesis rate , and is a sigmoidal regulation-expression function . The first term in the argument of represents the transcriptional cross regulation between the gap genes and the genetic interconnectivity is specified by the matrix . Positive elements of imply activation while negative ones imply repression . The regulation of the gap genes by Bcd is represented in the second term and is the regulatory strength . The regulation of the gap genes by upstream time-varying inputs is represented in the third term and is the number of such inputs . There are two such inputs in this model , Cad and Tll , and the elements of the matrix have the same meaning as those of . The last term , , represents the effect of ubiquitous transcription factors and sets the threshold of activation . The initial conditions for Hb are specified using cleavage cycle 12 data . Cycle 12 data are a good approximation to the maternal Hb gradient since the zygotic expression of hb appears to begin in cleavage cycle 13 [17] . The initial conditions for Kr , Gt , and Kni are taken to be zero , since their protein expression is first detected in cycle 13 [14] , [61] , [66]–[68] . The gene circuit's parameters were determined by performing a least-squares fit to a time series of averaged gap gene data [14] using the Parallel Lam Simulated Annealing algorithm ( see Methods ) . This time series has nine points ( time classes; see Table S1 ) , one in cycle 13 and the rest in cycle 14A . The output of the gene circuit ( Fig . 2E , F ) fits the data ( Fig . 2D ) well and its network topology ( Fig . 2K ) is consistent with previous results ( see [18] for discussion and parameters ) . In order to characterize the stability of the trajectories of the gap gene system in terms of qualitatively robust features like attractors , we apply the tools of dynamical systems theory [34] , [69] . Since the gene circuit has variables ( Gap Gene Circuits section ) its state is represented as a point in an -dimensional concentration space , or phase space . In general the concentrations of gap proteins change with time , and hence , a solution of the gene circuit is a curve in this phase space . The gene circuit can also have solutions which do not change with time . Such a solution , called an equilibrium or steady state solution , is represented as a single point in phase space . The positions of the equilibrium solutions in phase space and their stability properties determine the stability of a general time varying solution of the gene circuit . The reader not familiar with linear analysis near an equilibrium point should see Protocol S2 for a pedagogical description of equilibria and their stability in two dimensions . Based on the analysis in the previous section the region of interest , from 35% EL to 71% EL , can be divided into an anterior and a posterior region ( Fig . 2J ) having distinct modes of canalization and pattern formation . The two regions are separated by a saddle-node bifurcation that occurs at 53% EL ( Fig . 3A ) , that is , at the peak of the central Kr domain . We next demonstrate that in the anterior region ( Fig . 2J ) , which extends from the peak of the third anterior gt domain to the peak of the central Kr domain , the state of a nucleus at gastrulation is close to a point attractor . The trajectories are stable by virtue of being in the basin of attraction of the nucleus's attractor state and hence canalize . Pattern formation occurs by the selection of one state from many in a multistable phase space . The concentrations of the Bcd and Cad gradients control pattern formation in the anterior by determining the sizes of the basins and the positions of the attractors , while maternal Hb concentration selects a particular attractor by setting the starting point in its basin . Previous experimental [73] and theoretical [19] work suggested that Bcd and maternal Hb patterned the anterior of the embryo synergistically; our results identify specific roles for Bcd and Hb in anterior patterning . The posterior region extends from the peak of the central Kr domain to the peak of the posterior gt domain ( Fig . 2J ) and its nuclei have phase spaces with very different properties . In this region , the state of the nucleus is far from any attractor state at gastrulation . Instead the state of a nucleus is close to a one-dimensional manifold and canalization is achieved due to attraction by this manifold . Even though the phase space is multistable , the biological range of maternal Hb concentrations in the posterior region place all nuclear trajectories in one basin of attraction . As a consequence , the modes of pattern formation operative in the anterior cannot function in the posterior . Maternal Hb patterns the posterior by determining the position on the attracting manifold which a particular trajectory reaches by the time of gastrulation . These results reveal the mechanism by which maternal Hb acts as a morphogen in the posterior [73]–[75] and also explain the dynamical shifts of gap gene domains [14] , [21] , a significant biological property of the posterior region . We begin the presentation of detailed results by describing the phase spaces of typical nuclei in the two regions , highlighting mechanisms for canalization and pattern formation . An equilibrium is labeled by either ( point attractor ) or ( saddle equilibrium ) , denoted by a superscript , with subscripts denoting the number of eigenvalues having positive or negative real parts . For example denotes the second saddle equilibrium in the modeled region which has one eigenvalue with positive real part and three with negative real parts . Equilibria are also given descriptive names based on which proteins are at high levels ( on ) ignoring the proteins that are at low levels . For example , if a point attractor is at hb-on , Kr-off , gt-on , and kni-off , it is referred to as the “hb , gt-on” attractor .
A discrete [6] , [7] and buffered response to perturbations is the hallmark of a canalized developmental system . Without recourse to molecular data , Waddington sought to explain these two properties of the response by postulating certain favored stable developmental trajectories which he called chreods . Our results ( see Fig . 7 for summary ) show that dynamical systems with multiple attracting states possess both of these properties . Small perturbations are damped because of phase space volume contraction driven by attractors . A discrete response to larger perturbations is a consequence of the discontinuous boundaries between the basins of attraction of a multistable system or of bifurcations . Using a model based on gene expression data , we can conclude that the trajectory of the gap gene system is a chreod . The initial high variation of gap gene expression may arise from early events governed by stochastic laws . Previous observations indicate that the first nuclei in which gap gene transcription is activated are selected probabilistically [68] , [76] . Moreover , the gradients of Hb and Cad proteins are formed by translational repression from the Nanos and Bcd gradients respectively [57] , [77] , [78] under conditions of relatively low molecular number [23] , which is likely to lead to intrinsic fluctuations [79] . Our results show that a deterministic description of gap gene dynamics is sufficient to account for the reduction of initial variation regardless of its source . It is evident however that there are at least two other types of variation that the system might be subject to . First , a natural population will have genotypic variation which , in the framework of the model , would be reflected in the variation of its parameters . Second , gap gene expression itself is likely to be a stochastic process rather than a deterministic one . Notwithstanding this fact , there is no evidence in Drosophila for the coupling of molecular fluctuations to phenotypic fluctuations as seen in prokaryotes [80] , suggesting that molecular fluctuations are buffered in some sense . We emphasize that an attractor is stable against small perturbations of the model itself [29] , and hence is a model property that is preserved to an extent if there is genotypic variation in a population or if errors are introduced by stochastic gene expression . However , further study of both of these aspects of canalization is required in order to more fully understand their role . With regard to pattern formation in the blastoderm , the prevailing theory is that the border positions of downstream genes are determined at fixed values or thresholds of the Bcd gradient [23] , [52] . This idea cannot , however , account for either the low variability of downstream gene border positions [14] , [17] , [18] , or the dynamical shifts of domains in the posterior [14] , [21] . Fixed threshold specification also cannot explain precise placement of the borders in the posterior since the low molecular number of Bcd in the nuclei implies a high level of molecular noise [23] , [81] . In the dynamical picture ( Fig . 7 ) , contrary to the threshold view , Bcd ceases to have a role in positional specification posterior to the peak of the Kr domain since , posterior to this position , the geometry of the phase space does not change qualitatively with A–P position . Instead , maternal Hb acts as a morphogen , obviating the problems arising from a low molecular number of Bcd . Maternal Hb has long been recognized as a morphogen [74] , [75] for the posterior region but the mechanism with which it specifies the posterior region pattern was not clear . As is the case with Bcd , a threshold-based theory for positional specification by Hb [82] is incomplete and requires the postulation of thresholds that can be modified by their targets . The qualitative dynamics provides a viable mechanism for posterior patterning . The attracting manifold is the geometric manifestation of asymmetric repression between the gap genes in reverse order of gap gene domains , . The initial Hb concentration determines which neighborhood of the manifold the trajectory traverses as it is reaches the manifold: Kr-on , kni-on , or gt-on . In other words , posterior patterning works by triggering particular feedback loops in the gap gene network based on maternal Hb concentration . This mechanism also accounts for domain shifts , a property particular to the posterior region , since the trajectories mimic the geometry of the manifold as they approach it . The dynamical analysis of the gap gene system provides a simple and integrative view of pattern formation in the blastoderm ( Fig . 7 ) . The existence of distinct anterior and posterior patterning systems was inferred from the effect of maternal mutations on larval cuticle phenotype and was subsequently characterized in terms of the effects of the Bcd [52] , [73] , [83] , [84] and maternal Hb gradients [56] , [57] , [77] . But where and how is the control of patterning transferred from Bcd to maternal Hb ? Our analysis shows that the hand-off occurs at the A–P position where the Kr-on attractor is annihilated through a saddle-node bifurcation , implying a sharp rather than gradual transfer . With knowledge of the two dynamical regimes , the complex spatiotemporal dynamics of the gap gene system can be understood in the simple terms of three mechanisms: movement of attractors through phase space , selection of attractors by initial conditions , and the selection of states on an attracting manifold ( Fig . 7 ) . Finally , we mention the advantage of having the unexpected mechanism of a one dimensional manifold for canalization and patterning . The Bcd concentration is a bifurcation parameter of the dynamical equations . If there were specific attractors corresponding to each gap gene state , with bifurcations creating and annihilating them successively as the Bcd concentration is varied , the molecular noise in Bcd [23] would give rise to “jitter” or rapid switching between attractors . The manifold with its smooth dependence on maternal Hb is qualitatively robust to such fluctuations . In a connectionist model of cognition [85] , one dimensional unstable manifolds connecting a sequence of saddle points have been proposed as a means of representing transient brain dynamics . The gap gene phase space is a low dimensional projection of the high dimensional phase space of all the molecular determinants in the blastoderm . It may well be that the attractors found in our analysis are actually saddle points in the high dimensional phase space and are way points , with manifolds connecting them , rather than final end points .
The methods used to obtain and characterize the quantitative data are as described in earlier work [14] . All gene expression levels are on a scale of 0–255 chosen to maximize dynamic range without saturation . The numerical implementation of the gene circuit equations is as described [18] , [21] . The gap gene circuit was fit to integrated gap gene data [14] using Parallel Lam Simulated Annealing ( PLSA ) [86] , [87] . PLSA minimizes the root mean squared ( RMS ) difference between model output and data . For each nucleus , data were available at nine time points ( Table S1 ) . Search spaces , penalty function , and other annealing parameters were as described [22] , [88] . The circuit analyzed in detail had an RMS score of 10 . 76 , corresponding to a proportional error in expression residuals of about 4–5% . Equilibria were determined by the Newton-Raphson method as described in Protocol S3 . One-dimensional unstable manifolds of hyperbolic equilibria were calculated by solving the ODEs using the Bulirsch-Stoer [71] method with starting points in the unstable eigenspace of the equilibria [72] . The basin boundaries on the Hb axis were calculated by finding starting points for trajectories that reach saddle points with one positive eigenvalue ( Protocol S3 ) . The time evolution of volume phase space was calculated as described ( Protocol S8 ) . The methods used to calculate the equilibria branches and to determine the type of bifurcations are described in Protocol S4 .
|
C . H . Waddington predicted in 1942 that networks of chemical reactions in embryos can counteract the effects of variable developmental conditions to produce reliable outcomes . The experimental signature of this process , called “canalization , ” is the reduction of the variation of the concentrations of molecular determinants between individuals over time . Recently , Waddington's prediction was confirmed in embryos of the fruit fly Drosophila by observing the expression of a network of genes involved in generating the basic segmented body plan of this animal . Nevertheless , the details of how interactions within this genetic network reduced variation were still not understood . We use an accurate mathematical model of a part of this genetic network to demonstrate how canalization comes about . Our results show that coupled chemical reactions having multiple steady states , or attractors , can account for the reduction of variation in development . The variation reduction process can be driven not only by chemical steady states , but also by special pathways of motion through chemical concentration space to which neighboring pathways converge . These results constitute a precise mathematical characterization of a healing process in the fruit fly embryo .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"developmental",
"biology",
"computational",
"biology/transcriptional",
"regulation",
"mathematics",
"developmental",
"biology/pattern",
"formation",
"computational",
"biology",
"evolutionary",
"biology",
"computational",
"biology/systems",
"biology"
] |
2009
|
Canalization of Gene Expression and Domain Shifts in the Drosophila Blastoderm by Dynamical Attractors
|
Understanding the mechanisms of cell function and drug action is a major endeavor in the pharmaceutical industry . Drug effects are governed by the intrinsic properties of the drug ( i . e . , selectivity and potency ) and the specific signaling transduction network of the host ( i . e . , normal vs . diseased cells ) . Here , we describe an unbiased , phosphoproteomic-based approach to identify drug effects by monitoring drug-induced topology alterations . With our proposed method , drug effects are investigated under diverse stimulations of the signaling network . Starting with a generic pathway made of logical gates , we build a cell-type specific map by constraining it to fit 13 key phopshoprotein signals under 55 experimental conditions . Fitting is performed via an Integer Linear Program ( ILP ) formulation and solution by standard ILP solvers; a procedure that drastically outperforms previous fitting schemes . Then , knowing the cell's topology , we monitor the same key phosphoprotein signals under the presence of drug and we re-optimize the specific map to reveal drug-induced topology alterations . To prove our case , we make a topology for the hepatocytic cell-line HepG2 and we evaluate the effects of 4 drugs: 3 selective inhibitors for the Epidermal Growth Factor Receptor ( EGFR ) and a non-selective drug . We confirm effects easily predictable from the drugs' main target ( i . e . , EGFR inhibitors blocks the EGFR pathway ) but we also uncover unanticipated effects due to either drug promiscuity or the cell's specific topology . An interesting finding is that the selective EGFR inhibitor Gefitinib inhibits signaling downstream the Interleukin-1alpha ( IL1α ) pathway; an effect that cannot be extracted from binding affinity-based approaches . Our method represents an unbiased approach to identify drug effects on small to medium size pathways which is scalable to larger topologies with any type of signaling interventions ( small molecules , RNAi , etc ) . The method can reveal drug effects on pathways , the cornerstone for identifying mechanisms of drug's efficacy .
Target-based drug discovery is a predominant focus of the pharmaceutical industry . The primary objective is to selectively target protein ( s ) within diseased cells in order to ameliorate an undesired phenotype , e . g . , unrestrained cell proliferation or inflammatory cytokine release . Ideally , other pathways within the diseased cells , as well as similar phenotypes in other cell types , should remain unaffected by the therapeutic approach . However , despite the plethora of new potential targets emerged from the sequencing of the human genome , rather few have proven effective in the clinic [1] . A major limitation is the inability to understand the mechanisms or drug actions either due to the complex signaling transduction networks of cells or due to the complicated profile of drug potency and selectivity . Finding drug's targets is traditionally based on high-throughput in vitro assays using recombinant enzymes or protein fragments [2] . The main goal is to characterize the drug's biochemical activity ( binding affinities that describe potency and selectivity ) and depict them in drug-interaction maps [3] . In most cases , once the target ( s ) is known , the in vivo effect on the signaling pathway is validated by measuring the drug's efficiency to inhibit the activity ( usually measured as phosphorylation level [4] ) of the downstream protein . However , beyond that measurement , little is know on how the rest of the signaling network is affected . In addition , in vivo drug effects can hardly be calculated from in vitro assays for several reasons: most kinase inhibitors are promiscuous [5] , there is discrepancy between in vivo and in vitro binding affinities of drugs [6] , and there is an additional discrepancy between in vivo binding affinities and in vivo inhibitor activity for the phosphorylation of downstream signals . To address drug effects in more physiological conditions , novel genomic and proteomic tools have recently been developed [7] . In the genomic arena , large-scale mRNA analysis ( e . g . , [8] , [9] ) enhanced by computational approaches for drug target deconvolution ( e . g . , [10] , [11] ) have been developed . Despite the holistic advantages that genomic approaches have to offer , proteomic-based discovery is a step closer to the function of the cell . Towards this goal , affinity chromatography offers a viable strategy for in-vivo target identification . This approach utilizes a solid support linked to a bait ( usually the drug ) to enrich for cellular binding proteins that are identified by mass spectrometry ( MS ) [12] . However , such experiments usually require large amounts of starting protein , are biased toward more abundant proteins , and result in several hits due to nonspecific interactions [13] , [14] . In order to circumvent the non-specific interaction problem , another bait-based strategy uses quantitative MS with “dirty” inhibitors for baits to immobilize the kinome [15] , [16] . While this approach significantly reduces the non-specific interaction problem , it also limits the target-searching space to those kinases with the highest affinity to the bait . More recently , quantitative MS-based proteomics using SILAC technology [14] extends the search space to all targets that do not bind covalently to the drug . However , incorporation of the SILAC's isotopes requires 5 population doublings and thus , excludes the application on primary cells with limited replication capabilities . Taken together , all techniques listed above can -in the best case scenario- list the affinities of all targets to the drug but no information is provided whether this binding affinity is capable of inhibiting the transmission of the signal to the downstream protein or how those preferential bindings can collectively affect the signaling network of the cell . Here , we describe a significantly different approach to identify drug effects where drugs are evaluated by the alterations they cause on signaling pathways . Instead of identifying binding partners , we monitor pathway alterations by following key phosphorylation events under several treatments with cytokines . The workflow is presented in Figure 1 . On the experimental front , using bead-based multiplexed assays [17] , we measure 13 key phosphorylation events under more than 50 different conditions generated by the combinatorial treatment of stimuli and selective inhibitors . Based on the signaling response and an a-priori set of possible reactions ( i . e . generic pathway ) , we create a cell-type specific pathway using an efficient optimization formulation known as Integer Linear Programming ( ILP ) . This approach builds upon the Boolean optimization approach proposed in [18] . The ILP is solved using standard commercial software packages to guaranteed global optimality ( within a user-defined , numerically small tolerance ) . To evaluate drug effects , we subject the cells with the same stimuli in the presence of drugs and we tract the alterations of the same key phosphorylation events . Then , we reapply the ILP formulation without a-priori assumption of the drug target , and we monitor the changes in the pathway topology with and without drug presence . To demonstrate our approach , we construct a generic map and optimize it to fit the phosphoproteomic data of the transformed hepatocytic cell lines HepG2 . Then , we identify the effects of four drugs: the dual EGFR/ErbB-2 inhibitor Lapatinib [19] , two potent EGFR kinase inhibitors Erlotinib [20] and Gefitinib [21] , and the “dirty” Raf kinase inhibitor Sorafenib [22] . When our method is applied on those 4 drugs we find their main target effect and we also uncover several unknown but equally active off-target effects . In the case of Gefitinib , we find a surprising inhibition of cJUN in the IL1α pathway . In contrast to previously developed techniques , our method is based on the actual effect on phosphorylation events carefully spread into the signaling network . Theoretically , it can be applied on any type of intracellular perturbations such as ATP-based and allosteric kinase inhibitors , RNAi , shRNA etc . On the computational front , our ILP-based approach performs faster and more efficient than current algorithms for pathway optimization [18] and can identify the main drug effects as well as unknown off-target effects in areas of pathways constrained between the activated receptors and the measured phosphorylated proteins . Our fast and unbiased characterization of modes of drug actions can shed a light into the potential mechanisms drug's efficacy and toxicity .
High-throughput bead-based ELISA-type experiments using xMAP technology ( Luminex , Texas , USA ) are performed as briefly described in the Materials and Methods section and in [17] . We create two datasets: one for the construction of cell-type specific topology and another for the identification of the mechanisms of drug actions . To do that , HepG2s are stimulated in 10 different ways with combinatorial treatments with a diverse set of 5 ligands ( TNFα , IL1α , HGF , INS , TGFα , and no stimuli ) and either 4 highly selective inhibitors ( PI3K , MEK , p38 , cMET , and no inhibitor ) or 4 commercial drugs ( EGFR inhibitors Lapatinib , Erlotinib and Gefitinib , and the “dirty” inhibitor Sorafenib ) ( Figure 1b and 1d ) . For the purpose of this paper , we refer to “inhibitors” as the compounds for which we know the target and we use them in a concentration capable to block ∼95% of the downstream protein . Conversely , we refer to “drugs” as the compounds for which we assume no a-priori knowledge of their target . For each combination of cytokine and drug/inhibitor we collect cell lysates at 5 and 25 minutes . The two time points are pooled together in 1∶1 ratio and the mixed lysates are used as an indicator of the “average early signaling response” . For each treatment we measure 13 protein phosphorylations that we consider “key protein activities” ( raw data in Figure S1 ) . The key phosphorylation signals ( listed in Materials and Methods ) are chosen based on the availability of the reagents and quality controls performed at the early phases of the experimental setup [17] . The raw data ( arbitrary fluorescent intensities ) are normalized to fit logic models as described in [18] using a non-linear transformation that converts raw data into values between 0 and 1 where 1 corresponds to the fully activated state and 0 to no-activation . It has to be noted that logic-transformed data depends on what should be considered “protein activation” ( transformed value >0 . 5 ) , a criterion that is embedded in the transformation function and accounts for signal-to-noise limits , saturation of the detection scheme , and eliminates biases that could have been introduced by the variability of antibody affinities [18] . The generic pathway map is constructed in the neighborhood of the 5 stimuli and the 13 measurements . The ubiquitous presence of conflicting reports on pathway maps and alternative protein names makes this step a highly nontrivial one . We explored several pathway databases including STKE , Pathway Interaction Database , KEGG , Pathway Commons , Ingenuity , and Pathway Studio [23] , [24] . Our limited intracellular protein coverage makes impractical the reduction of very large pathway datasets such as those found in Pathway Commons . Here , we create the initial topology from the union of canonical pathways found in Ingenuity ( Redwood City , California ) with subsequent manual curation . A detailed description of Boolean representation of pathways can be found elsewhere [18] , [25]–[29] . In the present manuscript as opposed to [18] , the connectivity in our pathway ( Figure 2 , left panel ) is represented with OR gates and only few connections ( represented with small black circles in Figure 2 ) require an AND gate . We are therefore not comparing OR vs . AND gates , but rather assuming our pathways to be ‘causal’ graphs , and since there are a few AND gates we refer to it as Boolean model . The formulation for the optimal pathway identification is a 0–1 Integer Linear Program , i . e . , an optimization problem with binary variables and linear constraints ( see Materials and Methods ) . The optimizer picks values for the decision variables , such that the logical constraints are satisfied and the objective ( s ) optimized . The primary objective is to find an optimal pathway , i . e . , a pathway that best describes a set of phosphoproteomic data under a given model ( e . g . Boolean ) . A secondary objective is that the pathway is as small as possible , i . e . , has as few connections as possible , such that the best-possible fit of the experiments is maintained ( see Materials and Methods ) . It is shown that some of the binary variables can be relaxed to continuous , without changing the feasible set . The ILP is solved with the state-of-the-art commercial code ( CPLEX [30] , [31] ) that guarantees minimal error between experimental data and the Boolean topology . The goodness of fit ( percent error as described in Materials and Methods ) was decreased from 36 . 7% on the generic map to 8 . 3% on the optimized map ( Figure 2 ) . The main source of error is the inability of TGFα to activate the IRS1_s ( serine residue of IRS1 ) ( see the red background on the IRS1 row at the bottom panel of Figure 2 ) . This is a result of the infeasibility of the generic pathway to satisfy the activation of IRS1_s in a TGFα/IL1α-dependant but HGF/INS-independent manner: TGFα activation of IRS1_s requires mTOR activation via AKT which the optimization algorithm removes to satisfy the inactivation IRS1_s by INS that shares the same path with TGFα . This example highlights the importance of multi-perturbations to better constrain the optimization formulation . Figure 2 shows the optimized topology of HepG2s . Our ILP formulation uses two subsequently-imposed objective functions to remove reactions that do not fit the experimental data . During the optimization of the first objective the ILP formulation ( A ) keeps reactions that lead to phosphorylations of the key proteins and ( B ) removes reactions that lead to false protein activations . An example of the first case is the Insulin ( INS ) -induced AKT activation that is maintained via the INS→IRb→IRS1t→PI3K→PIP3→PDK1→AKT path ( see INS to AKT path in Figure 2 ) . An example of a removed reaction is the TNFR→PI3K reaction which is removed because there is no TNFα induced AKT activation ( see TNFR→PI3K→…→AKT in Figure 2 ) . During the optimization of the secondary objective ( see Materials and Methods ) , several reactions with no evidence of their existence ( no downstream measurements , or no stimuli ) are removed . In this step , the overall goodness of fit is not improved , but the size of the topology is reduced . To illustrate this case , we add to the initial topology the receptor IL6R but the associated stimulus IL6 is not introduced on the experiments . After the secondary optimization , all downstream reactions of IL6 are removed because no data are present ( see reaction arrows downstream for of IL6 in Figure 2 ) . Similarly , all reactions downstream of the bottom-of-the-network key proteins are removed ( e . g . CJUN→CFOS reaction in Figure 2 ) . All those reactions might be present in reality and could have been kept if the secondary objective was not present . Here , we apply the secondary objective and follow a network trimming which removes all reactions that might be present in the cell but due to the lack of measured signals or experimental conditions cannot be verified . The resulting network is significantly smaller but contains only elements for which there are solid experimental evidence that explain the topology . To validate our model , we also examine three scenarios where we remove 20% of our experimental data , and then we try to predict them . Specifically , we create three training datasets , each time by removing all cases where one inhibitor is present ( either MEKi , PI3Ki , or p38i ) and then we calculate how well our ILP-optimized map can predict each of the inhibitor cases ( see Figure S2 ) . For the MEKi , PI3Ki , and p38i scenarios the goodness of fit is 8 . 22% , 9 . 46% , 7 . 05% respectively and our ILP-formulation converges on the same or slightly less optimal solutions compared to the solutions obtained when the whole dataset is used for training ( 4 . 47% , 7 . 76% , and 7 . 05% respectively ) - See Figure S2 . Note that the errors given refer only to the subset considered in each case , not the entire dataset . More extensive validations for Boolean-type models on similar phospho-proteomic dataset can also be found in Saez-Rodriguez et al . [18] . In order to compare the ILP algorithm with the previously published genetic algorithm ( GA ) we use the same initial topology and the same normalized dataset [18] . The two algorithms reached almost identical results ( see Figure S3 ) . For the ILP , the computational requirements are manageable , in the order of a few seconds ( 14 . 3 seconds for this example ) on an Quad Core Intel Xeon Processor E5405 ( 2 . 00GHz , 2X6M L2 , 1333 ) running Linux 2 . 6 . 25 . 20 ( using only one core ) . In comparison , the same optimization problem using GA requires approximately 1 hour on a similar power computer . The optimal pathway furnished by the ILP matches all but 98 out of 880 experimental data , as opposed to 110 mismatches in the topology furnished by the GA . It has to be noted that GA does not provide termination criteria , and it is conceivable that after even larger CPU times the GA would have achieved the same fit as the ILP . In contrast the deterministic solution of the ILP guarantees that an optimal fit ( not necessarily unique ) has been identified within a user-specified tolerance ( 10−3 in our case ) . In addition to the guaranteed optimal solution , commercial ILP solvers are fast , robust and reliable . Note that open-source ILP solvers also exist , but in our experience are not yet adequate . Note also that for larger network topologies , the differences in CPU time will become even more dramatic , rendering the GA intractable . The notable differences between the proposed method and the method used in [18] is mainly due to fundamental algorithmic differences: the technology behind deterministic ILP solvers ( branch-and-bound , branch-and-cut ) is more sophisticated than genetic algorithms , it employs the inherent linearity of the problem , and makes use of the good scalability of linear programs ( sub-problems in branch-and-bound tree ) . In contrast , GA treats the model as a black-box and does not exploit the problem structure . Another point is that herein we used a well-established commercial solver , whereas Saez-Rodriguez et al . [18] used their own implementation of GA . Commercial deterministic ILP solvers , such as CPLEX , rely on several decades of research and development , and have extremely powerful features such as pre-processors and node selection heuristics . Thus , they typically become the default choice for ILPs . For the identification of the drug effects we make use of the second dataset in HepG2s where drugs are applied together with the same set of ligands . In this case , the ILP formulation is being used with the HepG2 specific topology ( topology obtained from the previous step ) and not the generic map . We also do not impose inhibitor constrains the way we do for pathway optimization ( e . g . , PI3K inhibitor blocks the signal downstream of PI3K ) but we let the optimization algorithm decide which reaction ( s ) should be removed in order to fit the drug-induced data . The effect of Lapatinib ( Figure 3a ) , the most selective and specific EGFR inhibitor [32] , is the complete removal of the downstream reactions of the TGFα branch: TGFα→GRB2→SOS→RAS→PI3K and RAS→RAF1→MEK1/2→ERK1/2 . This resulted from the fact that Lapatinib blocks the TGFα induced MEK1/2 , ERK1/2 , and AKT phosphosignals ( Figure 3e ) . Note that the PI3K→…→AKT branch is not removed because it is being used by the HGF and INS path for the activation of AKT that cannot be blocked by Lapatinib ( Figure 3e ) . Gefitinib , an EGFR tyrosine kinase inhibitor , alters the topology in a very similar pattern as Lapatinib , but , interestingly enough , it also results in the removal of the JNK→c-JUN branch ( Figure 3b ) . Closer examination of the raw data ( Figure 3f ) shows a potent inhibition of IL1α- and ( IL1α+TGFα ) -induced cJUN activity upon Gefitinib treatment . To follow up this interesting off-target effect , we did a dose-response experiment where Gefitinib shows that it can reduce the activation of cJUN signal induced by the IL1α stimuli ( Figure 3i ) . We believe that the inhibition of cJUN is not due to the binding of Gefitinib in the upstream molecule JNK but a collective effect of signaling inhibitions in several species that take part in the path between IL1α and cJUN . For this reason , a fitting with a typical dose response curve has been avoided and a simple linear equation has been used instead ( Figure 3i ) . Erlotinib , another EGFR inhibitor , has the same effects as Gefitinib ( Figure 3c ) but at the same time shows an effect in the TRAF6→MAP3k7 reaction . This effects is probably because IκB-α is inhibited in an IL1α -dependent but TNFα-independent manner ( see IκB-α signals upon IL1α and TNFα stimuli in Figure S1 ) ; the only way for the ILP to satisfy this behavior is to remove the transmission of signal before the merging of TNFα and IL1α paths which can be done through the TRAF6→MAP3K reaction . The “dirty” Raf inhibitor Sorafenib shows a very different profile: it also blocks the JNK→c-JUN branch ( Figure 3d ) and in addition affects the p38 path ( see complete HSP27 inhibition upon IL1α treatment in Figure 3h ) . An interesting observation is that network optimization does not remove the RAF→ERK1/2 reaction despite the fact that RAF is the main target of Sorafenib . Close inspection of the data shows that Sorafenib reduces but does not block the MEK1 phosphorylation ( see MEK phosphorylation in Figure 3h ) . This is in agreement with previous published results where Sorafenib does not inhibit activation of the RAF/MEK/ERK pathway in all human tumor cell lines [33] a finding that highlights the importance of in-vivo assays for the quantification of drug effects .
In this article , we present an unbiased phosphoproteomic-based approach and an optimization formulation to construct cell-type specific pathways and to identify drug effects on those pathways . For the pathway construction , we track 13 key phopshorylation signals in 55 different conditions generated by the combinatorial treatment of stimuli and inhibitors . Using Integer Linear Programming ( ILP ) for pathway optimization we take a generic network of 74 proteins and 105 reactions and construct a cell-type specific network of 49 proteins and 44 reactions that spans between the 5 stimuli and the 13 measured phosphorylated proteins . In this network , we monitor 4 cases of drug-induced pathway alterations using a similar computational scheme . In comparison to all other protein-based target identification approaches , our method is not based on measurements of drug affinities either by in vitro or in vivo assays . Instead , we use an “operative” signaling network and rely on key phosphorylation events and a-priori knowledge of possible connections to reveal the topology and monitor its alterations under the presence of the drug . Thus , our method is expandable to any type of intracellular perturbations such as ATP-based and allosteric inhibitors , RNAi , shRNA etc . Since no bait or MS is required , we have simple ELISA-type experimental procedure with minimal requirements of cell starting protein ( ∼30 , 000 cells per condition ) , without affinity immobilizations , protein fractionations , or carefully optimized wash conditions . With our current semi-automated procedures in our lab ( robotic liquid handlers ) , we can achieve total experimental and computational time for a similar size experiment in less than a week . On the other side , our approach can only detect signaling alterations in topologies bounded between the applied stimuli and the measured phosphorylated proteins and it misses off-target effects outside the constructed network . The expansion of the constructed network depends primarily on three factors: highly curated generic topology , multiplex assay availability for “key” phosphorylation measurements , and experimental cost . We believe that the explosive growth of multiplexed phosphoproteomic assays , the rapid reduction of the cost per datapoint , and the significant improvement in quality of several pathways databases will significantly increase the searching space for drug effects using our proposed methodology . However , our search space will always be significantly smaller compared to whole-genome based approaches [8]–[11] because it requires ( a ) the input of a generic pathway which is available only in well-studied pathways and ( b ) good quality antibodies for the detection scheme . By merging our phosphoproteomic method with genome-wide screening techniques , we might be able to combine the strengths of both approaches and increase the searching space for off-target drug effects . An important aspect of the current approach is the construction of pathway maps . Pathway construction is a major endeavor in biology and a variety of experimental [34]–[38] and computational approaches that span from data-driven methodologies ( e . g . , statistical , unsupervised machine learning ) to topology-based methods ( e . g . , kinetic models based on ordinary differential equations-ODEs ) [17] , [35] , [38]–[41] have been developed . Our approach , which is based on Boolean ( logical ) modeling [26]–[28] , [42] , represents a simplified topology-based method . Compared to ODE-based methods , a logic model has limited abilities to model kinetic behavior [25] ( especially when modeling feedback loops in single-step logic models ) or even to model the protein activity in a continuous fashion . On the flip side , logic models do not require parameter estimation ( sometimes ill-defined from lack of experimental data ) and thus can be applied for the simulation of large topologies . A refinement of the model formalism into multistep logic [28] , fuzzy logic [43] , or ODE-based logic systems [44] may provide a more precise simulation of the activity and time-dependency of the signaling network . Taking into account the current limitations of experimental assays ( throughput , sensitivity , reliability , cost ) we believe that Boolean modeling is the method of choice with high predictive power when large topologies are studied . Optimizing pathway topologies is a relatively new approach for the construction of cell-type specific pathways . Using Boolean topology and Genetic Algorithm ( GA ) for an optimization scheme , Saez-Rodriguez et al . [18] are able to fit a generic map to cell-type specific map from phosphoprotein data . Here we present an alternative method of optimal pathway identification based on ILP . Compared to GA , our algorithm gives guaranteed globally optimized map ( the solution identified is guaranteed to be no worse than 0 . 001 than any other possible solution ) . Additionally , the computational cost has cut down dramatically and allows pathway optimization with ∼70 species to be performed on a desktop computer in a matter of few seconds . Due to minimal computational requirements ILP can be used for the construction of large pathways ( assuming that experimental capabilities can by matched ) and for the exploration of alternative reactions beyond the generic topology to further improve the optimal fit . However , several factors should be addressed before expanding our formulation to larger topologies . Although our formulation is able to identify a globally optimal solution , additional optimal solutions might exist [18] in the same generic network and further more solutions might arise when the optimization formulation is relaxed . Larger and more interconnected networks increase the number of solutions that are equally ( or near equally ) optimal . A possible way to circumvent this problem is to reduce our network using techniques that have been described previously in graph theory or in [18] . Being aware of those limitations in the present manuscript we described a “simple” and not highly interconnected network in order to minimize redundancy of solutions . To address the issue of finding a both unique and optimal solution we are currently working on two complementary approaches: ( a ) instructing the ILP solver to furnish a pool of near-optimal solutions and ( b ) devising “clever stimulations” by taking into account experimental limitations ( i . e . , combination of inhibitors , stimuli , and key protein measurements ) that maximally constrains the optimization scheme and gives smaller number of unique solutions . When applied in HepG2s , our approach identifies both known and unanticipated results . As a positive control , it removes the TGFα branch upon EGRF drug treatments . Another easily understandable effect is Sorafenib's inhibition of the pathway downstream of p38 which can be explained by the drug's target affinity to p38α and p38β [32] , [45] . A surprising effect is the removal of the JNK→cJUN reaction under the influence 3 out of 4 cancer drugs Erlotinib , Gefitinib and Sorafenib . Interestingly , kinase profiles of those drugs [32] shows no medium or high affinity for the directly upstream JNK1/2 kinases . Despite that , Gefitinib shows a significant reduction of the cJUN activity upon IL1α treatment . A possible explanation is that the signaling propagation can collectively be attenuated from the low or medium off-target inhibitions of several kinases upstream of JNK and cJUN . This also might explain the inhibition curve in Figure 3i , where Gefitinib inhibition of cJUN activation does not follow a typical dose-response curve . In this context , sensitivity analysis in ODE-based pathway models [46] have shown that slight changes of reaction constants can have significant attenuations on protein activities several steps downstream the network and thus inhibitory curves cannot be simulated by simplified dose-response models . Our findings also highlight a unique feature of our approach: we find effects of drug's promiscuity that cannot be identified by the direct binding of the drug to the upstream target but are the result of a collective effect of drug's interactions with several upstream molecules . Bait-based analysis cannot reveal those effects since there is no binding involved between the drug and the protein . Understanding the interplay between cell function and drug action is a major endeavor in the pharmaceutical industry . Here , we provided a methodology to construct cell type specific maps and identify drug effects on those maps . Our ILP formulation was able to build the best possible topology from a set of a-priori determined reactions and choose those , where their presence is confirmed from high throughput phosphoprotein data . Since phosphorylation events are the ultimate reporters of protein/drug function the use of high-throughput phosphoproteomic datasets gave an advantage in data quality for modeling signaling network . We believe our approach complements standard biochemical drug profiling assays and sheds new light into the discovery of possible mechanisms for drug's efficacy and toxicity .
HepG2 cells were purchased from ATCC ( Manassas , VA ) , and seeded on 96-well plates coated with collagen type I-coated ( BD Biosciences , Franklin Lakes , NJ ) at 30 , 000 cells/well in DME medium containing 10% Fetal Bovine Serum ( FBS ) . The following morning , cells were starved for 4 hours and treated with inhibitors and/or drugs . Kinase inhibitors were used at concentrations sufficient to inhibit at least 95% the phosphorylation of the nominal target as determined by dose-response assays ( presented in [17] ) . AKT was chosen as the nominal target for Lapatinib , Erlotinib , and Gefitinib . The following saturated concentrations were used: p38 ( PHA818637 , 20 nM ) , MEK ( PD325901 , 100 nM ) and cMET ( JNJ38877605 , 1µM ) , PI3K ( PI-103 , 10 µM ) , Lapatinib at 3uM [47] , Erlotinib at 1 uM [47] , Gefitinib at 3uM [47] , and Sorafenib at 3 uM ( based on its inhibitory activity on ERK1/2 phosphorylation [33] ) . Following incubation for 45 minutes with inhibitors and/or drugs cells were treated with saturated levels of 5 ligands: Tumor Necrosis Factor alpha ( TNFα ) at 100ng/ml , Interleukin 1 alpha ( IL1α ) at 10ng/ml , Insulin ( INS ) at 2uM , Transforming Growth Factor ( TGFα ) at 100ng/ml , and Hepatocytes Growth Factor ( HGF ) at 100 ng/ml . Each ligand was added alone or in pairs and cell lysates were collected at 0 , 5 , and 25 minutes following the cytokine stimulation . The 5 and 25 minutes lysates were mixed together in 1∶1 ratio and the mixed lysate was measured as an indicator of the “average early signaling response” . The 5 and 25 minute time points were identified in a preliminary experiment as the optimal time points that maximally captured early phosphorylation activities [17] . A major improvement in the present dataset as compared to [17] was the “in-vitro” averaging of the signals from 5 and 25 minutes rather than “in-silico” averaging ( i . e . , first both time points are measured , then we take the average ) . Three are the main advantages using such approach: 1 ) two signals are used instead of one and thus very early signalling responses can be captured , 2 ) the experimental cost is reduced by 50% ( or more for averaging multiple time points ) , and 3 ) we achieved the averaging of some signals that could not be measured independently because their “active” state is reaching the saturation limits of our measuring instrument . From each lysate we measured 13 phosphorylation activities that we considered “key phosphorylation events” using a Luminex 200 system ( Luminex Corp , Austin , TX ) . The 13-plex phospho-protein bead set from Bio-Rad was used to assay p70S6K ( Thr421/Ser424 ) , CREB ( Ser133 ) , p38 ( Thr180/Tyr182 ) , MEK1 ( Ser217/Ser221 ) , JNK ( Thr183/Tyr185 ) , HSP27 ( Ser78 ) , ERK1/2 ( Thr202/Tyr204 , Thr185/Tyr187 ) , c-JUN ( Ser63 ) , IRS-1 ( Ser636/Ser639 ) , IκB-α ( Ser32/Ser36 ) , Histone H3 ( Ser10 ) , Akt ( Ser473 ) , and IR-β ( Tyr1146 ) . Data were normalized and plotted using with DataRail [48] . For the construction of the dose response curve in Figure 3i , HepG2 were starved for 4 hours and then incubated with Gefitinib ( from 20uM down to 27nM – 3 fold dilution ) for 45 minutes followed by incubation with IL1α at 10ng/ml final concentration for 30 minutes . Duplicate lysates were analyzed using the c-JUN ( Ser63 ) beads in the Luminex 200 system . Here , we describe how the Boolean model described in [18] can be reformulated as an ILP . Note that such a transformation was recently performed for a different problem , namely the satisfiability , by [49] . A pathway is defined as a set of reactions and species . Each reaction has three corresponding index sets , namely the index set of signaling molecules , inhibitors , and “products” ( “product” can also correspond to the phosphorylation level of the protein ) . These sets are all subsets of the species index set ( ) . Typically , these subsets have very small cardinality ( few species ) , e . g . , ; ; ; . A reaction takes place if and only if all reagents and no inhibitors are present . If a reaction takes place , all products are formed . Note that reactions without products as well as reactions with neither reagents nor inhibitors will be excluded here . While typically the set of species is known , the set of reactions is not known . Rather , only a superset of potential reactions is postulated . The goal of the proposed formulation is to find an optimal ( in some sense ) set of reactions out of such a superset . To that extent binary variables are introduced , indicating if a reaction is possible or not ( connection not present , connection present ) . A set of experiments is performed , indexed by the superscript . In each experiment a subset of species is introduced to the system and another subset is excluded from the system . These are summarized by the index sets and respectively ( two for each experiment ) . In the proposed formulation , constants are introduced for all such species , respectively and . In the following it will be assumed that these species do not appear as products in any reaction; this assumption is not limiting , since in the experiments performed only extracellular species and inhibitors are manipulated . In the experiments a third subset of the species is measured ( index set ) and for the remaining species no information is available . In the proposed formulation for each of the experiments and each such species a binary decision variable is introduced indicating if the species is present ( ) or not ( ) in the experiment according to the model predictions . It is proved that in the absence of loops , can be used for species that are not input species ( see Text S1 ) . This has some computational advantages . The last group of variables introduced indicate if reaction will take place ( ) or not ( ) in the experiment according to the model predictions . It is proved that a real variable can be used equivalently ( see Text S1 ) . This reformulation has some computational advantages . For the case that a species is measured , the measurement is defined as . For Boolean measurements ; otherwise ( assuming a scaling as afforementioned ) . The primary objective function is formed aiming to minimize the weighted error between model predictions and measurements . The absolute value is reformulated as . It can be easily verified that for binary and for this reformulation is valid: Note also that alternative norms , such as least-squares errors , could be also used . The resulting optimization problem would still be an ILP , since the objective function involves only integer variables . For instance for the least-square error objective function the following linear reformulation is valid: The secondary objective is to minimize the weighted number of possible reactions . In multiobjective optimization typically the concept of Pareto-optimal or noninferior solution is introduced , i . e . , a set of decision variable values , such that if one tries to improve one objective , another will be degraded [50] . The set of Pareto points forms the Pareto-optimal curve . Here , however , the primary objective is considered much more important than the secondary objective . Therefore , a single Pareto-optimal point is obtained , by first minimizing the primary objective and then the secondary objective by requiring that the former ( more important ) objectives are not worsened , see also [51]–[53] . The ILP proposed can be summarized as: ( 1 ) ( 2 ) ( 3 ) ( 4 ) ( 5 ) ( 6 ) ( 7 ) ( 8 ) ( 9 ) ( 10 ) ( 11 ) where the objectives are separated by a semi-colon . Note that for the elements of the matrices and , the row index ( experiment ) is indicated as superscript , and the column index ( species and reactions respectively ) is indicated as subscript . In formulation ( 1 ) – ( 11 ) for the manipulated species binary decision variables along with the constraints ( 9 ) and ( 10 ) are introduced . This simplifies notation . In the implementation , these variables are replaced by constants . Alternatively the preprocessor of the optimization solver can be used to exclude these trivial variables . In the following the reasoning for the formulation is given . The first set of constraints , i . e . , ( 2 ) allow the modeler to limit the combinations of connectivities considered . For instance , suppose that two reagents , form a product , but it is not known if both reagents ( AND ) or either ( OR ) are required . This can be modeled as three potential reactionswith the additional constraint that excludes and , which can be modeled as two linear inequalities:The constraints ( 3 ) indicate that a reaction can only take place if it is possible ( ) . This can be seen easily , since , gives and together with we obtain . Similarly , the constraints ( 4 ) and ( 5 ) ensure respectively that a reaction can only take place if all reagents and no inhibitors are present . If for instance a reagent is absent , is enforced , and the other constraints are redundant . On the other hand , the constraints ( 6 ) enforce that if a reaction is possible , all reagents are present , and no inhibitors are present , then the reaction will take place ( ) . The constraints ( 7 ) ensure that a species will be formed if some reaction in which it is a product occurs . Note that multiple reactions can give the same species; mathematically this will result in redundant constraints . In contrast , the constraints ( 8 ) enforce that a species will not be present if all reactions in which it appears as a product do not occur . Recall that manipulated species are not considered as products in reactions . Note also , that it would be possible to combine the constraints ( 7 ) into a single constraint for each species , e . g . , but this would result in weaker LP-relaxations . Also the reformulation of to would no longer be exact . In the present study , our ILP formulation was utilized in two different circumstances . For the creation of the cell-type specific pathway using combinations of inhibitors and stimuli our ILP formulation included 27887 constraints and 9732 variables . For each drug case , where the reduced and optimized pathway was utilized , we had 2477 constraints and 947 variables . For the goodness of fit , we calculated the percentage error as: Note that for binary and the percentage error cannot be 0% even when there is no mismatch between model and experiment data . Another way to quantify the goodness of fit is by counting the number of mismatches: the cases where the rounded experimental value ( 0 or 1 ) is not the same with the computational value , or in other words , when experimental – computational error is more than 0 . 5 .
|
Cells are complex functional units . Signal transduction refers to the underlying mechanism that regulates cell function , and it is usually depicted on signaling pathways maps . Each cell type has distinct signaling transduction mechanisms , and several diseases arise from alterations on the signaling pathways . Small-molecule inhibitors have emerged as novel pharmaceutical interventions that aim to block certain pathways in an effort to reverse the abnormal phenotype of the diseased cells . Despite that compounds have been well designed to hit certain molecules ( i . e . , targets ) , little is known on how they act on an “operative” signaling network . Here , we combine novel high throughput protein-signaling measurements and sophisticated computational techniques to evaluate drug effects on cells . Our approach comprises of two steps: build pathways that simulate cell function and identify drug-induced alterations of those pathways . We employed our approach to evaluate the effects of 4 drugs on a cancer hepatocytic cell type . We were able to confirm the main target of the drugs but also uncover unknown off-target effects . By understanding the drug effects in normal and diseased cells we can provide important information for the analysis of clinical outcomes in order to improve drug efficacy and safety .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"pharmacology/drug",
"development",
"computer",
"science/numerical",
"analysis",
"and",
"theoretical",
"computing",
"pharmacology/drug",
"interactions",
"oncology/gastrointestinal",
"cancers",
"computational",
"biology/signaling",
"networks",
"computational",
"biology",
"computational",
"biology/systems",
"biology"
] |
2009
|
Identifying Drug Effects via Pathway Alterations using an Integer Linear Programming Optimization Formulation on Phosphoproteomic Data
|
Vaccinia virus interferes with early events of the activation pathway of the transcriptional factor NF-kB by binding to numerous host TIR-domain containing adaptor proteins . We have previously determined the X-ray structure of the A46 C-terminal domain; however , the structure and function of the A46 N-terminal domain and its relationship to the C-terminal domain have remained unclear . Here , we biophysically characterize residues 1–83 of the N-terminal domain of A46 and present the X-ray structure at 1 . 55 Å . Crystallographic phases were obtained by a recently developed ab initio method entitled ARCIMBOLDO_BORGES that employs tertiary structure libraries extracted from the Protein Data Bank; data analysis revealed an all β-sheet structure . This is the first such structure solved by this method which should be applicable to any protein composed entirely of β-sheets . The A46 ( 1–83 ) structure itself is a β-sandwich containing a co-purified molecule of myristic acid inside a hydrophobic pocket and represents a previously unknown lipid-binding fold . Mass spectrometry analysis confirmed the presence of long-chain fatty acids in both N-terminal and full-length A46; mutation of the hydrophobic pocket reduced the lipid content . Using a combination of high resolution X-ray structures of the N- and C-terminal domains and SAXS analysis of full-length protein A46 ( 1–240 ) , we present here a structural model of A46 in a tetrameric assembly . Integrating affinity measurements and structural data , we propose how A46 simultaneously interferes with several TIR-domain containing proteins to inhibit NF-κB activation and postulate that A46 employs a bipartite binding arrangement to sequester the host immune adaptors TRAM and MyD88 .
Viral infection depends not only on the rate and precision of viral reproduction , but also requires a simultaneously efficient inhibition of host immune responses . Viruses have evolved varied strategies to interfere with immune responses of the host , including production of secreted molecules that mimic innate immune receptors , molecules that trap cytokines as well as the shut-off of the cellular transcription and translation machinery [1 , 2] . Vaccinia virus ( VACV ) , the virus used to eradicate smallpox , has been extensively studied as a model of virus-host interaction because of its plethora of anti-immune strategies and its large arsenal of immunomodulator tools [3] . Further interest in VACV stems from its role as a vaccine vector against important infectious diseases and its potential role against cancer [4 , 5] . Amongst approximately 200 genes in the VACV genome , only half encodes for the viral replication machinery; many of the remaining gene products have roles as extra- and intracellular modulators of the host immunity [6] . The VACV intracellular immunomodulators form a family of Bcl-2-like ( B-cell lymphoma 2 like ) proteins with low sequence identity but high structural similarity to the eukaryotic Bcl-2 protein family [7] . Eukaryotic Bcl-2 proteins present a diverse group of pro- and anti-apoptotic regulators that share α-helical BH domains [3 , 8] . To date , 11 Bcl-2-like proteins encoded by VACV have been identified . Those such as A46 , A49 , A52 , B14 , N1 , K7 and F1 have an experimentally confirmed Bcl-2 fold [9–16]; others such as C1 , C6 , C16/B22 and N2 are predicted to have such a fold [10 , 17 , 18] . NF-κB is a transcriptional factor that responds to the stimulation of Toll-like-receptors ( TLRs ) and Interleukin-like-receptors ( IL-1R ) by inducing expression of effector molecules . In the uninfected cell , inactive NF-κB is located in the cytoplasm as a precursor or in a complex with its inhibitor ( IκB ) . Upon stimulation of TLRs by pathogens , a signaling cascade is initiated through the recruitment of adaptor proteins ( e . g . MyD88 , MAL/TIRAP , TRIF , TRAM ) by the cytoplasmic domains of TLRs , consequent stepwise activation of IRAK2-IRAK6-IRAK4 kinases followed by activation of TRAF6 ubiquitin ligase and activation of the IKK ( IκB kinase ) complex . Finally , the release of the active form of NF-κB results from processing of the precursors or degradation of IκB . Nuclear migration of the free NF-κB permits expression of a range of cytokines allowing the development of both innate and adaptive immune responses [19] . VACV Bcl-2-like immunomodulators disrupt NF-κB activation pathways at different stages by targeting various components [3 , 7] . The A46 protein acts close to the plasma membrane by binding numerous TIR-domain containing adaptor proteins such as MyD88 , MAL/TIRAP , TRAM and TRIF as well as TLR4 to prohibit further signal propagation [20] . We recently determined the structure of the Bcl-2 domain of A46 comprising residues 87–229 [9] . However , structural information on the N-terminal domain ( residues 1–86 ) , its position relative to the Bcl-2-like domain and a plausible function were lacking . Here , we report the crystal structure of the A46 N-terminal domain comprising residues 1 to 76 and demonstrate that this domain binds fatty acids . Further , small-angle X-ray scattering ( SAXS ) was employed to derive a structural model of full-length of A46 ( 1–240 ) . Using a SAXS-derived model of A46 together with biochemical data , we postulate a mechanism explaining the biological function of this unusual VACV immunomodulator protein .
Members of the VACV Bcl-2-like family whose structure has been determined mainly comprise a single Bcl-2-like domain with an N- or a C-terminal extension ( ranging from 5 to 80 amino acids ) or both ( Fig 1A ) . At present , structural information is only available for the Bcl-2-like domains and a short unstructured N-terminal region of F1L [21] but not for the rest of extensions . However , the N-terminal extension of A46 spanning residues 1–80 was predicted by PSIPRED ( 31 ) to comprise exclusively β-strands . Previous studies using limited proteolysis on the full-length A46 protein confirmed the presence of a structured N-terminal domain in the first 80 residues , suggesting that it would be amenable to crystallography ( Fig 1A ) [9] . To examine the structure and function of the N-terminal domain of A46 , we designed two constructs for expression in E . coli . Both protein expression constructs contained the first methionine of the full-length protein and comprised 73 or 83 A46 residues , as constructs with fewer than 73 residues were either insoluble when His6-tagged or could not be removed from the MBP expression tag . Both variants contained an additional four amino acids ( MAQQ , Fig 1B ) to improve solubility as observed with full-length A46 [9] . Thus , both fusion proteins had the following structure: His6-TRX-TEVsite-MAQQ-A46 ( 1-73/83 ) . The average yield of both proteins was approximately 2 . 5 mg highly purified protein per L of bacterial culture . However , as we only obtained diffraction quality crystals with A46 ( 1–83 ) , we performed all subsequent work with this variant ( Fig 1C ) . We first examined the ability of A46 ( 1–83 ) to bind the TIR domains of its proposed cellular binding partners such as MyD88 and MAL . Using microscale thermophoresis , we previously demonstrated that the C-terminal domain of A46 binds in the low micromolar range to these TIR domains; the KD values were slightly lower than those observed with the full-length protein ( Table 1 ) [9] . In contrast , the N-terminal A46 ( 1–83 ) binds to TIR/MyD88 but not to TIR/MAL . The KD value was 8 . 8 μM , compared to that of 0 . 52 μM for A46 ( 1–229 ) . We also examined the binding of the TIR/TRAM domain , another proposed A46 in vivo binding partner , to A46 [20 , 24] . The interaction of full-length A46 and its C-terminal domain with TIR/TRAM shows KD values of 2 . 39 μM and 3 . 62 μM , respectively . However , under the conditions used , A46 ( 1–83 ) did not bind to TIR/TRAM ( Table 1 ) . Given the binding of A46 ( 1–83 ) to MyD88 , we next examined whether this fragment was sufficient to prevent IL-1β induction of NF-κB-mediated transcription using similar cell-based assays to those described previously [9] . Plasmid amounts were adjusted so that approximately the same amounts of each A46 variant were expressed; the total amount of transfected DNA ( 500 ng ) was kept constant by the addition of empty pCAGGS vector . Unlike full-length A46 and the truncated variant A46 ( 87–229 ) , the N-terminal domain exhibits no appreciable inhibition of the IL-1β driven induction of NF-κB ( Fig 2; see figure legend for statistics ) . Thus , binding of A46 residues 1–83 is insufficient to independently fulfil an immunomodulatory role . We initiated structural studies of the functional form of the N-terminal domain of A46 ( 1–83 ) by setting up crystallization trials with commercial screens . Small single crystals of around 20 μm in size were observed after 1 week of incubation at 22°C . They failed , however , to grow larger; nevertheless , several datasets were collected using the beam line for high throughput macromolecular data-collection MASSIF at ESRF , rendering the highest resolution between 1 . 8 and 2 . 3 Å . With no known close homologue in the PDB database , we were unable to solve the phase problem by molecular replacement . Thus , we labelled the protein with selenomethionine; diffraction quality crystals grew in the conditions used for the native protein . Data sets for SAD were collected using the MASSIF beamline up to 1 . 55 Å resolution . However , we were unable to phase the structure using the anomalous signals , most likely because all three methionines in the protein lie in the very N- and C-termini of the A46 ( 1–83 ) construct and , consequently , are located in flexible regions . Finally , the phases were obtained by ARCIMBOLDO_BORGES [25] crystallographic software . The program exploits tertiary structure libraries extracted from the Protein Data Bank for ab initio phasing . A library of 7650 superimposed polyalanine models , representing 925300 variations on the fold of three stranded antiparallel β-sheets totalling 20 amino acids , was used as fragment hypotheses . This three strands arrangement is most frequently found in β-sheets . Computations were executed on the Gordon supercomputer at the San Diego Supercomputer Center in California . A partial solution was obtained upon location with PHASER [26] of 4 models extracted from the unrelated PDB structures 2QLG , 2GSK , 2EFU , 4DCB as indicated by SHELXE [27] trace correlation coefficients above 40% . The root mean square deviation ( rmsd ) of the solving library models against the final structure was in the range of 0 . 35 Å ( model from 4DCB ) and 0 . 61 Å ( model from 2EFU ) . A46 ( 1–83 ) crystallized with two molecules in the asymmetric unit; electron density for a bound ligand , later identified as myristic acid , was found inside one of the molecules . The two A46 molecules comprise two β-sheets arranged head to head as an extended β-sandwich ( Fig 3A , Table 2 ) . A tetramer is formed over a crystallographic twofold axis continuing the β-sandwich with the second dimer rotated approximately 90° relative to the first ( S1A and S1B Fig ) . The PISA server [28] estimates both association interfaces to be present in solution , burying 1276 and 941 Å2 . The A and B independent subunits show marked differences , with a Cα rmsd of 1 . 3 Å for the 45 common β-strand residues ( Fig 3B ) ; the tetramer can be described as an A/B/B/A arrangement . Subunit A has 7 β-strands whereas subunit B presents only 6 , lacking the most C-terminal one ( Fig 3A ) . No electron density is seen for either residues 77–83 in subunit A or 67–83 in subunit B , suggesting that these regions may constitute a flexible linker between N- and C-terminal domains in the full-length molecule . A striking feature of the external β1-β7 face of the A subunit is a partially hydrophobic tunnel , spanning the whole subunit A and reaching into subunit B ( Fig 4A ) . A length of 22 Å , an average radius of 2 . 5–3 Å and an overall cavity volume of 1150 Å3 ( S2 Fig ) were calculated with the software MOLE 2 . 0 [30 , 31] . The tunnel is occupied by an extended well-defined electron density , reminiscent of a myristic acid molecule ( Fig 4A ) . The omit electron density map for the ligand is presented in S3 Fig . Mass spectrometry and gas chromatography ( GC ) analysis of the lipids extracted from purified protein identified the fatty acids C14:0 , C16:0 and C16:1 in complex with A46 ( 1–83 ) ( Fig 4B ) . Repetition of the experiment with a separate A46 ( 1–83 ) preparation revealed the same three fatty acids but in different ratios , indicating that the relative amounts may be preparation dependent . However , in all preparations so far examined , the C14:0 fatty acid was highly enriched compared to its overall representation in E . coli cells ( Fig 4E ) . Further , the lipid extraction and identification by mass spectrometry was also done with three independently purified samples of the full-length A46 as well as two purified samples containing the C-terminal domain of A46 ( 87–229 ) . For the full-length A46 , we also identified the three co-purified fatty acids , C14:0 , C16:0 and C16:1; in contrast , purified A46 ( 87–229 ) lacked any complexed lipids ( Fig 4C and 4D ) . Hence , only samples containing the N-terminal domain of A46 , either purified independently or as a part of the full-length protein , are capable of binding fatty acids . Subunit B , being partially penetrated by the fatty acid , cannot therefore lodge a second molecule . The cavity , present in subunit A , is collapsed in subunit B , bringing both β-sheets 3 . 5Å nearer ( Fig 3B ) . Tyr37 adopts a dual conformation in the two subunits , suggesting a gate-keeper role as it folds back in subunit A to make room for the myristic acid ( Fig 3B ) . The side chain of the preceding His36 , pointing to the outside and located in the loop displaying highest differences between both subunits , also has two conformations . A single loop at each side of the sandwich joins both sheets , allowing the displayed flexibility . One loop ( β1 to β2 ) is unchanged; the other ( β4 to β5 ) , containing four charged residues DRDK , differs between the subunits , altering its hydrogen bond pattern ( Fig 3B ) . Together with His36 , these electrostatic interactions may provide a lever for myristic acid binding . The absence of bound fatty acid as well as the lack of the β7 strand in subunit B results in a quite different interaction interface to that in subunit A , allowing association of two B subunits , with β1 occupying the position vacated by β7 , and thus the assembly of the symmetric tetramer ( S1 Fig ) . We examined the lipid-binding properties of A46 ( 1–83 ) by structure-based site-directed mutagenesis . We introduced the single mutations F3D , H36L , Y37A , Y37W , I72A into the expression plasmid for A46 ( 1–83 ) and successfully expressed and purified protein from all variants . Analysis of their lipid content showed that all variants contained C14:0 , C16:0 and C16:1 fatty acids . Furthermore , only the variant Y37A had a wild-type amount of lipids; all of the others had less bound lipid than the wild-type , with the variant I72A having the lowest value of 29% ( S1 Table ) . To investigate whether the level of bound lipids influence the function of A46 , the I72A mutant of full-length A46 was examined in a NF-κB transcriptional assay in TLR4-expressing HEK293 cells . The A46 I72A mutant was reproducibly expressed at higher levels , both in mammalian cells ( Fig 5A ) and bacteria . The A46 I72A variant could achieve similar levels of inhibition of NF-κB mediated signalling as the wild-type ( Fig 5B ) ; however , this level could only be reached when a 2-to-3-fold excess of A46 I72A was expressed compared to the wild-type variant ( Fig 5A and 5B ) . Thus , the lower lipid binding capacity of A46 I72A impairs its ability to inhibit TLR4 signalling . In the light of the crystal structure , we analysed the oligomeric state of A46 ( 1–83 ) using SAXS ( Table 3 ) . The theoretical scattering curve of the A46 ( 1–83 ) tetramer in crystals presents a good fit to experimental data with Chi2 ( Crysol [32] ) of 0 . 66 ( Fig 6A ) versus a very poor fit for the possible dimer found in the asymmetric unit with Crysol Chi2 of 13 . 52 ( Fig 6A ) . How are the N-terminal and C-terminal domains of A46 oriented relative to one another ? To address this question , we performed SAXS experiments on full-length A46 ( 1–240 ) ( Fig 6B , Table 3 ) . The envelope is shown in Fig 6C , together with the fitting of the N- and C-terminal structures . This arrangement agrees with the tetrameric nature of the A46 ( 1–240 ) and with proteinase digestion of the linker leading to the production of two domains with almost all proteinases tested [9] .
We have determined the first structure of a structured N-terminal extension of a VACV Bcl-2-like immunomodulator; additionally , we also show that it is complexed with myristic acid . The A46 ( 1–83 ) domain crystallized , forming regular continuous strands in a simple β-sandwich structure with few disordered residues ( Fig 3A; S1 Fig ) . Nevertheless , the solution of the X-ray structure was complex . Due to the crystals' small size , automatic beam focussing at the MASSIF beam line was essential . Additionally , selenium anomalous signals could not be used because of the position of the methionine residues . Molecular replacement also failed due to lack of a known protein structure to be used as search model . However , the regular crystal packing allowed high-resolution data sets to be obtained that were initially processed at 1 . 55 Å resolution . This high resolution data , together with the short length of the protein , allowed the phases to be solved using ab initio methods [25] . In this method , which has been used successfully for numerous α-helical structures [34] , fragments of known structures are employed as small search models . In our study , phases could be solved by a protein fragment of three β-strands that resembles part of the structure of A46 ( 1–83 ) , revealing two molecules in the asymmetric unit . Refinement of the structure allowed the determination of electron density for residues 1–76 of subunit A and 1–66 in subunit B . The electron density showed clearly that both subunit A and subunit B were comprised entirely of β-sheets , confirming previous bioinformatic predictions that the N-terminus of A46 has a β-sheet arrangement . Unexpectedly , in the subunit B , the C-terminal strand β7 is disordered and not visible in the electron density . We propose that this difference allows A46 ( 1–83 ) to form tetramers via the subunit B interfaces whilst interacting with ligands through the subunit A interfaces . A further wholly unexpected feature of A46 ( 1–83 ) is a partially hydrophobic cavity which spans the entire subunit A and part of subunit B . The cavity is open on the side of the A interface and accommodates long chain fatty acids that were co-purified from the E . coli cell lysate . In the X-ray structure of A46 ( 1–83 ) , clear electron density for C14:0 myristic acid was found ( Fig 4A ) , with the hydrophobic tail buried in the cavity whereas the carboxyl group is open to the solvent . Such an orientation leads us to hypothesize that the cavity might serve as a specific binding pocket for myristoylated binding partners . To this end , TRAM is the only binding partner of A46 known to be myristoylated; myristoylation is indeed essential for its innate immune function , providing correct location of TRAM to the membranes [35] . Binding of A46 to the myristate of TRAM would prevent the insertion of TRAM into the membrane and thus circumvent intracellular signalling . An acceptable alternative hypothesis would be that the bound fatty acids induce asymmetry of the A46 ( 1–83 ) dimer , as they block a polymerization interface equivalent to B/B and prevent binding of a second fatty acid copy in subunit B . In such manner , using the same primary sequence , a dimer of heterodimers is formed that allows utilization of different interfaces for distinct functions such as tetramerization ( interface B with 6 β-strands only ) or binding of cellular targets ( interface A with 7 β-strands ) . The I72A mutant of A46 , which binds lower amounts of fatty acids , indeed showed a reduced ability to inhibit TLR4-stimulated NF-κB-driven transcription compared to the wild-type protein ( Fig 5B ) . The inhibitory level of the wild-type A46 was achieved by the I72A mutant when higher quantities of the mutant protein were expressed . This is not unexpected , as the C-terminal domain alone ( A46 ( 87–229 ) ) can bind TIR/TRAM with KD of 3 . 6 μM ( Table 1 ) . Presumably , at higher concentrations , the C-terminus of the A46 I72A mutant can compensate for the loss of binding of the lipid-containing N-terminal domain . Pertinently , we have shown that the C-terminus of A46 alone is capable of efficiently inhibiting MyD88-mediated NF-κB activation when IL-1β stimulation system is used [9] . TLR4-stimulated activation of the NF-κB transcription factor involves both TRAM and MyD88-dependent cascades [36]; taken together , our data suggest that the N-terminal domain of A46 may play a more appreciable role in the inhibition of the TRAM pathway than the MyD88 pathway . To find similarities of A46 ( 1–83 ) to other known folds , we searched the PDB database with PDBeFold [37] , using subunit A of A46 ( 1–83 ) as search query . The highest match corresponded to the nuclear movement protein from E . cuniculi GB-M1 ( PDBID 2O30 , chain B ) . Six secondary structure elements were aligned involving 57 residues at an rmsd of 3 . 39 Å for 45 Cα; however , the mutual orientation of both sheets is markedly different and the CS domain seen in NudC does not show oligomerisation . Therefore , we searched for similar local folds of the same connectivity using the same core of 45 residues with the program BORGES [25] . The closest match for the strands of ligand bound subunit A was extracted from 2XN2 , with 3 . 09 Å rmsd , whereas for subunit B , a fold extracted from 2OQE gave 2 . 42 Å . No instance could be identified of an equivalent fold showing the same structural change upon ligand binding; nevertheless , a survey of the hits revealed recurring instances of carbohydrate binding proteins , proteins forming pores and participating in the proper insertion of periplasmic proteins into membranes . These include proteins such as YidC ( PDBID 3BLC ) , located in the periplasmic space of E . coli that could , theoretically , bind lipidated proteins . The geometry of the local fold described by the 6 sheets is also close to a part found in pore-forming hemolysins and leucodines . Indeed , the S-F heterodimer in the latter ones achieves asymmetry through the association of two components of very different sequence but very close geometry , with up to one C-terminal strand present in only one of the copies [38] . The structure of the A46 ( 1–83 ) protein illuminates the oligomerization state of both the N-terminal extension and the full-length protein . Previous data had indicated the presence of a tetramer in solution for the full-length A46 and a dimer for the Bcl-2-like C-terminal domain [9] . The structure of the N-terminal extension shows a tetramer formed by the association over a crystallographic twofold axis of the two copies present in the asymmetric unit ( S1 Fig ) , evaluated to be persistent under physiological conditions . SAXS analysis confirmed that that A46 ( 1–83 ) is tetrameric in solution ( Fig 6A ) . For full-length A46 in solution , structural information on the separated N- and C-terminal domains allowed interpretation of the envelope generated by SAXS . The full-length molecule has an elongated shape , with the N- and C- domains linked by a flexible , proteolytically sensitive linker that allows movement of the two domains relative to each other ( Fig 6C ) . Rigid body fitting of the structures in the envelope of the full-length A46 using CORAL software [39] indicated a movement of 90 degrees between the two domains . What are the implications of this structural data for the function of A46 in inhibiting signalling through the TRAM and MyD88 linked pathways ? We note that both the N-terminal and C-terminal domains of A46 can bind the TIR domains of MyD88 ( Fig 7A ) , although the binding of the N-terminal domain is tenfold lower and the expression of this domain alone does not inhibit IL-1 induced NF-κB mediated signalling in cells ( Table 1 , Fig 2 ) . However , we suggest that this bipartite binding enables A46 to generate a chain around the TIR domain of MyD88 that would prevent the association of its death domain to assemble the Myddosome , an important structure in the development of the inflammatory response [40] . Additionally , we propose the binding of the myristate post-translational modification of TRAM by the N-terminal domain of A46 , with the remainder of the TIR domain of TRAM being bound by the C-terminal Bcl-2 domain of A46 ( Fig 7A ) . For the TIR domain of MAL , an interaction was only observed with the C-terminal domain of A46; the binding site on A46 for the TIR domain of TRIF has not yet been determined . We propose here that the interaction is only with the C-terminal domain ( Fig 7A ) . The above model assumes binding of only one single TIR domain to the A46 tetramer . However , as depicted in Fig 7B , each tetramer can theoretically present four binding sites for TIR domains . We speculate therefore that A46 could form complexes with multiple binding partners . Indeed , it can even be envisaged that one molecule of A46 could bind one molecule each of MyD88 , MAL , TRAM and TRIF ( Fig 7B , right side ) . Thus , even with low initial concentrations of A46 , this arrangement would serve to strongly inhibit the inflammatory response by keeping MyD88 death domains apart , preventing proper cellular localization of TRAM and sequestering the other signalling and adaptor molecules . Future experimentation will show the accuracy of these predictions .
The cloning of the plasmid containing full-length sequence of the A46R gene from the VACV Western Reserve strain plasmids ( NCBI Gene ID:3707702 ) as well as of those encoding TIR domains of mammalian MAL and murine MyD88 was described previously [9] . The N-terminal portion of A46 was amplified from the plasmid containing full-length A46 [9] at different length using following primers for the indicated fragments: F: 5’-CGCAAGCCATGGCACAGCAAATGGCGTTTGATATATC-3’ and R: 5’-GCCCGGATCCTTAACT ATACTTATTATACAAGTAAGTC-3’ for the fragment A46 ( 1–90 ) ; F: 5’-CGCAAGCCA TGGCACAGCAAATGGCGTTTGATATATC-3’ and R: 5’-GCCCGGATCCTTAAGTCATACTAA CCGGCGTATTAAC-3’ for the fragment A46 ( 1–83 ) ; and F: 5’-CGCAAGCCATGGCACAGCA AATGGCGTTTGATATATC-3’ and R: 5’-GCCCGGATCCTTAACCAATATTAGTTTCCTCTG-3’ for the fragment A46 ( 1–73 ) . Obtained fragments were digested with NcoI and BamHI restriction enzymes and ligated into the pET-TRX1 containing HIS6-TEV-thioredoxin as an expression tag . To generate variants of A46 ( 1–83 ) to examine their lipid-binding properties , we performed PCR mutagenesis using the pTRX-A46 ( 1–83 ) plasmid as a template and the following primers: F3D , F: 5’-GCAAATGG CGGATGATATATCAG -3’ and R: 5’- CTGATATATCAT CCGCCATTTGC -3’; H36L F: 5’- GTTAATGATACACTCTACACTGTCG -3’ and R: 5’- CGACAGTGTAGAGTGTATCATTAAC -3’; Y37A , F: 5’-GATACACACGCCACTGTCGA-3’ and R: 5’-TCGACAGTGGCGTGTGTATC-3’; Y37W , F: 5’-GATACAC ACTGGACTGTCGAATTTG -3’ and R: 5’-CAAATTCGACAGTCCAGTGTGTATC-3’; I72A , F: 5’- GAAACTAATGCTGGTTGCGCGG -3’ and R: 5’- CCGCGCAACCAGCATT AGTTTC -3’ . The generated PCR products were digested with DpnI and subsequently transformed in E . coli TOP10 competent cells . For expression in mammalian cells , the plasmids coding for the full-length A46 and C-terminal portion of A46 with the respective tags were cloned previously [9] . For the cloning of the N-terminal domain of A46 ( 1–83 ) , the gene was obtained by amplification from the plasmid carrying the full-length A46 with the primers F: 5’-GCCCGAATTCCGAGAATGGAGCAGAAACTCA TCTCTGAAGAGGATCTGGCGTTTGATATATC-3’ and R1: 5’-CCGCTCGAGTTACTTA TCGTCGTCATCCTTGTAATCAGTCATACTAACCGGCG-3’ or R2: 5’- CCGCTCGAGTTA AGTCATACTAACCGGCG-3’ to yield myc-A46 ( 1–83 ) -FLAG or myc-A46 ( 1–83 ) , respectively . The amplified DNA fragments were digested with XhoI and EcoRI restriction enzymes and ligated into the pCAGGS vector [41] . The plasmid encoding a GST-fusion of the TIR domain of human TRAM ( amino acid residues 66–235 ) was a kind gift from Dr . H . Tochio [42] . Expression and purification of full-length A46 , TIR/MyD88 and TIR/MAL were performed as described previously [9] . E . coli BL21 ( DE3 ) competent cells were transformed with the plasmids coding for the variants of the N-terminal domain of A46 . The expression was performed in 2 liters of LB medium containing kanamycin ( 50 mg/liter ) . The cells were grown at 37°C until the mid-log phase ( A600 = 0 . 6 ) . Expression was induced with 0 . 25 mM isopropyl 1-thio-β-D-galactopyranoside at 23°C . After 4 hours , cells were harvested and resuspended in 20 mM Tris-HCl , pH 8 . 5 , 100 mM NaCl , 25 mM imidazole , 5% glycerol and 10 mM β-mercaptoethanol . An EmulsiFlex C3 homogenizer ( Avestin ) was used for cell lysis . The soluble phase was cleared from insoluble material by centrifugation at 18000 rpm for 30 min . Recombinant proteins were bound to Ni-NTA agarose ( 5 Prime ) charged with 300 mM NiCl2 and pre-equilibrated with lysis buffer . Resin was washed with five column volumes of lysis buffer and proteins of interest were eluted in three column volumes of 20 mM Tris-HCl , pH 8 . 5 , 300 mM NaCl , 200 mM imidazole and 15 mM β-mercaptoethanol . Recombinant TEV protease was added to release A46 domains by proteolysis during overnight dialysis against 20 mM Tris-HCl , pH 8 . 5 , 150 mM NaCl , 10 mM imidazole and 15 mM β-mercaptoethanol . The protein of interest was separated from the protease and the tag by four passages through Ni-NTA resin pre-equilibrated with the dialysing buffer . The resulting protein solution was dialysed against 20 mM Tris-HCl , pH 8 . 5 and 2 mM DTT for 2 hours . SEC with a HiLoad 16/60 Superdex 75 ( GE Healthcare ) was performed as final purification step in 20 mM Tris-HCl , pH 8 . 5 and 10 mM DTT . The concentration of the protein of interest was measured by NanoDrop ND-1000 ( Thermo Scientific ) . The accuracy of NanoDrop measurements was confirmed by additional measurement of the concentration of two samples from independent purifications using BCA Protein Assay Reducing Agent Compatible kit ( Thermo Scientific ) as described by the manufacturer . Microscale thermophoresis protein-protein interaction studies were performed on the Monolith NT . 115 ( Nanotemper Technologies , Munich ) using fluorescently labeled proteins as described [43 , 44] . For the TIR/MAL , TIR/MyD88 and TIR/TRAM protein labeling , the standard labeling kit for the fluorescent dye Alexa Fluor 647 from Nanotemper was used . Solutions of unlabelled A46 ( 1–229 ) , A46 ( 87–229 ) and A46 ( 1–83 ) were serially diluted from 150–450 μM to 8–20 nM in the presence of 30–70 nM of one of the labeled TIR/MAL , TIR/MyD88 or TIR/TRAM proteins . Measurements were performed at 25°C in 20 mM TrisHCl pH 8 . 5 , 100 mM NaCl , 5% glycerol , 1 mM TCEP , 1 mM EDTA , 0 . 05% Tween 20 using 50% LED power and 60% or 80% IR-laser power . Data analysis was performed with Nanotemper analysis software , v . 1 . 2 . 101 . Crystals of A46 ( 1–83 ) and A46 ( 1–73 ) were initially obtained at protein concentration of 6 . 75 and 3 . 5 mg/ml , respectively , in 20 mM TrisHCl pH 8 . 5 and 10 mM DTT in multiple buffer formulations of the PACT Premier crystallization screen ( Molecular Dimensions , Suffolk , UK ) using the sitting-drop vapour diffusion technique and a nanodrop-dispensing robot ( Phoenix RE; Rigaku Europe , Kent , United Kingdom ) . We obtained crystals of both protein constructs; however , for A46 ( 1–73 ) the crystals were not amenable for diffraction experiments . For A46 ( 1–83 ) , the largest crystals grown in 100 mM HEPES 7 . 0 , 20% PEG6000 and 0 . 2 M of one of following salts NaCl , LiCl or NH4Cl were mounted in the loop and flash-cooled in liquid nitrogen . Crystals with selenium methionine labeled A46 ( 1–83 ) were obtained in the same buffer formulations . The diffraction data set was collected at 100K at the peak of Se at λ = 0 . 979 Å at the beamline MASSIF-1 ID30A-1 at the European Synchrotron Radiation Facility ( Grenoble , France ) to 1 . 55 Å resolution and processed using the XDS package [45] . Crystals belonged to the space group C2 ( a = 65 . 79 Å b = 59 . 5 Å c = 47 . 26 Å ) . The structure was solved by ARCIMBOLDO_BORGES ab initio phasing software [25] combining fragment search with Phaser [26] and density modification with SHELXE [46] on the supercomputer Gordon at the SDSC . Autobuilding was carried out using the program AutoBuild from the Phenix package [47] . The structure was refined using the program Phenix Refine [48] and manual adjustments with the software Coot [49] . Stereo-chemistry and structure quality were checked using the program MolProbity [50] . Data collection and refinement statistics are reported in Table 2 . The coordinates of the A46 ( 1–83 ) X-ray structure have been deposited in the Protein Data Bank ( PDB ) database , accession number 5EZU . The experimental SAXS data and derived models of the either full-length A46 or its N-terminal domain have been deposited in small angle scattering biological data bank ( SASBDB ) with the deposition codes SASDBL7 and SASDBK7 . SAXS experiments for the A46 ( 1–83 ) and full-length A46 ( 1–229 ) were performed at 0 . 9918 Å wavelength ESRF at BioSAXS beamline BM29 coupled to the Superdex 200 10/300 exclusion column ( Grenoble , France ) and equipped with PILATUS 1M detector at 2 . 867 m distance from the sample , 0 . 04 < q < 0 . 5 Å-1 ( q = 4π sin θ/λ , 2θ is the scattering angle ) . The data were collected using protein concentrations of 15 . 5 and 4 . 4 mg/ml for the A46 ( 1–83 ) and A46 ( 1–240 ) , respectively . The samples were in a buffer containing 20 mM Tris-HCl pH 8 . 5 , 10mM DTT and the measurements were performed at 20°C . The data were processed and analyzed using the ATSAS program package [51] . The radius of gyration Rg and forward scattering I ( 0 ) were calculated by Guinier approximation . The maximum particle dimension Dmax and P ( r ) function were evaluated using the program GNOM [52] . To demonstrate the absence of concentration dependent aggregation and interparticle interference in the both SAXS experiments , we inspected Rg over the elution peaks and performed our analysis only on a selection of frames in which Rg was most stable ( S4 Fig ) . Overall , such stability of Rg over the range of concentrations observed in the SEC elution indicates that there were no concentration-dependent effects or interparticle interference . The data collection and structural parameter from SAXS analysis are summarized in Table 3 . The ab initio models were derived using DAMMIF [53] . 40 individual models were created for each run , which were then overlaid and averaged using DAMAVER . For the oligomeric state assessment , the theoretical scattering from either theoretical dimer or tetramer using the high-resolution structure ( 5EZU ) was performed . First , the residues missing in the crystal structure were added by CORAL modelling; later the theoretical scattering curves were generated using CRYSOL and compared to the SAXS experimental data for A46 ( 1–83 ) . To obtain a pseudo-atomic model of the full-length A46 , CORAL [39] software was used with the structures for A46 ( 1–83 ) ( 5EZU ) connected by dummy residue linkers to A46 ( 87–229 ) ( 4LQK ) ; the C-terminal domain A46 ( 87–229 ) was extended by 16–19 dummy residues to imitate the full length of the A46 protein . Cell culture and reporter gene assays were performed as reported previously [9] . The following expression plasmids were used: the full-length myc-A46 ( 1–240 ) -FLAG ( amounts 200 , 150 ng ) , the N-terminal domain A46 ( 1–83 ) -FLAG ( 100 , 50 ng ) , the C-terminal domain A46 ( 87–229 ) -FLAG ( 400 , 300 ng ) . The amount of DNA per well was kept constant at 500 ng by supplementation with pCAGGS empty vector . Human embryonic kidney cells 293 stably transfected with TLR4 or MD2 were kind gifts from Dr . Sylvia Knapp . HEK293-TLR4 and HEK293-MD2 were maintained in DMEM supplemented with 10% fetal calf serum , 1% penicillin/streptomycin and 0 . 5 mg/ml geniticin G418 . To perform a reporter assay with a wild-type or lipid-binding mutant of A46 , HEK293-TLR4 cells were grown in 24-well plates and transfected with 80 ng pNF-κB-luc reporter plasmid ( Firefly luciferase ) , 20 ng of pRL-TK ( Renilla luciferase ) internal control and 300 ng of the respective A46 containing plasmid . The supernatant from HEK293-MD2 cells was filtered and added in a ratio of 1:4 with DMEM to HEK293-TLR4 in the stimulation assay . 40 hours post transfection , HEK293-TLR4 cells were stimulated by addition of MD-2 supernatant , DMEM and 500 ng/ml of LPS . After 7 h , cells were collected , lysed in Passive Lysis Buffer ( Promega ) and whole cell lysates were analyzed for luciferase activity using the Dual-Luciferase Reporter Assay ( Promega ) . Firefly luciferase activity was normalized by Renilla luciferase activity . Expression levels of myc-A46-FLAG and myc-A46-FLAG I72A in HEK293T cells were estimated by western blotting . Myc-tagged A46 variants were detected with monoclonal anti-myc 4A6 antibody at a dilution of 1:1000 ( Millipore ) , ϒ-tubulin was used as a loading control and detected with monoclonal anti-tubulin GTU-88 antibody at a 1:5000 dilution ( Sigma ) . Lipid extractions from purified recombinant proteins were achieved by two different methods . Method A: 3 successive vigorous extractions with 10 volumes of diethyl ether after treatment with 2 volumes of 6M HCl overnight . The ether extracts were evaporated under nitrogen and analysed by electrospray mass spectrometric and tatty acid methyl ester analysis as described below . Method B: 3 successive vigorous extractions with ethanol to fully denature proteins ( final 90% v/v ) [54] . The pooled extracts were dried by nitrogen gas in a glass vial and analysed by electrospray mass spectrometry . For electrospray mass spectrometry analysis , extracts were analyzed on a Absceix 4000 QTrap , a triple quadrupole mass spectrometer equipped with a nanoelectrospray source as described previously [55] . Quantification of the fatty acids from method A were done by conversion to the corresponding fatty acid methyl esters ( FAME ) followed by GC-MS analysis as described previously [56] using the following GC temperature program: 70°C for 12 min followed by a gradient to 220°C at 4°C/min and held at 220°C for a further 10 min . Mass spectra were acquired from 50–500 amu . The identity of FAMEs was carried out by comparison of the retention time and fragmentation pattern with mixtures of FAME standards .
|
Viruses possess mechanisms to interfere with the host immune system to enhance their replication . Vaccinia virus , the viral vaccine used to eradicate smallpox , synthesizes many such proteins . The vaccinia virus protein A46 is one of a series of proteins preventing expression of host proteins that induce an anti-viral state . A46 acts early to inhibit anti-viral state induction by specifically binding to certain host adapter proteins such as MyD88 and TRAM . Here , we extend our knowledge of the A46 structure by determining the structure of the protein's N-terminal domain to be an unusual lipid binding fold . In addition , the full-length A46 molecule has a novel quaternary structure that can both bind proteins and lipids , indicating that A46 uses a variety of interactions to sequester host proteins , thus impairing the activation of the anti-viral state and improving the efficiency of viral replication .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"crystal",
"structure",
"condensed",
"matter",
"physics",
"crystals",
"materials",
"science",
"protein",
"structure",
"crystallography",
"materials",
"by",
"structure",
"lipids",
"computer",
"and",
"information",
"sciences",
"solid",
"state",
"physics",
"protein",
"structure",
"determination",
"proteins",
"molecular",
"biology",
"physics",
"biochemistry",
"scattering",
"computer",
"software",
"biology",
"and",
"life",
"sciences",
"fatty",
"acids",
"physical",
"sciences",
"small-angle",
"scattering",
"macromolecular",
"structure",
"analysis"
] |
2016
|
Vaccinia Virus Immunomodulator A46: A Lipid and Protein-Binding Scaffold for Sequestering Host TIR-Domain Proteins
|
We captured , ectoparasite-combed , and blood-sampled cave-roosting Madagascan fruit bats ( Eidolon dupreanum ) and tree-roosting Madagascan flying foxes ( Pteropus rufus ) in four single-species roosts within a sympatric geographic foraging range for these species in central Madagascar . We describe infection with novel Bartonella spp . in sampled Eidolon dupreanum and associated bat flies ( Cyclopodia dubia ) , which nest close to or within major known Bartonella lineages; simultaneously , we report the absence of Bartonella spp . in Thaumapsylla sp . fleas collected from these same bats . This represents the first documented finding of Bartonella infection in these species of bat and bat fly , as well as a new geographic record for Thaumapsylla sp . We further relate the absence of both Bartonella spp . and ectoparasites in sympatrically sampled Pteropus rufus , thus suggestive of a potential role for bat flies in Bartonella spp . transmission . These findings shed light on transmission ecology of bat-borne Bartonella spp . , recently demonstrated as a potentially zoonotic pathogen .
The role of bats as reservoirs for viral pathogens—including several responsible for severe human disease—has received increasing attention in recent years [1] . The extent to which this pattern is mirrored by bats’ abilities to host and transmit other zoonotic agents , including bacteria , is less widely acknowledged . Bats have been confirmed as asymptomatic reservoirs for several species of gram-negative Bartonella bacteria in localities as wide-ranging as the United Kingdom [2] , Kenya [3] , Guatemala [4] , Peru [5] , Taiwan [6] , Nigeria [7] , and Puerto Rico [8] . Bartonella spp . infect erythrocytes and epithelial cells of predominantly mammalian hosts , and some are known to cause zoonotic disease ( bartonellosis ) in humans . Most recently , bats in the Northern Hemisphere have been implicated as hosts for the human pathogen , Bartonella mayotimonensis , although the mechanism of transmission between bats and humans remains unclear [9] . Bartonella spp . are frequently transmitted via arthropod vectors [10] and have been identified in several bat ectoparasites , including 19 bat fly species ( Diptera: Hippoboscoidea: Nycteribiidae and Streblidae ) [11 , 12] . However , the presence of Bartonella spp . within these arthropods may simply reflect their ingestion of host blood , and vector transmission of Bartonella between bats has yet to be confirmed via experimental trial or controlled field study . Nonetheless , phylogenetic analyses of global bat fly-Bartonella-bat associations demonstrate Bartonella spp . similarities across bat hosts and ectoparasites [12] , suggesting that bat flies might play a vector role in transmission . To elucidate this relationship , we examined Bartonella prevalence in two sympatric Madagascar fruit bat species—one containing bat flies and fleas and one in which ectoparasites were conspicuously absent . Here we report the presence of closely related Bartonella genotypes in Madagascan fruit bats ( Eidolon dupreanum ) and their associated bat flies ( Nycteribiidae ) in Madagascar . We simultaneously report the absence of Bartonella spp . in bat fleas ( Thaumapsylla sp . ) of E . dupreanum , in addition to the concomitant absence of both ectoparasites and Bartonella in sympatric Madagascan flying foxes ( Pteropus rufus ) .
In November 2013 , 57 E . dupreanum and 32 P . rufus were mist-netted , sampled for pathogens , and live-released from four single-species roost sites in central Madagascar . E . dupreanum bats were captured from two cave roosts ( Angavobe -18 . 918050S , 47 . 94360E; and Angavokely 18 . 932450 S , 47 . 7574170 E ) and P . rufus bats netted from two tree roosts ( Marovitsika -18 . 842180S , 48 . 033630E; and Ambakoana -18 . 511280S; 48 . 171120E ) in the District of Moramanga . All four roost sites are within a 35km radius of one another and a 5km radius of neighboring human communities , distances well within the nightly foraging ranges of these flying foxes ( Fig . 1 ) [13] . This highland region is dominated by savannah grassland interspersed with non-native plantation and mid-elevation ( ~1100m ) humid forest . Both E . dupreanum and P . rufus feed on a range of fruits and nectars and are known to share feeding sites [13] . Upon capture , bats were thoroughly examined for ectoparasites , and all observed flies , fleas , and mites were removed and collected into vials of absolute ethanol with a comb ( fleas ) or tweezers ( mites and bat flies ) . Blood ( 1 . 0ml ) was collected from the brachial vein of adult bats ( forearm >100mm ) and robust juveniles ( 29 P . rufus , 47 E . dupreanum blood-sampled ) . Serum and blood cells were separated by centrifuging and stored in liquid nitrogen in the field , then transferred to -80°C freezers at the Institut Pasteur-Madagascar . This study was carried out in strict accordance with guidelines posted by the American Veterinary Medical Association . All field protocols employed were pre-approved by the Princeton University Institutional Animal Care and Use Committee ( IACUC Protocol # 1926 ) , and every effort was made to minimize discomfort to animals . Bat flies . Ectoparasite samples were processed at the University at Buffalo ( Buffalo , NY , USA ) . Ectoparasite DNA was extracted from a subset of samples ( 19 bat flies and 6 fleas ) using the Qiagen Animal Tissue kit ( QIAGEN , Valencia , CA , USA ) . Ectoparasite voucher specimens were slide-mounted and identified using available taxonomic keys . Blood pellets . Blood pellet samples were processed at the CDC’s Division of Vector-Borne Diseases ( Fort Collins , CO , USA ) . DNA was extracted from blood samples using a Qiagen QIAamp tissue kit ( QIAGEN , Valencia , CA , USA ) according to the manufacturer’s instructions . Bartonella spp . assay . All DNA extractions ( ectoparasites and blood ) were examined for Bartonella spp . by conventional PCR targeting multiple genes employed in previous research: gltA , ftsZ , and nuoG genes for arthropod bartonellae , and gltA and ITS sequence for blood samples [3 , 11 , 12] . Only samples with sequences that unequivocally BLASTed to Bartonella spp . and nested within known Bartonella sequences by phylogenetic analysis were considered positive ( RAxML 7 . 7 ) [14] . Samples positive by PCR with inconclusive sequence data were thus considered negative for Bartonella spp . in our analysis . We compared the frequency of bat fly ( C . dubia ) and bat flea ( Thaumapsylla sp . ) infections , as well as Bartonella spp . prevalence in both bat hosts and in ectoparasite arthropods . Differences were examined between species and across sampling sites using chi-squared and Fisher exact tests in the statistical program R [15] . We used a p-value threshold of 0 . 01 to assess whether observed ectoparasite burden and Bartonella spp . prevalence were independent of species and sampling site .
Seven of 24 ( 29 . 2% ) Eidolon dupreanum sampled from Angavobe cave and 20 of 23 ( 87% ) E . dupreanum sampled from Angavokely cave were found to host Cyclopodia dubia ( Nycteribiidae ) bat flies . Ten of those 23 ( 43 . 5 . 1% ) Angavokely E . dupreanum also hosted Thaumapsylla sp . fleas ( Table 1 ) . Both frequency of bat fly and flea hosting varied significantly by roosting site , via analysis by chi-squared tests of independence ( bat fly: X2 = 13 . 768 , df = 1 , p = 0 . 0002; flea: X2 = 10 . 7863 , df = 1 , p = 0 . 001 ) and Fisher’s exact tests ( bat fly: p = 8 . 828e-05; flea: p = 0 . 0002 ) . Two of 2 ( 100% ) bat flies processed from Angavobe and 15 of 17 ( 88 . 2% ) bat flies processed from Angavokely were considered positive for Bartonella DNA by sequence , although all bat flies processed were Bartonella spp . positive by PCR alone . None of the six Thaumapsylla fleas processed were positive for any Bartonella target gene . The presence of Thaumapsylla sp . at the Angavokely site represents the first geographic record for Madagascar; this genus is known from Eidolon spp . elsewhere [16] . Blood samples from eight of 24 ( 33 . 3% ) Angavobe E . dupreanum and thirteen of 23 Angavokely ( 56 . 5% ) E . dupreanum were positive for Bartonella DNA by PCR confirmed with sequence for one or more genes ( Table 1 ) . Bartonella spp . prevalence did not vary significantly between Angavobe and Angavokely roosting sites as indicated by a chi-squared test for independence ( X2 = 1 . 7029 , df = 1 , p-value = 0 . 1919 ) and Fisher’s exact test ( p-value = 0 . 1468 ) . In Angavobe , bats demonstrated both singular infections with Bartonella spp . and with bat flies , as well as simultaneous co-infection with bat flies and Bartonella spp . ( Fig . 2 ) . In Angavokely , E . dupreanum individuals hosted every possible combination of bat fly/flea/Bartonella spp . infection and co-infection save for singular flea infestations in the absence of other pathogens ( Fig . 2 ) . It should be noted that , prior to processing , bats were housed together with others from the same sample site in wooden transport cages , and ectoparasite sharing among individuals was easily facilitated . No ectoparasites were recovered from either the 12 Pteropus rufus examined at the Marovitsika site or the 17 P . rufus sampled at the Ambakoana site ( Table 1 ) . As with ectoparasites , none of the 29 P . rufus samples ( 12 from Marovitisika , 17 from Ambakoana ) were positive for Bartonella spp . by either molecular target ( Table 1 ) . All Bartonella spp . sequences from E . dupreanum bats and associated C . dubia bat flies nested within or close to known major Bartonella lineages ( Fig . 3 ) [17] . Although sequence data retrieved are insufficient to reach final Bartonella species identification , novel genotypes are present . Sequences ( gltA ) from sampled bats group with those retrieved from Cyclopodia bat flies .
The recent identification of bats as reservoirs for human pathogenic Bartonella mayotimonensis [9] validates further investigation of the zoonotic potential of Bartonella spp . in Chiropteran reservoirs . In Madagascar , insectivorous bats are known to roost in human residences , and both P . rufus and E . dupreanum are widely consumed as bushmeat , highlighting the extent of human-wildlife interface in the region [13] . In keeping with trends of persistent bacterial infection exhibited by bat-borne Bartonella elsewhere [3–5] , we report high Bartonella spp . prevalence ( 57 . 4% ) in a long-lived , cave-roosting E . dupreanum host ( lifespan 10–20 years [13] ) . We correspondingly report no Bartonella infections in sympatric P . rufus , though our current sample size is too small to determine whether this absence is universal across the Madagascar population . Additionally , further study is needed to address whether these Bartonella spp . prevalence patterns are an artifact of phylogeny or ecology . The ability of Eidolon bats to serve as hosts for Bartonella spp . has now been documented in both tree-roosting [3] and cave-roosting environments—consistently in association with bat flies . P . rufus does not seem to host bat flies in Madagascar [18] , although sampling has not been exhaustive enough to consider this absence a certainty . Tree-roosting Pteropus spp . are known to host bat flies throughout southeast Asia [19] and Australia [20] , and investigation of Bartonella spp . infections in these populations will help address the relative influence of host genetic predisposition for Bartonella infection versus vector ecology . In addition to pathogen prevalence in the host , we report Bartonella spp . infection in bat flies ( Cyclopodia dubia ) of E . dupreanum , simultaneous with Bartonella DNA absence in flea ectoparasites ( Thaumapsylla sp . ) of those same bats . Fleas are the confirmed vector for Bartonella henselae , the causative agent in cat scratch fever [21] , and fleas of bats have been previously reported in association with Bartonella DNA [9 , 22] . In our study , both bat flies and fleas were host-specific and likely consumed host blood , although only flies tested positive for Bartonella spp . , suggesting that the mechanisms by which arthropods host and transmit pathogens vary and impact their functionality as vectors . Sampling of flea ectoparasites was not extensive enough to assess the true extent of their ability , or lack of ability , to transmit Bartonella spp . , and further experimental studies of the vector potential of both bat flies and fleas for Bartonella is warranted . Finally , observed differences in the frequency of ectoparasite burden between sample sites for E . dupreanum indicated significantly higher rates of ectoparasite infection with both flies and fleas in Angavokely vs . Angavobe . These differences could result from ecological variation in both host density and/or climate between the two cave roosts . More extensive spatial sampling , in conjunction with climactic monitoring , in other E . dupreanum roost sites of varying size , temperature , and humidity across Madagascar will help elucidate habitat thresholds for ectoparasite invasion . In particular , sampling of E . dupreanum in reported tree roosts in central Madagascar will shed light on the extent to which roosting behavior limits bats’ abilities to support ectoparasites in this system [13] . If Bartonella spp . are , indeed , transmitted by bat fly vectors , such findings will have important implications for our understanding of the distribution , prevalence , and transmission dynamics of a potentially zoonotic pathogen .
|
Bartonella spp . are bacteria that inhabit the red blood cells of both human and animal hosts . Among humans , Bartonella spp . are known to cause several febrile illnesses , including Carrion’s disease ( Bartonella bacilliformis ) , trench fever ( Bartonella quintana ) , and cat scratch fever ( Bartonella henselae ) , all of which are transmitted via arthropod vectors—respectively sand flies , lice , and fleas . Bats are known to host multiple Bartonella spp . , including some capable of infecting humans . Some bat species are also known to host obligate ectoparasites known as bat flies ( Diptera: Hippoboscoidea ) , which also sometimes support Bartonella spp . infections . The role of bat flies and other bat ectoparasites as vectors for Bartonella spp . transmission has been suggested , but not fully explored . We demonstrate Bartonella spp . infection in one species of Madagascar fruit bat , which hosts bat flies , simultaneously with the absence of Bartonella in a fruit bat species of overlapping range that appears not to support these ectoparasites . In light of ongoing trends of zoonotic emergence of human diseases from bat reservoirs , further understanding of the transmission dynamics of bat-borne pathogens is paramount .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[] |
2015
|
Bartonella spp. in Fruit Bats and Blood-Feeding Ectoparasites in Madagascar
|
The RIG-I-like RNA helicase ( RLR ) -mediated interferon ( IFN ) response plays a pivotal role in the hepatic antiviral immunity . The hepatitis A virus ( HAV ) and the hepatitis C virus ( HCV ) counter this response by encoding a viral protease that cleaves the mitochondria antiviral signaling protein ( MAVS ) , a common signaling adaptor for RLRs . However , a third hepatotropic RNA virus , the hepatitis E virus ( HEV ) , does not appear to encode a functional protease yet persists in infected cells . We investigated HEV-induced IFN responses in human hepatoma cells and primary human hepatocytes . HEV infection resulted in persistent virus replication despite poor spread . This was companied by a type III IFN response that upregulated multiple IFN-stimulated genes ( ISGs ) , but type I IFNs were barely detected . Blocking type III IFN production or signaling resulted in reduced ISG expression and enhanced HEV replication . Unlike HAV and HCV , HEV did not cleave MAVS; MAVS protein size , mitochondrial localization , and function remained unaltered in HEV-replicating cells . Depletion of MAVS or MDA5 , and to a less extent RIG-I , also diminished IFN production and increased HEV replication . Furthermore , persistent activation of the JAK/STAT signaling rendered infected cells refractory to exogenous IFN treatment , and depletion of MAVS or the receptor for type III IFNs restored the IFN responsiveness . Collectively , these results indicate that unlike other hepatotropic RNA viruses , HEV does not target MAVS and its persistence is associated with continuous production of type III IFNs .
The hepatitis E virus ( HEV ) causes significant morbidity and mortality worldwide [1 , 2] . Although HEV is known for causing acute hepatitis in developing countries , cases of chronic HEV infection have been reported in recent years in industrialized countries in persons with an immune system compromised by treatment with suppressive therapies or HIV co-infection . Patients chronically infected with HEV can rapidly progress to liver fibrosis and cirrhosis if left untreated . The majority of chronic cases are in developed countries and caused by genotype 3 HEV , the most prevalent HEV genotype in those countries . There are no HEV-specific treatments available at present . Ribavirin ( RBV ) alone or in combination with pegylated-interferon ( PegIFN ) has been used to treat chronic HEV infection with some success . However , not all patients can be treated with RBV and resistance has been described [3] . Mechanisms for immune control of HEV particularly during chronic infection are poorly understood . The hepatitis C virus ( HCV ) induces a strong baseline IFN-stimulated gene ( ISG ) expression that is associated with a persistent infection outcome and poor responsiveness to IFN-based therapy [4 , 5] . In contrast , the hepatitis A virus ( HAV ) does not persist and induces only limited type I IFN responses [6 , 7] . Relatively little is known about the IFN response or evasion mechanisms in HEV infection . Elevated ISG expression was detected in patients with chronic HEV infection and HEV-infected mice engrafted with human hepatocytes [8 , 9] . In experimentally infected chimpanzees , HEV also induced ISG expression , although the levels were lower than those measured after HCV infection [10] . Interestingly , recent studies have shown that HEV is more resistant to the antiviral effect of IFNs than HCV [11 , 12] , but the underlying mechanism is not clear . Despite the differences in early IFN responses and infection outcomes , both HAV and HCV target the mitochondria antiviral signaling protein ( MAVS ) , thereby blocking IFN production in virus-infected cells [13 , 14] . A recent study demonstrated that the capacity of HAV to evade MAVS-mediated type I IFN responses defines its host species range [15] . These studies involving HAV and HCV suggest that MAVS inactivation is a requirement for successful infection of the liver by small hepatotropic viruses . Whether HEV also targets MAVS is unknown . A sole member of the Hepeviridae family , HEV has a 7 . 2 kb single-stranded positive-sense RNA genome encoding three open reading frames ( ORF1-3 ) [16] . ORF1 is a large polyprotein that contains several functional domains essential for virus replication , whereas ORF2 and ORF3 are both translated from a 2 . 2 subgenomic RNA generated during virus replication and involved in virus assembly and egress , respectively [17 , 18] . It has been shown that the putative papain-like protease ( PCP ) domain and the macro domain of the HEV ORF1 protein block RIG-I and Tank-binding kinase ( TBK ) -1 ubiquitination , thereby suppressing IFN production [19] . The HEV ORF3 protein , on the other hand , has been reported to enhance IFN production [20] . HEV also induced ISG expression in PLC/PRF/5 human hepatoma cells [21] and in A549 human lung epithelial cells [22] . Relevance of these observations to natural infection is uncertain , however , as most of studies were not conducted in hepatocyte cell lines and/or relied on overexpression of viral proteins . In this study , we investigated HEV-induced IFN responses in HepG2 human hepatoma cells and primary human hepatocytes . We found that unlike HAV and HCV , HEV does not cleave MAVS , leading to a sustained IFN response in persistently infected cells . Moreover , the JAK/STAT1 pathway was persistently activated and poorly responded to exogenous IFNs , potentially explaining the relative IFN resistance of this virus . These results provide insights into the interactions between HEV and innate signaling during persistence .
To investigate cellular responses to HEV , we used a cell culture-adapted genotype 3 HEV strain ( Kernow C1/p6 ) that replicates efficiently in cell culture [23] . As described elsewhere [23] , the particle to FFU ratio of HEV is extremely low ( ~15 , 000 ) , and only a small fraction of cells ( 5–10% ) were HEV-positive despite a high dose of inoculum ( 1x103 HEV genome equivalents per cell ) . The percentage of HEV-positive cells was stable over 11 days of culture and the majority of HEV foci contained only single infected cells ( Fig 1A ) . This suggested establishment of persistent infection , but with poor , if any , virus spread . The percentage of infected cells slightly decreased after 20 days , but the majority remained singly infected ( S1 Fig ) . Quantification of HEV ORF1 and ORF2 RNA levels by qRT-PCR showed that HEV replication peaked at around 5 days after infection , then declined to a relatively stable level ( Fig 1B ) . Reduction of HEV RNA replication in HepG2 cells after 5 days may have been caused by activation of the IFN pathways . To test this hypothesis , we measured the concentrations of different types of IFNs in the culture supernatants over the course of HEV infection . Neither IFN-α nor IFN-β proteins were detected . However , IFN-λ protein was readily detected ( Fig 1B ) . Persistent HEV replication was observed despite continuous IFN-λ production , even at 30 days after infection ( S1 Fig ) . To corroborate this result , we measured the mRNA expression of different types of IFNs ( and subtypes for IFN-α ) as well as IFN-stimulated genes ( ISGs ) by RT-qPCR at various times following HEV infection . IFN-λ1 and IFN-λ2/3 were found to be increased at the mRNA levels ( Fig 1C ) . IFN-λ4 , a recently discovered type III IFN associated with the control of HCV [24] , did not increase ( Fig 1C and S2 Fig ) . In contrast , little increase in the mRNA levels of IFN-β and multiple IFN-α subtypes was detected , consistent with the absence of IFN-α/β protein production . Increased IFN-λs were associated with increased expression of a number of ISGs ( e . g . , ISG15 , IFIT1 , RSAD2 , and CXCL10 ) . The IFN-λ and ISG response to viral replication peaked at 4–6 days post-inoculation then declined slightly before stabilizing . Importantly , increased IFN-λ , but not IFN-α/β , protein expression was also detected upon HEV infection of primary human hepatocytes ( Fig 1D and 1E ) . Similar results were obtained in two independently generated HepG2 cell clones harboring an HEV subgenomic replicon RNA ( Fig 1F–1H ) . Replication of HEV RNA was required for the induction of IFNs and ISGs since their expression was reduced to a basal level when the replicon was eliminated by treatment with IFN-α and ribavirin . Collectively , these results demonstrate that HEV persisted in the cells despite the continuous production of IFN-λs . To assess the role of released IFN-λs in regulating HEV replication , we depleted IFNLR1 , a component of the type III IFN receptor , by transducing HepG2 cells with a lentivirus expressing an IFNLR1-specific short hairpin RNA ( shRNA ) . For comparison , we also depleted IFNAR1 , receptor for type I IFNs . The knockdown efficiency was examined by western blotting ( Fig 2A ) . Depletion of these receptors greatly reduced the cellular responsiveness to IFN-α and IFN-λ , respectively ( Fig 2B ) . Notably , the magnitude of HEV-induced ISG expression was significantly reduced in cells depleted of IFNLR1 when compared to the parental cells or cells depleted of IFNAR1 ( Fig 2C ) . Moreover , depletion of IFNLR1 , but not IFNAR1 , resulted in a 3-fold increase in the HEV RNA abundance and infectious virus production ( Fig 2D and 2E ) , suggesting that HEV replication is restricted by the type III IFNs induced by infection . Both genomic and subgenomic HEV RNAs are capped and polyadenylated [25] , raising the question of how HEV is detected by the cells . To determine the signaling pathway ( s ) involved in type III IFN production by HEV , we depleted cytoplasmic RNA sensors , namely RIG-I and MDA5 , as well as their downstream adaptor protein MAVS by transducing HepG2 cells with lentiviruses expressing gene-specific shRNA . Western blots showed that the respective protein levels in the transduced cells were substantially reduced when compared to control cells transduced with a lentivirus expressing GFP ( Fig 3A ) . Cells were then challenged with Sendai virus ( SeV ) , a potent agonist for RIG-I , or poly IC , which primarily stimulates MDA5 when delivered intracellularly by transfection . Depletion of RIG-I led to a substantial reduction in the IFN response to SeV , and to a lesser degree poly IC , whereas depletion of MDA5 reduced the cellular response to poly IC , but not to SeV , confirming the functional knockdown of respective innate sensing pathways ( Fig 3B and 3C ) . As expected , depletion of MAVS also reduced the IFN response . Upon HEV infection , cells depleted of MDA5 , MAVS , and to a lesser degree RIG-I , displayed reduced type III IFN production ( Fig 3D ) , suggesting that both RIG-I and MDA5 are involved in HEV-induced IFN response . Notably , the level of intracellular HEV RNA and the number of infected cells were both increased in cells depleted of MDA5 , MAVS , and to a less degree RIG-I ( Fig 3E and 3F ) . Depletion of RIG-I , MDA5 , or MAVS in the replicon cells also resulted in increased HEV RNA levels ( S3 Fig ) . Transcription factors interferon regulatory factor ( IRF ) -3 and IRF-7 have been implicated for the induction of IFN-λ [26] . Depletion of IRF-3 , but not IRF-7 , resulted in a similar reduction in IFN-λ mRNA expression in HEV-infected cells ( Fig 3G and 3H ) . Accordingly , HEV replication was increased in IRF-3-depleted cells , but not in IRF-7 depleted cells ( Fig 3I and 3J ) . IRF-3 was found in the nucleus of ~5% of HEV-infected cells , but in 0 . 1% of uninfected cells ( S4 Fig ) , indicating that IFN-λ was produced primarily , if not exclusively , from HEV-infected cells . Taken together , these data demonstrated that both RIG-I and MDA5 participated in sensing HEV genomes , resulting in a MAVS-dependent type III IFN response . The critical dependence of HEV-induced IFN response on MAVS indicated that the function of MAVS remained intact in HEV-infected cells . To test this , we determined the MAVS protein abundance in cells infected with either HEV or HAV . MAVS was largely absent in HAV-infected cells , as expected . However , its expression was not reduced in cells infected with HEV ( Fig 4A ) , and its mitochondrial localization was not altered in cells harboring the HEV replicon ( Fig 4B ) . A small fraction of MAVS colocalized with peroxisomes . However , no gross difference in this rare colocalization was found between the parental and the replicon cells ( S5 Fig ) . In addition , the MAVS protein size was identical between cells with or without HEV replicon ( Fig 4C ) . These data indicate that MAVS is not degraded by HEV . Upon activation , MAVS forms “prion-like” polymers [27] . In HepG2 cells , MAVS polymers only became detectable following poly IC transfection . However , MAVS polymers were present in the replicon cells even without poly IC transfection , consistent with elevated ISG expression in these cells . Transfection with poly IC in the replicon cells further increased the amount of MAVS polymers ( Fig 4D ) , indicating that a portion of MAVS were not polymerized and capable of responding to stimulation . The above analyses suggested that MAVS is activated to mediate IFN-λ production in HEV-infected cells . To further investigate the impact of HEV on MAVS-mediated signaling , we compared cells with or without the HEV replicon for their responsiveness to poly IC ( an MDA5 agonist ) , or HCV 3’ untranslated region ( UTR ) RNA , an agonist of RIG-I [28] . Although unstimulated replicon cells had a higher baseline IFN response than the parental cells , poly IC transfection stimulated IFN expression to a similar extent in both ( Fig 4E ) . Likewise , more IFN-λ proteins were released from the replicon cells during the 6 h period of poly IC treatment ( Fig 4F ) . Moreover , Western blots revealed that the protein levels of several ISGs including RIG-I , MDA5 and ISG56 were higher in the replicon cells than in the parental cells or in the replicon cured cells , and their expression was further increased following poly IC stimulation ( Fig 4G ) . However , the HCV 3’UTR RNA failed to induce more IFNs in the replicon cells ( Fig 4H ) and a similar amount of IFN-λ protein was released from HepG2 cells and HepG2/replicon cells during the 6 h period of treatment ( Fig 4I ) . Moreover , the protein levels of ISGs ( RIG-I , MDA5 , and ISG56 ) , which were expressed at higher levels in unstimulated replicon cells , remained unchanged following HCV 3’UTR RNA stimulation ( Fig 4J ) . The poor response to HCV 3’UTR RNA in the replicon cells was not due to impaired transfection since the same method was used for poly IC transfection . Similar results were obtained when cells were infected with SeV , which also stimulates the RIG-I pathway ( S6 Fig ) . These results demonstrated that MAVS remained functionally intact in the HEV-replicating cells , and that HEV specifically blocks RIG-I signaling , likely at a step ( s ) upstream of MAVS . HEV is considered relatively resistant to exogenous IFN treatment when compared to HCV [11 , 12] . The mechanism for the IFN resistance is currently not understood . We obtained similar results when recombinant IFNs were added one day after transfection of viral RNA ( Fig 5A ) . The antiviral effect of IFNs waned when added at later times ( Fig 5B ) , suggesting that HEV replication became relatively resistant to IFNs once virus replication has been established . Although type I and type III IFNs utilize different receptors , both signal through the JAK/STAT pathway [29] . To address the question of whether constant activation of the JAK/STAT signaling renders the infected cells more resistant to exogenous IFNs , a luciferase reporter driven by an IFN-stimulated response element ( ISRE ) promoter was used . Activity of the ISRE was measured in cells with or without an HEV replicon after treatment with recombinant IFN-α or IFN-λ . The basal level of luciferase expression was higher ( ~6-fold ) in the replicon cells , consistent with an elevated IFN response in these cells . Both IFN-α and IFN-λ dose-dependently induced luciferase expression in cells without replicon , ranging from 4–16 fold ( Fig 5C ) . By comparison , the ISRE reporter gene response increased less than 2-fold in replicon cells treated with the highest dose of IFNs , indicating an establishment of IFN resistance in these cells . Induction of several ISGs ( e . g . , ISG15 and IFIT1 ) by IFN-α or IFN-λ was also impaired in the replicon cells ( S7 Fig ) . Importantly , the responsiveness to IFNs was largely restored after MAVS or IFNLR1 depletion ( Fig 5C and S8 Fig ) , suggesting that HEV-induced activation of the endogenous IFN pathways plays a critical role in developing resistance to further IFN treatment . Consistent with this , MAVS depletion also enhanced the antiviral effects of IFN-α or IFN-λ in the replicon cells ( Fig 5D ) . Both tyrosine phosphorylation and serine phosphorylation of STAT1 is required for its maximal activity [30] . Whereas phosphorylation at the tyrosine 701 ( pY701 ) is essential for efficient DNA binding , phosphorylation at the serine 727 ( pS727 ) is thought to augment ISG expression in a gene-specific manner [30 , 31] . To further investigate the mechanism underlying HEV-induced IFN refractoriness , we compared the levels of the total and phosphorylated STAT1 proteins in cells with or without HEV replicon either before or after IFN treatment . The basal level of the total STAT1 protein was much higher in the replicon cells than in the parental cells ( Fig 5E and 5F ) , consistent with STAT1 itself being an ISG . In addition , the basal level of pS727 , and to a lesser extent pY701 , were also higher in the replicon cells . Treatment with IFNs led to a significant increase in the levels of both pY701 and pS727 in the parental cells . However , the increase was much less in the replicon cells , despite higher levels of the total STAT1 proteins ( Fig 5F ) . A close examination of STAT1 nuclear/cytoplasmic distribution revealed that although high levels of pS727 were present in the replicon cells , the majority of it retained in the cytoplasm and failed to translocate into the nucleus even after high dose IFN treatment ( S9 Fig ) . IFN-induced nuclear translocation of pY701 appeared to be normal in the replicon cells , but the extent of phosphorylation was much less than in the parental cells since the level of the total STAT1 proteins was much higher in the replicon cells . Notably , MAVS depletion led to a reduced basal level of total and phosphorylated STAT1 , and IFN-induced STAT1 phosphorylation was largely restored ( Fig 5G ) . Since MAVS depletion led to a 3-fold increase in the HEV RNA abundance ( S3 Fig ) , this result indicated that HEV does not interfere with the JAK/STAT1 signaling directly .
The purpose of this study was to compare host IFN responses against HEV with those elicited by other hepatitis viruses that persist ( HCV ) or not ( HAV ) . Both HAV and HCV target MAVS for proteolysis , suggesting that this mechanism for inactivating the IFN response is generally important for infection of the liver by hepatotropic viruses . Our studies with HEV indicate this is not the case . Several lines of evidence indicate that HEV does not target MAVS . First , MAVS abundance , protein size as well as its mitochondria localization were not altered in HEV-infected cells or in cells harboring an HEV replicon . Second , HEV-infected cells and replicon cells produced a sustained IFN response in a MAVS-dependent manner . Third , MAVS-mediated MDA5 signaling remained intact in the replicon cells . Lastly , MAVS was required for elevated STAT1 phosphorylation in the replicon cells . These results provide strong evidence that in contrast to HAV and HCV , HEV neither cleaves MAVS nor interferes with its function . The lack of MAVS cleavage in infected cells is likely due to an absence of HEV protease activity directed against this signaling protein . Although the HEV ORF1 protein contains a putative papain-like cysteine protease ( PCP ) domain , convincing evidence for protease activity is lacking . Proteases encoded by positive-stranded RNA viruses are generally involved in the processing of viral polyproteins . However , while apparently truncation of HEV ORF1 proteins was observed in some early studies [32–34] , only full-length ORF1 proteins have been detected using robust mammalian overexpression systems and HEV replicon cells [35 , 36] . ORF1 may be present at low abundance in infected cells and development of high quality antibodies against ORF1 will likely be needed to firmly address if the PCP domain possesses protease activity towards the HEV polyprotein and cellular substrates . Consistent with the preservation of MAVS function , HepG2 cells containing the HEV replicon produced a sustained type III IFN response , as evidenced by the elevated expression of IFN-λs and multiple ISGs . Why only type III , but not type I IFNs were produced is not clear , but similar results were reported in studies with hepatitis B virus ( HBV ) and HCV [37–42] . Interestingly , polymorphisms in type III IFNs have been linked to HCV clearance [43] , suggesting that type III IFNs play a critical role in regulating antiviral responses in liver . Since the receptors for type I IFNs are broadly expressed while the receptors for type III IFNs are restricted to epithelial cells such as hepatocytes , one hypothesis is that hepatocytes mainly produce type III IFNs in response to infections to limit host antiviral responses to only locally infected cells . Another finding from this study is that the activation of the IFN pathway rendered HEV-infected cells more resistant to further stimulation with exogenous IFNs . Our results suggest that this is likely due to persistent activation of the JAK/STAT1 signaling , rather than virus-mediated inhibition . Both the total and phosphorylated STAT1 proteins were elevated in cells harboring the HEV replicon , but STAT1 phosphorylation was minimally increased after treatment with exogenous IFN-α or IFN-λ . Furthermore , although present at a high level in the replicon cells , serine phosphorylated STAT1 ( pS727 ) primarily located in the cytoplasm and did not enter the nucleus even after treatment with high doses of IFN-α or IFN-λ . Thus , these cells became highly resistant to both type I and type III IFNs ( likely type II IFN as well ) provided exogenously . Such a strategy likely favors the virus in vivo , where infected cells are exposed to IFNs produced from different cell types ( e . g . plasmacytoid dendritic cells , natural killer cells , and T cells ) [44 , 45] . Importantly , MAVS depletion effectively reversed the IFN responsiveness of these cells and enhanced the antiviral effects of IFNs , indicating a host feedback mechanism rather than a direct viral antagonism is responsible . Several ISGs , such as USP18 and members of the suppressor of cell signaling ( SOCS ) , have been shown to negatively regulate JAK/STAT activity [46 , 47] . USP18 was significantly increased in HEV-infected cells . However , SOCSs were not ( S10 Fig ) . While additional work is needed to elucidate the mechanism for HEV-induced IFN resistance , this data may provide an explanation for the relative resistance of HEV to exogenous IFN treatment observed in this and two other recent studies [11 , 12] . Despite the elevated expression of IFN-λs and ISGs , the IFN level was insufficient to eliminate HEV and the virus persisted in culture . Elimination of HEV was possible , but only when cells were treated with high doses of IFNs for an extended period ( S11 Fig ) . How HEV is able to replicate in the presence of multiple ISGs remains a question for future studies . It is worth noting that certain ISGs ( e . g . , ISG15 , PKR , and ADAR ) facilitate virus replication by countering the antiviral actions of IFNs [48 , 49] . For example , ISG15 is a well-known factor that is associated with IFN resistance in hepatitis C patients [50] , and was highly induced in HEV-infected cells as well as in HEV-infected patients and chimpanzees [9 , 10] . More work is needed to define the roles of different ISGs in the life cycle of HEV . This study sheds new light on HEV persistence . HEV infection is typically self-limited , but it frequently establishes persistence when the host immune system is compromised . Elevated ISG expression has been detected in both acute and chronic HEV infections [9 , 10] , as well as in HEV infected humanized chimeric mice where no human immune cells are involved [8] . Thus , the type III IFNs produced by HEV infected cells could be an important source of IFNs that drive ISG expression in vivo , although involvement of other cell types cannot be ruled out . In the case of HAV and HCV , pDCs are recruited to infected liver and produce copious IFNs after contacting infected cells [51–53] . pDCs may be similarly activated in HEV infection and contribute to the overall IFN response . In this regard , creation of an IFN refractory state in infected cells would favor persistent HEV replication . Thus , the production of type III IFNs by HEV infected cells may be essential for HEV persistence in immunosuppressed patients where adaptive immunity is compromised and the virus is not cleared from persistently infected cells . In summary , we have shown that HEV induced a sustained type III IFN response in infected cells . This is in sharp contrast to HAV and HCV , both of which cleave MAVS and ablate IFN production in cells they infect . We show that although HEV induced-type III IFNs restricted HEV replication , the IFN level was insufficient to eliminate the virus . Instead , it rendered infected cells refractory to high doses of exogenous IFNs . Our data provide insight into the mechanisms of HEV persistence and the relative IFN resistance of this virus .
Huh-7 cells ( a gift from Stanley Lemon at the University of North Carolina ) and HepG2 cells ( CRL-10741 , ATCC ) were maintained at 37°C in 5% CO2 in Dulbecco Modified Eagle medium ( DMEM ) containing 10% fetal bovine serum ( FBS ) , 100 U/ml penicillin and 100 μg/ml streptomycin . Cryopreserved primary human hepatocytes were purchased from the In Vitro ADMET Laboratories ( Columbia , MD ) and maintained according to provider’s instructions . HepG2 cells harboring HEV replicon p6/neo , HepG2/p6neo , were generated by transfecting HepG2 cells with in vitro transcribed p6/neo RNA , and selected with 700 μg/ml G418 sulfate ( Invivogen ) starting 3 days post-transfection . To eliminate the HEV replicon , HepG2/p6neo cells were treatment with IFN-α ( 100 ng/ml , Sigma ) and ribavirin ( 10 μM , Sigma ) for 5 weeks . These resulting cells ( replicon-cured ) did not contain detectable HEV RNA . Stable short hairpin RNA ( shRNA ) knockdown cells were generated by transducing cells with shRNA-expressing lentiviral particles ( MISSION shRNA lentiviral system , Sigma ) , and selected with 2 ug/ml puromycin ( Invivogen ) . Pooled cells were used in all experiments . The target sequences are the following: RIG-I: CCAGAATTATCCCAACCGATA; MDA5: CCAACAAAGAAGCAGTGTATA; MAVS: GCATCTCTTCAATACCCTTCA; IRF-3: GCCAACCTGGAAGAGGAATTT; IRF-7: CCCGAGCTGCACGTTCCTATA; IFNAR1: GCTCTCCCGTTTGTCATTTAT; IFNLR1: CCCTAGTTAGGCCCAGATAAA . HEV stock was generated by transfecting Huh-7 cells with in vitro transcribed HEV Kernow C1/p6 RNA , as previously described [54] . Hepatitis A virus stock ( HM175/18f ) and hepatitis C virus infectious clone ( H77S3/Gluc ) were kindly provided by Stanley Lemon ( University of North Carolina at Chapel Hill ) . Sendai virus ( Cantell strain ) was purchased from Charles River Laboratories . The infectious cDNA clone of the HEV genotype 3 Kernow-C1 p6 strain , was kindly provided by Suzanne Emerson ( National Institutes of Health , Bethesda , MD ) [23 , 55] . The HEV Kernow-C1 p6 replicon construct harboring a neomycin resistant gene ( neo ) was constructed by overlapping polymerase chain reaction ( PCR ) . The fragments from nt 4767 to 5359 of the HEV p6 genome and the full-length neo gene were amplified from p6 construct and pcDNA3 . 1 ( Invitrogen ) by using the following primer pairs: AflII-p6-F ( 5'-CACCCTTAAGGGTTTCTGGAAGAAGCATTCTG-3' ) and M-p6/neo-R ( 5'-CACCCTTAAGGGTTTCTGGAAGAAGCATTCTG-3' ) , as well as M-p6/neo-F ( 5´-TGTTTGTTGCATCGCCCATTGGATCACCATGATTGAACAAGATGGATTGCA-3´ ) and HpaI-p6/neo-R ( 5´-CACCGTTAAC TCAGAAGAACTCGTCAAGAAGGCGAT-3´ ) . The resulting PCR fragments were joined and cloned into the intermediate plasmid pMD18T-p6/3´ which was generated by cloning the fragment of HEV kernow-C1/p6 from nt 4767 to the 3′ end into pMD18-T Simple vector ( Clontech ) , yielding pMD18T-p6/3´-neo . The fragment digested from this construct with AflII and HindIII was subsequently cloned into the parental p6 backbone , yielding plasmid p6/neo . Reporter plasmids IFN-β-luc ( firefly luciferase under the human IFN-β promoter ) , TK-RLuc ( Renilla luciferase under the human thymidine kinase promoter ) , ISRE-luc ( firefly luciferase under the human IFN-stimulated responsive element promoter ) were kindly provided by Stanley Lemon ( University of North Carolina , Chapel Hill ) . High molecular weight ( HMW ) poly IC was obtained from InvivoGen and reconstituted in PBS at 5 mg/ml . HepG2 cells were transfected with poly IC using DMRIE-C or Lipofectamine 3000 reagent ( Invitrogen ) for 6 h unless otherwise indicated . HCV 3'-UTR RNA was kindly provided by Takeshi Saito ( University of Southern California ) . HepG2 cells were transfected with 3 . 6 μg/ml of HCV 3'-UTR RNA using the Lipofectamine 3000 reagent . IFNL4-Halo and Halo-control plasmids as well as recombinant human IFN-λ4 were kindly provided by Ludmila Prokunina-Olsson ( National Institutes of Health , Bethesda , MD ) . Recombinant human IFN-α2a ( Sigma , H6041 ) was used to treat cells for 24 h at 100 ng/ml unless otherwise indicated . Human IL-29/IFN Lambda 1 ( 11725–1 ) was purchased from PBL and was used to treat cells for 24 h at 200 ng/ml unless otherwise indicated . Chimpanzee anti-HEV convalescent-phase serum ( ch1313 ) was kindly provided by Suzanne Emerson ( National Institutes of Health , Bethesda , MD ) . Rabbit anti-pORF2 antibody was a gift from XJ Meng ( Virginia Tech ) . Mouse monoclonal antibody K24F2 to HAV was a gift from Stanley Lemon ( University of North Carolina , Chapel Hill ) . Other antibodies were obtained from: RIG-I ( Enzo , ALX-210-932-C100 ) , MDA5 ( Enzo , ALX-210-935-C100 ) , MAVS ( Enzo , ALX-210-929-C100 ) , HA ( Sigma , H9658 ) , PMP70 ( Sigma , SAB4200181 ) , IFN-λ4 ( Millipore , MABF227 ) , STAT1 ( Cell Signaling , 14994S ) , pSTAT1 ( Ser727 ) ( Cell Signaling , 8826 and 9177S ) , pSTAT1 ( Tyr701 ) ( Cell Signaling , 9167S ) , IRF3 ( Cell Signaling , 11904S ) , IRF-7 ( Santa Cruz , sc-74472 ) , Sendai virus ( MBL , PD029 ) , ISG56 ( Thermo , PA3-848 ) , IFNLR1 ( R&D , AF5260-SP ) , IFNAR1 ( Santa Cruz , SC9391 ) , SOCS1 ( Cell Signaling , 3950T ) , SOCS2 ( Cell Signaling , 2779T ) , SOCS3 ( Cell Signaling , 2932T ) , USP18 ( Cell Signaling , 4813 ) , Lamin A/C ( Cell Signaling , 2032S ) , GAPDH ( Millipore , MAB374 ) , and β-actin ( Sigma , A2228 ) . Total RNA was extracted from HepG2 cells with the RNeasy Kit ( Qiagen ) in accordance with the manufacturer’s instructions . Real-time qRT-PCR was performed to quantify the HEV RNA with the iTaq Universal Probes One-Step kit ( Bio-Rad ) using the primer pair that specifically target ORF2 gene: forward primer: HEV-F ( 5´-GGTGGTTTCTGGGGTGAC-3´ ) , reverse primer: HEV-R ( 5’- AGGGGTTGGTTGGATGAA-3´ ) , and probe: HEV-P ( 5´-FAM-TGATTCTCAGCCCTTCGC–TAMRA-3´ ) or the primer pair that specifically target ORF1 gene: forward primer: p6/ORF1-F ( 5´- AAGACCTTCTGCGCTTTGTT-3´ ) , reverse primer: p6/ORF1-R ( 5’- TGACTCCTCATAAGCATCGC-3´ ) , and probe: p6/ORF1-P ( 5´-FAM- CCGTGGTTCCGTGCCATTGA–TAMRA-3´ ) . A synthetic full-length HEV Kernow C1/p6 RNA was used as standards . The endogenous IFN and ISG expression levels were measured by real-time RT-PCR using an iTaq Universal SYBR Green One-Step Kit ( Bio-Rad ) with specific primers: IFN-α , [56] [42] , IFN-λ1 , IFN-λ2/3 [42 , 57] , IFN-λ4 [58] , CXCL10 , GAPDH [42] , IFN-β [42] [59] , ISG56 [59] , ISG15 , RSAD2 [39] , USP18 [60] , SOCS1 , SOCS2 and SOCS-3 [60] . The mRNA levels of glyceraldehyde-3-phosphate dehydrogenase ( GAPDH ) were determined in the same samples for normalization . ΔΔCT was used to calculate the fold changes relative to the controls [61] . The detailed primer sequences are provided in the supplementary S1 Table . SDD-AGE was performed as described previously [27] . Briefly , crude mitochondria isolated from cells transfected with or without poly IC were resuspended in 1x sample buffer ( 0 . 5× TBE , 10% glycerol , 2% SDS , and 0 . 0025% bromophenol blue ) , loaded onto a 1 . 5% agarose gel ( Bio-Rad ) in the running buffer ( 1x TBE and 0 . 1% SDS ) , and subjected to electrophoresis for 2 h with a constant voltage of 100 V at 4°C . The proteins were transferred to a polyvinylidene difluoride ( PVDF ) membrane followed by immunoblotting with a rabbit antibody to MAVS . Cellular lysates were collected on ice in the lysis buffer ( 100 mM Tris-HCl ( pH 7 . 5 ) , 50 mM NaCl , 5 mM EDTA , and 1% Triton X-100 ) in the presence of protease inhibitor cocktail ( Roche ) . Samples were separated by SDS-PAGE , and transferred to a PVDF membrane ( Bio-rad ) . Membranes were incubated overnight at 4°C with primary antibody diluted in the Odyssey® Blocking Buffer ( LI-COR Biosciences ) . After washing with PBS-T for three times , the membranes were incubated for 1 h with the appropriate secondary antibodies . Protein bands were detected with an Odyssey Infrared Imaging System ( LI-COR Biosciences ) . For the detection of IFNLR1 , for which appropriated secondary antibodies were not available at LI-COR , membranes were wetted with SuperSignal West Pico Chemiluminescent Substrate ( Thermo ) and exposed to X-ray film ( RPI ) . HepG2 cells ( 4×104 ) were seeded onto eight-well Lab-Tek II CC2 slides ( Nunc ) one day before infection . IFA for detection of HEV-infected cells was performed as described [62] . To examine the subcellular localization of MAVS and IRF3 , cells were fixed with 4% paraformaldehyde for 20 minutes and permeabilized with 0 . 2% Triton X-100 for 15 minutes . Cells were stained with pre-absorbed ch1313 serum and a rabbit anti-MAVS , mouse anti-PMP70 , or rabbit anti-IRF3 for 1 h , and subsequently incubated with Alexa Fluor 488/594-conjugated goat-anti-rabbit IgG , Alexa Fluor 488/594-conjugated goat-anti-mouse IgG , or Alexa Fluor 488-conjugated goat-anti-human IgG ( Invitrogen ) for 1 h . After adding antifade-4 6-diamidino-2-phenylindole ( DAPI ) mounting solution ( Sigma ) , slides were viewed with a Zeiss LSM 510 confocal microscope with a 63x ( NA1 . 2 ) apochromatic water objective . Images were acquired using the ZEN 2009 software . Cells in 96-well plates were transfected with IFN-β-Luc ( 100 ng/well ) or ISRE-Luc ( 100 ng/well ) , together with TK-RLuc ( 10 ng/well ) by using the TransIT-X2 Dynamic Delivery System ( Mirus Bio ) . Transfected cells were then infected with 100 hemagglutinin units/ml of Sendai virus for 20 h or treated with recombinant IFN-α or IFN λ1 for 24 h at the indicated concentrations . Luminescence assays were performed in opaque 96-well plates with a Dual-Luciferase Reporter Assay System ( Promega ) according to the manufacturer’s instructions . Luminescence was measured using a FLUOStar Optima ( BMG Labtech ) plate reader . Each experiment was performed in triplicate wells . The full-length HCV construct containing a Gaussia luciferease ( Gluc ) gene inserted between core and p7 ( H77S3/Gluc2A ) was linearized with XbaI and subjected to in vitro transcription using the T7 In Vitro Transcription kit ( Ambion ) . The subgenomic HEV construct containing a Gluc gene , kindly provided by Suzanne Emerson ( NIH , Bethesda ) , was linearized with MluI and subjected to in vitro transcription using the mMachine mMessenger Transcription kit ( Ambion ) . Huh-7 cells were transfected with in vitro transcribed viral RNA using the TransIT mRNA transfection reagent ( Mirus Bio ) . One day after transfection , cells were split into 96-well plates and treated with IFN-α or IFN-λ at indicated concentrations . In the time-of-addition experiment , IFN-α was added at different days after transfection and replaced with fresh medium on the following days . Luciferase activity in the culture supernatant was measured on day 5 after transfection by a Gaussia luciferase kit ( Promega ) . Supernatant IFN-α , IFN-β or IFN-λ concentrations were measured by the human IFN-α ( 41100 ) , human IFN-β ( 41410 ) or IFN-λ ( 61840 ) ELISA kits ( PBL Interferon Resources , Piscasaway , NJ ) following manufacturer’s instructions . Values are shown as mean ± SD . Statistical significance between groups was determined with unpaired student’s t-test using GraphPad Prism 6 . 0 ( GraphPad , San Diego , CA ) .
|
HEV infection is a common cause for acute viral hepatitis worldwide . Approximately 20 millions of people are infected annually . In immunocompetent hosts the infection is self-limited and mostly asymptomatic , but the virus frequently persists when immunity is compromised leading to increased risk for cirrhosis . Currently there are no FDA-approved diagnostics or treatments for HEV . Understanding how HEV induces and manipulates host innate immune responses will help elucidate the mechanism ( s ) of HEV persistence and identify potential targets for therapy . Our results show that unlike other hepatotropic RNA viruses , HEV did not cleave MAVS and stimulated a sustained type III IFN response in persistently infected cells . Furthermore , the JAK/STAT pathway was persistently activated in HEV-replicating cells and responded poorly to exogenously added IFNs . This study uncovers a unique interplay between HEV and the host IFN pathway and provides insight into the mechanism of HEV persistence in patients .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"transfection",
"phosphorylation",
"medicine",
"and",
"health",
"sciences",
"luciferase",
"pathology",
"and",
"laboratory",
"medicine",
"hepacivirus",
"pathogens",
"enzymes",
"microbiology",
"enzymology",
"viruses",
"rna",
"viruses",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"proteins",
"medical",
"microbiology",
"oxidoreductases",
"microbial",
"pathogens",
"hepatitis",
"c",
"virus",
"hepatitis",
"viruses",
"viral",
"replication",
"molecular",
"biology",
"hepatitis",
"a",
"virus",
"biochemistry",
"flaviviruses",
"post-translational",
"modification",
"virology",
"viral",
"pathogens",
"interferons",
"biology",
"and",
"life",
"sciences",
"organisms",
"hepatitis",
"e",
"virus"
] |
2017
|
Hepatitis E virus persists in the presence of a type III interferon response
|
Genome mosaicism in temperate bacterial viruses ( bacteriophages ) is so great that it obscures their phylogeny at the genome level . However , the precise molecular processes underlying this mosaicism are unknown . Illegitimate recombination has been proposed , but homeologous recombination could also be at play . To test this , we have measured the efficiency of homeologous recombination between diverged oxa gene pairs inserted into λ . High yields of recombinants between 22% diverged genes have been obtained when the virus Red Gam pathway was active , and 100 fold less when the host Escherichia coli RecABCD pathway was active . The recombination editing proteins , MutS and UvrD , showed only marginal effects on λ recombination . Thus , escape from host editing contributes to the high proficiency of virus recombination . Moreover , our bioinformatics study suggests that homeologous recombination between similar lambdoid viruses has created part of their mosaicism . We therefore propose that the remarkable propensity of the λ-encoded Red and Gam proteins to recombine diverged DNA is effectively contributing to mosaicism , and more generally , that a correlation may exist between virus genome mosaicism and the presence of Red/Gam-like systems .
Bacterial viruses ( bacteriophages ) are the most abundant and diverse life form and exhibit high levels of evolvability and adaptability [1] . Moreover , bio-informatic studies suggest that they contribute substantially to bacterial genome evolution . For example , in γ Proteobacteria , most genes unique to a particular bacterial species or to a taxonomic group of species , are relatively short and AT-rich–two hallmarks of phage genes [2] . A particularity of temperate virus genome evolution is their extensive sequence mosaicism [3] due to exchange of DNA sequences , facilitated by the frequent encounter inside the same bacterial host , for example between an invasive and a resident virus [4] , [5] . However , most of the time , this mosaicism does not perturb the general gene order ( synteny ) , probably due to counterselection of suboptimal gene combinations [6] . Little is known about the precise molecular processes underlying this viral genome mosaicism . In the case of fully sequenced lambdoid viruses isolated from enterobacteria , genomes are on average 50% identical , except for DNA sequence patches showing more than 90% identity . The apparent absence of any particular signals at the borders of sequence-similar patches has led to the proposal that they have probably been acquired by illegitimate recombination [7] , [8] . In some cases however , exchange of sequence modules can be explained by homologous recombination involving flanking , short and conserved sequences shared by a subset of related viruses [9] . But it is also possible that some regions flanking the most similar shared sequences have undergone homeologous recombination , i . e . recombination between related but diverged DNA sequences [10] . The temperate virus λ has been a major model system in classical molecular genetics , including in the study of homologous recombination , which occurs at high rates in the λ genome ( reviewed by [11] ) . λ encodes its own homologous recombination genes redα , redβ ( the Red system ) and the gam gene , all belonging to the pL-operon . Redα is a double strand specific 5′ to 3′ exonuclease [12] and Redβ mediates strand annealing and exchange reactions starting from DNA extremeties [13] . The λGam protein inactivates the E . coli exonucleaseV ( RecBCD ) , thereby protecting the ends of its linear genome from degradation ( reviewed by [14] ) . Furthermore two other genes in the nin region participate in Red-mediated recombination: the orf gene product can replace the three proteins RecFOR involved in the E . coli RecF recombination pathway , and the rap gene codes for a Holliday junction resolvase [15]–[18] . Intracellular λ DNA is substrate for both virus-encoded and E . coli host recombination machineries , i . e . , λ+ recombines well in a recA host and so does λ red gam in the Rec+ host if it contains a Chi site to resist RecBCD degradation . In both cases , most events are non-reciprocal [19] , [20] . For both RecA-dependent and Red-dependent recombination , the required minimal homology is around 30 bp [21] , [22] . To test the efficiency of recombination between diverged sequences in viruses , we have investigated the capacity of λ to recombine pairs of homeologous oxa genes , starting from a λ strain initially described by the group of Kleckner [23] and later examined in greater detail by Ennis et al . [24] . In this system , recombination between inverted repeats framing the pL promoter leads to its inversion , which is accompanied by a phenotypic switch . We observed that the Escherichia coli RecABCD pathway recombined 22% diverged genes with a frequency of 10−6 per virus generation . Interestingly , the λ Red pathway showed a 100-fold higher efficiency . The recombination editing proteins MutS [25] , UvrD [26] and RecQ [27] had only marginal , if any , effect on λ recombination . Sequences of genes resulting from homeologous recombination revealed a broad spectrum of hybrids , and some differences between the products generated by the Rec and the Red proteins which may reflect intrinsic properties of the two recombination pathways . Therefore this λ system provides an efficient “gene machine” to create large libraries of hybrid sequences for biotechnology applications . In an attempt to assess the contribution of homeologous recombination between diverged sequences to phage mosaicism , we undertook a systematic bioinformatic analysis of mosaic flanking sequences , in three families of lambdoid phages . We found that half of them had at least one moderately diverged flanking region . This suggests that homeologous recombination within such flanking sequences may facilitate the reshuffling of phage genome modules and underlines the important role of virus recombination proteins in their genome evolution .
To test the efficiency of homeologous recombination in virus genomes , we studied recombination between pairs of sequence diverged genes inserted into the genome of λ . The experimental system is based on a genetic switch in λ resulting from homologous recombination between two identical inversely oriented IS10 sequences flanking the promoter pL [23] . Recombination between such inverted repeats is accompanied by the inversion of the pL promoter , leading to a phenotypic switch used to score recombinants ( Figure 1 ) . In the normal pL orientation the red and gam genes are transcribed such that λ grows on a recA mutant host , but not on a P2 lysogen . In the opposite orientation of pL , red and gam are not expressed , and λ grows on a P2 lysogen , but not on a recA strain . Starting from the construct with the inverted pL orientation , the two IS10 sequences were replaced with approximately 800 bp oxa genes having different levels of divergence [28] . A properly oriented Chi site was introduced rightward from the recombination cassette , to allow for stimulation of RecBCD-promoted recombination ( see Figure 1 , alignments of oxa genes are shown in Figure S1 ) . Recombinant frequencies at 0% , 4% , and 22% divergence were measured during single step growth on C600 ( P2 ) . In this background , homologous recombination occurs via the host RecABCD pathway only , because the λ-encoded pathway mediated by Red is repressed , due to the inverted pL promoter , and the Chi site protects λ rolling-circle forms from pure RecBCD degradation . Similarly , using λ with the native pL orientation , frequencies were measured during single step growth on a recA host . This time , recombination occurs via the phage-mediated Red pathway only , as the recA gene is mutated , and RecBCD is inactivated by Gam . We verified that in both backgrounds , λ replicated by a rolling circle mode ( see Methods ) . In the RecABCD pathway , maximum inversion frequency was 3×10−4 for identical sequences , whereas the minimum measured was 3×10−6 for 22% divergence ( Figure 2 ) . Unexpectedly , the recombinant frequency was as high between 4% diverged sequences as between identical sequences , and this effect persisted in the mutS background ( see Table 1 ) . A 4% divergence was reported to reduce by 1000 fold homologous recombination in the E . coli chromosome [29] . No obvious sequence stimulating recombination , such as a Chi site , is present in the oxa11 sequence used to construct the 4% diverged substrate . Rather than being a stimulation of recombination between 4% diverged sequences , it could be that some process inhibits recombination between the strictly identical sequences in λ . Recombination by the phage Red pathway was more efficient than recombination by the RecABCD pathway , especially for 22% divergence ( 1 , 5×10−4 versus 3 , 5×10−6 , Figure 2 ) . However , no recombinant was obtained by the Red pathway within 52% diverged sequences ( less than 10−8 ) . In order to measure recombinant frequencies in various genetic backgrounds , a protocol involving growth of bacteria on agar plates rather than single cycle liquid growth was chosen . Bacteria were infected with phages at a multiplicity of infection of 0 . 1 and grown to confluence , in the non-permissive host for growth of recombinants . This counteracted selection effects and revealed recombinants produced at the last generation . Recombinant frequencies in the wild type hosts were found to be consistent with the single step experiments ( compare Table 1 with Figure 2 ) . The methyl-directed mismatch repair ( MMR ) MutL and MutS proteins , and to a lesser extent MutH and UvrD , inhibit homeologous recombination by preventing DNA exchange between diverged repeated chromosomal sequences [30]–[32] and among entire genomes of related species [25] , [33] . In our system , mismatch repair deficiency ( mutS ) had an eight-fold stimulating effect on RecABCD promoted recombination for 4% diverged sequences ( Table 1 , 4% divergence set , lane ‘mutS’ , RecABCD pathway ) . This effect was less pronounced ( two-fold ) for 22% divergence . No stimulating effect of the mutS mutation was detected for recombination catalyzed by the phage Red system ( Table 1 , Red pathway , ‘mutS’ lanes ) . Thus in this virus assay , mismatch repair operates a modest control on the fidelity of the bacterial , RecABCD pathway , and not at all on the phage Red pathway . In addition to its role in MMR , UvrD helicase has a distinct activity in preventing homologous recombination , such that in a uvrD mutant , recombination between identical sequences is increased , generally by a factor of 10 [25] , [34]–[36] . UvrD appears to act directly as an “antirecombinase” by dismantling RecA nucleoprotein filaments [26] . In yeast , homeologous recombination is increased in a sgs1 mutant , a member of the RecQ helicase family [37] , [38] , and in E . coli , RecQ prevents illegitimate recombination [27] . We therefore tested E . coli uvrD and recQ mutants for a hyper-recombination phenotype both in RecA-mediated and Red-mediated events . The uvrD mutation had no effect on the recombination between identical sequences . However , similarly to mutS , it conferred a four-fold increase in recombinant frequency only at 4% divergence and only in RecA-dependent recombination . This suggests that it does not exert its distinct anti-recombinase activity on the λ substrates ( Table 1 , lanes ‘uvrD’ ) . RecQ did not prevent recombination in any of our substrates . Rather , recombination appeared slightly decreased in the recQ mutant , on 22% diverged sequences ( Table 1 , lanes ‘recQ’ ) . In our λ constructs , the set of chosen diverged sequences were pairs of oxa genes , encoding different beta-lactamases . Depending on recombination end-points , different gene combinations should form . A total of 152 phages scored as recombinants were used for sequencing the hybrid oxa copies . In all 304 oxa genes sequenced , a hybrid was found . This indicates that recombination indeed took place within the 800 bp of partial homology . Among these , a total of 136 new gene combinations were obtained . The presence of 32 and 176 sites of polymorphism for the oxa7-oxa11 and oxa7-oxa5 pairs , respectively , allowed us to map precisely strand exchanges and to class recombination events into two main categories: the “non-symmetrical” ones , for which the two joints are present in different intervals , and the “symmetrical” ones , for which the two joints occur in the same interval . Category “complex” includes more complex sequence patterns . Bacteriophage λ recombines essentially in a non-reciprocal mode , but in our recombination assay , only the events that terminate as reciprocal at the DNA level can yield viable recombinants . However , such ‘final’ reciprocity can be reached by two successive non-reciprocal events [39] , [40] , as shown on Figure 3 , left panel . The two events being independent , most products are expected to be of the non-symmetrical category . If , under some conditions , λ recombines in a reciprocal mode at the molecular level , by a simple crossing-over , as shown in Figure 3 , right panel , approximately half of the products , those derived from the RuvC-cut strand , are expected to be of the symmetrical category ( see Discussion ) . For the RecABCD promoted recombination between 4% diverged sequences , most events were non-symmetrical ( 81% ) , whereas only 17% were symmetrical ( Table 2 ) . Similarly , for the Red promoted events between 22% diverged sequences , a majority ( 81% ) of all events were non-symmetrical and only 17% were symmetrical . In contrast , for the RecABCD promoted recombination between 22% diverged sequences , 55% were symmetrical events , whereas 40% were non-symmetrical events . The difference in the proportions of non-symmetrical events promoted by RecABCD between 4% and 22% diverged sequences was statistically significant as determined by a Chi2 test ( p<0 . 0001 ) . Precise positions of the joints in each pair of oxa sequence for the 22% diverged DNA are given in Table S1 . Complex recombination products , involving ( formally ) more than two non-reciprocal events , were observed at similar but low frequencies under all conditions tested . To test whether the symmetrical events were processed by the RuvABC enzymes , that resolve Holliday junction in a symmetrical way , recombination frequencies were measured in a ruvABC mutant strain ( Table 1 , lanes ‘ruv’ ) . Efficiency of recombination between 22% diverged DNA via the RecABCD pathway , was decreased by a factor of 50 in the ruv mutant . In contrast , this mutation had no effect on the Red-mediated events for 22% diverged DNA , nor did it affect 4% diverged , RecABCD mediated recombination . Therefore , most of the recombination events observed between 22% diverged DNA in the RecABCD pathway are resolved by Ruv . Inspection of the location of all recombination joints relative to the length of shared identical sequence blocks revealed , for the 22% diverged sequences , that the joints can occur in regions of homology as small as two bp , but in most cases they were located in the longer identical blocks ( Figure 4 , A and C ) . Positions of joints along the oxa gene were inspected ( Figure 4 , B and D ) , and revealed two preferential blocks for the RecABCD pathway . The first hot spot ( nt 266–281 , 28% of joints ) is 16 nt long and contains two RuvC cutting sites ( one on each strand ) . It may correspond to a preferred resolution site . The second ( nt 661–677 , 18% of joints ) is 17 nt long , does not contain RuvC cutting site , but it is separated by only one mismatch from a 12 bp interval , so that the sum of the two segments is 30 nt , with a 60% GC content , which may help stabilizing the recombination intermediate . In contrast , the Red pathway did not exhibit such marked hot spots ( Figure 4D; the maximal occurrence of a joint was 10% ) . In both pathways , an overall deficit of joints in the first 260 bp of the gene was observed . It is most likely due to its higher divergence ( 30% in this segment , versus 18% for the remaining part of the gene , the curve reporting local % identity is drawn above the joints locations in Figure 4 B and D ) . In summary , the characteristics of homologous recombination promoted by λ suggest that it may constitute an ideal vector for in vivo gene shuffling . To explore the potential role of homeologous recombination in the evolution of virus genomes , we looked for hallmarks of such events by a comparative bioinformatics analysis of a variety of lambdoid phage genomes . Consider ancestral viruses A and B sharing overall 60% identity except for two 80% identical segments ( I , in Figure 5 ) . Homeologous recombination within the 80% identity segments would give rise to phage C consisting of the A sequence with a patch of B . If so , one would expect to find , in the virus C to B alignment , two regions of 80% identity , called hereafter “shoulders” , flanking a patch of 100% identity , called “hit” ( II , in Figure 5 ) . Subsequent divergence between ancestral phages B and C would finally lead to 90% identical hits , flanked by 70% shoulders , over a background of 50% identical sequences ( III , in Figure 5 ) . An analysis of ten lambdoid bacteriophages from enterobacteria was performed . It showed that of 83 hits sharing more than 90% identity between any two members of the family , six had two flanking shoulders and 35 a single shoulder . For the remaining 42 hits there was no detectable shoulder ( Table 3 , first series of data , see Table S2 for the complete data set ) . To determine the significance of the observed number of shoulders , an estimate of their number expected at random was made . Only seven should have been detected under the random hypothesis , which is six-fold lower than observed and highly significant ( p<0 . 001 ) . The average identity of shoulders was 64% ( +/−6 . 9% ) and their lengths were unevenly distributed , with the median of 200 bp ( Table 3 ) . To extend the analysis , hits and shoulders were looked for in 15 lambdoid phages from lactic acid bacteria and 20 lambdoids from Staphylococcus aureus ( Table 3 , last two series , see Tables S3 and S4 for complete data sets ) . Shoulders were found again in approximately 50% of all hits tested , with a frequency significantly greater than expected at random ( p<0 . 0001 in both cases ) .
The remarkable efficiency of the Red promoted recombination between 22% diverged sequences ( 10−4 ) , in contrast with RecABCD promoted events ( 10−6 ) , can be interpreted in two ways: ( i ) Redα and Redβ may be less sensitive to sequence divergence during heteroduplex DNA formation than RecABCD , and ( ii ) Redα and Redβ may escape host factors that prevent RecA-mediated recombination . In support of the first option , Redβ promotes efficient annealing and integration of single strand oligonucleotides containing mismatches , a technique known as recombineering [41] . Redβ is a single strand annealing protein , and has no ATPase activity [42] . It appears therefore as a simpler form of pairing protein as compared to RecA , which may explain its greater tolerance for sequence divergence . It could be also that the two-strand annealing process in Red-promoted events generates mismatched intermediates more readily than the three-strand RecA-promoted D-loops , due to the competition in the latter case with the displaced , and perfectly matched , strand . Interestingly , the related RecE-RecT recombination proteins of prophage rac ( in a recBC sbcA host background ) were used successfully to recombine 30% diverged recA sequences [43] . Furthermore , recombination between 32% diverged DNA during virus crosses was reported [44] . Finally , RecET promotes recombination between very short sequences ( 5–13 bp ) , in a process that may not be very different from the homeologous recombination reported here , albeit less efficient ( 10−8 ) [45] . Interestingly , in yeast , microhomology-mediated end-joining ( MMEJ ) depends on Rad52 [46] , a protein that has definitely some structural and functional similarities with Redβ and RecT [47] . In support of the other alternative , i . e . the escape from the host recombination editing systems , we have observed that the MutS protein , which prevents RecA-mediated homeologous recombination , is ineffective in the Red pathway . However , MutS can act on Red-mediated single strand annealing [41] excluding the possibility that MutS simply does not detect mismatches generated by Redβ . Actually , even the inhibition by MutS of RecA-mediated homeologous recombination in λ was low ( eight fold effect for 4% diverged sequences ) . In a different but comparable assay , where 4% diverged sequences are recombining in the E . coli chromosome , a much more profound , 60 fold inhibiting effect of MutS was reported [29] . It may be that some of the unknown gene products encoded within the λ genome ensure “immunity” against mismatch repair proteins , for instance by inhibiting MutS or MutL . Alternatively the high copy number of λ during the lytic cycle might titrate MutS and/or MutL . Neither of the helicases UvrD and RecQ showed inhibitory effects on homologous or homeologous recombination , in either the RecABCD or the λ Red pathway ( Table 1 ) . Whereas bacterial editing systems act to prevent promiscuous recombination events that cause genome instability , λ virus , and perhaps other lambdoids , appear to evade such editing thereby accelerating the rate of their genome evolution . Decades of work and careful analysis of the recombination products in λ crosses have led to the conclusion that in most cases , recombination is non-reciprocal at the molecular level , whether it occurs by the RecABCD pathway , or by the Red pathway [19] , [20] , [48] . This means in molecular terms that most often , recombination intermediates are not double Holliday junctions resolved by a break-join , RuvABC-dependent process , giving the classical crossover product ( as depicted Figure 3 , right panel ) , but are rather one of the three following cases: i ) half crossovers resolved by break-join , using either RuvABC or the λ encoded Rap protein [49] , ii ) D-loops dealt by a break-copy , replication-dependent process , also called BIR [49] , [50] , or iii ) single-strand annealing ( SSA ) intermediates . The two first situations are compatible with the sketch depicted Figure 3 , left panel . The last situation is mostly described for the Red pathway [51] , [52] . In the case of our present study , where the recombining sequences are present in λ in inverted orientation in the same molecule , two non-reciprocal events are needed to produce a viable inverted product ( Figure 3 , left ) . When the two recombining sequence are diverged , the position of the junction can give a hint of the underlying recombination process . Sequence analysis of pairs of recombinant genes revealed that , in most cases , the junction between the two partner sequences is not at the same position . Thus the hybrid sequence of the two recombined genes is called non-symmetrical , something expected for λ which recombines essentially non-reciprocally . However , half of all RecABCD-promoted recombination between 22% diverged genes showed symmetrical products , i . e . the junctions occurred in the same interval in both copies ( Table 2 ) . Because of the abundance of nucleotide polymorphism that define 137 possible intervals for strand exchange , it appears unlikely that two successive events occurred by chance in the same place ( probability of 1/137 = 0 . 7% ) , and suggests rather that in these cases , recombination occurred by a single crossing-over event . Holliday junction resolution is not expected to give more than 50% symmetrical products in our assay , because the progeny of the two strands of the recombination product is slightly different , due to the difference between the invasion step ( not necessarily strictly symmetrical ) and the resolution step ( symmetrical due to the RuvC action ) . The high proportion of symmetrical products , combined with the 100-fold lower efficiency of recombination for 22% as compared to 4% diverged sequences , may suggest the existence of two recombination mechanisms inherent to the RecABCD pathway: one being prominent at low sequence divergence ( non-reciprocal ) , and the other at high levels of divergence ( crossover ) . In line with this , we found that RecABCD-promoted recombination was independent of RuvABC at low divergence , but depended on RuvABC for the 22% diverged DNA . The prevalence of crossovers at high divergence might result from a combination of two favouring conditions: i ) the requirement of a single event , rather than two for the non-reciprocal recombination , ii ) the relative higher stability of highly mismatched heteroduplexes within Holliday junctions , as compared to the non-reciprocal recombination intermediates . This reasoning , in turn , underlines again the different activity of Red proteins , which produce mainly ( 81% of cases ) non-symmetrical recombinants between 22% diverged sequences . Still , the probability that the observed 17% symmetrical products were generated by chance during two successive non-reciprocal exchanges occurring in the same of the 137 possible intervals is very low . We propose that a fraction ( ∼2×17% = 34% ) of all Red promoted events in our experimental set up are indeed cross-overs . Biochemical studies of the RecT protein , which belongs to the same family as Redβ , have suggested that it might be able to generate three-strand intermediates [53] . In vivo , both reciprocal and non-reciprocal events are promoted by Red enzymes , and the balance is given by the length of homology available at the broken extremity: the longer the homology , the more non-reciprocal events are made [52] . Furthermore , the detailed analysis of joints produced via the Rec and Red pathways between 22% diverged DNA suggests again mechanistic differences which are compatible with the biochemical properties of the two systems: two hot spots are observed for the Rec products . RecA-promoted homologous recombination is expected to act more or less equally on all DNA sequences , but the absence of any single identical interval large enough to accommodate a MEPS ( minimal efficient pairing sequence , [22] ) may force the appearance of preferred regions where the three strand intermediate had a better stability . Indeed , a detailed study of the effect of mismatches on RecA-mediated joint molecule formation has shown that the position of mismatches relative to the identical regions can have different effects , depending on the stability of the heteroduplex progressively formed as exchange proceeds [54] . In contrast , no such hot spot is seen with Red products , which may well correspond to a ‘sandwich-like’ mode of action of single-strand annealing enzymes , rather than the progressive invasion process mediated by RecA . We demonstrate that , starting from pairs of similar genes , irrespective of their origin , phage genetic promiscuity can be exploited to generate large new gene families creating potentially interesting new biochemical entities . Even at 22% divergence , the Red recombination pathway can routinely create 105 to 106 recombinant genes ( and viruses ) per single Petri dish , 40% of which represent different new genes . The yield of recombinant genes is orders of magnitude higher than when the same genes were carried in E . coli plasmids [55] . It is also possible to lead this system through unlimited iterative cycles of inversion recombination , which should yield even more diverse gene products . As such , λ is therefore a convenient genetic vector for evolutionary biotechnology . Can we relate our experimental results on homeologous recombination in λ to the evolutionary history of lambdoid virus genomes ? Our bioinformatic analysis showed that among all detected blocks of highly similar sequences ( hits ) , about one half showed no flanking “shoulder” of moderate divergence , about 40% showed only one shoulder and the remaining hits were clearly framed by two shoulders . Even a single shoulder is compatible with an involvement of homeologous recombination . For example , a sequence block can be acquired by an homeologous recombination event ( shoulder ) at one junction , accompanied by an homologous event between identical sequences [9] or an illegitimate event at the second junction ( no shoulder , [45] ) . When shoulders were detected , their identity was in the range of 64% to 68% , and the hit sequences were on the average 94% identical . The 6% divergence of the hit sequence suggests that , at the time of recombination , the shoulders identity was about 70 to 74% ( Figure 5 ) . This is close to the 78% identity that was tested in our assay and found as substrate for homeologous recombination . Because never more than 50% of the detected hits were flanked by at least one recognizable shoulder , illegitimate recombination and homeologous recombination appear to contribute to phage genomic mosaicism to a similar extent . If virus mosaicism is really related to the presence of Redβ-like recombination enzymes , it should be possible to verify that all virus genomes exhibiting mosaicism encode such a function . Among the ten lambdoids from enterobacteria that were analysed here , only two encoded a Redβ ortholog . However at least one other family of virus recombinases , of which Erf is the best studied member , has been described [47] . It may act similarly to Redβ , as it forms similar ring structures [56] , and cross-complementation has been observed [57] . Four among the ten lambdoids from enterobacteria encode an Erf ortholog , and eight among the fifteen lambdoids from lactic acid bacteria as well . Whether this type of recombinase promotes efficient homeologous recombination remains to be tested . None of the S . aureus virus analysed encode either a Redβ or Erf ortholog . It may be that one or more virus recombinase families remain unknown at present . Interestingly , viruses belonging to the family of T4 , composed exclusively of virulent members , appear not to have a mosaic structure , but to consist rather , like bacteria , in a common backbone genome , interrupted by a few large variable regions [58] . These viruses do not encode proteins of the Redβ nor Erf family , but a UvsX protein which has ATPase activity like RecA . Also , among dairy viruses , a different genomic structure for virulent and temperate viruses has been reported [59] . This scattered evidence is therefore compatible with the possibility that the mosaicism of lambdoid genomes is connected with the particular type of homologous recombination enzymes they encode , which may be fit to provide , in a short time , large gene repertoires and therefore bring about an extraordinary evolvability .
All Escherichia coli and λ strains used in this study are described in Table 4 . Lysogenization was performed as described by Cromie and colleagues [60] . Primary phage stocks , which all contained the thermosensitive cI857ts mutation , were obtained by shifting cultures of lysogenic bacteria at a OD600 0 . 4 for 10 minutes to 45°C , followed by further incubation ( up to 4 hrs ) at 37°C . These primary stocks usually contained 1010 plaque forming units per ml . The λ 366 described by N . Kleckner [23] contained a copy of IS10 inserted into the ea10 gene ( our unpublished observation ) , and a copy of Tn10 inserted into the rexA gene . A derivative obtained by G . Smith , λ 1390 , in which the pL promoter is inverted , was used as the starting material for our constructions [24] . Our goal was to replace the IS10 and Tn10 copies by a set of related oxa genes which diverge by 4% ( between oxa7 and oxa11 ) , 22% ( between oxa7 and oxa5 ) , or 52% ( between oxa7 and oxa1 , [55] ) . A fragment of the λ1390 genome was cloned onto plasmid pACYC184 , and successive cloning steps allowed to substitute part of the IS10 with oxa7 , and the totality of Tn10 with three elements: i ) either oxa7 , oxa11 , or oxa5 , inverted relative to the copy of oxa7 inserted into ea10 , ii ) the chloramphenicol-resistance ( cmR ) gene of pACYC184 , and iii ) a Chi site [24] . Integration of these cassettes into λcI857ts was done using the protocole of Datsenko and Wanner [61] , with strain JTM146 as a recipient . This permitted to obtain λNec1 , 2 , 3 , in which the pL promoter is inverted . To get the inversed orientation of pL , recombinants obtained starting from λ Nec1 , 2 , and 3 constructions were selected , and a clone in which the recombinant product was symmetrical was kept . The construction to test 52% diverged sequences in the Red pathway ( λNec8 ) was done by replacing the rightward oxa7-CmR cassette of λNec4 by an oxa1-phleoR cassette . Construction details are available in Text S1 and Figure S3 . The construction to test 52% diverged sequences in the Red pathway ( λNec8 ) was done by replacing the rightward oxa7-CmR cassette of λNec 4 by an oxa1-PhleoR cassette . To do this , a plasmid containing the cI to N region of λ , in the native orientation of the pL promoter , interrupted by the oxa5-CmR cassette ( pMAP189 ) was used to substitute a different cassette , made of the oxa1 gene flanking a PhleoR gene , giving plasmid pMAP195 . The 3 . 2 kb AvaII-SapI fragment of pMAP195 was then gel purified and used to transform a C600 derivative lysogenic for λ Nec4 and containing pKD46 , and selecting phleomycin resistant transformants ( 1 µg/ml ) , in which the rightward oxa7-CmR cassette had been replaced by the oxa1-phleoR cassette . Single plaques of recombinants were purified by streaking , purified plaques were toothpicked and resuspended in SM . These crude phage particles were directly used for PCR amplification with oligonucleotides flanking the oxa gene to be sequenced . When the same pairs of oligonucleotides were used on the starting , non-inverted phages , no PCR product was obtained , ensuring that the recombinants analysed were not generated during the PCR itself .
|
Temperate bacterial viruses alternate between a dormant state , during which viral DNA remains integrated in the host genome , and a lytic state of phage multiplication . Temperate viruses have a characteristic genome organisation known as ‘mosaic’ – they contain ‘foreign’ segments that originate from related viruses . In pairwise alignments between a given virus and its relatives , the overall nucleotide sequence identity is around 50% . In contrast , the mosaic segments are 90% to 100% identical . How mosaics are generated is largely unknown , but it is likely that related viruses meet in the same bacterium and undergo random recombination , with emergence of the most robust recombinatory viruses . The prevalent hypothesis is that mosaics are formed by illegitimate recombination . We propose and demonstrate that an alternative driving mechanism , homologous recombination , is used for mosaic formation between similar but diverged viral sequences . Using the well known Escherichia coli λ virus as a paradigm , we show that such homeologous recombination is remarkably efficient . This finding has important implications in the field of virus genome evolution , as it may explain the high plasticity of viral genomes . It is also applicable to the field of biotechnology , and reveals viruses to be promising vectors for shuffling genes in vivo .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"molecular",
"biology/bioinformatics",
"molecular",
"biology/recombination",
"microbiology/microbial",
"evolution",
"and",
"genomics"
] |
2008
|
The λ Red Proteins Promote Efficient Recombination between Diverged Sequences: Implications for Bacteriophage Genome Mosaicism
|
During active somatosensation , neural signals expected from movement of the sensors are suppressed in the cortex , whereas information related to touch is enhanced . This tactile suppression underlies low-noise encoding of relevant tactile features and the brain’s ability to make fine tactile discriminations . Layer ( L ) 4 excitatory neurons in the barrel cortex , the major target of the somatosensory thalamus ( VPM ) , respond to touch , but have low spike rates and low sensitivity to the movement of whiskers . Most neurons in VPM respond to touch and also show an increase in spike rate with whisker movement . Therefore , signals related to self-movement are suppressed in L4 . Fast-spiking ( FS ) interneurons in L4 show similar dynamics to VPM neurons . Stimulation of halorhodopsin in FS interneurons causes a reduction in FS neuron activity and an increase in L4 excitatory neuron activity . This decrease of activity of L4 FS neurons contradicts the "paradoxical effect" predicted in networks stabilized by inhibition and in strongly-coupled networks . To explain these observations , we constructed a model of the L4 circuit , with connectivity constrained by in vitro measurements . The model explores the various synaptic conductance strengths for which L4 FS neurons actively suppress baseline and movement-related activity in layer 4 excitatory neurons . Feedforward inhibition , in concert with recurrent intracortical circuitry , produces tactile suppression . Synaptic delays in feedforward inhibition allow transmission of temporally brief volleys of activity associated with touch . Our model provides a mechanistic explanation of a behavior-related computation implemented by the thalamocortical circuit .
Thalamocortical circuits represent model systems for multi-area computations [1] . Sensory information enters the cortex through the thalamus . Transformations in thalamocortical circuits have mostly been studied in anesthetized animals with passive sensory stimuli [2–8] or with artificial whisking [9] . In the somatosensory system these studies have revealed subtle differences in receptive field structure across neurons in the thalamocortical circuit [2–4] . However , active sensation in awake animals involves dynamic interactions with the world , such as saccades [10] , palpation with the digits of the hand [11] , or movements of the whiskers on the face of rodents [12–14] . During active sensation , movement of the sensors produces ‘reafferent’ signals , whereas interactions with the world generate ‘exafferent’ signals . During haptic exploration , movement activates peripheral sensors to produce reafference and touch generates exafference [15–21] . The brain needs to parse these different signals for perception [13] . During active sensation , movement attenuates the transmission of certain sensory signals to the cortex [22–24] . Tactile suppression is thought to enhance perception of salient events that cannot be predicted based on movement . Tactile suppression is an example of adaptive filtering [25 , 26] , which is critical for low-noise encoding of relevant sensory stimuli . Here we identify the mechanisms of adaptive filtering in the thalamocortical circuit of the mouse whisker system . Whisker touch and movement are transduced by mechanosensory afferents in the whisker follicle . Information then flows through the trigeminal ganglion , to the brainstem , thalamus ( barreloids in the ventral posterior medial thalamic nucleus , VPM ) and terminates in the primary somatosensory cortex ( vS1 ) . The main target of VPM axons is the Layer 4 ( L4 ) barrels in vS1 . The microcircuit of each L4 barrel is mostly contained within the barrel , and the connections between specific cell types within the barrel have been mapped: L4 excitatory neurons and L4 fast-spiking , parvalbumin ( PV ) -expressing GABAergic interneurons ( FS ) are connected within type and across types [27 , 28] . Apart from neuromodulation , the only known long-range input to L4 originates in VPM [29–31] . VPM excites all L4 neuron types , and L4 FS neurons inhibit the excitatory neurons to implement feedforward inhibition [32–35] . Cell type-specific recordings from VPM , L4 excitatory [36] and L4 FS neurons [37] uncovered a fundamental computation performed by L4 barrels . Neurons in VPM respond to touch , but they also increase their activity during whisker movement ( ‘whisking’ ) [17–21] . L4 FS neurons have nearly identical dynamics as VPM neurons . In contrast , L4 excitatory neuron spikes are strongly coupled to touch , but respond only weakly to whisking [36] . L4 microcircuits therefore transmit touch signals and suppress reafferent signals generated by whisking . These observations indicate that the thalamocortical circuit accentuates salient tactile information by suppressing signals related to self-movement . Multiple observations regarding the L4 circuit remain to be explained [37] . First , the circuit suppresses self-movement signals but transmit touch signals . From a theoretical perspective , this selective filtering violates the linear response expected from theories of strongly coupled cortical networks [38 , 39] . Second , whereas the baseline spike rates of VPM and L4 FS neurons are substantial during baseline conditions and more than double during whisking , the spike rate of L4 excitatory neurons is very low at baseline and does not increase during whisking . Third , when an inhibitory opsin ( halorhodopsin ) is photostimulated in L4 FS neurons , the spike rates of FS neurons decrease and the spike rates of L4 excitatory neurons increase [37] . Although this result is naively expected , it contradicts the "paradoxical effect" predicted from models of neural circuits that are stabilized by inhibition and from models of strongly-coupled networks [40–42] . According to these models , both inhibitory and excitatory neurons should respond with increased spike rates at modest levels of inactivation of inhibitory neurons [41] . To understand the mechanisms underlying these experimental observations we constructed and analyzed a conductance-based model of the L4 circuit . The synaptic circuits of L4 of the barrel cortex have been studied in great detail in brain slices [27–29 , 43–46] . These measurements allowed us to constrain the numbers of neurons , neuronal properties , the patterns of connectivity , and the average synaptic strengths in the model . We set parameter values near values extracted from these measurements . This "reference set of parameters" was adjusted over a restricted range , such that the model circuit displayed dynamics similar to the actual L4 circuit . We then varied parameters to explore their roles in controlling system dynamics . We studied how the model network responds to thalamic input at baseline , during whisker movement and during touch . We used the model to disentangle the roles of feedforward synaptic connections from the thalamus and recurrent intracortical connections in shaping L4 dynamics . The model revealed an important role for synaptic delays , fast synaptic kinetics , and inhibitory and excitatory conductance strengths in shaping the L4 responses during behavior .
We first summarize recordings made in VPM and in L4 from excitatory and FS GABAergic interneurons in mice performing an object location discrimination behavior ( Fig 1a and 1b ) . Head-restrained mice had to localize a vertical pole with their whiskers for a water reward [47 , 48] . Whisker movement and touch were tracked on millisecond time scales with high-speed videography [49 , 50] ( Fig 1a ) . Recordings were made with extracellular silicon probe recordings in VPM [37] , and loose-seal cell-attached and whole cell recordings in L4 [36 , 37 , 51] . VPM neurons have significant spike rates ( mean , 5 . 1 Hz ) , even in the absence of whisker movement and touch ( Table 1; Fig 1c and 1d; see Figs 2 and 3 for representative examples ) [37] . VPM neurons were also highly sensitive to whisking ( Fig 2b and 2c ) [17 , 18 , 21] . Spike rates increased after whisking onset ( average , 3-fold; Fig 1c and 1d ) . Given that mice whisk at approximately 15 Hz under these conditions [36] , the spike rates during whisking correspond to approximately one spike per whisking cycle , on average . The modulation depths with whisking amplitude and phase ( Supplementary Figure 2 in [37] ) were similar to published studies in rats [18] . VPM neurons also responded to active touch with a brief increase in spike rate ( 0 . 6 spikes per touch ) ( Fig 1e ) . The peri-stimulus time histogram ( PSTH ) aligned to touch onset shows a sharp peak in spike rate , with short latency after touch ( 3 . 1±0 . 6 ms ) and brief duration ( 2 . 9±1 . 7 ms ) ( Fig 2i , 2n and 2s ) . Exafferent touch signals and reafferent whisking signals were multiplexed in individual VPM neurons ( Fig 2 ) . The modulation of L4 FS neurons to touch and whisker movement were similar to VPM neurons ( Figs 1c–1e , 2 and 3 ) . L4 FS neurons increased their spike rate after onset of whisking ( average , 3-fold ) . L4 FS neurons responded reliably to touch . The response onset had slightly longer latency ( 5–15 ms ) and longer duration ( 2–15 ms ) than for VPM neurons . The increased latency is expected from the propagation time delay between VPM and L4 ( 2 ms ) [37 , 53] . The responses of L4 excitatory neurons differed profoundly from those of VPM and L4 FS neurons [36] ( Figs 1c–1e and 3 ) . The baseline spike rate of L4 excitatory neurons was much lower and did not increase significantly after onset of whisking . The response after touch onset occurred with longer latency and longer duration than for the VPM neurons , and similar to L4 fast-spiking neurons ( Fig 1e ) . The transformation performed by L4 circuits can be summarized by three main findings ( Fig 1c–1e; Table 1 ) . First , baseline spike rates are high in VPM and L4 FS neurons , and low in L4 excitatory neurons . Second , VPM and L4 FS neurons elevate their spike rates further during periods of whisker movement , whereas L4 excitatory neuron spike rates remain low . The spike rates of VPM and L4 FS neurons during periods of whisking are more than one order of magnitude larger than those of L4 excitatory neurons . Third , all three neuron types respond to touch reliably ( Table 1; Fig 1e ) . The touch responses of the three neuronal populations are the integrals under the curve in Fig 1e . The average touch response of L4 excitatory neurons ( 0 . 3 spikes/touch ) is about half of the average touch response of VPM neurons , whereas the spike rates of L4 excitatory neurons are lower than those of VPM neurons by a factor of 20–30 . The VPM response is more transient than the L4E response . The L4 circuit therefore performs a behaviorally relevant computation by propagating information related to external stimuli and suppressing responses to whisker movement , a predictable stimulus . The dynamics of FS neurons during whisker movement suggest that these neurons suppress whisker movement-related activity in L4 excitatory cells . Consistent with this view , optogenetic reduction of activity in a subset of L4 FS neurons expressing the light-gated , inhibitory chloride pump eNphHR3 . 0 ( L4I-Hr+ ) unmasks movement-related activity in L4 excitatory neurons [37] ( Fig 4a and 4b ) . Photostimulation of L4I-Hr+ neurons decreases the activity in L4 FS neurons on average and increases the activity in L4 excitatory neurons . The suppression of L4I-Hr+ neurons causes an increase in response in L4 excitatory neurons during whisker movements ( Fig 4c ) and an overall increase in the number of spikes per touch ( Fig 4d ) . This result implies that suppression of whisking response in L4 excitatory neurons involves inhibition from FS neurons . It remains unclear , however , why touch responses in L4 excitatory neurons are not suppressed as well . Membrane potential measurements with whole-cell recordings provide additional clues about mechanisms [37] . L4 excitatory neurons depolarize substantially ( 6 mV ) and briefly after touch . Touch-related inhibitory input ( from FS neurons ) to L4 excitatory neurons is delayed by approximately 0 . 5 ms with respect to excitatory input ( Fig 1e ) . It has been proposed that this short ‘window of opportunity’ [2 , 35] allows L4 excitatory neurons to spike after touch , typically with one spike [36] , before inhibition suppresses L4 excitatory activity . During whisker movement , excitation is matched by inhibitory input , keeping the L4 excitatory neuron membrane potential well below spike threshold . Suppression of self-movement signals is therefore implemented by inhibition within L4 . Yu et al . [37] also showed that activating halorhodopsin in L4I-Hr+ neurons during whisker movement suppresses activity in these neurons . This result poses a new question . The major cellular effect of activation of halorhodopsin is hyperpolarization [54 , 55] . Theoretical investigations of cortical circuits , consisting of recurrently connected excitatory neurons stabilized by inhibitory neurons , have investigated the effects of hyperpolarization of inhibitory neurons [41] . Over a large range of conditions , hyperpolarization of inhibitory neurons leads to increased spike rates in both excitatory and also inhibitory neurons . This "paradoxical effect" is common to multiple network regimes , including networks with strong synaptic conductances that fire in an asynchronous manner [40] and networks with moderate synaptic conductances , if the excitatory-to-excitatory synaptic coupling gEE is sufficiently strong [41] . Does the L4 circuit lie outside of the modeled parameter regimes , or can other factors explain the lack of paradoxical effect ? To explain the dynamical response of the L4 circuits to thalamic input , and to understand the roles of feedforward and recurrent connections , we constructed and analyzed a detailed computational model of the L4 circuit . L4 excitatory and L4 FS neurons make connections within type and across types with high connection probability ( Fig 5a; Table 2 ) . The thalamocortical circuit of the rodent whisker system is one of the most extensively studied mammalian circuits . In vitro and in vivo studies have measured many fundamental parameters that are required to construct a realistic computational model of the thalamocortical circuit . Our model circuit consists of 1600 L4 excitatory ( E ) neurons and 150 L4 FS ( I ) neurons [28] receiving input from 200 VPM ( T ) neurons ( Fig 5a ) . When referring to modeling results we denote VPM neurons as T , L4 excitatory neurons as L4E , and GABAergic interneurons as L4I ( Fig 5b and 5c ) . Since we know little about non-FS GABAergic interneurons during behavior we model only one inhibitory neuronal population . Cortical neurons were simulated using a conductance-based model with a single compartment per neuron [58 , 59] . Synaptic conductances are denoted as gαβ , where β is the presynaptic population and α is the postsynaptic population ( Fig 5a ) . Model parameters , including numbers of neurons , unitary synaptic conductance , connection probability , resting potential , and membrane time constants are based on neurophysiological measurements in brain slices [27 , 28 , 56 , 60] with small adjustments ( see Methods ) . VPM provides the only external input to the model circuit , with connectivity estimates based on in vivo and brain slice measurements [34 , 53] . Spike trains of T neurons were modeled as independent inhomogeneous Poisson processes with a generating function FT representing thalamic activity during quiescence , whisking or whisking and touch ( Fig 5d; see Methods ) . FT ( t ) =AT[1+BTsin ( 2πtτw+ϕ ) ]+CTτcΘ ( t−nτw−tc ) Θ ( nτw+tc+τc−t ) ( 1 ) where AT is the spike rate averaged over a whisker movement cycle , BT is the modulation depth , and CT is the number of spikes per touch , τw is the whisking period , tc is the time of touch onset within a whisking cycle , n is the cycle number , and Θ is the Heaviside function . During whisker movements AT increases above a baseline , with sinusoidal modulation phase-locked to a single preferred phase ϕ . Touch is represented by adding a rectangular function at touch onset t = tc ( with respect to the whisking cycle ) , stretched over τc = 3 ms with an integral of CT = 0 . 6 spikes per touch ( Eq 1 ) . Correlations between thalamic neurons beyond those generated by Eq 1 have not been measured and were thus not modeled . The population-average spike rate of neurons within a population over whisking cycles is denoted by να ( α = T , E , I ) , and the population-average spiking response to touch is denoted by Rα . Note that without touch , νT = AT . A ‘reference parameter set’ is described in Methods and used unless otherwise stated . We start by simulating the model for whisking without touch ( thalamic spikes per touch , CT = 0 ) . L4E spike rates were low ( νE < 1 Hz ) compared to those of L4I neurons ( Fig 6a and 6b ) . The average rates νE and νI depend on AT ( the population- and time-average thalamic spike rate during whisking only ) but are almost independent of the modulation depth BT ( Eq 1 ) . Except near threshold , νI was proportional to AT , in consistent with theoretical results for large and sparse neuronal networks with strong synapses , in which strong excitation is compensated by inhibition ( balanced networks ) [38] . The L4E spike rate νE increased linearly with AT , but with a much smaller slope compared to νI . The linearity of the νI-AT curve fit empirical observations ( Table 1 , Fig 1c; spike rates of both VPM and L4 FS neurons more than double in the transition from non-whisking to whisking ) . The average spike rate of L4 excitatory neurons , νE , is low [37] . During transition from non-whisking to whisking , the average spike rate of L4 excitatory neurons barely changes ( from 0 . 4 Hz to 0 . 6 Hz; Table 1 ) . The spike rates of L4E neurons in our model show similar dynamics . L4E neurons fire at most one spike each after touch , whereas L4I neurons fire more than one ( up to two ) spikes per touch ( Fig 6c and 6d ) . The synaptic delay , τdelayEI , is a critical factor in determining the strength of the touch response ( Fig 7a ) . For τdelayEI=0 , the average normalized response of L4E neurons after touch , RE , is small ( RE = 0 . 01 spikes/touch ) , and L4I neurons fire RI = 0 . 64 spikes/touch . Strong inhibition overwhelms excitation before L4E neurons have a chance to spike . The delay between the appearance of excitation and feed-forward inhibition in the L4E neuron has been termed ‘window of opportunity’ [2 , 35 , 61–63] . This ‘window of opportunity’ increases with τdelayEI , but is not equal to it because it is affected by the durations of thalamocortical synaptic process and the time needed for the L4I to fire in response to the brief and strong thalamic input . For a more realistic value , τdelayEI=0 . 85ms , L4E and L4I fire 0 . 34±0 . 24 and 1 . 3±0 . 07 spikes/touch respectively , similar to experimental measurements ( Table 1 ) . The touch responses for L4E and L4I ( Fig 6d ) are narrower than those seen in real data ( Fig 1e ) . The detailed shapes of the modeled responses depend on network parameters ( e . g . , see Fig 8d–8i below ) . In addition , heterogeneity among neurons , which was not modeled here , is expected to broaden the responses . The response of the network to touch depends on the time-course of synaptic conductances . Without touch signals , L4E neurons are inhibited by L4I neurons and spike at low rates ( Fig 6a and 6b ) . Touch produces strong , brief and synchronous thalamic excitation ( Fig 1e ) , which depolarizes L4E neurons and enables them to fire before inhibition terminates the response . This mechanism demands that excitatory synaptic conductance changes at thalamocortical synapses are brief . Touch responses of L4E and L4I neurons decrease with tAMPA , the decay time of AMPA-mediated EPSCs ( Fig 7b ) . Substantial touch responses in L4E neurons require tAMPA < 2–3 ms , consistent with the brief excitatory conductances measured at thalamocortical synapses [60] . Touch responses of L4E and L4I neurons increased with the strength of the thalamic touch signal ( Fig 7c ) , consistent with graded responses to touch strength measured in L4E neurons [36] . The response saturates at one and two spikes/touch for L4E and L4I respectively; this is in part due to the intrinsic properties of these neurons , which preclude them from firing more spikes in response to brief thalamic input . We have shown that a L4 network with parameters similar to experimentally determined values can replicate major experimental findings . L4E excitatory neurons exhibit low baseline activity , low activity in response whisker movement , and significant response to touch . However , during behavior the responses of L4 excitatory neurons is not all-or-none: L4 excitatory neurons are tuned to multiple sensory features including touch direction , intercontact interval , strength of touch , and likely other factors [36] . Our simulation results indicate that L4E and L4I neurons respond to touch in a graded manner and robustly respond to touch at different whisking amplitudes . Specifically , our simulations show that RE and RI decrease with AT ( Fig 7d ) for all parameter values consistent with the data . Mechanistically , increasing thalamic input causes larger inhibition in the cortical circuit that decreases touch response . Future experiments could test this model prediction . The model allows us to explore how specific synaptic connections within L4 contribute to fine-tuning L4 function . These connections currently cannot be specifically manipulated , making them difficult to evaluate experimentally . Recurrent excitation ( gEE ) can only be tuned over a limited range before runaway excitation is triggered ( for about gEE~0 . 4 mS/cm2 ) , even during baseline or whisking ( Fig 8a and 8b ) . The conductance gEE does not modify the slope of the νE-AT curve ( Fig 8b ) , but shifts this curve upward . gEE amplifies touch responses RE and RI ( Fig 8c ) [64] , mostly by increasing the duration of touch responses ( Fig 8d–8i ) . The dynamics of L4E neurons are shaped in subtle ways by L4 inhibition . Fig 9 shows how parameters involving inhibition ( gII , gEI , gIE ) affect circuit responses , while holding the other parameters at their reference values . Intracortical excitation of inhibition ( gIE ) is necessary to prevent runaway excitation , and thus keeps the spike rates of neurons in the network moderate ( Fig 9a and 9b ) . Reducing gIE shifts the νE-AT curve during whisking upward , without modifying its slope . For touch responses , gIE shifts the touch-response RE-CT curve to the right while decreasing the slope of the linear section of this sigmoid ( Fig 9c ) . Inhibition to L4E neurons ( gEI ) is also required to prevent runaway excitation ( Fig 9d and 9e ) . For low values of gEI ( but above values for runaway excitation ) and large values of gII ( Fig 9g and 9h ) , νE scales linearly with thalamic input AT . This linear scaling is known from balanced networks with strong synaptic coupling [38 , 40] . The situation is different for sufficiently strong inhibition ( relatively high gEI , moderate gII ) , where the response of L4E neurons to slowly-varying thalamic input during whisking , νE , is independent , and even decreasing , with AT ( Fig 9e and 9h ) . For small enough gII , L4E are quiescent except of near firing threshold . In addition , low gET and large gIT values are critical to keep νE low for all whisker movement amplitudes ( Fig 10a , 10b , 10d and 10e ) . Furthermore , these low values νE are nearly independent of AT . This regime is consistent with experimental data , where L4 FS neurons increase their spike rate with whisker movement , whereas L4 excitatory neuron spike rates remain low ( Fig 1d , Table 1 ) . Increasing values of gEI and gII shifts the touch-response RE-CT curve to the right and to the left respectively ( Fig 9f and 9i ) . In addition , gEI , but not gII , decreases the slope of the linear section of this sigmoid ( Fig 9f ) . Similarly , gET and gIT shift the RE-CT curve to the left and to the right respectively ( Fig 10c and 10f ) . Yu et al . ( [37] , Fig 8 , supplementary Fig . 13 ) expressed halorhodopsin in FS neurons in L4 and explored how the responses of excitatory and FS neurons in L4 under baseline conditions , whisking and touch vary under halorhodopsin activation . L4 excitatory neurons increase their spike rates and L4 FS neurons decrease their spike rates , both during baseline and whisking conditions . Similarly , L4 excitatory neurons increase their spiking responses to touch and L4 FS neurons decrease their responses . The halorhodopsin expression levels in FS neurons likely varied widely across individual neurons [65] . We therefore divided the L4I neurons in the model into hr expressing neurons and non-expressing neurons ( Hr+ and Hr- respectively ) . The fraction of Hr+ neurons among L4I neurons is denoted by fhalo . Manipulation of the components of neural networks can cause complex and counterintuitive change in network dynamics ( ‘paradoxical’ response; [40–42 , 66]: injecting negative current to all inhibitory neurons in a network , that includes spiking excitatory neurons , increases the average spike rates in inhibitory neurons ) . If the synaptic conductance strengths are moderate and the excitatory-to-excitatory synaptic conductance is above a certain level , injecting negative current to the inhibitory neurons causes the spike rates of excitatory neurons νE to increase , and as a result the network dynamics causes the spike rates of inhibitory neurons νI to increase as well [41 , 42 , 66] . Alternatively , if synaptic conductances are strong and the excitatory population is active , the condition that the activity of excitatory neurons should be moderate ( non-zero and not epileptic ) causes νI to increase under negative current injection to inhibitory interneurons [40] . The major effect of halorodopsin is to hyperpolarize FS neurons ( see Methods ) . Therefore , if all inhibitory neurons in the model are assumed to express halorodopsin equally ( fhalo = 1 ) , halorhodopsin activation will cause both L4E and L4I neurons to increase their spike rates . We simulated the response of L4E and L4I neurons to whisking for fhalo = 1 without and with halorhodopsin activation , and replicated the 'paradoxical effect': the average responses of L4E and L4I neurons to whisking , νE and νI , increase with simulated light activation . The simulation result for L4I neurons is in contrast to experimental observations showing that most L4 FS neurons increase their spike rates . This discrepancy is resolved if we assume fhalo = 0 . 5 . For this value , spike rates during whisking increase , on average , for L4E neurons ( Fig 11a ) , decrease for almost all L4I-Hr+ neurons ( Fig 11b ) , and increase for inhibitory neurons that do not express halorhodopsin ( L4I-Hr- ) . The spike responses of L4E neurons to touch increase moderately for fhalo = 0 . 5 , especially for neurons with low baseline response ( Fig 11c and 11e ) , whereas the spike response of L4I-Hr- inhibitory neurons remains about the same and that of L4I-Hr+ increases somewhat ( Fig 11d and 11f ) . These dynamics stem from the fact that the initial response to touch , before feedforward inhibition hyperpolarizes the neuron , is determined mainly by feedforward excitation . While halorhodopsin activation hyperpolarizes inhibitory neurons and increase the difference between spiking threshold and their membrane potentials before touch onset , the driving force of excitation ( the difference between the reversal potential of APMA-mediated excitation and their membrane potential ) increases as well and enhances excitation . The reduction in response in L4I-Hr+ neurons partially disrupts whisker movement suppression in L4E neurons , and increases the slope of the νE-AT curve ( Fig 11g ) . During photostimulation of L4I-Hr+ neurons , their response to increasing whisker movements becomes shallower ( Fig 11h ) . This change in response of L4-Hr+ neurons in turn causes a steeper response to whisker movements in L4E neurons ( Fig 11g ) . Finally we explored how the previous results depend on fhalo . The mean touch responses for all neuronal populations is non-monotonic: it increases when fhalo increases from zero , and then decreases for higher values ( Fig 12a ) . For our parameter set , simulated halorhodopsin activation increases touch response of L4I-Hr+ neurons ( in comparison with no activation ) for fhalo < 0 . 78 . During whisker movements , the spike rates of all neuronal populations increases with fhalo , and the average spike rate of L4I-Hr+ neurons during simulated halorhodopsin activation is lower than that with no activation for fhalo < 0 . 74 ( Fig 12b ) . Interestingly , L4-Hr+ neurons reduce their spiking response to touch with halorhodopsin activation for large fhalo , but reduce their spike rates in response to whisking for small fhalo . For small fhalo , the spike rates of L4-Hr+ neurons in response to whisking are low ( Fig 12b ) , because this small group of neurons is both suppressed by halorhodopsin activation and inhibited by the majority of L4-Hr- neurons . In contrast , the spiking responses to touch of L4-Hr+ and L4-Hr- neurons are similar and to the responses of L4I neurons without halorhodopsin activation . This behavior is obtained because touch responses are transient and are reduced by the global level of inhibition before touch . For low fhalo , population inhibition is similar to inhibition without halorhodopsin activation .
We compared our model to recordings from VPM projection neurons ( Figs 1–3 ) , L4 FS neurons , and L4 excitatory neurons [37] during performance of an object localization task , while whisker movements and touches were tracked with millisecond time scale precision [36 , 48 , 51 , 67] . VPM and L4 FS neurons respond to whisker movement and touch , whereas L4 excitatory neurons responded almost exclusively to touch . During whisking , excitation to L4 excitatory neurons from VPM is only slowly modulated in time and is matched by feedforward and feedback inhibition from L4 FS neurons , which cancels self-movement signals . In contrast to the self-movement input , touch-related inputs are brief and synchronous ( Fig 1e ) . Our model establishes the conditions for a ‘window of opportunity’ , in which L4 excitatory neurons can fire before inhibition catches up [35 , 62 , 64 , 68 , 69] . The I-to-E synaptic delay τdelayEI must be sufficiently large ( say , ~1 ms ) and the AMPA-mediated synaptic conductances should be brief ( Fig 8a and 8b ) . The brief duration of this window diminishes the chance of L4 excitatory neurons to fire multiple spikes upon touch , producing low trial-to-trial variability in spike count after touch [36 , 70] . During baseline and whisking , L4 FS neurons spike on average at tens of Hz , while L4 excitatory neurons spike on average below 1Hz . The average spike rates of VPM and L4 FS excitatory neurons , νT and νI , more than double after whisking onset , while the average spike rates of L4 excitatory neurons remains below 1 Hz . Low spike rates of L4E neurons require particular combinations of synaptic parameters: gIE , gEI and gIT need to be strong , and gEE , gII and gET should be weak ( Figs 8c , 9 , 10 ) . If gET is too small or gIT is too large , L4E neurons will be quiescent . Similarly , L4E neurons will not spike if gII is too small ( Fig 9h ) . For moderate gII , νE can remain about constant with AT . The response of L4E neurons to touch , RE , increases with CT ( the thalamic response to touch ) in a sigmoid manner , because the intrinsic properties of L4E neurons in the model impose a refractory period that in general precludes rapid firing of more than one spike . Synaptic parameters that increase νE during whisking shift the sigmoid function leftward , and those that decrease νE shift that function rightward ( Figs 9 and 10 ) . In our model , the inhibitory neuronal population spikes at significant rates ( 20–40 Hz ) even at baseline . This allows the number of spikes/touch RE to be smaller than one and vary gradually with input strength ( Fig 9f ) . As a result , stronger thalamic responses to touch , for example as a result of stronger touch , generates proportionally stronger L4E responses , consistent with experimental results ( Fig 1e ) [36] . In response to halorhodopsin activation , L4 excitatory neurons increase their average spike rates and L4-Hr+ FS neurons reduce their spike rates [37] . The model does not replicate this behavior if halorhodopsin is "expressed" in all inhibitory neurons , because νI increases with simulated photostimulation . This discrepancy is resolved if a sufficient number of L4I neurons lack halorhodopsin ( or express halorhodopsin at very low levels ) ( Figs 11 and 12b ) . With photostimulation , L4-Hr+ I neurons decrease their spike rates if their fraction among I neurons , fhalo , is below a certain value ( 0 . 74 in Fig 12a ) , while L4-Hr- FS neurons increase their spike rates . Future work is needed to determine whether including more populations of inhibitory interneurons , for example somatostatin-positive neurons [71] will generate parameter regimes in which L4-Hr+ neurons decrease their responses to both whisking and touch upon light activation . The development of powerful computing resources has enabled simulations of entire cortical columns of diverse neurons with realistic morphologies and biophysically plausible membrane conductances [72 , 73] . Because of the profusion of parameters , detailed models are difficult to analyze to extract the underlying principles . So far these detailed models have failed to explain any neural computation [74] . Here we took a different approach . We implemented a computational model based on hard won numbers for connection probability , connection strength , neuron numbers and basic cellular parameters for the whisker thalamocortical circuit . The neurons themselves were generic single compartment models . Given the reduced nature of the model , it is possible to analyze the model in detail . This analysis yields testable predictions . For example , the dependence of average spike rates of neuronal populations on the average spike rates of thalamic neurons ( Fig 6b ) . These predictions could be tested in experiments where stimuli , applied for example with a magnetic stimulator , are applied during whisker movement at different amplitudes . The main characteristics of the thalamocortical circuit are: ( i ) Strong external inputs to fast-spiking inhibitory neurons and excitatory neurons . ( ii ) Strong inhibition within L4 , implementing feedforward and lateral inhibition . ( iii ) Recurrent excitation . iv ) A brief but nonzero delay τdelayEI between the spike of an inhibitory neuron and the onset time of the inhibitory post-synaptic conductance in a post-synaptic neuron . Parameters based from neurophysiological measurements produce a model circuit with behavior that is in qualitative agreement with in vivo measurements over a large range of conditions . However , it was necessary to adjust the model parameters slightly ( i . e . within a factor of 2–3; Methods ) to best match experimental data ( reference parameter set , Methods ) . Experimental inaccuracy , sampling errors , and differences across biological conditions ( such as in vitro brain slices versus behaving brain ) likely preclude more accurate estimation of synaptic parameters . Changing even a single parameter could lead to quantitative , and sometimes even qualitative changes in behavior ( Figs 8–10 ) , as has been noted in other contexts [75] . This means that the detailed parameters matter . For example , if gII is too strong then L4I neurons do not respond sufficiently briskly to reduce whisking signals in L4E neurons ( Fig 9g and 9h ) ; on the other hand , if gII is too weak then inhibition shuts down L4E neurons . Similarly , gEE > 0 is required to amplify touch signals , but elevating gEE by a factor of two beyond the optimal level causes run-away excitation ( Fig 8b ) . Moreover , the kinetics matter , such as the durations of the I-to-E synaptic delay τdelayEI ( Fig 7a ) , axonal delays , the time-courses of synaptic conductances ( Fig 7b ) , and intrinsic neuronal properties . Rate models have been used to study cortical responses to fast-rising stimuli in the whisker system [8 , 61] and elsewhere [71] . These previous models of L4 aimed to account for fast cortical responses to passive whisker deflections [8] , but rate models cannot reliably describe brief responses to rapidly-varying stimuli [76] . Moreover , in important aspects these models have opposite behavior to our experimental and modeling results and to the known anatomy and physiology of L4 circuits . First , the modeled L4 neurons show activity in the absence of input , in contradiction to recent measurements [37 , 77] . Second , L4 neurons exhibit strong and brief response to touch even for τdelayEI=0 , in contrast to our model . The propagation of synchronous and brief activity in cortical circuits has been investigated [2 , 63 , 64 , 78 , 79] . In these models , excitatory inputs produce strong inhibition in local circuits , which is slightly delayed with respect to the excitation . In L4 of the barrel cortex , the convergence of thalamocortical input onto L4 FS neurons is higher than for L4 excitatory neurons ( probabilities of connections are 0 . 75 vs 0 . 4 ) [34] and thalamic stimuli produce larger synaptic potentials in L4 FS than L4 excitatory neurons [35] . These features allow propagation of brief synchronous activity across network layers . These models can respond to brief , strong stimuli by brief responses of the L4 E and I neuronal populations . Our model goes beyond this effect by showing how this brief touch response can be obtained together with large νI and small νE in response to baseline and whisking . For example , simple feedforward models such as [35] cannot exhibit small νE for a wide range of AT , corresponding to both baseline and whisking . That model , however , explores the development of cortical responses of L4E and L4I neurons over the long time scale of short-term synaptic plasticity . Conducting such a study in a model with recurrent connections and comparing its outcome to in vivo measurements remains to be carried out . Models of networks of strongly-coupled neurons compensated by inhibition have been studied extensively [38–40 , 58 , 80] . If inhibition on the excitatory neurons is not too large , the population-average response of excitatory and inhibitory neuron ( νE and νI ) scales linearly with the external ( here , thalamic ) input AT [39 , 81] . The slope of the νE vs . AT increases with gII and decreases with gIE . If inhibition is too strong , νE = 0 and νI scales linearly with AT . In L4 and our model network , the numbers of neurons and the strengths of synaptic conductances are not very large . As a result , when νE is small ( νE ≲ 1 Hz ) , similar to measured values [36 , 37] , νE can ( except of near firing threshold , namely the AT values for which cortical neurons start firing ) increase weakly with AT , be independent of AT , or even decrease with AT ( Figs 9e , 9h , 10b and 10e ) . Similar behavior was obtained in rate models of V1 [82 , 83] . This result mimics the empirical observation that during transition from non-whisking to whisking , the spike rates of VPM neurons more than double , the spike rates of L4I neurons increase proportionally , but the spike rates of L4E neurons remains low , with no significant increase with thalamic input strength . For larger νE ( > 2 Hz ) , νE depends linearly on AT . Larger νE could be obtained with larger gII or smaller gEI ( Fig 9b , 9e , 9h ) , or by suppressing a fraction of L4I neurons using halorhodopsin activation ( Fig 11g ) . Based on the linearity of strongly-coupled networks , one would expect that if responses of L4E neurons to whisking is an order of magnitude smaller than that of VPM neurons , the same proportions will maintain for the responses to touch , whereas experimentally the touch responses of L4 excitatory and VPM neurons differ by only a factor of two . We explain this effect by the fact that , with τdelayEI , inhibition lags excitation in response to brief and strong thalamic input . Indeed , for τdelayEI=0 , the response to L4E neurons is significantly reduced ( Fig 7a ) . In this paper we emphasize the suppression of tactile reafference signals in the somatosensory cortex . However , reafference signals could still contribute to cortical computation . Consistent with previous studies [16 , 18 , 84] , we observed that spike rates were modulated by the phase of the whisking cycle [37] . Spike rates were larger for certain phases of the whisker motion , and individual neurons had different preferred phases . Neurons in VPM and L4 all showed modulation with whisking phase . Excitatory and inhibitory inputs to L4 excitatory neurons showed phase-tuning . Across the population , the phase-tuning of inhibition from L4 fast-spiking neurons is biased towards protractions . This suggests that the touch-evoked responses in L4 excitatory neurons may be modulated by whisking phase , as has been previously suggested [84] . This kind of phase-dependent modulation could play a role in localizing objects by whisker touch during active sensation .
The experimental methods were described in [37] . In brief , we performed chronic multi-electrode silicon probe recordings from VPM and cell-attached recordings from L4 excitatory and L4 fast spiking neurons . To search for VPM neurons we lightly anesthetized the mice ( 0 . 6–1 . 2% isoflurane ) and stimulated individual whiskers during extracellular recordings . We mapped the principal whisker ( PW ) to assess whether the PW was one of the large whiskers that can be tracked reliably during behavior . We stimulated several individual whiskers around the PW with a piezoelectric stimulator at multiple frequencies ( i . e . 5 , 10 , 20 , 40 Hz ) and recorded the neural activity . Before and/or after each day of behavioral recording we confirmed that neurons respond to stimulation of the PW ( we re-positioned the electrode drive daily ) . In addition , after placing the animal in the behavioral apparatus , we usually maintained the anesthesia for several minutes to check that neurons still responded to the PW ( by manual stimulation and/or by contacting the whiskers with the pole ) . Animals performed the behavioral task shortly after anesthesia was withdrawn . Comprehensive neuroanatomical and neurophysiological data sets are revealing the connectivity between defined cell types over multiple spatial scales [28 , 43 , 44 , 46 , 86 , 87] . But links between neural representations , computation and detailed anatomy are rarely achieved [6 , 88] . One challenge is a lack of knowledge about strengths and dynamics of synapses between specific cell types during behavior . Most of such studies were carried out in slices , and there are differences between intrinsic and synaptic properties measured in slices and in vivo and between different in vivo states . Therefore , our strategy is to set parameter values ( synaptic conductances , connectivity ) close to measured in vitro values . However , differences between in vitro and in vivo conditions and experimental errors ( e . g . estimates of unitary synaptic strengths ) currently preclude exact quantification of the synaptic and connectivity properties ( e . g . , [53] ) . We define a set of parameter values , named "reference parameter set" [89] , close to measured values when known ( Fig 5 ) , allowing for adjustment over a restricted range ( mostly within a factor of 2 relative to empirical values ) , such that the circuit displays dynamics similar to experimentally-observed behavior . The reference parameter set specified below is used unless otherwise stated . Then , we vary one or two parameters to explore the role of those parameters on the system dynamics ( e . g . , Figs 8–10 ) .
|
We study how information is transformed between connected brain areas: the thalamus , the gateway to the cortex , and layer 4 ( L4 ) in cortex , which is the first station to process sensory input from the thalamus . When mice perform an active object localization task with their whiskers , thalamic neurons and inhibitory fast-spiking ( FS ) interneurons in L4 encode whisker movement and touch , whereas L4 excitatory neurons respond almost exclusively to touch . To explain these observations , we constructed a computational model based on measured circuit parameters . The model reveals that without touch , when thalamic activity varies slowly , strong inhibition from FS neurons prevents activity in L4 excitatory neurons . Brief and strong touch-induced thalamic activity excites both excitatory and FS neurons in L4 . FS neurons inhibit excitatory neurons with a delay of approximately 1 ms relative to ascending excitation , allowing L4 excitatory neurons to spike . Our results demonstrate that cortical circuits exploit synaptic delays for fast computations . Similar mechanisms likely also operate for rapid stimuli in the visual and auditory systems .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"action",
"potentials",
"medicine",
"and",
"health",
"sciences",
"neural",
"networks",
"nervous",
"system",
"membrane",
"potential",
"social",
"sciences",
"electrophysiology",
"neuroscience",
"animal",
"anatomy",
"interneurons",
"zoology",
"computer",
"and",
"information",
"sciences",
"animal",
"cells",
"neural",
"pathways",
"touch",
"animal",
"physiology",
"cellular",
"neuroscience",
"psychology",
"neuroanatomy",
"cell",
"biology",
"vibrissae",
"anatomy",
"physiology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"sensory",
"perception",
"neurophysiology"
] |
2017
|
Mechanisms underlying a thalamocortical transformation during active tactile sensation
|
Patterns of trait distribution among competing species can potentially reveal the processes that allow them to coexist . It has been recently proposed that competition may drive the spontaneous emergence of niches comprising clusters of similar species , in contrast with the dominant paradigm of greater-than-chance species differences . However , current clustering theory relies largely on heuristic rather than mechanistic models . Furthermore , studies of models incorporating demographic stochasticity and immigration , two key players in community assembly , did not observe clusters . Here we demonstrate clustering under partitioning of resources , partitioning of environmental gradients , and a competition-colonization tradeoff . We show that clusters are robust to demographic stochasticity , and can persist under immigration . While immigration may sustain clusters that are otherwise transient , too much dilutes the pattern . In order to detect and quantify clusters in nature , we introduce and validate metrics which have no free parameters nor require arbitrary trait binning , and weigh species by their abundances rather than relying on a presence-absence count . By generalizing beyond the circumstances where clusters have been observed , our study contributes to establishing them as an update to classical trait patterning theory .
Competition is a driving force in nature , and a central question in ecology is how it shapes community structure . Species traits mediate their interactions with the environment and each other , and therefore determine how they compete . As such , patterns of trait distribution among co-occurring species can give insights into the underlying coexistence mechanisms [1–5] . Theory shows that coexistence is only stable if species differ in their ecological needs and impacts—i . e . they must display niche differences ( see Glossary in S1 Text ) [6 , 7] . These differences reduce competition between species compared to competition within species , thus allowing each species to grow from low abundance in the presence of the others . If species are arranged on niche axes so that proximity on those axes indicates similarity in their niche strategies , they are expected to display limiting similarity–greater-than-chance differences on those axes [8] . Ecologists use trait axes as proxies for niche axes , and look for similar patterns therein , usually in the form of overdispersion or even spacing between species on these axes [9–17] . This classical idea remains the dominant paradigm , despite mixed empirical support [2 , 3 , 5] . In contrast , recent studies suggest that competition causes the spontaneous emergence of transient clusters of species with similar traits [18–20] . They find that limiting similarity results only if competition plays out to the final stage , whereby all species are excluded except for those with optimal niche strategies ( see Glossary in S1 Text ) . Those species are niche-differentiated enough from one another ( i . e . have enough space between them on the niche axis ) to all stably coexist together , having emerged as dominant over other strategies through the competitive process . When species outnumber optimal niches , e . g . phytoplankton communities with more species than resources [21] , competitive exclusion will ensue . However , species near the optimal niche strategies are excluded more slowly than those further away from these optimal strategies [22–24] . As a result , the community temporarily self-organizes into clusters of similar species near optimal strategies , with gaps in between . Clustering has traditionally been associated with environmental filters [25] , and more recently with one-sided competition without a balancing tradeoff , such as competition for light where taller is better [26] . However , those will tend to produce a single big cluster around the favored trait , whereas partitioning of a niche axis will lead to multiple clusters , one per optimal niche strategy . Patterns of multiple clusters have in fact been reported in empirical studies of body size in phytoplankton communities and certain animal taxa [18 , 27–29] . This phenomenon has been interpreted to bridge coexistence through differences ( niche theory ) and similarity ( neutral theory ) [30] , and as such potentially represents a unification of classical ideas and a generalization of limiting similarity . Scheffer and van Nes [18] first demonstrated the emergence of transient clusters in Lotka-Volterra dynamics . Later work further suggested that clusters are a generic outcome of Lotka-Volterra dynamics [20] except for special shapes of the function connecting competition to traits [31 , 32] . Several mechanisms can make these transient clusters persist , such as specialist enemies [18] and periodic environments [33 , 34] . However , the generality of clusters as a signature of competition cannot be established without showing that they emerge in communities subject to stochastic processes . In particular , immigration and ecological drift are intrinsic to most communities , and have been amply demonstrated to be important players in nature [30 , 35–37] . In fact , models ignoring all of biology but drift and immigration successfully describe observed macroecological patterns [38] ( which of course does not mean deterministic forces are unimportant ) . Yet , clustering remains unseen in competition models incorporating these processes [39–41] . In plankton models where clustering occurs , immigration has as a negative impact [21] . Moreover , clusters have not been widely demonstrated beyond Lotka-Volterra dynamics . Lotka-Volterra competition equations are a heuristic description which does not specify a niche mechanism . They are a special limit [42] of MacArthur’s consumer-resource dynamics model [43] , and cannot describe all types of competition [44] . While species clusters have been shown to emerge in explicit consumer-resource dynamics [21 , 34] , these studies ignored the possibility of resource depletion effects . The latter have been shown to greatly affect competitive relations between consumers , for example by violating the assumption that competition always decreases with trait differences [45] . It is also not clear that clustering should emerge among species competing along environmental gradients; indeed , early studies of stochastic niche dynamics [39–41] focused on competition of this type and found no clustering pattern ( although these studies , which predate [18] , were not specifically looking for clustering ) . Finally , it is not known whether clusters emerge in communities characterized by hierarchical competition among species that coexist via life-history tradeoffs . These mechanisms , such as the competition-colonization tradeoff [46] and the tolerance-fecundity tradeoff [47] , have been chiefly studied in terrestrial plants , but may enable coexistence among other sessile organisms with a dispersive stage , such as coral , coral fishes , and microbes [47] . To determine whether clusters are a general outcome of competition as opposed to an artifact of specific models , and to verify their robustness to stochastic forces , here we use a stochastic niche simulation approach to investigate the emergence of clustering by traits in species assemblages undergoing competitive dynamics and open to immigration . We start with Lotka-Volterra dynamics , where clusters are known to emerge in the deterministic model [18] , to address the question of their robustness to stochastic forces . We test this robustness with and without the confounding influence of environmental filters , which we hypothesize might mask any clustering caused by niche partitioning . We then see how regional diversity and immigration rate influence this robustness . Clustering intrinsically involves many species , and should in principle be stronger , or at least more detectable , under a high species-to-niches ratio . Therefore , we expect more clustering under higher regional diversity . As for immigration , we expect that while it may contribute to the persistence of clusters by keeping weak competitors from being excluded , too much immigration will drown any pattern caused by competition . Next , to determine whether clusters are a general outcome of competition , we analyze models spanning three key niche mechanisms: resource partitioning , habitat partitioning , and a competition-colonization tradeoff . Under resource partitioning , we further explore the differences between scenarios with low and high resource depletion . Under competition for habitat , we also examine the impact of dispersal limitation as opposed to global dispersal . We expect that larger scale dispersal relative to the scale of the environmental gradient can influence cluster emergence by lessening the dominance of emergent optimal strategies , since their propagules often spread to locations with suboptimal habitat . If clusters are a signature of competition and niche differentiation , we expect clustering to emerge generally under all scenarios of all three niche mechanisms . To determine whether communities are clustered beyond patterns that could arise by chance , we present two metrics: the first uses the k-means algorithm [48] , which assigns species to clusters by minimizing trait differences within clusters . The second is based on Ripley’s K [49] , and quantifies clustering based on the sparseness of the regions between clusters . Both metrics apply the gap statistic method for comparing data with null models [50] . These metrics improve over existing tools to detect clustering [18 , 27–29 , 51] as they contain no free parameters and do not arbitrarily bin traits . Furthermore , rather than reducing species to presence-absence counts , our metrics weigh them by their abundance . We validate the metrics by confirming that they detect clusters in Lotka-Volterra communities but not in neutral communities .
We implemented a stochastic analogue of classical Lotka-Volterra competition , which assumes species influence each other directly via competition coefficients . We tied those coefficients to similarity in traits ( Fig 1A ) , so that trait differences can enable stable coexistence ( niche differentiation ) . In its deterministic form , this type of model is known to produce clusters [18] ( except for special cases where competition coefficients form a positive-definite matrix [32 , 52] ) . We thus used this model to consider the robustness of clusters to stochastic forces . We also used it to consider whether that robustness of clustering is influenced by the presence of environmental filtering , and by regional diversity and immigration . Clustering was strong enough in the Lotka-Volterra communities to be easily distinguishable by eye ( Fig 2A and 2B ) , with over 80% of all 100 replicates testing significant at the p < 0 . 05 level by either metric ( Fig 3 ) . We note that the transient clusters produced by deterministic Lotka-Volterra competition are being maintained in our stochastic model by immigration; when immigration is turned off , communities begin to lose species and tend towards a limiting similarity pattern [23] . Assuming fast resource dynamics relative to consumers , one can derive the Lotka-Volterra model from resource-consumer dynamics [42 , 53] , with trait similarity indicating similar resource preferences [8 , 43] ( Fig 1B ) . However , this approach ignores the effects of stochastic fluctuations in resource availability , and particularly resource depletion . The latter has been shown to impact coexistence outcomes [54] and even competition-similarity relationships [45] . These relationships are drivers of pattern [31] , and we therefore tested whether incorporating stochastic resource dynamics and depletion affected the emergence of clusters . In the case of low resource depletion , we found weak clustering ( Fig 3 ) . This is because when resource depletion is low , consumers undergo approximate Lotka-Volterra dynamics with competition coefficients following a Gaussian function of trait separation [8] . This function is positive-definite , a property that has been shown to lead to weak or no pattern ( see Fig S8 in [23] and [32 , 52] ) . Species sorting under Gaussian competition is slow ( see Fig S8 in [23] ) , and therefore easily overpowered by immigration . Indeed , clustering was stronger under lower immigration pressure ( S4B and S5 Figs ) . Resource depletion breaks the link to Gaussian Lotka-Volterra dynamics , thereby fundamentally changing the competitive interactions between consumers [45 , 54] . That strengthened clustering in our model ( compare low- and high-depletion scenarios in Fig 3 ) . As resources were extirpated , species ended up clustering based on specialization to the remaining ones ( S6 Fig ) . Resource depletion thus sets the optimal niche strategies ( i . e . specialization onto remaining resources ) , thereby strengthening the pattern . Habitat is a critical resource for which species compete , and as such the environment is thought of as a key axis of niche differentiation [55–57] . Traits reflect adaptations to different environments , and hence competition for habitat must shape trait distribution . To test whether niche-differentiation based on environmental preference leads to clustering , we used the model introduced by [40] . We assume a linear landscape on a habitat gradient , e . g . an elevation gradient . Different species are optimally adapted to different habitats , and competition arises from overlap in environmental preference ( Fig 1C ) . Competition occurs at the recruitment stage , where the probability of recruiting is based on tolerance to the local environment . Because dispersal can play a central role in competition for space [57 , 58] , we consider two scenarios: global and local dispersal , which differ by whether individuals are more likely to disperse shorter distances from their parents . Since this is an individual-based model , we used a smaller community size ( 1 , 000 individuals rather than 21 , 000 ) for computational expedience . We found that clusters also emerge under competition for habitat . Switching dispersal from global to local had a strong impact on cluster shape , and reduced the number of species per cluster ( compare Fig 2E and 2F ) . This occurred as dispersal limitation effectively decreases immigration . Moreover , the attending reduction of the diluting effects of immigration substantially strengthened clustering ( Fig 3 ) . All models examined so far describe symmetric competition , whereby the competitive impact of species A on B is similar to that of B on A . Competitive hierarchies stand in contrast to this . That is the case of the competition-colonization tradeoff [46 , 59] , where propagule production , or colonization ability , trades off with the ability to displace individuals of other species , or competitive ability ( Fig 1D ) . Even though in this case some species are better competitors than others , the tradeoff is a niche mechanism because it allows for the stable coexistence of multiple species ( see Glossary in S1 Text , S7 Fig ) . In fact , the model can be cast in Lotka-Volterra form , and one can show that the net competitive impact , while asymmetric , is stronger between more similar species ( S7 Fig ) . Clusters also emerged under this niche mechanism . The asymmetry in species interactions was reflected in its asymmetric clusters ( Fig 2G ) . The k-means metric picked up on just three clusters in most replicates , even though without stochasticity and immigration the model produces about 13 transient clusters ( S7 Fig ) . This is perhaps because species in the first cluster so strongly dominated the community ( e . g . , the most abundant species in the first cluster in Fig 2G had 4 , 798 individuals , compared with 395 in the second and 392 in the third ) . This indicates that species adopting the high-competitiveness strategy ( left side of the trait axis ) outperform both those who invest in high fecundity and the intermediate group . Dispersal limitation could reduce the asymmetry by augmenting the benefits of high fecundity . We hypothesize that a spatially explicit formulation of this mechanism would produce more similarly sized clusters . While our metrics performed equally well in the Lotka-Volterra communities , they differed in the other niche mechanisms . Our Ripley’s K metric fared better than the k-means metric in scenarios where species partition resources and habitat ( Fig 3 ) ( though performance was similar for resource partitioning under low immigration , see S5 Fig ) . The Ripley’s K metric focuses on identifying a scale of interspecific trait difference with particularly low representation ( i . e . by which very few species pairs are separated ) . As such , it relies on regular spacing between clusters and intercluster gaps . Ripley’s K is good at identifying clusters when that spacing is regular , even in cases where the overall pattern is noisy ( the resource partitioning cases ) , or involves low-occupancy clusters ( the habitat partitioning cases ) . On the other hand , the k-means metric found the clusters in the competition-colonization tradeoff while Ripley’s K missed them ( Fig 3 ) . This is because the k-means algorithm is less sensitive to strong asymmetries between the clusters . In these cases , the k-means metric is a better choice .
Ecologists have long sought to understand how competition shapes community structure . While competing species are usually expected to be more different than predicted by chance [25 , 60] , recent studies suggest that competition may cause species to cluster by traits , such that the community self-organizes into groups of similar species [18] , a phenomenon which has been interpreted to bridge coexistence through differences–niche theory–and similarity–neutral theory . Our study verified that clustering transcends Lotka-Volterra dynamics , occurring under a number of niche mechanisms . Further , we showed that clustering is robust to stochastic drift , an intrinsic property of real-life communities . Immigration maintains clusters that are otherwise transient , and the strength of clustering has a modal relationship with immigration pressure . We showed that clustering may be detectable under the confounding influence of environmental filters , and is enhanced by regional diversity . Finally , we provided metrics for detecting and quantifying clusters in nature . Why do clusters arise ? Different niche mechanisms share the common property that competition is stronger between species with more similar strategies . It thus seems paradoxical that clusters should emerge . However , it is precisely because species with similar niches compete more strongly that clusters appear [61] . While similar pairs compete more strongly , they experience similar competitive pressure ( or relief ) from the rest of the community . If a given niche strategy is favored because it minimizes competition with the rest of the community or capitalizes on greater resource supplies , then similar strategies are similarly favored . This hilly fitness landscape causes exclusion to be slower near the center of the niches than in the gaps between them , making it easy for immigration to permanently maintain the clusters [23] . Modern coexistence theory [7] splits coexistence-promoting processes into those that reduce competition among species relative to competition within species ( stabilizing mechanisms , here referred to as niche mechanisms ) , and those that reduce differences in average fitness between species ( equalizing mechanisms ) . It is thus tempting to interpret clusters as reflecting a harmonious combination of stabilizing and equalizing forces: species within a cluster are equalized , while those in different clusters are stabilized . However , this interpretation is problematic . The equalization-stabilization dichotomy is based on applying invasibility criteria to closed communities regulated by a small number of limiting factors [62] , an approach which does not extend easily to multispecies communities under immigration and a continuum of resources . In diverse communities with complex competitive interactions , it is difficult to calculate equalizing and stabilizing terms and tie them to trait differences , let alone interpret how specific patterns such as clustering connect with equalization and stabilization forces [62] . One approach is to assume all pairs of species compete with equal intensity [7 , 63] , but this assumption is strongly violated in all models where clustering has been observed so far . While competition is responsible for clusters’ emergence , immigration is responsible for their persistence . Immigration joins other mechanisms that have been previously shown to sustain clusters , namely specialist enemies [18 , 64] and environmental fluctuations [33 , 34] . We found that clustering appears generally under different immigration regimes , especially if the number of species far exceeds the number of available niches . Studies of stochastic competitive preceding Scheffer and van Nes 2006 found little impact of immigration on resulting trait pattern [39] , [40] . However , successful immigration was highly infrequent in those models . In [39] , resources made available through deaths were assumed to be redistributed broadly , so that resource supply remained low everywhere , making recruitment of new individuals highly unlikely . In [40] , immigration took the form of a single immigrant seed being added to a large pool of local seeds competing for the site . As such , immigration rates were effectively much lower than ours . [41] tested both very high and very low immigration , and also saw no clusters ( in no small part because the authors were not looking for them ! ) . This could be due to their use of a Gaussian competition kernel ( i . e . competition coefficients are a Gaussian function of trait difference ) , which leads to weak niche sorting dynamics , easily overwhelmed under high immigration [32] . The fact that immigration maintains clusters seemingly defers the question of coexistence to the regional scale . The problem dissipates by considering mass effects in the metacommunity framework [65] . The regional pool is a combination of local communities , and species are selected for different traits at different sites due to their own local niche dynamics . Therefore , each community receives immigrants which may be dominant elsewhere despite being disfavored locally . However , this does not mean that pattern is expected regionally , as the sum total of heterogeneous communities , each with a different trait pattern , may result in no discernible pattern at a regional scale . Clusters caused by partitioning of a niche axis are often distinct from clustering due to other processes . Environmental filters favoring a single best trait and one-sided competitive dynamics where a particular trait outperforms all others without a balancing tradeoff [26] will produce a single cluster as opposed to multiple clusters . Where evolutionary rates are commensurate with ecological dynamics , small mutations and sympatric speciation may also generate species or genotype clusters without niche differentiation [66 , 67] . Ruling out these alternative sources of clustering could require sampling at a larger scale than applicable to the niche mechanism [5 , 60] , or directly verifying competitive effects and frequency dependence [4] . While clusters may have more than one source , observation of multiple clusters in specific functional traits can help identify potential drivers of niche differentiation in a community . We presented and validated two nonparametric abundance-weighted tools for detecting and quantifying clusters . Our metrics successfully distinguished clustering in niche-differentiated communities from no clustering in neutral communities and a single cluster in communities under environmental filters without a niche mechanism . We note that none of these results would appear without considering species abundances , as presence-absence counts do not reveal clustering in our communities . Also , our metrics did not require arbitrarily binning traits , nor fitting parameters to the data . Although our study focused on one-dimensional trait axes , competitive interactions may often be mediated by variation in multiple traits [68] . Theoretical work on simple models indicates that multidimensional niche space leads to multidimensional clustering [20] . Our metrics can quantify clusters in any dimension , using generalized measures of niche separation . The main challenge is to connect multidimensional phenotypes to competitive relations , in order to define the correct measure of distance in high-dimensional trait space [61] . However , even if species cannot be arranged on linear trait axes , our Ripley’s K metric can still detect clustering by similarity as long as a measure of species differences can be defined , e . g . Hamming distance in genetic sequences [66] . Clustering as a signature of coexistence under competition is an update to the still dominant paradigm that competing species will display greater-than-chance differences [3–5 , 16 , 17] . Our finding that clusters appear under various niche mechanisms and can be easily maintained by immigration , even when confounding forces are at play , suggests that clusters are a likely feature of nature beyond the instances where it is currently known to occur [18 , 28 , 29] . For example in tropical forests , where both competition and dispersal are recognized as key drivers of community assembly [69 , 70] , clustering could help explain the high diversity and seemingly continuous phenotypic variation .
We used a lottery model framework [71] to implement stochastic niche dynamics in a fixed-size community open to immigration from a regional pool . We start with a random draw of offspring from the regional pool , and then alternate death and recruitment events until species abundance distributions are stationary . A proportion m of deaths are replaced by immigrants , and the remainder by local offspring . This is analogous to Hubbell’s neutral model [38] , except here the niche mechanism sets the probabilities of birth and death across different species . Schematic illustrations of our niche models are shown in Fig 1 . We used the 50-hectare plot of tropical forest on Barro Colorado Island , Panama , as a reference point for our community size ( 21 , 000 individuals >10 cm dbh [38] ) and immigration rate ( 0 . 08 immigrant recruits per recruitment event [72] ) . Our regional pool is a fixed neutral metacommunity with biodiversity parameter θ = 50 [38] and 150 , 000 individuals , leading to c . 400 species ( bigger metacommunities did not change results ) . All simulations and statistical analyses were done in R [73] . We apply a variation of the gap statistic method [50] , using two different clustering measures: k-means dispersion [48] and Ripley’s K function [77] . In general terms , the metric takes in the list of species traits and abundances , and returns the number of clusters ( k-means ) or average trait separation between them ( Ripley’s K ) , as well as a z-score and a p-value . In S1 Box we summarize the k-means version and give a step-by-step recipe for its implementation , and illustrate it at work on two example communities . Both versions are described in detail in S1 Appendix . The code for the k-means version is available on GitHub [78] . We assess statistical significance and degree of clustering by comparing a community against 100 null communities where we randomly shuffle local abundances across all species in the regional pool . This allows us to test specifically for a nonrandom association between traits and abundances , while keeping the observed abundance distribution fixed . One alternative null model is neutrality , whereby abundances follow a characteristic dispersal-limited multinomial distribution [79] . Because our models are stochastic , we run 100 replicates to account for variation within the same scenario . For each niche scenario we report the average z-score across 100 replicates , as well the percentage of replicates that were significantly clustered at level p < 0 . 05 . To assess if a given run is significantly clustered we compared with the distribution across 100 nulls . We check for false positives by testing our metric on neutral communities and communities where differences in species performance are due strictly to environmental filtering . From the former we expect significant clustering in circa 5% of runs when using a p = 0 . 05 cutoff , and from the latter we expect that species will cluster around the favored trait . Thus we distinguish niche differentiation from neutrality and pure environmental filtering by the presence of multiple clusters as opposed to none or a single one .
|
Species traits determine how they compete with each other . As such , patterns in the distributions of traits in a community of competing species may reveal the processes responsible for coexistence . One central idea in theoretical ecology is that the strength of competition relates to similarity in species needs and strategies , and therefore if competition plays out at short timescales , coexisting species should be more different than expected by chance . However , recent theory suggests that competition may lead species to temporarily self-organize into groups with similar traits . Here we show that this clustering is a generic feature of competitive dynamics , which is robust to demographic stochasticity and can be indefinitely maintained by immigration . We show that clustering arises whether species coexist by partitioning resources , environmental preferences , or through tradeoffs in life-history strategies . We introduce and validate metrics that , given species traits and abundances , determine whether they are clustered , and if so , how many clusters occur . By showing the generality of self-organized species clusters and providing tools for their detection , our study contributes to updating classical ideas about how competition shapes communities , and motivates searches for them in nature .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"community",
"ecology",
"population",
"metrics",
"ecological",
"metrics",
"ecology",
"and",
"environmental",
"sciences",
"ecological",
"niches",
"species",
"diversity",
"ecology",
"species",
"delimitation",
"community",
"structure",
"speciation",
"habitats",
"biology",
"and",
"life",
"sciences",
"population",
"biology",
"evolutionary",
"biology",
"evolutionary",
"emergence",
"evolutionary",
"processes",
"fecundity"
] |
2019
|
Generalizing clusters of similar species as a signature of coexistence under competition
|
Pentameric ligand-gated ion channels are activated by the binding of agonists to a site distant from the ion conduction path . These membrane proteins consist of distinct ligand-binding and pore domains that interact via an extended interface . Here , we have investigated the role of residues at this interface for channel activation to define critical interactions that couple conformational changes between the two structural units . By characterizing point mutants of the prokaryotic channels ELIC and GLIC by electrophysiology , X-ray crystallography and isothermal titration calorimetry , we have identified conserved residues that , upon mutation , apparently prevent activation but not ligand binding . The positions of nonactivating mutants cluster at a loop within the extracellular domain connecting β-strands 6 and 7 and at a loop joining the pore-forming helix M2 with M3 where they contribute to a densely packed core of the protein . An ionic interaction in the extracellular domain between the turn connecting β-strands 1 and 2 and a residue at the end of β-strand 10 stabilizes a state of the receptor with high affinity for agonists , whereas contacts of this turn to a conserved proline residue in the M2-M3 loop appear to be less important than previously anticipated . When mapping residues with strong functional phenotype on different channel structures , mutual distances are closer in conducting than in nonconducting conformations , consistent with a potential role of contacts in the stabilization of the open state . Our study has revealed a pattern of interactions that are crucial for the relay of conformational changes from the extracellular domain to the pore region of prokaryotic pentameric ligand-gated ion channels . Due to the strong conservation of the interface , these results are relevant for the entire family .
During activation of a pentameric ligand-gated ion channel ( pLGIC ) , the binding of agonists promotes the opening of a selective ion conduction pore at a distance of more than 50 Å away from the binding sites [1 , 2] . This process has been described by means of the Monod Weyman Changeux ( MWC ) mechanism of allosteric proteins , where activation can be broken down into distinct steps defining ligand binding and the shift in the equilibrium between the open and closed state of the pore [3–5] . pLGICs constitute a large family of membrane proteins that are expressed in animals and certain prokaryotes [6] . In mammals , the family encompasses ionotropic neurotransmitter receptors for acetylcholine , serotonin , GABA , and glycine , which are key players in electrical signal transduction at chemical synapses [7] , whereas prokaryotic pLGICs are potentially involved in pH resistance [8 , 9] . All family members share a conserved molecular architecture composed of five either identical or closely related subunits . Over recent years , insight into the structural properties of pLGICs has been obtained from different sources . Electron microscopy studies of the nicotinic acetylcholine receptor ( nAChR ) from Torpedo electric ray have shed light on the structure of a heteropentameric receptor at medium resolution [10 , 11] . A recent study by single-particle electron cryomicroscopy revealed agonist and antagonist bound views of the glycine receptor ( GlyR ) [12] . Structures at higher resolution have been provided by X-ray crystallography for various pro- and eukaryotic family members [13–21] . Although these structures show different conformations of the channels , whose assignment to defined functional states is in certain cases still ambiguous [22] , they closely resemble each other with respect to their general architecture . Each subunit consists of a predominantly β-stranded extracellular domain and an α-helical transmembrane pore , which interact via an extended interface . Both domains constitute independent folding units that , in certain cases , can be expressed as isolated proteins , thereby maintaining their respective structure observed in the full-length receptors [23–25] . The acetylcholine binding protein , which resembles the extracellular domain , even is an independent soluble protein [26] . Besides their close structural relationship , family members also share a common gating mechanism . Whereas the probability for channel opening in the ligand-free state is very low , it is increased by several orders of magnitude following agonist binding to sites in the extracellular domain located at the boundary between two adjacent subunits [1 , 5] . Since conformational rearrangements in this part of the protein are transduced via the domain interface to the transmembrane pore [27] , it is not surprising that the residues at this interface belong to the most conserved parts of the protein . In this study , we were interested in the role of interactions at the domain interface for the transduction of conformational changes in pLGICs . For that purpose , we have characterized mutants of ELIC and GLIC , two prokaryotic family members , by electrophysiology , calorimetry , and X-ray crystallography . These prokaryotic channels are ideal targets for mechanistic investigations: Their detailed structures have been determined in different conformations and show compact proteins that contain the main features of pLGICs [28] . Moreover , unlike many eukaryotic pLGICs , they form functional homopentamers that have been characterized on a macroscopic and a single channel level and that exhibit a functional behavior that closely resembles family members of higher organisms [8 , 9 , 29 , 30] . Whereas ELIC forms a cation-selective channel with high conductance that is activated with high efficacy by the primary amines cysteamine , propylamine , and GABA [9] , the cation-selective GLIC is activated by protons and inhibited by bulky positively charged compounds that also act as open channel blockers of the nAChR [8 , 31] . Our study has identified a cluster of interacting residues located at the β1-β2 turn , the β6-β7 loop and the pre-M1 region of the extracellular domain , and the M2-M3 loop of the pore that exert a strong influence on channel function . These residues face a tightly packed core of the subunit , suggesting that their mutual interactions are critical for the transduction of signals underlying channel activation . Our results are generally consistent with previous investigations on eukaryotic receptors , which underlines the conservation of the activation mechanism throughout the family .
To investigate the role of interactions between the ligand-binding and the pore domain of pLGICs , we have selected residues in ELIC and GLIC that are either part of the domain interface or that are located in close proximity ( Fig 1 and S1 Fig ) . In an initial screen , we have mutated these residues to alanine and expressed them in Xenopus laevis oocytes . Surface expression was quantified by ELISA with an antibody that recognizes a tag fused to the extracellular N-terminus of the respective protein . Most constructs showed robust expression , with few exceptions where the truncation of the side chain has led to a strong reduction of the ELISA signal ( S2 Fig ) . To probe whether the mutated proteins would still be activated by ligands , we have measured the current response upon application of agonist by two-electrode voltage clamp electrophysiology . Whereas the majority of the investigated constructs showed response at high agonist concentration , the mutation of certain positions , although well expressed , resulted in either an apparent loss of activation or very low currents ( Fig 2 and S2 Fig ) . In general , equivalent positions in ELIC and GLIC exhibited a similar pattern , which underlines the role of conserved residues at the domain interface for channel activation , but there were also some differences observed . In both cases , mutants with strongly compromised activation properties cluster at the loops connecting β-strands 6 and 7 ( the cys-loop of eukaryotic receptors that contains the region of highest conservation ) and α-helices M2 and M3 of the pore domain ( the M2-M3 loop ) . A nonactivating phenotype was also found for certain residues of the β8-β9 loop that connect to the neighboring subunit and in GLIC , for an aspartate in the β1-β2 turn . Finally , no activation in case of ELIC and no expression in case of GLIC was observed for a strictly conserved arginine at the end of β-10 ( the pre-M1 region ) . None of the investigated mutations showed detectable basal activity in the absence of agonists . When mapped on the structure , most mutations resulting in nonactivation by agonist point into a tightly packed core of the protein , irrespectively of their position in the sequence , thus suggesting that any disruption of this core may interfere with channel activation ( Fig 2A and 2B and S3A Fig ) . To exclude that these mutations cause misfolding of the protein , we have expressed several of them in Escherichia coli . Most mutants showed wild type ( WT ) -like expression levels and were stable in detergent solution . For two cases , the ELIC mutants F116A of the β6-β7 loop and Y258A of the M2-M3 loop , we have grown crystals and determined structures at 3 . 5 and 3 . 2 Å , respectively ( Table 1 ) . Both mutants crystallized in the same nonconducting conformation that has been observed in all known ELIC structures . Small structural differences in the vicinity of the respective mutations indicate local rearrangements of protein interactions due to the loss of the bulky aromatic side chains ( Fig 3A and 3B , S4A and S4B Fig ) . The data suggests that the mutations , despite their severe phenotype on channel activation , have only a local effect on the protein structure . We have also investigated whether both mutants would at least show residual activity , and we have thus expressed them in X . laevis oocytes and HEK-293 cells and studied excised patches in the outside-out configuration upon fast application of agonist . Neither of the two mutants display ligand-induced channel activity in any of numerous independent recordings ( Fig 3C–3E ) . To exclude that the two nonactivating mutations have compromised the ability of the protein to recognize its ligand , we have studied agonist and antagonist binding to the detergent solubilized protein by isothermal titration calorimetry ( ITC , S5 Fig ) . WT ELIC binds the agonist propylamine and the competitive antagonist acetylcholine with an effective dissociation constant ( Keff ) of 8 and 2 . 5 mM , respectively ( Fig 3F ) . Whereas the affinity for the antagonist , which stabilizes the closed state of the channel , is similar in calorimetry and electrophysiology experiments [30 , 32] , Keff of the agonist is about 20-fold higher than its EC50 measured in two-electrode voltage clamp recordings ( EC50 of 450 μM in the absence of Ca2+ ) [9] , which suggests that the channel may not be fully activated in detergent solution . It is also noteworthy that the measured value is very close to the dissociation constant for propylamine to the resting state of 7 . 1 mM that was obtained from a detailed kinetic analysis of single channel recordings of ELIC [29] . To enhance binding of the agonist , we also carried out calorimetry experiments in the background of the mutant R91A located in the ligand binding site , which was previously shown to increase the potency of cysteamine , propylamine , and acetylcholine [9] . In accordance with electrophysiology , ITC experiments show that the Keff values of ligands are decreased in the mutant R91A , although this effect is stronger for the agonist than the antagonist ( Fig 3G ) . In the background of the mutant R91A , both nonactivatable mutants F116A and Y258A bind agonist and antagonist with similar Keff as the single mutant R91A , thus emphasizing that the mutation has likely not affected ligand binding but instead interfered with channel activation ( Fig 3H and 3I ) . Similar to the phenylalanine in the β6-β7 loop , the mutation of an equally conserved aspartate in the same region ( Asp122 in ELIC and Asp121 in GLIC ) results in an apparent loss of activation in both proteins ( Fig 2E and 2F ) . This residue is located just above the interface and forms a salt bridge with a conserved arginine ( Arg199 in ELIC and Arg191 in GLIC ) at the end of β-strand 10 at the boundary to the pore domain ( Fig 4A and 4B and S3B Fig ) . The mutation of the respective arginine to alanine also causes a nonactivating phenotype in ELIC , whereas no expression of this mutant was observed in GLIC ( Fig 2C–2F ) . In GLIC , Arg191 also interacts with Asp31 on β-strand 2 , a position that , with respect to its negative charge , is conserved in many pLGIC subunits but not in ELIC where the respective residue is a threonine and the glutamate-gated chloride channel ( GluCl ) from C . elegans where it is a valine ( S3B Fig ) . In GLIC , the mutant D31A is well expressed but shows no activity at pH4 ( S2B Fig ) . In ELIC , the equivalent mutant T28A can be activated but with a 3 . 5-fold higher EC50 of agonist than WT ( Fig 4C and S6A Fig ) . When mutating Thr28 in ELIC to aspartate , thereby introducing a negative charge that is present in many family members , the EC50 of channel activation shifts to a 45 times lower agonist concentration ( 18 μM , Fig 4C and S6B Fig ) . A similar increase in the affinity was observed in ITC experiments , where the agonist binds with a Keff of 90 μM , a 90-fold decrease in concentration compared to WT , while the binding of the antagonist acetylcholine was unchanged ( Fig 4D and S5E Fig ) . Patch clamp recordings already show considerable basal activity of this mutant in the absence of ligand and increased currents upon ligand application ( Fig 4E ) , whereas no basal activity is observed in WT . The single channel conductance of the mutant is similar to WT , but the current density is lower in both two-electrode voltage clamp and patch clamp experiments ( Fig 4E , S6B Fig ) . All experiments suggest that this mutant stabilizes a high affinity state with respect to ligand binding , and we were thus interested whether it would be sufficient to change the crystallization behavior and allow us to determine the structure of ELIC in a different conformation . While we succeeded in obtaining crystals of the T28D mutant in the same conditions , they exhibited poorer diffraction than WT and only allowed us to collect data at 4 . 5 Å in the absence and 9 . 5 Å in the presence of ligand ( Table 1 ) . The structures indicate that , despite the drastic impact on the potency of the ligand , the mutant crystallized in the familiar nonconducting conformation , which further underlines the stability of this state in a crystalline environment ( S4C Fig ) . Collectively , our studies emphasize the importance of ionic interactions between three conserved residues located in the β1-β2 turn ( GLIC Asp31 , ELIC Thr28 ) , the β6-β7 loop ( GLIC Asp121 , ELIC Asp122 ) , and the pre-M1 region ( GLIC Arg191 , ELIC Arg199 , Fig 4A and 4B ) for channel activation and the stabilization of the open state . At the tip of the GLIC β1-β2 turn , immediately adjacent to Asp31 , Lys32 interacts with a strictly conserved proline in the M2-M3 loop ( Pro254 in ELIC and Pro246 in GLIC , Fig 5A and 5B ) . This interaction is also observed in the presumably open structures of GluCl , the GlyR , and the structures of a homopentameric GABAA receptor . Conversely , this interaction is not formed in the structures of ELIC , the ligand-free GluCl , the antagonist-bound GlyR and the structure of the 5-HT3 receptor , where the contact is broken ( S3C Fig ) . We thus suspected that this interaction might play an important role for the relay of conformational changes from the extracellular domain to the pore [14 , 28] . In all pLGICs of known structure , the interaction between the tip of the β1-β2 turn and the pore domain is mediated by the protein backbone , whereas the side chain of the respective residue points towards the channel lumen ( Fig 5A and 5B and S3C Fig ) . In contrast to other residues in the interaction interface , this position is not conserved and contains a lysine in GLIC and the 5-HT3 receptor and a hydrophobic amino acid in ELIC , GluCl and the GABAA receptor . When expressed in X . laevis oocytes , the respective alanine mutants in ELIC and GLIC can still be activated with similar EC50 values as the respective WT ( Fig 5C and 5D , S6C and S6D Fig ) , although with a lower maximal current response ( S2 , S6C and S6D Figs ) . Remarkably , even a deletion mutant of L29 in ELIC showed a comparable activation pattern ( Fig 5C and S6E Fig ) . Mutations of the conserved proline in the M2-M3 loop to alanine ( P246A in GLIC and P254A in ELIC ) resulted in channels that , apart from a small shift in the EC50 in the case of GLIC , show robust activation with similar properties as WT ( Fig 5E and 5F , S6F and S6G Fig ) . In line with the comparably small effect in functional experiments , the crystal structure of the GLIC mutant P246A is virtually identical to WT ( S7A and S7B Fig , Table 2 ) . To avoid residual side chain interactions with the ligand-binding domain that may still be present after replacing the proline by alanine , we next investigated the mutation of the respective proline residue in both channels to glycine . In ELIC , the mutation P254G has a small but significant effect on the structure , which overall shows the frequently observed nonconducting conformation . The reorganization of the well-structured M2-M3 loop indicates a change in the conformational properties of this region ( Fig 5G , S4D Fig , Table 2 ) . The equivalent mutation P246G in GLIC results in large rearrangements when compared to WT . The structure determined at 3 . 2 Å shows a molecule with small differences throughout most of the protein except for the M2-M3 loop and the pore-lining helix M2 , which both have undergone major conformational changes ( Fig 5H , S7C Fig , Table 2 ) . The introduced conformational freedom upon replacement of the restrained amino acid proline with the flexible glycine has resulted in the rearrangement of the M2-M3 loop and the unfolding of the C-terminal part of M2 . The remainder of the helix has collapsed towards the pore axis as to maximize the hydrophobic interactions of residues at the extracellular part leading to a structure , which , most probably , prevents the permeation of ions ( Fig 5H , S7D and S7E Fig ) . In contrast to the large conformational change in the extracellular part of the helix , its conformation at the intracellular half remained unchanged . Remarkably , this pore conformation is very similar to structures of GLIC obtained from cysteine crosslinking of residues at the domain interface , which were previously assigned to a locally closed conformation of the ion conduction path [33] , the structures of two nonactivating mutants in the M2-M3 loop [34] and a recent structure of GLIC at neutral pH [20] . In this locally closed conformation , an interaction of a histidine residue in the pore-forming helix M2 with the backbone of the neighboring helix M3 established in the low pH crystal structure of GLIC is broken ( S7F Fig ) . It is noteworthy that the histidine is in a position that in other family members of known structure is predominantly hydrophobic , and that it was previously proposed to be involved in the pH-dependent activation of GLIC [35 , 36] . Despite the strong impact of the mutation on the structure , both proteins can still be activated ( Fig 5E and 5F , S6G–S6J Fig ) . When studied by two-electrode voltage clamp electrophysiology , the GLIC mutant P246G shows dose-dependent channel activation with an EC50 that is shifted by 0 . 5 pH units towards higher proton concentrations ( Fig 5F and S6H Fig ) . In ELIC , the mutant P254G shows agonist-induced currents with a similar EC50 as WT but with a slower activation and an unusually slow deactivation of the channel upon washout of the ligand ( Fig 5E , S6I Fig ) . This behavior can be observed in two-electrode voltage clamp recordings , where the activity of the protein after a change to ligand-free solution decays slowly ( S6J Fig ) , and it becomes even more pronounced in macroscopic recordings of excised outside-out patches ( Fig 5I , S8A–S8C Fig ) . The cause for this unusual deactivation phenotype was revealed in the structure of the P254G mutant of ELIC . In this structure , electron density between Arg255 on the M2-M3 loop and Glu155 located in the β8-β9 loop of the extracellular domain of an adjacent subunit indicates the formation of a strong ionic interaction that is absent in WT and that may stabilize the open conformation of the pore ( Fig 5G , S4D Fig ) . In the background of the mutation R255A , the kinetics of channel deactivation of the P254G mutant becomes similar to WT , thus confirming the role of the interaction for the unusual functional behavior ( Fig 5I and S8D Fig ) . The single mutation of R255A behaves similar to WT but shows faster activation and deactivation kinetics and an increased rate of desensitization ( S8E–S8G Fig ) . Thus , to our surprise , the mutation of a conserved proline in the M2-M3 loop to glycine still promotes activation in both channels , ELIC and GLIC . These results are in contrast to the much more drastic effects that are observed in mutants of other residues at the domain interface , including several positions in the same region , where even the truncation of the side chain to alanine has apparently prevented channel activation .
We have used a mutagenesis approach to investigate the role of the domain interface of two prokaryotic pLGICs for the transduction of conformational changes from the extracellular to the pore domain . Our study has revealed a pattern of corresponding residues in ELIC and GLIC that , if mutated to alanine , had a similar effect on either channel . Remarkably , whereas mutations in several positions have apparently prevented activation in both proteins , none of the investigated alanine mutations showed detectable basal activity in the absence of agonists . Our results suggest that the respective side chain truncations may have either stabilized a closed conformation of the channel , where the energy of ligand binding is no longer sufficient for activation , or alternatively , that they have interfered with the coupling of both domains and that the pore region that is uncoupled from the extracellular domain resides in a stable nonconducting conformation . In all cases , the mutations likely did not interfere with ligand binding but instead impeded gating , as suggested by calorimetry experiments of two nonactivating mutants of ELIC , which showed WT-like binding properties of agonists and antagonists ( Fig 3H and 3I ) . In case of WT , the observed agonist binding affinity is lower than expected from the EC50 value measured by electrophysiology [9] and instead matches the binding affinity to the resting state obtained from single channel analysis [29] ( Fig 3F ) . It thus appears that in detergent solution , ELIC resides in a single conformation that , with respect to ligand binding , resembles a resting state . Assuming that the conformation that is probed by calorimetry is also observed in the crystal , since in both cases the protein is solubilized in the same detergent , it is unlikely that the structure of ELIC represents a desensitized state with high affinity for the ligand . The conformational rigidity of solubilized ELIC is in accordance with the fact that all currently available structures show the same nonconducting conformation of the protein , irrespectively of whether agonist is bound or mutations were introduced that have stabilized the open state as it is the case for the mutations T28D or P254G . This behavior is in line with previous observations for ELIC [9 , 37] . Additionally , a reduced conformational freedom in detergent solution was observed in an electron paramagnetic resonance ( EPR ) spectroscopy study of GLIC [38] , which suggests that both proteins may require a lipid environment for full activation . Whereas the predominantly local effects of most mutations , which modulate the open to closed equilibrium of the channel , resembles the behavior of eukaryotic receptors [39] , the mutation of a conserved proline ( Pro254 ) to glycine in ELIC resulted in the formation of novel interactions between residues apart from the site of mutation that were not present in WT , causing an unusual functional phenotype that would have been difficult to explain in the absence of structure ( Fig 5G ) . Our studies show that mutations with a nonactivating phenotype predominantly cluster in two regions of the protein , the β6-β7 loop of the extracellular domain and the M2-M3 loop of the pore . The results are in accordance with a previous investigation of two mutations in the M2-M3 loop of GLIC [34] , and they overall mirror the functional behavior of eukaryotic pLGICs [40–43] . Differences between pro- and eukaryotic channels may originate from an altered energetic relationship between distinct states , and effects may generally be less pronounced if mutations only concern one or two subunits of a heteropentameric receptor . An early study has identified a mutation in the M2-M3 loop of the homopentameric α7 nAChR , which prevents activation but not ligand binding [44] . A similar phenotype was found for a mutation of the equivalent residue of the GlyR causing Startle disease , as well as by an additional mutation located two residues upstream [45] . Based on these observations , an involvement of the M2-M3 loop in gating has been proposed [44–46] . Similarly , our study has shown that mutations of the corresponding positions in the two prokaryotic channels ( T248A , Y250A in GLIC and L256A , Y258A in ELIC ) have interfered with activation but not ligand binding , as suggested by the calorimetry experiments of the ELIC mutant Y258A . The same region was also investigated in the hetero-pentameric muscle nAChR . Based on the effect of mutations on the equilibrium and kinetics of the open to closed transition , a prominent role of the M2-M3 loop of the α-subunits on the activation of the channel was postulated , but in this case the positions with the strongest phenotype differ from the residues identified in this study [39 , 47 , 48] . Whereas the kinetics of ELIC and GLIC is comparably slow [9 , 29 , 49] , we found that the point mutation R255A in the M2-M3 loop of ELIC not only accelerated activation and deactivation but also increased the rate of desensitization ( S8E–S8G Fig ) . In that respect , it is interesting to note that alanine is found in the same position of the fast desensitizing α7 nAChR and that a mutation of the corresponding residue of the GlyR causes Startle disease [50] . Since in a different study the domain interface was shown to influence the desensitization rate of homomeric pLGICs [51] , it appears that the same region determines activation and desensitization of the channels . Like the M2-M3 loop , the β6-β7 loop of eukaryotic pLGICs has also been proposed to play a critical role in channel activation . Mutations in equivalent positions that interfered with activation in both prokaryotic channels decreased the agonist response of the α1 GlyR and abolished the potentiation of currents by general anesthetics [52] . In a different study , the activation in chimeras of the α7 nAChR and the GlyR was enhanced by point mutations of residues of the β6-β7 loop that are in equivalent positions as nonactivating mutants in ELIC and GLIC [53] . A strong effect of mutations on the gating equilibrium constant was also found in the α-subunit of the nAChR [39 , 43 , 47] , and strong energetic coupling of two conserved phenylalanines of the β6-β7 loop to the M2-M3 loop was proposed based on mutant cycle analysis [54] . In ELIC and GLIC , a nonactivating phenotype was found for mutations of a conserved aspartate of the β6-β7 loop ( Asp122 in ELIC and Asp121 in GLIC ) that interacts with an equally conserved arginine in the pre-M1 region ( Arg199 in ELIC and Arg191 in GLIC ) . Mutations of the corresponding aspartate also interfered with activation in the nAChR [43 , 55] , the 5-HT3 receptor [56] or the GlyR [57] . In GLIC , the pre-M1 arginine bridges the β6-β7 loop with the β1-β2 turn by interaction with a negatively-charged residue ( Asp31 ) that is found in most pLGIC subunits ( Fig 6A , S3B Fig ) . The mutation of this residue to alanine prevents activation of GLIC , whereas the mutation of a threonine residing at the equivalent position in ELIC ( T28A ) appears to remain functional , although with decreased potency of the agonist ( Fig 4C , S2 Fig ) . The role of this interaction in channel activation and the destabilization of the resting state is underlined by a mutation of the respective threonine to aspartate in ELIC , which strongly increases the potency for the agonist and where the channel shows increased basal activity , which is not observed in WT ( Fig 4 ) . This basal activity demonstrates that ELIC can , in principle , also open in the absence of agonist and thus underlines the validity of the MWC model also for this channel . Equivalent ionic interactions have previously been investigated in the nAChR and other pLGICs [54–56 , 58 , 59] . In one study , they were postulated to be part of a molecular pathway that plays an important role in the relay of signals from the extracellular domain to the pore region , thereby connecting ligand binding to gating [54 , 59] . Since similar but smaller effects were found in a different study , the central importance of this ionic interaction for channel activation was questioned [58] , and it was instead proposed that the total charge of the interface rather than specific pairwise interactions may govern channel activation in the nAChR [55] . Contrary to our previous expectation [14 , 28] , conformational changes appear not to be predominantly transduced via an interaction of the tip of the β1-β2 turn to a conserved proline in the M2-M3 loop of the pore domain ( Pro254 in ELIC and Pro246 in GLIC ) , as mutations of this residue to alanine and glycine still permit channel activation . This is remarkable since the equivalent proline was proposed to play a prominent role in the early events of gating in the nAChR [39 , 48] , and since in several structures of different pLGICs assigned to potentially open conformations , this interaction is present , whereas it is broken in presumably nonconducting conformations ( S3C Fig ) . A coupling of the β1-β2 turn to a proline of the M2-M3 linker in the nAChR was previously proposed based on a model of the receptor from electron microscopy data at 4 Å [10 , 54 , 59] . However , due to a mismatch in the structural interpretation of this region , this position does not coincide with the residue investigated in this study . The structures of different pLGICs at high resolution [12–21] provide a framework for the comprehension of the results of this study . With respect to the transmembrane domain , most known structures cluster around three distinct conformations: A presumably conducting state of the pore , which has initially been observed for a low pH crystal form of GLIC [14 , 15] , is shared by GluCl and the GlyR in complex with their agonists and the allosteric modulator ivermectin [12 , 16] , as well as the GABAA receptor [17] ( S9 Fig ) . It is still debated whether these structures correspond to conducting , partially conducting , or even desensitized states [22] . A structure of the GlyR in complex with glycine has a larger pore diameter at the intracellular part of the transmembrane domain but shares a very similar pattern of interactions at the domain interface [16] . Conformations resembling the nonconducting state of ELIC were later observed for GluCl crystallized in the absence of ortho- and allosteric agonists , a conformation of the ligand-free protein in complex with a bound lipid [19] , and for the GlyR bound to its competitive antagonist strychnine [12 , 21] ( S9 Fig ) . This is remarkable in light of the controversy concerning the relationship of the ELIC structure to a resting conformation of the receptor [22 , 37 , 38 , 60] . The structure of the 5-HT3 receptor is in between the two previously described states but closer to the GLIC-like conformation [18] ( S9 Fig ) . Another distinct nonconducting conformation of the pore region was observed in a high pH crystal form of GLIC [20] and in several mutants of the same channel [33 , 34] , including the structure of the mutant P246G determined in this study ( S9 Fig ) . This third conformation has thus far only been observed in GLIC , a member of the family that is activated by protons , and it remains to be shown whether a similar conformation of the pore region can also be adopted by other family members . The fact that side chains that are truncated in nonactivating mutants point into a common core suggests that the mutation may have disrupted a critical interaction ( Fig 2A and 2B and S3A Fig ) . These interactions appear to be most extended in the presumably open GLIC-like conformations ( Fig 2A and Fig 6 ) . In these cases , the interface between the extracellular domain and the pore region is tightly packed , residues from the M2-M3 loop are in close contact with residues of the β6-β7 loop , and in several structures a negatively charged residue in the β1-β2 turn interacts with a conserved arginine in the pre-M1 region and an equally conserved proline in the M2-M3 loop ( Fig 6A and 6B , S3B and S3C Fig ) . This network is partially disrupted in the two nonconducting conformations , which might explain why several mutations in the interaction interface stabilize a closed state of the channel . In the locally-closed high pH structure of GLIC , the distance between the β1-β2 turn and the arginine in the pre-M1 region has increased and , due to a change of the conformation of the M2-M3 loop leading to a collapse of the pore-forming helix M2 , the contact to the proline in the respective region is broken ( Fig 6C , S3C Fig ) . This conformational change also causes an interruption of interactions between residues in the N-terminal part of the M2-M3 loop with the β6-β7 loop , whereas the interactions of the residues in the region preceding M3 appear less affected ( Fig 6C ) . In the ELIC-like conformations , a similar pattern of interactions is observed , but in this case , the disruption of interdomain contacts is due to a concerted move of helices M2 and M3 while preserving the conformation of the M2-M3 loop ( Fig 6D ) . The accompanying rearrangement of the β1-β2 turn of the extracellular domain weakens its interaction with the pre-M1 region and disrupts the contact to the proline in the M2-M3 loop . In both nonconducting conformations , the interactions with the C-terminal part of the M2-M3 loop with the β6-β7 linker remain intact , and the observed nonactivating phenotype of respective mutations ( F116A and Y258A in ELIC and F115A and Y250A in GLIC ) could thus originate from an interruption of the coupling between the extracellular domain and the pore region . Despite the plethora of structural information , definitive assignments of observed conformations to functional states of the receptors and the resulting activation mechanisms are still controversial [22] . It is thus interesting to observe that , regardless of the difference of agonists , highly conserved and closely interacting residues at the domain interface of ELIC and GLIC exert similar effects on activation in both prokaryotic ion channels . Critical interactions involve residues of the M2-M3 loop , the pre-M1 region and the β1-β2 turn that all contact the β6-β7 loop , whereas a direct interaction between the β1-β2 turn and the M2-M3 loop appears expendable . As in eukaryotic receptors , mutations at the interface predominantly affect the close to open equilibrium of the pore . Our results thus suggest that there is a common pathway for signal transduction in both proteins that , regardless of differences in the detailed energetic relationships between pro- and eukaryotic receptors , appears to be conserved within the entire family .
All expression constructs were cloned into vectors that were modified to be compatible with FX cloning [61] . For expression in X . laevis oocytes , WT and mutant open reading frames of ELIC and GLIC preceded by the signal sequence of the chicken α7 nAChR were cloned into a modified pTLN vector [62] . For surface expression analysis , the constructs contained an additional hemagglutinin-tag ( HA-tag ) attached to the N-terminus of the respective protein . For expression in human embryonic kidney 293 ( HEK293 ) cells , the respective genes preceded by the signal sequence of the chicken α7 nAChR were cloned into a modified pcDNA3 . 1 vector ( Invitrogen ) . For expression and purification in E . coli , the respective genes were cloned into a modified pET26b vector ( Novagen ) as constructs of the respective channels preceded by a fusion protein consisting of a pelB signal sequence , a His10-tag , maltose-binding protein and a human rhinovirus ( HRV ) 3C protease cleavage site . X . laevis oocytes were obtained either from Ecocyte or from an in-house facility . Animal procedures and preparation of oocytes followed standard procedures and were in accordance with the Swiss Cantonal and Federal legislation relating to animal experimentation . Plasmid DNA containing the genes coding for the respective constructs for expression in X . laevis oocytes were linearized by MluI , and capped mRNA was transcribed with the mMessage mMachine kit ( Ambion ) and purified with the RNeasy kit ( Qiagen ) . 10–200 ng of mRNA was injected into defolliculated X . laevis oocytes , which were subsequently incubated in Barth’s solution ( 88 mM NaCl , 1 mM KCl , 1 mM CaCl2 , 0 . 33 mM Ca ( NO3 ) 2 , 0 . 82 mM MgSO4 , 10 mM Na-Hepes ( pH 7 . 4 ) and 50 μg / ml Gentamycin ) and stored at 16°C . One to three days after injection , two-electrode voltage clamp measurements were performed at 20°C ( OC-725B , Warner Instrument Corp . ) . For ELIC , maximal currents were recorded in a bath solution containing 10 mM Hepes ( pH 7 ) , 130 mM NaCl , 2 mM KCl , 0 . 5 mM CaCl2 and either 5 mM or 25 mM Cysteamine . For GLIC , the maximal currents were recorded in a bath solution containing 10 mM Citrate ( pH 4 ) , 130 mM NaCl , 2 mM KCl , 1 . 8 mM CaCl2 and 1 mM MgCl2 . Dose-response experiments were carried out at agonist concentrations indicated in the respective figures . Voltage was clamped at −40 mV , and data was filtered at 20 Hz unless stated otherwise . Surface expression in constructs containing an HA-tag was assayed after electrophysiological characterization as described [63] . For that purpose , the oocytes were placed in a 96-well plate ( TPP ) and incubated in ND96 solution ( 93 . 5 mM NaCl , 2 mM KCl , 1 . 8 mM CaCl2 , 2 mM MgCl2 and 10 mM Hepes , pH 7 . 4 ) containing 1% BSA for at least 30 min . All steps were carried out at 4°C with the same buffer unless mentioned otherwise . Oocytes were subsequently transferred into buffer containing 1 μg/ml rat monoclonal anti-HA antibody ( 3F10 , Roche ) for 1 h , washed 3 times and incubated with buffer containing 0 . 16 μg/ml horseradish peroxidase ( HRP ) coupled to a secondary antibody ( HRP-conjugated goat anti-rat F ( Ab ) 2 fragments , Jackson ) for 30–60 min . The oocytes were washed 5 times with ND96 solution and subsequently transferred to a white 96-well plate ( flat bottom , Nunclon Delta Surface ) . The solution was aspirated , 30 μl of Super Signal ELISA femto solutions 1 and 2 ( Pierce ) was added , and luminescence was quantitated with a Tecan infinite M1000 plate reader . X . laevis oocytes were transferred to a hyperosmotic solution to manually remove the vitelline layer . Excised membrane patches were subsequently recorded in the outside-out configuration 3–5 d after injection of mRNA with an Axopatch 200B amplifier ( Axon Instruments ) at 20°C at −80 mV . Data was sampled at 20 kHz and filtered at 2 kHz and analyzed using Clampfit ( Axon Instruments , Inc . ) . Bath solutions contained 10 mM HEPES , pH 7 . 0 , 150 mM NaCl , 0 . 2 mM CaCl2 and indicated concentrations of ligands . Electrodes had a resistance of 3–5 MΩ . Pipette solutions contained 150 mM NaCl , 10 mM EGTA , 5 mM MgCl2 and 10 mM HEPES , pH 7 . 0 . Bath electrodes were placed in 1 M KCl solution connected to the bath solution by Agar bridges . Freshly prepared agonist solutions were applied to the patch using a stepper motor ( SF77B Perfusion fast step , Warner ) . HEK293 cells ( American Type Culture Collection-CRL-1573;LGC Promochem ) were maintained at 37°C in a 95% air/5% CO2 incubator in DMEM supplemented with 0 . 11 g/l sodium pyruvate , 10% ( v/v ) heat-inactivated fetal bovine serum , 100 U/ml penicillin G , 100 μg/ml streptomycin sulfate , and 2 mM L-glutamine ( Invitrogen ) . Cells ( passaged every 2 d , up to 30 times ) were plated and transfected by calcium phosphate-DNA coprecipitation [64] , with a total amount of DNA of 3 μg/dish ( 82% ELIC and 18% eGFP DNA , both subcloned in pcDNA3 . 1 ) . Cells were bathed in an extracellular solution containing 150 mM KCl , 0 . 2 mM CaCl2 and 10 mM HEPES , pH 7 . 4 . Patch pipettes were pulled from thick-walled borosilicate glass ( GC150F; Harvard Apparatus ) and fire polished to a resistance of 8–12 MΩ . Intracellular solution contained 150 mM KCl , 0 . 5 mM CaCl2 , 5 mM EGTA and 10 mM HEPES , pH 7 . 4 . Agonist-evoked currents were recorded at 20°C with an Axopatch 200B amplifier ( Molecular Devices ) from outside-out patches at −50 mV . No correction for junction potential was applied ( calculated value 0 . 2 mV ) . Data was sampled at 10 kHz and filtered at 1 kHz and analyzed using Clampfit ( Axon Instruments , Inc . ) . All concentration jumps were performed using a piezo stepper ( Siskiyou ) with an application tool made from theta tube glass ( Hilgenberg; final tip diameter , 150 μm ) . Agonist solutions were freshly prepared before measurements . ELIC WT and point mutants were expressed and purified as described [9 , 13] . BL21-DE3 cells transformed with a pET26b vector carrying the respective expression constructs of ELIC were grown in M9 minimal medium containing 50 mg/l kanamycin at 37°C to an OD600 of 1 . 0 and subsequently cooled to 20°C . Expression was induced by addition of 0 . 2 mM IPTG and carried out overnight . BL21-DE3 cells transformed with a pET26b vector carrying the respective expression constructs of GLIC were grown at 37°C in TB medium containing 50 mg/l kanamycin to an OD600 of 1 . 6–1 . 8 . Expression was induced by addition of 0 . 2 mM IPTG overnight at 20°C . All following steps were performed at 4°C . ELIC was extracted from isolated membranes in a buffer containing 1% n-Undecyl-β-D-Maltoside ( UDM , Anatrace , Inc . ) and further purified in buffers containing 0 . 145% UDM . GLIC was extracted from isolated membranes in a buffer containing 1% n-Dodecyl-β-D-Maltoside ( DDM , Anatrace , Inc . ) and further purified in buffers containing 0 . 044% DDM . Both proteins were purified by Ni-NTA chromatography ( Qiagen ) and digested with HRV 3C protease to cleave the His10-MBP fusion tag . His10-MBP and 3C protease were subsequently removed from solution by binding to Ni-NTA resin and the flow-through was concentrated and subjected to gel-filtration on a Superdex 200 column ( GE Healthcare ) . The protein peak corresponding to the ELIC pentamer was pooled , concentrated to 10 mg/ml and used for crystallization and ITC . The protein peak corresponding to the GLIC pentamer was pooled and concentrated to 10 mg/ml , and used for crystallization . Both proteins were crystallized in sitting drops at 4°C as described [9 , 13] . ELIC containing additional 0 . 5 mg/ml E . coli polar lipids ( Avanti Polar Lipids , Inc . ) was mixed in a 1:1 ratio with reservoir solution composed of 200 mM ( NH4 ) 2SO4 , 50 mM ADA , pH 6 . 5 and 10%–13% ( w/v ) PEG4000 . GLIC containing additional 0 . 5 mg/ml E . coli polar lipids was in mixed in a 1:1 ratio with reservoir solution composed of 225 mM ( NH4 ) 2SO4 , 50 mM sodium acetate , pH 4 . 0 and 9%–12% ( w/v ) PEG 4000 . The crystals were cryoprotected by transfer into solutions containing additional 30% ethylene glycol . All data sets were collected on frozen crystals on the X06SA beamline at the Swiss Light Source ( SLS ) of the Paul Scherrer Institute ( PSI ) on a PILATUS detector ( Dectris ) . The data were indexed , integrated , and scaled with XDS [65] and further processed with CCP4 programs [66] ( Tables 1 and 2 ) . The structure of mutants was determined by molecular replacement in PHASER [67] using either the ELIC pentamer in a P43 crystal form ( 2YN6 ) or the GLIC pentamer ( 3EHZ ) as search model . The models were rebuilt in Coot [68] and refined maintaining strong NCS restraints in PHENIX [69] . R and Rfree were monitored throughout . Rfree was calculated by selecting 5% of the reflection data in thin slices that were selected for the initial datasets of ELIC and GLIC and that were omitted in refinement . For low resolution data of the ELIC mutant T28D , refinement was restricted to rigid body refinement followed by few cycles of restrained positional and group b-factor refinement . The pore radii were calculated with HOLE [70] . Binding of the agonist propylamine and the antagonist acetylcholine to ELIC was measured by ITC with a MicroCal ITC200 system ( GE Healthcare ) . The syringe was loaded with agonist solution containing between 30–37 mM propylamine or acetylcholine dissolved in measurement buffer ( 25 mM Hepes , pH 7 . 0 , 150 mM NaCl and 0 . 9 mM UDM ) . The sample cell was loaded with 300 μl of purified ELIC in measurement buffer at a concentration between 80–110 μM . Agonist was applied by sequential injections of 2 μl aliquots followed by a 180 s equilibration period after each injection . The data was recorded at 4°C . For analysis , the heat released by each injection was integrated , and the background was subtracted with NITPIC [71] . The background-corrected data was analyzed by a fit to a single-site binding isotherm with the Origin ITC analysis package . ITC experiments were performed at least twice for each protein , with similar results .
|
The pentameric ligand-gated ion channels constitute a large family of membrane proteins that are expressed in animals and certain bacteria . Their molecular architecture and function is conserved throughout the family . In mammals , they operate as receptors of the neurotransmitters acetylcholine , serotonin , GABA , and glycine and play a key role in electrical signal transduction at chemical synapses . These receptors are called ionotropic because they open a selective ion conduction path across the membrane upon binding of the neurotransmitters to a site that is exposed to the extracellular medium . Ligand binding promotes a conformational change in the extracellular domain that is transmitted over more than 50 Å to the pore domain . Due to this long-range effect , pentameric ligand-gated ion channels have become important model systems for the study of allosteric processes , a mechanism that is of large importance for biology and entails the regulation of a protein activity by an effector that binds to a distant domain . In the present study , we investigated the role that residues in the contact region between the ligand-binding and the pore domains of two bacterial pentameric ligand-gated ion channels of known structure have in the transduction of conformational changes . Our study shows that single mutations severely influence the functional properties , with certain mutations preventing activation . The results underline the importance of highly conserved residues in the domain interface for the transmission of allosteric signals and thus likely apply also to other family members .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"crystal",
"structure",
"chemical",
"compounds",
"aliphatic",
"amino",
"acids",
"condensed",
"matter",
"physics",
"vertebrates",
"organic",
"compounds",
"animals",
"xenopus",
"germ",
"cells",
"animal",
"models",
"mutation",
"oocytes",
"model",
"organisms",
"amphibians",
"amino",
"acids",
"protein",
"structure",
"crystallography",
"research",
"and",
"analysis",
"methods",
"alanine",
"nicotinic",
"acetylcholine",
"receptors",
"solid",
"state",
"physics",
"animal",
"cells",
"proteins",
"chemistry",
"cyclic",
"amino",
"acids",
"transmembrane",
"receptors",
"proline",
"molecular",
"biology",
"physics",
"acetylcholine",
"receptors",
"biochemistry",
"signal",
"transduction",
"point",
"mutation",
"cell",
"biology",
"ova",
"organic",
"chemistry",
"genetics",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"frogs",
"organisms",
"macromolecular",
"structure",
"analysis"
] |
2016
|
Signal Transduction at the Domain Interface of Prokaryotic Pentameric Ligand-Gated Ion Channels
|
Natural isolates of Burkholderia pseudomallei ( Bp ) , the causative agent of melioidosis , can exhibit significant ecological flexibility that is likely reflective of a dynamic genome . Using whole-genome Bp microarrays , we examined patterns of gene presence and absence across 94 South East Asian strains isolated from a variety of clinical , environmental , or animal sources . 86% of the Bp K96243 reference genome was common to all the strains representing the Bp “core genome” , comprising genes largely involved in essential functions ( eg amino acid metabolism , protein translation ) . In contrast , 14% of the K96243 genome was variably present across the isolates . This Bp accessory genome encompassed multiple genomic islands ( GIs ) , paralogous genes , and insertions/deletions , including three distinct lipopolysaccharide ( LPS ) -related gene clusters . Strikingly , strains recovered from cases of human melioidosis clustered on a tree based on accessory gene content , and were significantly more likely to harbor certain GIs compared to animal and environmental isolates . Consistent with the inference that the GIs may contribute to pathogenesis , experimental mutation of BPSS2053 , a GI gene , reduced microbial adherence to human epithelial cells . Our results suggest that the Bp accessory genome is likely to play an important role in microbial adaptation and virulence .
Melioidosis is a potentially fatal infectious disease of humans and animals caused by the Gram-negative bacterium Burkholderia pseudomallei ( Bp ) [1] . An environmental saphrophyte found in South East Asia , Bp infections in endemic areas may be responsible for up to 20% of deaths due to septicemia [2] , [3] , and Bp has been designated a Category B biothreat agent [4] . A wide spectrum of disease symptoms are associated with melioidosis often leading to late diagnosis and treatment [5] . Commonly presenting as an acute septicemic illness , chronic Bp infection is also well recognized which can be confused with TB or malignancy [6] . Besides humans , Bp has a broad host range and can infect nematodes , amoebae , dolphins , birds , swine , sheep , and gorillas [7]–[11] . Bp can also be isolated from diverse environmental sources such as soil , water , and air [12]–[17] . Identifying the molecular factors responsible for this tremendous ecologic flexibility may improve our understanding of microbial survival and adaptation , and suggest novel diagnostic and treatment strategies for melioidosis . The phenotypic versatility of Bp is likely to be underpinned by the presence of a highly dynamic genome . For example , lateral gene transfer events may cause large-scale variations in genome content [18] . The portion of the genome that is variably present between individual strains is often termed the “accessory genome” , to distinguish these genes from genes common to all strains in a population and involved in essential functions ( the “core” genome ) . In several microbial species , accessory genes have been shown to play key roles in host adaptation and , in the case of Bp , the accessory genome may contribute to virulence and antibiotic resistance [19] . Interestingly , previous studies indicate that in Bp , gene loss , as well as gene acquisition events , can both cause phenotypic shifts towards virulence . For example , comparisons between Bp and B . thailandensis , an avirulent closely related species , have shown that an important evolutionary step in the development of Bp pathogenicity was the loss of an anti-virulence arabinose assimilation cluster [20] , [21] . Such findings thus raise a compelling need to accurately define the core and accessory genomes of Bp . In other γ proteobacteria genera ( E . coli , Pseudomonas , Vibrio ) , the accessory genome can encompass up to 20% of all genomic content , and similar percentages may also hold for Burkholderia spp . [22]–[25] . However , to date , comprehensive qualitative and quantitative studies of the core and accessory genome in Bp have not been carried out , and the full extent to which gene content differences contribute to virulence in Bp is still unclear . While some previous studies have attempted to explore these issues , they have not incorporated data from the whole genome [19] , [26]–[28] , or have used only a very small sample of strains [29] , [30] . In this study , we performed a detailed array-based comparative genomic hybridization ( aCGH ) analysis of close to 100 clinical , animal and environmental Bp isolates from South East Asia . To our knowledge , this is the first time a whole genome comparative study has been applied to such a large Bp strain cohort . We found that 86% of the reference Bp K96243 genome was present in all the strains , while the remaining 14% was variably present across the strain panel . Surprisingly , isolates associated with human melioidosis exhibited a tendency to harbor certain GIs compared to isolates from either animal or environmental sources , suggesting that genes on these mobile elements might facilitate colonization of the human host . Taken collectively , our results support the notion that the Bp accessory genome may play a central role in adaptation and virulence . Besides providing important evidence concerning genes likely involved in Burkholderia pathogenesis , this study also raises the possibility of targeting molecular diagnostics to specific Bp accessory regions for monitoring the presence of human-virulent variants in the environment .
Using a previously validated Bp K96243 DNA microarray [30] , [31] , we generated aCGH profiles for ninety-four Bp strains isolated from human patients , animals , and environmental soils in Singapore , Malaysia or Thailand ( Table S1 ) . We applied a Gaussian Mixture Model ( GMM ) to the aCGH data and identified 750 out of 5369 genes ( 14% ) as being variably present across the strain panel ( see Methods and Figure S1 ) . The variability of the 750 genes was experimentally validated by several independent methods , including bioinformatic comparisons to previously-known variable genes , comparisons against publicly available genome sequences , and experimental confirmation by targeted PCR assays ( Figure S2 and Table S2 ) . 86% of the Bp K96243 genes ( 4619 ) were found in all strains , representing the Bp core genome ( Figure 1 ) . Using pathway analysis , we found that the core genes were significantly over-represented in several functions necessary for basic bacterial growth and survival , including amino acid metabolism ( 1 . 52×10−3 ) , inorganic ion transport ( 3 . 96×10−3 ) , nucleotide metabolism ( 1 . 52×10−2 ) and protein translation ( 7×10−3 ) ( Table 1 ) . The core genes were also significantly enriched in genes conserved in other Burkholderia species ( Bp , B . mallei , B . thailandensis and B . cepacia ) ( p = 8 . 68×10−11 ) ( Text S1 and Table S3 ) ) , suggesting that a significant proportion of these Bp core genes may represent core genes in other related species as well [32] . Besides these basic housekeeping functions , the Bp core genes were also significantly enriched in commonly encountered virulence-related genes such as secretion proteins , capsular polysaccharides , exoproteins , adhesins , fimbriae and pili ( p = 1 . 8×10−3 ) ( Table 1 ) . For example , three Bp-specific fimbrial gene clusters ( BPSL1626-1629 , BPSL1799-1801 , BPSS0120-0123 ) were found in all strains . This finding suggests that most , if not all , Bp isolates are likely to possess a common ‘virulence machinery’ . Notably , many of these conventional virulence genes are also found in other related species such as B . thailandnesis that although non-infectious to mammals can kill other species such as nematodes [20] , [33] . This is consistent with the possibility that Bp might have descended from a pathogenic ancestor with a non-mammalian host . 14% of the Bp K96243 genome was variable across the strain panel , representing the Bp accessory genome . Since our analysis is confined to genetic elements present in the reference K96243 genome , the extent of genomic variability reported here should be regarded as a lower limit . The 750 variable genes were equally distributed between both Chromosome 1 and Chromosome 2 after normalizing for chromosome size differences . The accessory genes were significantly enriched in paralogous genes ( p = 2×10−7 ) and genes encoding hypothetical proteins ( p = 3×10−4 ) ( Table 1 ) . Approximately one-third ( 30 . 8% ) of the accessory genes were localized to a series of previously identified “genomic islands” ( GIs ) in the K96243 genome [34] . GIs are regions bearing unusual sequence hallmarks , such as atypical GC content and/or dinucleotide frequencies , and are likely to have been recently acquired by lateral gene transfer . Of sixteen GIs in the K96243 genome , fourteen GIs were represented by accessory genes . In contrast , two GIs ( 7 and 14 ) were found in all strains , suggesting that GIs 7 and 14 should be regarded as part of the Bp core genome . Besides the GIs , we also identified several novel regions of at least three contiguous probes that were absent in at least three strains . Henceforth referring to these regions as ‘indels’ , we identified eight indels on chromosome 1 , and twelve on chromosome 2 ( Table 2 ) . We experimentally validated two of these indels using PCR assays ( Figure S3 ) . The indels ranged in size from 1 . 3 to 7 . 5 kb , and were absent in 12 . 9% to 45 . 2% of strains ( Figure 2 ) . Three indels ( n1 , n4 and n11 ) were associated with atypical GC content ( 53 . 7–58 . 6% , compared to 68% for the Bp genome ) , and four ( n2 , n9 , n11 and n16 ) carried genes characteristic of mobile genetic elements such as integrases , transposases and bacteriophage-related genes , consistent with lateral transfer . These indels may therefore share similar dynamics to the larger genomic islands , and may be considered as genomic “islets” . In other species , analogous islets which are typically <10 kb long , have been shown to play a role in virulence ( e . g . the sifA islet in S . typhimurium ) [35] . Of note , n16 and n18 were flanked at both their 5′and 3′ends by tandem repeat sequences , while n4 , n6 , n8 and n19 possessed sequence repeats at either their 5′ or 3′ ends . In some cases , the islets in the Bp genomes may actually form part of the larger GIs . For example , n2 ( BPSL0741-BPSL0744 ) was located at the 5′ boundary of GI 4 ( BPSL0745-BPSL0772 ) , while n11 ( BPSS0395-BPSS0397 ) was located immediately 3′ to GI 13 ( BPSS0378-BPSS0391A ) . Three indel regions ( n6 , n12 and n19 ) contained genes associated with LPS metabolism . Lipolysaccharides ( LPS ) are macromolecular components on the outer membranes of Gram-negative bacteria composed of lipid A , core oligosaccharide , and O-antigen polysaccharides [36] . LPS molecules are commonly immunogenic and have been previously implicated in virulence for numerous microbes [37] , [38] . Region n6 ( BPSL2666-BPSL2668 ) contains a phosphoglucomutase ( BPSL2666 ) , a lipopolysaccharide LPS biosynthesis protein ( BPSL2667 ) and a glycosyltransferase ( BPSL2668 ) , and was located four genes away from a larger LPS biosynthesis cluster ( BPSL2672-BPSL2688 ) . Both regions n12 ( BPSS0427 - BPSS0429 ) and n19 ( BPSS2245-BPSS2255 ) contained two O-antigen related genes , including O-acetyltransferase and glycosyltransferase . While n12 corresponds to a previously identified type III O-PS polysaccharide gene cluster [39] , the contribution of n19 genes to Bp LPS biology is currently unknown . The identification of three physically unlinked indels related to LPS metabolism provides a mechanism by which high levels of LPS diversity may be maintained in the Bp population [40] . To explore if differences in accessory genome content might be associated with host adaptation or the propensity to cause disease , we applied unsupervised clustering to cluster the strains using the entire set of 750 accessory genes ( “accessory genome clustering” , AGC ) . We identified three large AGC clusters each containing 27 to 42 strains , with each cluster containing at least 4–6 sub-branches ( Figure 3 ) . Most strikingly , the majority of human clinical isolates ( 73 . 1% ) fell into one AGC cluster ( Clade C ) , another cluster contained 73 . 7% of the animal isolates ( Clade A ) , and a third cluster contained 45% of the environmental isolates ( Clade E ) . Similar results were obtained when the clustering was repeated using either Chromosome 1 or Chromosome 2 accessory genes ( Figure S4 ) . The over-representation of human clinical isolates in the C clade was highly significant ( P = 2 . 001×10−14 , Fisher's exact test ) , and of the remaining 13 clinical isolates nine segregated within the E clade and four in the A clade . This clustering pattern is unlikely to represent differences in geographical distribution , since the majority of the clinical ( 65% ) , animal ( 89% ) and environmental isolates ( 80% ) were isolated in Singapore within a ∼700 km2 region or from nearby islands . Furthermore , clinical isolates from Thailand clustered with the other clinical isolates , despite being geographically remote . This analysis therefore suggests that strains associated with human melioidosis may possess an accessory genome distinct from most animal and environmental strains . We also note that all three clades contained environmental isolates , which is consistent with the view that the environment represents a diverse reservoir from which human and animal adapted strains emerge . We then performed a supervised analysis to identify which of the 750 accessory genes were significantly different between the C and A/E clades . Of the 750 genes , 218 genes were commonly present in isolates in the C clade but absent from strains in the other two clusters ( Figure 4A ) . Strikingly , we found that almost all of these 218 genes ( 85% ) were localized to the GIs , with all fourteen GIs being represented . This figure ( 85% ) is significantly higher than the 31% of all accessory genes located on GIs , raising the possibility that GIs may play an important role in determining ecological niche and host adaptation . Is there any direct evidence that genes encoded on GIs , and which define the C clade , might play an important role in the biology or pathogenicity of Bp ? Unfortunately , almost 35% of the GI genes encode ‘hypothetical’ proteins ( Table S4 ) , meaning that their function is unknown . For those genes specific to the C clade where functions could be assigned , several broad functional classes were represented . For example , GI8 contains several genes spermidine/putrescine transport genes ( potB , potC , potG ) , which have been associated with biofilm formation and the regulation of Type III secretion genes [41] , [42] . Type I restriction-modification enzymes are found on GI5 and GI10 , and a glutathione S-transferase gene ( BPSS2048 ) on GI16 may impart resistance to oxidative stress . Also supporting their potential role in Bp biology , several GI genes exhibited distinct and complex gene expression patterns during Bp growth ( Text S2 ) . However , the role of such genes in pathogenesis remains speculative . In order to explore this further , we generated an experimentally mutated strain ( ATS2053 ) disrupted in BPSS2053 , a GI 16 gene encoding a hemagglutinin-related protein , and determined the adherence of the mutant strain to human buccal epithelial cells . A highly significant reduction in the adherence to buccal epithelial cells was noted between the 1026b clinical isolate and the isogenic ATS2053 mutant strain ( mean adherence: 1026b - 16 . 3±3 . 2 vs ATS 2053 - 4 . 4±1 . 7 , p<0 . 001 , Students t test ) . This finding provides evidence pointing both to the biological relevance of GI genes , but more specifically to a role of these genes in virulence . Finally , we examined the concordance between strain clusters defined on the basis of accessory gene content and the phylogenetic signal within the Bp core genome . We characterised 45 representative isolates by Multilocus Sequence Typing ( MLST ) , a typing scheme that indexes variation at seven core housekeeping genes [43] . Using the previously published Bp scheme [44] , we resolved the 45 isolates into 9 sequence types ( ST 46 , 51 , 54 , 84 , 169 , 289 , 414 , 422 and 423 ) . Seven of these STs ( ST51 , 54 , 84 , 46 , 169 , 289 , 414 ) have been previously observed in Malaysia , Thailand , and Singapore and two ( ST422 and 423 ) are specific to Singapore [44] , [45] . Previous analyses of MLST for Bp have highlighted the difficulties in building robust phylogenetic trees for this species , owing to a paucity of informative sites in the concatenated data and frequent homologous recombination [46] . We thus favored a categorical approach to comparing the AGC and MLST data by examining the distribution of sequence types across the three clades defined by the AGC data ( Table 3 ) . This analysis revealed that the STs are not randomly distributed between the three clusters , indicating some consistency between the MLST and AGC datasets . Most strikingly , of the 20 ST51 isolates , 17 clustered within the animal-associated clade ( A ) , three within the clinical C clade , and none in the environmental E clade . Of the other STs where at least 4 isolates were observed , all four ST422 isolates corresponded to the C clade , and all four ST84 isolates clustered within the E clade . Finally , of the nine ST423 isolates , five clustered within the C clade and four in the E clade . These data suggest that the animal-associated clade is likely to correspond to a single clone ( ST51 ) and provides some evidence for concordance between STs 422 and 84 with the AGC data , although the evidence in these latter cases is equivocal due to the small number of strains . In contrast , the “split” of the ST423 isolates between the clinical and environmental clades , and the 3 ST51 isolates belonging to the clinical clade , represent clear discrepancies between the two datasets . Possible explanations for these discrepancies may represent convergence of either the MLST or the AGC data , as discussed below .
In this report , we present a comprehensive aCGH analysis for a large series of natural Bp isolates . We found that the accessory ( variably present ) portion of the Bp genome corresponds to ∼14% of the whole genome content , which is broadly similar to other γ-proteobacteria . Since this approach is limited to the detection of elements present in the Bp K96243 genome , and novel elements in query genomes are not detected , this estimated fraction of the accessory genome should be regarded as a lower bound . In the only published study of a Bp genome sequence to date , Holden et al ( 2004 ) computationally identified 16 GIs comprising 6% of the K96243 genome [34] , and our data confirm that most of these islands are indeed highly variable between strains . However , two GIs ( 7 and 14 ) were found in all strains and should thus be regarded as part of the Bp core genome . Furthermore , our data also revealed the variable presence of several other small genomic islets/indels across the two chromosomes , which might contribute to the phenotypic diversity of Bp . Notably , we observed that several indels ( n6 , n12 and n19 ) were related to LPS biology . Currently , the exact contribution of LPS to Bp virulence is unclear . For example , DeShazer et al ( 1998 ) showed that Bp type II O-PS is essential for serum resistance and virulence [47] , and mice pre-immunized with Bp LPS displayed enhanced survival to a subsequent challenge [48] . In contrast , other groups have reported that Bp LPS exhibits a reduced ability to activate immune cells compared to E . coli LPS , suggesting that LPS might play only a minimal role in Bp virulence . It is possible that these conflicting results might reflect heterogeneity in LPS pathways resulting from the variable presence of these indels , and represent an important mechanism for host adaptation . Interestingly , while it was recently shown that type III O-PS mutants ( indel n12 ) do not appear to exhibit significant virulence attenuation in mouse infection assays [39] , we have found in preliminary work that Bp strains lacking the indel n19 LPS cluster generally exhibited lower levels of virulence compared to strains where this cluster was present ( SSH , data not shown ) . In the AGC tree , n19 was absent both from three strains segregating as a single branch in the A clade , and from 5 strains in the C clade that segregated across multiple branches . This suggests that n19 may have been recurrently lost in different Bp lineages . Further experiments are clearly required to understand the role of these LPS clusters in Bp virulence . We also found that the Bp strains could be clustered into distinct clades based on both the presence and absence of specific accessory genes . Of primary interest , strains belonging to the C clade of clinical isolates were largely defined by the presence of 218 genes , of which 85% are localized to the GIs . These findings provide evidence for a distinct repertoire of Bp genes that may cause a predisposition to human disease and that these genes tend to be located on GIs . Although many of the genes encoded on the GIs are of unknown function , we present experimental evidence that a strain mutated in one of these genes exhibited decreased adherence to human buccal endothelial cells , supporting a role in virulence potential . We also observed coordinated growth-associated expression of several GI genes , which is also consistent with the view that they play an important biological role . What might this biological role be ? At present , we consider it most likely that this “virulent” combination of genes has likely emerged for reasons other than to cause human disease , particularly since cases of human ( or animal ) infection are relatively rare compared to the density of Bp in the soil . In contrast to bacteria which are obligately associated with eukaryotic hosts , soil bacteria such as Bp commonly face extreme and unpredictable biotic and abiotic challenges including extreme temperature shifts , solar radiation , variable humidity , competition for nutrients , and the requirement to survive ingestion by predatory protozoa , nematodes , the production of bacteriocides from other bacteria and phage infection . It thus seems entirely plausible that genes facilitating survival against these environmental challenges might have also indirectly enhanced the microbe's ability to colonize and “accidently” infect a human host , particularly when the host is immunocompromised [49] . Another possibility that might explain the enrichment of GIs in the clinical isolates is that Bp is undergoing cryptic cycling through normal human hosts ( as opposed to the immunodeficient host ) , and that these GIs are selected during this host-pathogen interaction . In melioidosis-endemic NE Thailand , the majority of healthy individuals have antibodies to Bp by the age of 4 years , indicating a constant exposure to the bacterium that may occur by inoculation , inhalation or ingestion [50] . Within these normal hosts , Bp is likely to spend a period of time being exposed to the effects of the host immune response , after which the microbe may experience bacterial death , persistence , or expulsion from the host in a viable state and subsequent return to the environment . This latter process might occur through skin desquamation or urine and stool , since human excrement commonly finds its way back to the environment . Such cryptic cycling of Bp through the normal human host population could also lead to the selection of factors that promote survival in vivo . However , as we consider the human host to be a relatively minor component of Bp ecology , we argue that this scenario is , on balance , less likely . The availability of both MLST and aCGH data for a representative sub-sample of isolates also provided us the opportunity to compare clade distributions defined either by accessory genome content or allelic variation in the core genome . We found that the animal associated strains largely corresponded to a single MLST clone ( ST51 ) . These isolates were assembled from three distinct sources: the Singapore zoo , the University of Malaya and a pig abbatoir in Singapore . The soil isolates corresponding to ST51 ( which also clustered in the A clade ) were not isolated from soil samples in proximity to the animal ST51 isolates , which suggests that this genotype is also present in the environment . The homogeneity of these isolates is therefore striking and cannot be explained simply by sampling bias . The consistency between the microarray and MLST data strongly suggest that this clade is monophyletic , and that the strains harbour similar gene repertoires by virtue of common descent . In contrast , we also observed clear discrepancies between the MLST and aCGH clades . For example , three ST51 isolates clustered within the clinical aCGH clade , and ST423 was split between the clinical and environmental aCGH clades . There are three possibilities to explain these discrepencies: i ) The MLST data represents the ancestral state which is inherited by descent into two AGC-defined clades - this is unlikely for the animal cluster as the vast majority of isolates are ST51 , but might conceivably explain the ST423 split between the clinical and environmental clades . ii ) Convergence of the MLST alleles - this would imply that isolates with the same ST are not identical by descent but happen to share the same combination of alleles . The presence of a few very common alleles for each gene , combined with high rates of recombination in Bp make this possibility more likely . iii ) Independent convergence of gene content to one of the three clusters . Unless large numbers of genes can be transferred in single events , this possibility seems less parsimonious than ( ii ) . More data are required to examine which of these hypotheses is most likely . In summary , our study provides direct experimental confirmation that the Bp genome is highly plastic , and that gene acquisition and deletion are major drivers of this variability . This variability is far from random , and is functionally biased towards genes involved in mobile elements , hypothetical and paralogous genes , and LPS biosynthesis . Furthermore , genes on mobile elements may predispose individual strains , either directly or indirectly , towards causing human disease . We believe this latter result is significant in that most Bp research to date has focused on virulence components in the Bp core genome rather than genes on mobile elements . We conclude by noting that most of the Bp genome sequences currently available have been obtained from human clinical isolates . Given our results , it might be highly informative to subject a panel of animal and environmental Bp isolates to similar detailed genome analysis as well .
Ninety-four Bp isolates were used in this study . These include: a ) the K96243 reference strain , b ) 52 clinical isolates from melioidosis patients between 1996 and 2005 , c ) 19 animal isolates from various species ( eg monkeys , pigs , birds , and dogs ) diagnosed with melioidosis between 1996 and 2000 , d ) 20 soil isolates from 1994 to 2003 , and e ) two type strains ( ATCC23343 and ATCC15682 ) . All strains were isolated in Singapore , neighboring islands , or surrounding countries ( Malaysia , Thailand ) . The isolates were sampled from a diversity of locations and not a single site , supporting their unbiased nature ( Aw Lay Tin and Joseph Tong , personal communication ) . Further strain information is provided in Table S1 . Strains were cultured on Tryptone Soy Agar ( TSA ) ( Difco Laboratories , Detroit , Michigan ) at 37°C , and genomic DNA extracted using a genomic DNA purification kit ( Qiagen ) . The Bp DNA microarray has been previously described [29]–[31] and comprises approximately 16 , 000 PCR-amplified array probes representing all 5742 predicted genes in the K96243 genome printed in duplicate . Test genomic DNA ( 2 µg ) was fluorescently labeled with Cy3-dCTP ( Amersham Pharmacia Biotech ) using nick-translation and co-hybridized to the array with an equal quantity of Cy5-dCTP ( Amersham Pharmacia Biotech ) labeled reference K96243 DNA . The absence of significant dye-bias artifacts was confirmed by analyzing reciprocal dye-swap hybridizations for 10 isolates data not shown , also see ref [29] . Raw fluorescence data was acquired using an Axon scanner with GENEPIX v4 . 0 software ( Axon Instruments , Redwood City , CA ) . Individual arrays were internally normalized between the Cy3 and Cy5 channels by LOWESS normalization , and the entire dataset was cross-normalized by median-scaling each array to the same Cy3/Cy5 ratio . To filter the microarray data , we eliminated probes exhibiting a missing value score across >40% of samples ( indicating that they were not reliably measured ) , and probes whose genomic loci were redundant with other probes . This data filtering procedure generated a final high-quality data set of 5369 non-redundant probes . The entire microarray data set is available at the Gene Expression Omnibus database under accession number GSE9491 . A Gaussian mixture model ( GMM ) [51] was used to identify accessory and core genes in the data set . In concept , a GMM fits a test signal distribution ( such as microarray data ) to either a single or double gaussian curve , and the likelihood that the distribution corresponds to a single curve is computed . The GMM was applied in two stages . First , p-values were computed using the aCGH profiles of each individual array spot , following a chi-square distribution with 3 degrees of freedom under the null hypothesis that the data distribution of the spot follows a 1-gaussian distribution . Second , since each probe was spotted twice on the array , we obtained composite p-values of each array probe using Inverse Chi-square Meta-Analysis [52] , squaring the p-values of both spots belonging to the same probe . This latter statistic follows a chi-square distribution with 4 degrees of freedom . All p-values were corrected for multiple-hypothesis testing according to the Benjamini-Hocheberg procedure [53] . A cut-off of p≤1 . 83E-08 was selected to define the top 750 most highly variable probes , representing the accessory genome . All protein coding sequences in the Bp K96243 genome were queried by BLASTP against the Cluster of Orthologous group ( COGs ) database , a public bioinformatic database that groups protein sequences on the basis of phylogenetic similarity to various cellular functions , such as protein translation , DNA replication and transcription , nuclear structure and defense mechanisms ( accessible at http://www . ncbi . nlm . nih . gov/COG/new/ ) . Matches were defined as database hits with an e-value threshold of <10−6 . Based on the COG assignments , the K96243 proteins were assigned to functional categories . Fisher's exact tests were used to identify significantly overrepresented COG categories in either the core or accessory genes . To identify conserved genes ( metagenes ) across four Burkholderia species , we queried the 3460 Chr 1 and 2395 Chr 2 ORFs in the Bp K96243 genome against the B . cenocepacia ( Bc ) , B . mallei ( Bm ) , and B . thailandensis ( Bt ) genomes using tblastn [32] ( Text S1 ) . To minimize the number of ambiguous predictions including ORFs with matches to multiple genomic locations , we constrained the resulting matches to have I ) a minimum length of 50 amino acids , II ) a minimal e-value cut-off of 1e-6 and III ) a minimum percent identity of 50% . Homology assignments returned 2675 genes and were validated by a reciprocal blast assay resulting in 2590 genes . Control analyses using either Bc , Bm or Bt as starting reference genomes yielded similar metagene sets ( data not shown ) . Paralogous genes were identified using the CD-HIT program [54] as genes with >60% identity to one another , following established studies [55] , [56] . Tandem repeat regions in the K96243 genome were identified using the Tandem Repeats Finder program [57] . Phylogenetic trees based on aCGH profiles were constructed using MultiExperiment Viewer ( MeV ) version 4 ( http://www . tm4 . org/mev . html ) using an average linkage clustering algorithm with a Euclidean distance metric . Support trees were based on 1000 bootstrap samples . Neighbor-joining trees based on MLST sequence data were constructed by MEGA ver . 2 . 1 software using the Kimura-2-parameter method of distance estimation . eBURST v3 ( http://eburst . mlst . net ) was used to demonstrate relationships between closely related STs ( those differing at only a single locus ) [58] , [59] , with the tree files visualized using PhyloDraw [60] . The BPSS2053 ( fhaB ) gene was disrupted in strain DD503 , an isogenic derivative of wild-type 1026b . In DD503 , the amr locus , encoding a multidrug efflux system , has been experimentally deleted [61] . The increased antibiotic susceptibility of DD503 makes it a useful strain for allelic exchange experiments as it allows the use of currently available allelic exchange vectors . There is no significant difference in virulence between the1026b parent strain and DD503 [61] . A 1036-bp internal region of the BPSS2053 ( fhaB ) gene was amplified by PCR using primers 53F:TGGTGGTGCAAGAGAATGGC and 53R:ATCGTGACCGATTGCTTGCC from Bp 1026b chromosomal DNA as previously described [21] . The PCR product was cloned into pCR2 . 1-TOPO ( Invitrogen Life Technologies , Burlington , Ontario , Canada ) according to the manufacturer's instructions . The internal region from BPSS2053 was cloned as an EcoR1 fragment into pGSV3-lux , a suicide vector containing a promoterless lux operon as a reporter , to create pATS2053 . The recombinant plasmid pATS2053 was transformed into E . coli SM10λpir [62] . Transformed E . coli containing pATS2053 were conjugated with Bp DD503 , and transconjugants selected on LB-gentamicin-polymyxin B agar . The transconjugants were screened for lux-mediated light production by assaying 100 µl of overnight broth cultures of individual colonies . One of the light-producing transconjugant strains was designated as Bp ATS2053 . Adherence of BPSS2053 ( fhaB ) mutants ( Bp ATS2053 ) to human buccal epithelial cells in vitro were compared against wild-type parental Bp 1026b as previously described [63] . Briefly , buccal epithelial cells from healthy control individuals were isolated by vigorous scraping of the buccal mucosa with a cotton-tipped swab . The swabs were placed into phosphate buffered saline ( PBS ) , transported to the laboratory , and the epithelial cells were incubated in vitro with bacteria at a ratio of 100 bacteria to 1 epithelial cell for 1 h at 37C in a shaking water bath . Unattached bacteria were removed from the mixture by repeated washing with PBS and centrifugation . Bacteria per cell were counted following staining of the bacteria-cell mixture with methylene blue by counting the number of bacteria attached to each of 50 cells and obtaining a mean number of bacteria/cell . MLST on 45 strains was performed as described in Godoy et al ( 2003 ) [44] using primer pairs for seven housekeeping genes ( ace , gltB , gmhD , lepA , lipA , narK ndh ) on Bp chromosome 1 . A complete list of primer pair sequences and PCR conditions is provided in Table S5 . Alleles at each of the MLST loci were assigned using the B . pseudomallei MLST website ( http://bpseudomallei . mlst . net/ ) - each allele was assigned a different allele number and the allelic profile ( string of seven integers ) was used to define the sequence type ( ST ) . Sequences that were not in the database were checked by re-sequencing , assigned as new alleles and deposited in the MLST allele database .
|
Melioidosis is a serious infectious disease of humans caused by Burkholderia pseudomallei , a soil bacterium endemic to many areas in South East Asia . Besides humans , B . pseudomallei is also capable of infecting many other species and can be isolated from diverse environmental sources including soil , water , and air . In this study , we used DNA microarrays to probe the stability of the B . pseudomallei genome in a large panel of clinical , animal , and environmental strains . We found that evidence of a highly dynamic B . pseudomallei genome , with up to 14% being variably present across different strains . Surprisingly , strains recovered from human patients were significantly associated with the presence of “genomic islands” , corresponding to regions of DNA directly acquired from other microorganisms . Genes on these genomic islands may thus play an important role in the pathogenesis of human melioidosis .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"microbiology/environmental",
"microbiology",
"microbiology/microbial",
"evolution",
"and",
"genomics"
] |
2008
|
The Core and Accessory Genomes of Burkholderia pseudomallei: Implications for Human Melioidosis
|
Interleukin-4 receptor ( IL-4Rα ) is critical for the initiation of type-2 immune responses and implicated in the pathogenesis of experimental schistosomiasis . IL-4Rα mediated type-2 responses are critical for the control of pathology during acute schistosomiasis . However , type-2 responses tightly associate with fibrogranulomatous inflammation that drives host pathology during chronic schistosomiasis . To address such controversy on the role of IL-4Rα , we generated a novel inducible IL-4Rα-deficient mouse model that allows for temporal knockdown of il-4rα gene after oral administration of Tamoxifen . Interrupting IL-4Rα mediated signaling during the acute phase impaired the development of protective type-2 immune responses , leading to rapid weight loss and premature death , confirming a protective role of IL-4Rα during acute schistosomiasis . Conversely , IL-4Rα removal at the chronic phase of schistosomiasis ameliorated the pathological fibro-granulomatous pathology and reversed liver scarification without affecting the host fitness . This amelioration of the morbidity was accompanied by a reduced Th2 response and increased frequencies of FoxP3+ Tregs and CD1dhiCD5+ Bregs . Collectively , these data demonstrate that IL-4Rα mediated signaling has two opposing functions during experimental schistosomiasis depending on the stage of advancement of the disease and indicate that interrupting IL-4Rα mediated signaling is a viable therapeutic strategy to ameliorate liver fibroproliferative pathology in diseases like chronic schistosomiasis .
Schistosomiasis is a parasitic disease caused by blood-dwelling parasitic flatworms of the genus Schistosoma , mainly , Schistosoma mansoni ( S . mansoni ) , S . japonicum and S . haematobium that are infective to humans and the most clinically relevant [1] . Schistosomiasis is estimated to affect more than 200 million people worldwide and causes up to 200 , 000 deaths per annum in developing countries [1] . The disease is caused by parasite eggs trapped in the microvasculature of the host organs ( liver , intestine and bladder ) that induce a vigorous inflammatory response [1] . The kinetics of the ensuing immune responses induced by S . mansoni infection are well defined and characterized [2 , 3] . Briefly , the outcomes of disease persistence and progression are organ enlargement , fibrosis , scarring , portal hypertension or hematuria ( S . haematobium specifically ) that drive host morbidity and eventually death in severe cases [1] . The immune response to schistosomiasis , similarly to that against other tissue-dwelling helminth infections [4–6] , is highly polarized as it progresses , going from i ) an early Th1 response to ii ) a powerful Th2 response that culminates as the adult parasite-released eggs are trapped in the host tissues [2 , 3] and finally iii ) a chronic regulatory phase with a minimized but still dominant Th2 response [3 , 7 , 8] with a more clinically relevant tissue fibroproliferative pathology . Our current understanding of schistosomiasis pathology heavily relies on the use of experimental murine models [2] . Studies aimed at uncovering factors that drive host protection or susceptibility to schistosomiasis have been conducted using gene-deficient mice . The disease associates with the formation of granulomas and excessive collagen deposition ( fibrosis ) around tissue-trapped eggs [3 , 7 , 8] . An important role was defined for the host immune effector responses in these pathognomonic processes as nude mice [9] , T cell-depleted [10–13] or mice with severe combined immunodeficiency [14] failed to form proper fibrogranulomatous responses . Even though Th1 , Th17 and Treg responses have been shown to play major roles in regulating schistosomiasis pathogenesis , type 2 immune responses , which are typically induced by the disease-mediating eggs of the parasite [15–17] , have been ascribed a more dominant role [3 , 7 , 8 , 18] . Initiation and polarization of type 2 immune responses is orchestrated by interleukin-4 ( IL-4 ) and IL-13 signaling via a common IL-4Rα chain [2 , 19] . Signaling via this receptor drives the activation of the transcription factor STAT6 in hematopoietic cells , the proliferation of T and B cells , the production of immunoglobulins by B cells , the priming and chemotaxis of mast cells and basophils [2 , 19] . In non-hematopoietic cells , this receptor plays a central role in inducing airway hyper-responsiveness by enhancing contractions and mucus secretion by gut epithelial cells [20] and has been shown to play a role in STAT6-dependent fibroblast activation leading to collagen deposition that define fibro-proliferative diseases [21 , 22] . Understandably , mice deficient in this receptor show impaired granuloma formation , enhanced liver damage and augmented gut inflammation that leads to endotoxemia and septic shock during acute schistosomiasis [23–26] . Moreover , studies conducted in our laboratory have refined the requirement of IL-4Rα to a cell-specific level showing that IL-4Rα-responsive macrophages [25] , pan-T cells [27] and smooth muscles cells [28] are individually essential for driving host survival and limiting tissue pathology during acute schistosomiasis . In all these studies employing mice constitutively deficient in IL-4Rα , a critical role for IL-4Rα mediated signaling during acute [25 , 26] and chronic schistosomiasis [25 , 29 , 30] is suggested . However , the constitutive lack of IL-4Rα led such transgenic mice to succumb prematurely to experimental schistosomiasis with high ( acute model ) as well as low ( chronic model ) infection doses i . e . chronic model of infections succumb during the acute phase in the absence of IL-4Rα [30] casting an equivoque on the reliability of using models of constitutive deletion of IL-4Rα to assess the role of this receptor during chronic schistosomiasis . Moreover , congenital IL-4Rα deletion has now been shown to affect the development of animals [31] , challenging our current knowledge on the role of IL-4Rα throughout experimental schistosomiasis ( acute and chronic ) using mouse models of constitutive IL-4Rα deficiency . In this study , the role of IL-4Rα during acute and chronic schistosomiasis was investigated using a novel murine model that allows for inducible deletion of il-4rα gene at any time point during S . mansoni infection . Our findings further confirmed a protective role played by IL-4Rα mediated signaling during acute schistosomiasis . Contrastingly , we showed for the first time that partial deletion of the il-4rα gene , specifically , at the chronic stage of schistosomiasis ameliorates the tissue pathology by reducing type-2 immune responses , improving immune balance between T helper cytokines and skewing the diminished immune response towards a more regulatory profile without affecting animal viability .
Inducible IL-4Rα deficient C57BL/6 mice ( RosaCreERT2IL-4Rα-/lox mice , termed iCre-/+IL-4Rα-/lox mice ) were established using a modified cyclization recombinase ( Cre ) under the control of the ubiquitously expressed Rosa promoter . This modified Cre incorporated a mutated fragment of the ligand-binding domain of the estrogen receptor ( ERT2 ) , that makes the activity of Cre conditional to the specific presence of Tamoxifen , an estrogen ligand homologue [32] . RosaCreERT2 C57BL/6 mice were intercrossed with IL-4Rα-/- C57BL/6 mice [33] to generate RosaCreERT2IL-4Rα-/- mice ( Fig 1A ) and subsequently intercrossed with floxed IL-4Rα ( IL-4Rαlox/lox ) C57BL/6 mice ( exon 6 to 8 flanked by loxP ) ( Fig 1B , [25] ) to generate RosaCreERT2-/+IL-4Rα-/lox C57BL/6 mice ( Fig 1A ) . Tamoxifen feeding ( Fig 1C ) did not impair the fitness of naïve RosaCreERT2-/+IL-4Rα-/lox C57BL/6 mice , as judged by body weight change ( Fig 1D ) . In Tamoxifen-fed RosaCreERT2-/+IL-4Rα-/lox C57BL/6 mice , CreERT2-mediated deletion of the exon 6 to 8 of the il-4rα gene ( Fig 1B ) was identified by specific Cre- , loxp- and il-4rα- PCR genotyping from tail DNA ( Fig 1E ) , and real-time qPCR from liver ( Fig 1G ) and spleen DNA ( Fig 1H ) . Analysis of IL-4Rα surface expression on total cells from different organs by flow cytometry ( Fig 2A ) demonstrated that IL-4Rα was considerably depleted following administration of Tamoxifen to RosaCreERT2-/+IL-4Rα-/lox mice ( Fig 2A and 2B and S1A Fig ) . To rule out a non-specific toxic effect or bystander immune alteration in Tamoxifen-fed RosaCreERT2-/+IL-4Rα-/lox mice , spleen weights ( S1B Fig ) , organ cellularity ( Fig 2C and S1C Fig ) , seric liver enzymes ( S1D Fig ) , baseline IgE levels ( S1E Fig ) , IL-2-driven proliferative responses of total splenocytes ( S1F Fig ) , frequencies of major myeloid and lymphoid cells ( S2A Fig and S2B Fig ) and total CD4+ ( S2C Fig ) and CD8+ ( S2D Fig ) T cell numbers in spleens and mesenteric lymph nodes ( MLN ) were determined . This revealed that , amid a minimal cellular deficiency in Spleen CD4+ and CD8+ T cells at baseline in our murine model , organ cellularity , weight and baseline cellular and humoral immune responses were not generally affected in Tamoxifen-fed RosaCreERT2-/+IL-4Rα-/lox mice ( Fig 2C , S1 Fig and S2 Fig ) . Tamoxifen treatment of RosaCreERT2-/+IL-4Rα-/lox mice significantly reduced or even abrogated surface IL-4Rα expression on spleen CD4+ T cells , MLN CD19+ B cells , peritoneal macrophages as well as bone marrow-derived dendritic cells ( Fig 2D ) . Robustness of IL-4Rα knockdown in Tamoxifen-fed RosaCreERT2-/+IL-4Rα-/lox mice was monitored in white blood cells over a period of 16 weeks following Tamoxifen administration ( Fig 2E ) . A relative expression level of 0% was attributed at all times to blood B cells from global IL-4Rα-/- mice , whereas a relative expression level of 100% was attributed to IL-4Rα-/lox mice . Blood B cells from oil-fed RosaCreERT2-/+IL-4Rα-/lox mice oscillated around a level of IL-4Rα expression of 100% , Tamoxifen-fed RosaCreERT2-/+IL-4Rα-/lox mice expressed only 20% of IL-4Rα ( Fig 2E ) , which increased to a maximum of 30% in 16 weeks with no significant body weight changes throughout the monitoring period ( Fig 2F ) . To assess the cellular knockdown of IL-4Rα functionally , splenocytes from IL-4Rα-/lox littermate controls , oil-fed RosaCreERT2-/+IL-4Rα-/lox , Tamoxifen-fed RosaCreERT2-/+IL-4Rα-/lox and IL-4Rα-/- mice were cultured with or without recombinant IL-4 for 48h then harvested , stained for IL-4Rα expression and analyzed by flow cytometry ( Fig 2G ) . As expected , spleen T ( Fig 2H ) and B cells ( Fig 2I ) derived from IL-4Rα-/lox mice and oil-fed RosaCreERT2-/+IL-4Rα-/lox controls up-regulated IL-4Rα expression after the addition of IL-4 ( Fig 2G ) . In contrast , rIL-4 stimulated spleen T and B cells derived from Tamoxifen-fed RosaCreERT2-/+IL-4Rα-/lox or from IL-4Rα-/- mice showed no upregulation of IL-4Rα expression ( Fig 2H and 2I ) . This showed functional impairment of IL-4Rα mediated signaling on cells from Tamoxifen-fed RosaCreERT2-/+IL-4Rα-/lox mice , complementing the above-demonstrated genotypic and phenotypic impairments . Taken together , these results indicated that Tamoxifen administration to the RosaCreERT2-/+IL-4Rα-/lox mouse model leads to a timely , efficient , safe and stably induced IL-4Rα knockdown mouse model . A protective role for IL-4Rα mediated signaling has been established during acute schistosomiasis , where IL-4Rα deficient mice but not wild-type ( wt ) mice died around 6 to 8 weeks after natural infection with S . mansoni [25] . To determine whether IL-4Rα mediated signaling is required throughout the course of experimental schistosomiasis , IL-4Rα was knocked down in S . mansoni-infected RosaCreERT2-/+IL-4Rα-/lox mice at the early acute ( Tamoxifen administration at 2 weeks post-infection termed Tam2 ) , late acute ( Tamoxifen administration at 6 weeks post-infection termed Tam6 ) and chronic phase ( Tamoxifen administration at 16 weeks post-infection termed Tam16 ) ( Fig 3A ) , as previously defined [3] . As expected , most of IL-4Rα deficient mice ( 70% ) succumbed prematurely to infection with 35 S . mansoni cercariae as early as from 7 weeks post-infection ( Fig 3B ) . Similarly , the viability of Tam2- and Tam6-fed RosaCreERT2-/+IL-4Rα-/lox mice declined rapidly ( 60 and 50% respectively at week 8 post-infection ) . From Tam-2-fed , Tam-6 fed or IL-4Rα deficient mice , no death was further reported as from 12 weeks post-infection , at the chronic phase of the disease . This indicated that IL-4Rα is necessary for host survival during acute schistosomiasis , but not required for host survival at the chronic phase of the disease . Indeed , removal of IL-4Rα in Tam16-fed S . mansoni-infected RosaCreERT2-/+IL-4Rα-/lox mice failed to affect the morbidity ( as indicated by serum levels of alanine transaminase as a marker of liver disease ( S3 Fig ) and the mortality ( Fig 3B ) up to 24 weeks post-infection , further supporting a dispensable role of IL-4Rα mediated signaling during chronic schistosomiasis . Taken together our results suggest that IL-4Rα mediated signaling differentially regulates schistosomiasis disease depending on the stage of the infection . The findings above demonstrated an impaired viability of S . mansoni-infected mice following IL-4Rα knockdown at 2 weeks post-infection ( Tam2 ) . Hence , the immune and histopathological response of Tam2-fed RosaCreERT2-/+IL-4Rα-/lox mice ( termed iCre-/+IL-4Rα-/lox Tam2 , Fig 4A ) , which might associate with the host premature death during experimental schistosomiasis was dissected . A consistent reduction of MLN CD4+ ( S4A Fig ) and CD8+ ( S4B Fig ) T cell counts was observed in S . mansoni-infected iCre-/+IL-4Rα-/lox Tam2 animals at week 7 when compared to littermate controls , consistent with the S . mansoni-infected global IL-4Rα-/- animals ( S4A Fig and S4B Fig ) . Ex vivo stimulation with a cocktail of PMA/Ionomycin/Monensin for 4h at 37°C and subsequent intracellular FACS analysis of MLN cells from S . mansoni-infected iCre-/+IL-4Rα-/lox Tam2 animals at week 7 resulted in impaired IL-4 production , but similar IFNγ production when compared to control mice ( Fig 4B–4D ) , suggesting a type2 impairment . This was paralleled by a significantly higher rate of reduction in the number of IL-4-producing CD4+ T cells in S . mansoni-infected iCre-/+IL-4Rα-/lox Tam2 animals ( ~50% , S4C Fig ) when compared to the minimal reduction in IFNγ-producing CD4+ T cells reported ( ~20% , S4D Fig ) . This impaired IL-4 production was confirmed within the supernatant of anti-CD3-stimulated MLN cells from S . mansoni-infected iCre-/+IL-4Rα-/lox Tam2 animals by ELISA where a greatly diminished production of other type-2 cytokines as well , i . e . IL-13 , IL-5 and IL-10 , amid rather minimally altered IFNγ responses ( Fig 4E ) was observed . Subsequently , Type 2 antibody responses ( IgG1 and total IgE ) appeared markedly reduced , whereas Type 1 antibodies ( IgG2a ) were similar to control mice ( Fig 4F–4H ) . However , liver egg burden was similar between the different groups ( Fig 4I ) , ruling out a differential level of infection as the cause of the observed diminished type-2 responses in iCre-/+IL-4Rα-/lox Tam2 and IL-4Rα-/- mice . Together , these results demonstrate that knocking down IL-4Rα at the early acute phase of experimental schistosomiasis considerably diminishes host ability to subsequently mount a type-2 immune response . Liver granuloma size ( Fig 4J and 4K ) and fibrosis ( Fig 4L and 4M ) were reduced in S . mansoni-infected iCre-/+IL-4Rα-/lox Tam2 similar to global IL-4Rα-/- mice , translating into a significantly reduced level of hepato- ( Fig 4N ) and splenomegaly ( Fig 4O ) compared with IL-4Rα-responsive control mice . As expected , from our previous mortality studies [25 , 27] , body weights of S . mansoni-infected IL-4Rα-/- and iCre-/+IL-4Rα-/lox Tam2 mice rapidly declined starting 6 weeks post-infection ( Fig 4P ) that preceded the death of these animals ( Fig 4Q ) when compared to IL-4Rα-responsive control mice . Bleeding was visible in the gut of the animals that rapidly succumbed to infection following removal of IL-4Rα . Taken together , these results suggest that IL-4Rα knockdown at the early acute phase of experimental schistosomiasis considerably diminishes the host ability to mount a protective fibro-granulomatous response around the S . mansoni eggs and this was associated with gut bleeding , rapid weight loss and premature death . As impaired viability of S . mansoni-infected mice following IL-4Rα knockdown at 6 weeks post-infection ( Tam6 ) was observed ( Fig 3B ) , the associated immune and histopathological responses of Tam6-fed RosaCreERT2-/+IL-4Rα-/lox mice ( termed iCre-/+IL-4Rα-/lox Tam6 , Fig 5A ) was investigated . As expected , surface IL-4Rα protein on lymphocytes from S . mansoni-infected iCre-/+IL-4Rα-/lox Tam6 mice was abrogated as demonstrated by flow cytometry ( S5A Fig and S5B Fig ) . A significant reduction of T lymphocytes in the MLN of S . mansoni-infected iCre-/+IL-4Rα-/lox Tam6 animals at week 7 post infection ( S5C Fig and S5D Fig ) was observed . Ex vivo stimulation and subsequent intracellular FACS analysis of MLN cells from S . mansoni-infected iCre-/+IL-4Rα-/lox Tam6 animals at week 7 post infection revealed impaired IL-4 production , similar to global IL-4Rα-/- mice ( Fig 5B and 5C ) , whereas IFN-γ responses were similar compared to IL-4Rα-/lox and IL-4Rα+/+ control mice ( Fig 5B and 5D ) . This suggests an impairment of type 2 immune responses in S . mansoni-infected iCre-/+IL-4Rα-/lox Tam6 animals as confirmed by the drastic reduction of IL-4-producing CD4+ T cell numbers in the MLN ( ~83% , S5E Fig ) , that paralleled a significant but less important reduction of IFNγ-producing CD4+ T cell numbers ( ~60% , S5F Fig ) . This reduction of IL-4 production was confirmed by anti-CD3-stimulated MLN cells from S . mansoni-infected iCre-/+IL-4Rα-/lox Tam6 mice and analysis of the released cytokines by ELISA ( Fig 5E ) . Consistently , we observed a significant decrease in the production of type 2 cytokines , i . e . IL-4 and IL-10 , but minimally altered IFN-γ responses ( Fig 5E ) . As a result of reduced IL-4 , type 2 antibody responses ( IgG1 and total IgE ) were markedly reduced ( Fig 5F and 5G ) , whereas type 1 antibodies ( IgG2a ) were similar to control mice ( Fig 5H ) . Liver egg burden was similar between the different groups ( Fig 5I ) , ruling out a differential level of infection as the cause of the observed diminished type 2 responses in iCre-/+IL-4Rα-/lox Tam6 and IL-4Rα-/- mice . Together , these results demonstrate that knock down of IL-4Rα after egg deposition does diminish host ability to maintain the type 2 immune responses . Reduced type 2 responses decreases pathological features , including liver granuloma size ( Fig 5J and 5K ) and fibrosis ( Fig 5L and 5M ) , hepato- ( Fig 5N ) and splenomegaly ( Fig 5O ) compared with IL-4Rα-responsive control mice . However , the body weights of S . mansoni-infected iCre-/+IL-4Rα-/lox Tam6 mice rapidly declined following Tamoxifen-driven removal of IL-4Rα at 6 weeks post-infection similar to IL-4Rα-/- mice ( Fig 5P ) and culminated into the early death of these animals ( Fig 5Q ) , when compared to IL-4Rα-responsive control mice . Bleeding was visible in the gut of the animals that rapidly succumbed to infection following removal of IL-4Rα . No premature mortality was reported with S . mansoni-infected Tam6-fed IL-4Rα+/+ ( control for Tamoxifen side effects ) , RosaCreERT2-/+IL-4Rα+/+ ( control for activated CreERT2 ) and RosaCreERT2-/+IL-4Rα-/lox ( control for CreERT2 Tamoxifen-independent activity ) mice when compared to S . mansoni-infected IL-4Rα+/+ ( positive control ) mice ( S6 Fig ) ruling out any non-specific effect ( s ) of Tamoxifen or CreERT2 as mediator ( s ) of the impaired survival of S . mansoni-infected iCre-/+IL-4Rα-/lox Tam6 mice . Taken together , these results show that IL-4Rα knockdown after egg deposition during the acute phase of experimental schistosomiasis considerably diminishes the host ability to maintain a type 2 immune response around the S . mansoni eggs which associates with gut bleeding , rapid weight loss and premature death . S . mansoni-infected mice following IL-4Rα knockdown at 16 weeks post-infection , i . e . iCre-/+IL-4Rα-/lox Tam16 mice ( Fig 6A ) did not result in any weight loss ( Fig 6B ) or mortality ( Fig 6C ) , for up to 24 weeks post infection . Liver egg burden was similar between the control IL-4Rα-/lox ( Fig 6D ) , ruling out a differential level of infection between both groups of mice . IL-4Rα knockdown considerably reduced liver ( Fig 6E ) and spleen ( Fig 6F ) enlargement in chronically infected mice . Apparent scarification was visible on the liver lobes of control mice whereas IL-4Rα knockdown resulted in the removal/reversal/inhibition of liver scarification ( Fig 6G ) . Moreover , IL-4Rα knockdown considerably reduced granuloma size ( Fig 6H and 6I ) and collagen levels ( Fig 6J and 6K ) in the livers of chronically infected mice . These data indicated that IL-4Rα knockdown ameliorate granulomatous inflammation , hepato- and splenomegaly and liver fibrosis during chronic schistosomiasis further consolidating the idea of a deleterious role for IL-4Rα signaling in mediating fibroproliferative pathology during chronic schistosomiasis . To analyse the immune polarization and responses that are triggered by IL-4Rα knockdown during chronic schistosomiasis and associate with the amelioration of tissue disease , IL-4Rα was knockdown in mice chronically infected with S . mansoni at week 16 post infection and the immune response analyzed at week 18 post infection ( Fig 7A ) . A significant reduction of CD4+ ( S7A Fig ) and CD8+ ( S7B Fig ) T lymphocytes in the MLN of S . mansoni-infected iCre-/+IL-4Rα-/lox Tam16 animals at week 18 post infection was observed . Our analyses revealed a reduced Th2-mediated production of IL-4 and IL-13 , but present IFN-γ and IL-10 production by MLN T cells of IL-4Rα knockdown animals , as judged by frequencies ( Fig 7B , gated as per S7C Fig ) , total numbers ( Fig 7C ) and ratios ( Fig 7D ) of cytokine-producing MLN CD4+ T cells ( S6 Fig and Fig 7B ) . Canonical transcription factor analysis ( Fig 7E ) confirmed this conclusion with reduction of GATA3 but normal Tbet production in effector T cells ( Fig 7E and 7F ) . Interestingly , the frequencies of Foxp3+ regulatory T cell responses were increased in the MLNs of S . mansoni-infected iCre-/+IL-4Rα-/lox Tam16 mice ( Fig 7G and 7H ) , when compared with S . mansoni-infected IL-4Rα-/lox control mice . However , most likely as a result of total CD4+ T cell drop ( S7A Fig ) , Treg cell numbers were reduced following Tam16 treatment in S . mansoni-infected iCre-/+IL-4Rα-/lox Tam16 animals when compared to their littermate controls ( Fig 7I ) . Serum titers of type 2 antibodies ( IgG1 and total IgE ) were reduced ( Fig 7J and 7K ) but not type 1 antibodies ( IgG2a , Fig 7L ) , supporting reduced type 2 responses in S . mansoni-infected iCre-/+IL-4Rα-/lox Tam16 mice . Of interest , regulatory B cell frequencies increased during infection and particularly in infected iCre-/+IL-4Rα-/lox Tam16 mice ( Fig 7M and 7N ) amid a rather stable total count in S . mansoni-infected iCre-/+IL-4Rα-/lox Tam16 animals when compared to littermate controls ( Fig 7O ) . Innate type 2 immune effectors i . e . eosinophils [34 , 35] , ILC2 [36] and macrophages [37] have been positively linked to liver fibrosis , the pathophysiological process that drives the host morbidity during chronic schistosomiasis . Conversely , arginase expression by macrophages has been shown to counter tissue inflammation and fibrosis [38] . The analysis of the MLN cells of S . mansoni-infected iCre-/+IL-4Rα-/lox Tam16 mice for these cell types by flow cytometry ( S7D Fig–S7G Fig ) revealed that the pro-fibrotic innate effectors i . e . eosinophils ( Fig 7P , S8A Fig and S8B Fig ) , ILC2 ( Fig 7Q and S8C Fig ) and macrophages ( S8D Fig ) were significantly diminished . Conversely , the mean arginase expression by macrophages ( S8E Fig ) was not affected in S . mansoni-infected iCre-/+IL-4Rα-/lox Tam16 mice , when compared to S . mansoni-infected IL-4Rα-/lox control mice . This suggests that IL-4Rα knockdown in chronically infected mice does skew the MLN response away from a pro-fibrotic response . Taken together , these results suggest that knocking down IL-4Rα at the chronic phase of experimental schistosomiasis considerably skews the host immune response away from the type 2 arm of the immune response , fosters a qualitatively more regulatory , anti-inflammatory and anti-fibrotic profile with no deleterious effect on host survival .
Taking advantage of a newly established temporal inducible IL-4Rα deficient mouse model , we demonstrated that interrupting IL-4Rα mediated signaling prevents the onset and maintenance of egg-driven type 2 immune responses and its associated fibro-granulomatous inflammation during schistosomiasis . Whereas early knockdown of the receptor during the acute phase of the disease led to aggravated morbidity and mortality , late targeting at the chronic phase considerably ameliorated fibrogranulomatous inflammation and reduced hepato- and splenomegaly without impairing the animal viability . Amelioration of chronic schistosomiasis pathology was further associated with reduction of type 2 immune effector responses but expansion of regulatory T and B cells , suggesting that IL-4Rα mediated immune responses are detrimental in chronic schistosomiasis . Hence , therapeutic intervening of IL-4Rα mediated signaling to reduce type 2 responses might provide a strategy to ameliorate fibroproliferative pathology in diseases like chronic schistosomiasis . The group of Cheever et al . ( 1994 ) were the first to show that abrogation of type 2 immune responses in S . mansoni-infected mice resulted in impaired granulomatous inflammation around the trapped eggs in tissue . A subsequent study by Chiaramonte et al . [21] showed the importance of IL-13-mediated signaling in fibrogenesis through the blockade of IL-13 . Moreover , this group further reported a critical role for IL-13 in granuloma formation induced by S . mansoni eggs [39] . An independent group further reported on the achievement of significantly reduced tissue fibrosis by blocking type 2 responses in S . mansoni-infected mice with anti-IL-4 antibody treatment [40] . This indicated that IL-4-orchestrated type 2 responses , as well as IL-13-driven responses , are all causally linked to fibrogranulomatous pathology . More recently , using Schistosomiasis-infected IL-4Rα deficient mice , we and others demonstrated reduced fibrogranulomatous inflammation . However , these mice died during acute schistosomiasis due to cachexia [25] . Using inducible IL-4Rα deficient mice in the present study , we now dissected the role of IL-4/IL-13-mediated type 2 responses during acute and chronic murine schistosomiasis . IL-4Rα removal ( knockdown ) early during schistosomiasis infection led to impaired type 2 responses with reduced fibrogranulomatous inflammation around the trapped eggs of the parasites , as demonstrated before . This resulted in exacerbated morbidity and premature death of the animals , as demonstrated previously in IL-4Rα deficient mice [25] . Thus , IL-4Rα-elicited type 2 immune effector responses like granuloma formation and fibrosis are important for the host survival during acute schistosomiasis . This concept has been previously established where a tissue protective role of these responses against the toxic secretions of the parasite eggs was suggested [41] . Interestingly , liver integrity was not affected after acute knockdown of IL-4Rα in S . mansoni-infected animals . This finding argues against liver toxicity being the pathological event that drives death in these animals . A more likely explanation for their premature death would be the intensive gut bleeding reported in our previous study on IL-4Rα deficient mice [25] , and similarly observed in this study . As a result of the compromised gut integrity , bacteria would translocate to the blood stream and death by septic shock would ensue , as previously demonstrated [25] . Tamoxifen-induced knockdown of the IL-4Rα after egg deposition during the chronic phase ( 16 weeks post-infection ) uncovered a hitherto unappreciated facet of the IL-4Rα mediated type 2 responses . Indeed , IL-4Rα knockdown during chronic schistosomiasis did not lead to gut bleeding and did not affect animal viability but ameliorated liver pathology with reduced granuloma size and fibrosis in the liver and no visible scarification and reduced level of liver and spleen enlargement . This clearly suggests that IL-4Rα mediated type 2 responses are detrimental during chronic schistosomiasis and the cause for fibroproliferative liver pathology . Of interest , regulatory T and B cell compartments were significantly increased following IL-4Rα removal during chronic schistosomiasis . It is tempting to associate the beneficial effect of IL-4Rα blockade on tissue pathology during chronic schistosomiasis to the enhanced regulatory response observed . In fact , previous studies have reported an amelioration of the fibrogranulomatous inflammation during chronic schistosomiasis by Foxp3+ regulatory T cells [42 , 43] . Whether IL-4Rα mediated signaling causally dictates the anti-inflammatory and anti-fibrotic activities of these regulatory cells during chronic schistosomiasis is not known . As of now , a role for IL-4Rα signaling in the development of immune hyporesponsiveness after chronic exposure of host immune cells to schistosomal antigens has been demonstrated [44] . This further re-emphasizes the potential of IL-4Rα in modulating the dynamics of the host regulatory responses during chronic diseases such as schistosomiasis . Such a potential has already been widely reported with a negative regulation of Foxp3+ Tregs and a loss of their suppressive capacity suggested to occur when the IL-4Rα signaling was solicited [45–47] . Alternatively , however , the remnant IL-4Rα mediated signaling in Tam16 mice argues against the absence of Th2 responses as the sole driver of the ameliorated pathology observed . A rather noticeable finding is the upregulation of other cytokines i . e . IL-10 and IFNγ resulting in a better balanced cytokine profile between T cells producing IL-4 , IFN-γ and/or IL-10 . Consequently , the impairment of IL-4Rα mediated signaling during chronic schistosomiasis by inducing a more equilibrated and mixed Th profile might prevent untoward immune polarization and tissue immunopathology . This hypothesis is strongly supported by the recently demonstrated role for immune balance rather than strong immune polarization in controlling fibrogranulomatous pathology during experimental schistosomiasis [48] . Further experiments are now required to empirically disentangle these hypotheses . What remains clear and worth focus at present is the fact that targeting IL-4Rα mediated signaling for the management of non-communicable type 2-mediated diseases in humans is in advanced clinical trials [49–51] . Understandably , building on the present study , the translatability of targeting IL-4Rα mediated signaling during fibroproliferative diseases like chronic schistosomiasis is further supported . What do we add to the current knowledge on the control of fibroproliferative disease ? It should be recalled that our present report builds on the previous observations made during IL-13 blockade experiments where a key role for this cytokine , and the indication of the potential of the IL-4Rα signaling axis in driving fibroproliferative responses during experimental schistosomiasis was defined [21] . In as much as an efficient anti-fibrotic strategy already transpired from the sole blockade of IL-13 [21] , the noticeable and independent pro-fibrotic effect of IL-4 [21 , 40] altogether argues for the higher anti-fibrotic potential of dually targeting IL-4 and IL-13 by blocking IL-4Rα rather than IL-13 alone . The picture might not be that straightforward , however , as caution should also be exerted in dually targeting IL-4 and IL-13 via IL-4Rα given that IL-4 unlike IL-13 is critical for type 2 immune responsiveness . A state of immune deficiency might therefore arise from IL-4Rα targeting as opposed to IL-13 targeting where Th2 responses are optimally elicited [21] . Also , consistent with the observation that IL-13 targeting was not toxic for the host [21] , our present report shows that IL-4R targeting does not impair animal fitness . This strongly argues for the safety of our approach . Conclusively , as of yet , one could therefore speculate on an added value of targeting IL-4Rα rather than just IL-13 given the different profibrotic potentials of IL-13 [21 , 39 , 52 , 53] and IL-4 [40 , 54–58] as both cytokines signal through IL-4Rα . Clearly , such a conclusion would still need to be experimentally validated . In summary , we provide evidence on the role of IL-4Rα during experimental schistosomiasis whereby early signaling helps the host survive the acute phase of the disease whereas signaling at the late chronic phase mediate the morbidity . Targeting IL-4Rα might therefore represent a novel therapeutic strategy against the fibroproliferative pathology that drives the morbidity of fibrotic diseases like chronic schistosomiasis .
IL-4Rα-/- , IL-4Rα-/lox and CreERT2 mice on a C57/BL6 background were previously described [2 , 25 , 33] . We generated a novel inducible IL-4Rα deleting mouse strain ( RosaCreERT2-/+IL-4Rα-/lox ) by intercrossing transgenic RosaCreERT2-/+ mice with IL-4Rα-/- and IL-4RαLox/Lox mice . CreERT2 transgenic negative littermates ( IL-4Rα-/lox ) expressing functional IL-4Rα were used as controls in all experiments . Mice were maintained in the University of Cape Town specific pathogen-free animal facility in accordance with the guidelines established by the Animal Research Ethics committee of the Faculty of Health Science of the University of Cape Town and the South African Veterinary Council ( SAVC ) . All animal experiments were conducted under strict recommendation of the South African national guidelines and of the University of Cape Town practice for laboratory animal procedures as outlined in protocols 010/048 and 014/003 reviewed and approved by the Animals Research Ethics Committee of the Faculty of Health Science of the University of Cape Town . Both male and female mice aged 6–12 weeks were used for all experiments . Care was taken under these protocols to minimize animal suffering in accordance with the guidelines of the Animal Research Ethics committee of the Faculty of Health Science of the University of Cape Town and the South African Veterinary Council ( SAVC ) . Mice were infected percutaneously via the abdomen with 35 , 80 or 100 cercariae , as indicated , with a Puerto Rican strain of Schistosoma mansoni obtained from infected Biomphalaria glabrata ( a generous gift from Adrian Mountford , York , UK ) . Eggs were purified from digested sections of liver or ileum from infected animals and counted at 40× magnification as previously described [40] . To activate il-4rα gene excision by CreERT2 , Tamoxifen ( Sigma , Deisenhofen , Germany ) solubilized in vegetable oil was administered by oral gavage to mice for four consecutive days ( 2 . 5mg/day ) . Polymerase chain reaction was used to confirm the genotype of RosaCreERT2-/+IL-4Rα-/lox mice . PCR conditions were 94°C for 2 minute , 94°C for 20 seconds , 45°C for 30 seconds , and 72°C for 20 seconds for 40 cycles . To quantify the efficiency of deletion , real-time PCR was performed on genomic DNA from liver and spleen cells using primers specific for IL-4Rα exon 5 ( control ) and exon 8 ( deleted by CreERT2 activation ) as described previously [25] . Il-4Rα surface expression was detected on splenocytes , lymph node cells , lung cells , hepatocytes , bone marrow cells and peritoneal exudate cells by phycoerythrin ( PE ) anti-CD124 ( IL-4Rα , M-1 ) . Cell subpopulations were identified with Alexa Fluor 700 , BD Horizon V500 , BD Horizon V450 , PerCP-Cy5 . 5 , APC , APC-Cy7 , Fluoroscein isothiocyanate , PE , PE-Cy7 or biotinylated monoclonal antibodies against CD3 , CD4 , CD8 , CD19 , Lineage , CD1d , CD5 , Foxp3 , Gata-3 , T-bet , IL-4 , IL-13 , IFN-γ , IL-10 , F4/80 , Ly6G , CD11c , MHCII , SiglecF , T1/ST2 , ICOS , Arginase , CD11b . Biotin-labeled antibodies were detected by Allophycocyanin or PercP-Cy5 . 5 . For staining , cells ( 1x 106 ) were labeled and washed in PBS , 3% FCS and 0 . 1% NaN3 . Between each step of staining , cells were washed extensively . For intracellular cytokine staining , cells were restimulated with a cocktail of PMA/Ionomycin/Monensin for 4h at 37°C then fixed in 2% PFA , permeabilized and cytokine production was analyzed as previously described [25] . For intranuclear staining , a commercially available transcription buffer set ( BD Bioscience ) was used as per the manufacturer’s instructions . All antibodies were from BD Pharmingen ( San Diego , CA ) except where noted otherwise . Stained cells were then acquired on a LSR Fortessa machine ( BD Immunocytometry system , San Jose , CA , USA ) and data were analyzed using Flowjo software ( Treestar , Ashland , OR , Usa ) . Tissue samples were fixed in neutral buffered formalin , processed , and 5–7 μm sections stained with hematoxylin and eosin ( H & E ) . Granuloma diameter of 20–50 granulomas per animal was determined using an ocular micrometer ( Nikon NIS-Elements , Nikon Corporation , Tokyo , Japan ) . For fibrosis assessment , tissue sections were stained with chromotrope 2R and analine blue solution ( CAB ) and counterstained with Wegert's hematoxylin for collagen staining . Complementarily , a modified protocol of tissue hydroxyproline quantification was used [59] . In brief , weighed liver samples were hydrolyzed and the supernatant was neutralized with 1% phenolphthalein and titrated against 10 M NaOH . An aliquot was mixed with isopropanol and added to a chloramine-T/citrate buffer solution ( pH 6 . 0 ) ( Sigma ) . Ehrlich's reagent solution was added and measured at 570 nm . Hydroxyproline levels were calculated by using 4-hydroxy-L-proline ( Calbiochem ) as standard , and results were expressed as μg hydroxyproline per weight of liver tissue that contained 104 eggs . Statistical analysis was conducted using GraphPad Prism 4 software ( http://www . prism-software . com ) . Data were calculated as mean ± SD . Statistical significance was determined using the unpaired Student's t test , One-Way or Two-Way ANOVA with Bonferroni's post test , defining differences to C57BL/6 , IL-4Rα-/lox or oil-treated RosaCreERT2-/+IL-4Rα-/lox as significant ( * , p≤0 . 05; ** , p≤0 . 01; *** , p≤0 . 001 ) .
|
Liver fibroproliferative diseases drive a considerable fraction of the overall human mortality . This is closely linked to the absence of efficient control measures against such diseases . Schistosomiasis , a chronic disease that affects humans , preferentially causes liver fibrosis and is responsible for devastating economic losses in developing nations where the disease is still endemic . Using reverse genetics , loss-of-function mouse models have helped uncover a protective role for Interleukin-4 receptor ( IL-4Rα ) in the host survival to experimental schistosomiasis . However , given the contributing role for this receptor in the etiology of some models of tissue fibrosis , its role during chronic schistosomiasis where the highly fibrotic liver of the infected individuals mediate the morbidity had not been properly addressed hitherto . Taking advantage of a third generation mouse model of inducible loss of a gene , we found a debilitating role for IL-4 receptor during chronic schistosomiasis as signaling via this receptor supported both liver inflammation and fibrosis . These findings demonstrate that although the host requires IL-4Rα to survive the acute phase of schistosomiasis , the more clinically relevant morbid phase of the disease is driven by the excessive utilization of this receptor . A therapeutic potential of blocking IL-4Rα to ameliorate liver fibroproliferative disease is therefore suggested .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"blood",
"cells",
"schistosoma",
"invertebrates",
"schistosoma",
"mansoni",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"immune",
"physiology",
"enzyme-linked",
"immunoassays",
"helminths",
"spleen",
"immunology",
"tropical",
"diseases",
"parasitic",
"diseases",
"animals",
"animal",
"models",
"model",
"organisms",
"experimental",
"organism",
"systems",
"neglected",
"tropical",
"diseases",
"immunologic",
"techniques",
"research",
"and",
"analysis",
"methods",
"white",
"blood",
"cells",
"animal",
"cells",
"t",
"cells",
"immunoassays",
"mouse",
"models",
"immune",
"response",
"antibody-producing",
"cells",
"helminth",
"infections",
"schistosomiasis",
"cell",
"biology",
"b",
"cells",
"physiology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"organisms"
] |
2017
|
Host regulation of liver fibroproliferative pathology during experimental schistosomiasis via interleukin-4 receptor alpha
|
The treatment of leishmaniasis relies mostly on parenteral drugs with potentially serious adverse effects . Additionally , parasite resistance in the treatment of leishmaniasis has been demonstrated for the majority of drugs available , making the search for more effective and less toxic drugs and treatment regimens a priority for the control of leishmaniasis . The aims of this study were to evaluate the antileishmanial activity of raloxifene in vitro and in vivo and to investigate its mechanism of action against Leishmania amazonensis . Raloxifene was shown to possess antileishmanial activity in vitro against several species with EC50 values ranging from 30 . 2 to 38 . 0 µM against promastigotes and from 8 . 8 to 16 . 2 µM against intracellular amastigotes . Raloxifene's mechanism of action was investigated through transmission electron microscopy and labeling with propidium iodide , DiSBAC2 ( 3 ) , rhodamine 123 and monodansylcadaverine . Microscopic examinations showed that raloxifene treated parasites displayed autophagosomes and mitochondrial damage while the plasma membrane remained continuous . Nonetheless , plasma membrane potential was rapidly altered upon raloxifene treatment with initial hyperpolarization followed by depolarization . Loss of mitochondrial membrane potential was also verified . Treatment of L . amazonensis – infected BALB/c mice with raloxifene led to significant decrease in lesion size and parasite burden . The results of this work extend the investigation of selective estrogen receptor modulators as potential candidates for leishmaniasis treatment . The antileishmanial activity of raloxifene was demonstrated in vitro and in vivo . Raloxifene produces functional disorder on the plasma membrane of L . amazonensis promastigotes and leads to functional and morphological disruption of mitochondria , which culminate in cell death .
With an incidence of two million new cases per year , leishmaniasis is endemic in 98 countries and territories , representing an area where more than 350 million people are at risk of acquiring the infection [1] . The treatment of leishmaniasis relies mostly on parenteral drugs and involves high costs . Antimonial compounds were discovered nearly 100 years ago and remain the drug of choice for the treatment of leishmaniasis in many parts of the world , despite their high toxicity and consequent severe adverse effects . Amphotericin B and pentamidine are used as second-line drugs , but their administration also involves serious adverse effects [2] . Miltefosine has been approved for the therapy of visceral leishmaniasis in India , but its efficacy for the treatment of American leishmaniasis has been shown to be variable depending on the causative species [3] . Moreover , parasite resistance in the treatment of leishmaniasis was demonstrated for the majority of drugs available [4] . Therefore , the search for more effective and less toxic drugs and treatment regimens is a priority for the control of leishmaniasis [5] . Screening of drugs originally developed for another purpose , or drug repositioning , constitutes a promising strategy for the discovery of new compounds effective against neglected diseases . In the treatment of leishmaniasis , examples of this strategy are miltefosine , initially developed for the treatment of breast cancer , pentamidine , a hypoglycemic agent , as well as amphotericin B , used in the treatment of fungal infections [6] . The benefits of repositioning include the availability of materials and data such as toxicology studies , resulting in reduction of time and costs to bring the drug to the market [7] . Selective Estrogen Receptor Modulators ( SERMs ) are a class of therapeutic agents widely prescribed for the treatment and prevention of breast cancer , osteoporosis , and postmenopausal symptoms [8] . The most widely used SERM is tamoxifen , a triphenylethylene used in the management of breast cancer . Due to its mixed antagonist and agonistic estrogenic activity , long term use of tamoxifen has been associated with an increased risk of endometrial cancer in postmenopausal patients [9] . Drugs belonging to a second class of SERMs built on a benzothiophene scaffold are also in clinical use . One example of this second class is raloxifene , an oral SERM which has estrogen agonist effects on bone and cholesterol metabolism but behaves as a complete estrogen antagonist on mammary gland and uterine tissue [10] , which results in skeletal benefit , with little , if any , uterine stimulation [11] . Previous studies have shown that tamoxifen is active against different species of Leishmania in vitro and in vivo [12] , [13] , [14] . In infections caused by Leishmania amazonensis in BALB/c mice , treatment with tamoxifen resulted in significant and sustained improvement in both clinical and parasitological parameters [15] . The activity of this drug has also been demonstrated in experimental infections by Leishmania braziliensis , the main causative agent of cutaneous leishmaniasis in Brazil . In this model , tamoxifen was able to significantly reduce the size of lesions and parasite burden [13] . Tamoxifen was also effective in the treatment of infections caused by Leishmania major in a murine model and in the treatment of visceral leishmaniasis in hamsters infected with Leishmania infantum chagasi [13] , [14] . The activity of tamoxifen in the treatment of experimental leishmaniasis led us to investigate whether raloxifene also presents leishmanicidal effect . Here , we describe the antileishmanial activity of raloxifene in vitro and in vivo and investigate its mechanism of action against L . amazonensis , which is responsible for most cases of human cutaneous leishmaniasis in the Amazon region of Brazil [16] .
Animal experiments were approved by the Ethics Committee for Animal Experimentation ( Protocol 033/42/02 ) of the Biomedical Sciences Institute of the University of São Paulo . The research adhered to the Brazilian Guidelines for Care and Utilization of Animals from the Conselho Nacional de Controle e Experimentação Animal ( CONCEA ) . Raloxifene hydrochloride and miltefosine were purchased from Sigma-Aldrich ( St Louis , MO , USA ) . Stock solutions of raloxifene ( 10 mM ) and miltefosine ( 20 mM ) were prepared in DMSO and in sterile water , respectively , and kept at −20°C . Dilutions from the stock solutions were done in culture media . For in vivo experiments , fresh solutions of raloxifene were prepared in saline or Cremophor A25 ( Sigma-Aldrich ) . Promastigotes of Leishmania ( Leishmania ) amazonensis ( MHOM/BR/1973/M2269 ) , Leishmania ( Leishmania ) donovani ( LD-15/MHOM/SD/00 ) , Leishmania ( Leishmania ) infantum chagasi ( MHOM/BR/1972/LD ) , Leishmania ( Leishmania ) major ( MHOM/IL/1981/Friedlin ) , Leishmania ( Leishmania ) mexicana ( MHOM/BR/1974/M2682 ) and Leishmania ( Viannia ) braziliensis ( MHOM/BR/1975/M2903 ) were maintained in M199 medium ( Sigma-Aldrich ) supplemented with 10% heat-inactivated fetal calf serum ( FCS ) ( Invitrogen Corporation , NY , USA ) and 0 . 25% hemin at 25°C . L . braziliensis and L . infantum chagasi cultures were also supplemented with 2% sterile male human urine . Promastigotes of a L . amazonensis transgenic line expressing luciferase ( LaLUC ) were grown in the same medium supplemented with 32 µg/mL G418 [17] . Amastigotes were purified from lesions as described [18] . Briefly , amastigotes were obtained from lesions induced in BALB/c mice , about 8 to 12 weeks after intradermal inoculation of 1×106 parasites in the hind footpads . Lesions were removed and homogenized in phosphate-buffered saline ( PBS ) ; the suspension was cleared of cell debris by centrifugation at 50 g for 8 min; the supernatant was then washed three times in PBS and passed through a 25-gauge needle . Amastigotes recovered from tissue were resuspended in RPMI-1640 medium supplemented with 10% FCS , 2 mM glutamine and 50 mg/mL gentamicin and kept at 33°C in a 5% CO2 atmosphere . Bone marrow-derived macrophages ( BMDM ) were obtained from BALB/c mice as previously described [19] . J774 macrophages were maintained in RPMI-1640 medium supplemented with 10% FCS at 37°C in a 5% CO2-humidified incubator . Cell viability was evaluated in vitro by cultivating promastigotes ( 5×106 per well ) or lesion-derived amastigotes ( 1×107 per well ) in M199 or RPMI-1640 medium , respectively , supplemented with 10% FCS . Parasites were incubated in the presence of increasing concentrations of raloxifene ( 7 . 75 to 62 . 0 µM for promastigotes and 3 . 75 to 30 . 0 µM for intracellular amastigotes , assayed at a 1 . 5-fold dilution ) for 24 h . Miltefosine ( assayed in concentrations varying from 1 . 5 to 45 . 0 µM ) was used as a control drug . Quantification of viable cells was assessed either by cell counting or by measuring the cleavage of 3- ( 4 , 5-dimethylthiazol-2-yl ) -2 , 5-diphenyl tetrazolium bromide ( MTT; Sigma-Aldrich ) as previously described [20] . MTT cleavage was assessed in a microplate reader ( POLARstar Omega , BMG Labtech , Ortenberg , Germany ) with a reference wavelength of 690 nm and a test wavelength of 595 nm . The activity of raloxifene was also evaluated against L . amazonensis promastigotes ( 3×107/mL ) incubated in M199 medium or in Hank's balanced salt solution supplemented with 10 mM D-glucose ( HBSS+Glc ) ( 137 mM NaCl , 5 . 3 mM KCl , 0 . 4 mM KH2PO4 , 4 . 2 mM NaHCO3 , 0 . 4 mM Na2HPO4 , pH 7 . 2 , 10 mM D-glucose ) for 2 hours . Quantification of viable cells was assessed by measuring the cleavage of ( 3- ( 4 , 5-dimethylthiazol-2-yl ) -5- ( 3-carboxymethoxyphenyl ) -2- ( 4-sulfophenyl ) -2H-tetrazolium ) with the CellTiter 96 Aqueous One Solution Cell Proliferation Assay ( MTS , Promega ) , according to the manufacturer's instructions . The half maximal effective concentrations ( EC50 ) were determined from sigmoidal regression of the concentration-response curves using GraphPad Prism 5 . 0 software . Assays were performed in triplicate and results are expressed as the mean and standard deviation ( SD ) of at least three independent experiments . In vitro cytotoxicity was evaluated by cultivating J774 macrophages ( 5×105 per well ) in 24-well plates for 24 h in the presence of increasing concentrations of raloxifene . Cell viability was assessed by the MTT assay as described above and results are expressed as percentage reduction in cell viability compared with untreated control cultures . The half maximal cytotoxic concentration ( CC50 ) was determined as described above for EC50 values . The Selectivity Index of raloxifene was calculated as the ratio between the CC50 for J774 macrophages and the EC50 against Leishmania intracellular amastigotes . Killing of intracellular L . amazonensis amastigotes was assayed by analysis of parasite burden in BMDM monolayers . Macrophages were plated on 96-well plates ( 8×104 per well ) in RPMI-1640 medium supplemented with 10% FCS and allowed to adhere overnight at 37°C , at 5% CO2 . LaLUC stationary-phase promastigotes ( in a ratio of 20 parasites: 1 macrophage ) were added to the wells and the cultures were incubated at 33°C in a 5% CO2 atmosphere . After 3 h , parasites were removed by washing with RPMI-1640 medium and infected cultures were treated with increasing concentrations of raloxifene or miltefosine for 48 h . The monolayers were washed with PBS and luciferase detection was performed with the One Glo Luciferase Assay System ( Promega Corporation ) , according to the manufacturer's instructions . Briefly , 20 µL of reagent at room temperature were added to each well containing 100 µL of PBS followed by homogenization . Luminescence units were determined immediately after adding the substrate in a Polarstar Omega reader ( BMG Labtech ) . Drug activity against L . infantum chagasi intracellular amastigotes was determined in BMDM infected macrophages in 16-well chamber Slides ( NUNC ) ( 8×104 per well ) and infection was carried out at 37°C as described above for L . amazonensis . For the evaluation of parasite burden under light microscopy , 16-well chamber slides were fixed in methanol and stained with the Instant Prov kit ( Newprov , Pinhais-Paraná , Brazil ) . The percentage of infected macrophages was determined by counting 200 cells in each of the replicates . Logarithmic-phase L . amazonensis promastigotes ( 2×107/mL ) were incubated with 60 µM raloxifene for different periods of time ( 30 min , 2 h and 14 h ) at 25°C in M199 medium , supplemented with 10% FCS , in 24-well plates . Subsequently , promastigotes were centrifuged at 230 g for 10 min and the pellet was fixed in 2 . 5% glutaraldehyde : 4% paraformaldehyde in 0 . 1 M sodium cacodylate buffer ( pH 7 . 4 ) , rinsed in the same buffer and post-fixed in 1% osmium tetroxide , dehydrated in acetone series , and embedded in Epon resin . BMDM ( 4×105 cells per well ) were seeded in 24-well microplates on ACLAR film ( Electron Microscopy Sciences , USA ) cut into discs and allowed to adhere overnight at 37°C at 5% CO2 . Infection with L . amazonensis promastigotes was carried out at a ratio of 20∶1 parasites per macrophage . Infected macrophage cultures were kept at 33°C , 5% CO2 for 3 h in RPMI 1640 medium with 10% FCS and then washed twice with sterile PBS to remove free promastigotes . Infected cultures were treated with 9 µM raloxifene for 48 hours . The medium was removed and the monolayers were fixed as described above for promastigotes . Ultrathin sections were obtained in a Sorvall Ultramicrotome , stained with uranyl acetate and lead citrate and observed under a JEOL transmission electron microscope operating at 80 kV . Images were recorded with a Gatan 785 ES1000W Erlangshen camera . Parasites or infected macrophages without treatment were used as a control . Monodansylcadaverine ( MDC ) staining was used to label autophagic vacuoles as previously described [21] , [22] . L . amazonensis promastigotes ( 2×107/mL ) from early log phase cultures were treated with 30 and 60 µM raloxifene for 30 min in M199 medium supplemented with 10% FCS , at 25°C . As a positive control , parasites were subjected to starvation in PBS for 3 h [23] . Untreated parasites incubated in M199 medium for the same period were used as a negative control . Following the incubation period , cells were loaded with 100 µM MDC for 2 h . Cells were washed four times in PBS and fixed in 4% paraformaldehyde . One drop of fixed cells was added onto glass microscopic slides , covered by a coverslip and immediately visualized in a Zeiss AxioObserver microscope ( Oberkochen , Germany ) with excitation and emission wavelengths of 358 and 463 nm respectively and photographed using a digital camera ( AxionCam HRc , Zeiss ) . Twenty fields were randomly chosen from each sample ( magnification 630× or 1 , 000× ) . Assays were performed in two independent experiments . L . amazonensis promastigotes ( 5×106/mL ) were treated with 30 and 60 µM raloxifene for 20 min or 2 h in M199 medium , supplemented with 10% FCS at 25°C . Parasites treated with 25 µM digitonin were used as a positive control . Untreated parasites and parasites incubated with the highest volume of diluent ( DMSO 0 . 6% ) were used as negative controls . Parasites were stained with 10 µM propidium iodide ( PI ) and immediately analysed by flow cytometry using a Guava EasyCyte Mini Flow Cytometer System ( Millipore ) . A total of 5 , 000 events were acquired in the region previously established as corresponding to the parasites . Fluorescence was quantified using the CytoSoft 4 . 2 . 1 software ( Guava Technologies Inc . , Hayward , CA , USA ) . Histograms were drawn using FlowJo software , version 9 for Macintosh ( Tree Star , Inc . , Ashland , OR ) . Estimation of ΔΨp was monitored by measuring the increase in absorbance of bis- ( 1 , 3-diethylthiobarbituric acid ) trimethine oxonol [DiSBAC2 ( 3 ) ] ( Invitrogen ) as previously described [24] . Briefly , L . amazonensis promastigotes ( 2×107/mL ) were added into black polystyrene 96-well microplates in HBSS+Glc containing 0 . 2 µM DiSBAC2 ( 3 ) in a final volume of 100 µL per well . The plate was incubated at 25°C in a microplate reader ( POLARstar Omega , BMG Labtech ) and fluorescence was recorded ( λex = 544 nm; λem = 584 nm ) every 2 min . After signal stabilization raloxifene was added to final concentrations of 15 , 30 or 60 µM . Gramicidin D 8 µM ( Sigma-Aldrich ) was used as a positive control . Untreated parasites and parasites incubated with the highest volume of diluent ( DMSO 0 . 6% ) were used as negative controls . No interference in DiSBAC2 ( 3 ) fluorescence was observed when raloxifene was added to HBSS+Glc in the absence of cells . Three independent experiments were performed , each one with triplicate samples . Changes in ΔΨm were monitored by flow cytometry , using the fluorescent dye rhodamine 123 ( Rh123 ) ( Sigma-Aldrich ) as previously described [24] . L . amazonensis promastigotes ( 2×107/mL ) were resuspended in HBSS+Glc and incubated ( 1 mL final volume ) with increasing concentrations of raloxifene at 25°C for 20 min . After treatment , parasites were washed twice in HBSS+Glc , loaded with Rh123 ( 0 . 3 µg/mL , 10 min , 37°C ) , washed twice in HBSS+Glc and analysed by flow cytometry . Fluorescence emission was quantified using the CytoSoft 4 . 2 . 1 software as described above . Parasites depolarized with 100 µM carbonyl cyanide 4- ( trifluoromethoxy ) phenylhydrazone ( FCCP , Sigma-Aldrich ) were used as a positive control . Untreated parasites and parasites incubated with the highest volume of diluent ( DMSO 1 . 2% ) were used as negative controls . Results were obtained from three independent experiments . Female BALB/c mice ( 4 to 5 weeks-old ) were inoculated intradermally with 1×106 stationary-phase L . amazonensis promastigotes at the proximal end of the tail , as described [12] . Three weeks after infection , mice were randomly assigned into experimental groups . In a first series of experiments ( n = 2 ) , mice infected with wild type L . amazonensis were assigned to groups ( n = 10 ) that were either left untreated or received 40 mg/kg/day raloxifene by oral gavage in 100 µL final volume of saline . Treated animals received 10 doses of raloxifene , given on weekdays . A second series of experiments ( n = 2 ) was performed with mice infected as described above , except for the use of the luciferase expressing parasites , LaLUC . Treated groups ( n = 5 ) received 100 mg/kg/day raloxifene prepared with Cremophor A25 ( Sigma-Aldrich ) ( 100 mg/mL ) by oral gavage in 100 µL final volume . The control group received 100 µL of the vehicle used to dilute raloxifene . All animals received 10 doses of the assigned scheme , given on alternate days . In both series of experiments , disease progression was evaluated once a week by recording the average diameter of the tail measured as the mean of tail base diameters in horizontal and vertical directions . Measurements were taken with a caliper ( Mitutoyo Corp . , Japan ) . For the second series of experiments , parasite burden was evaluated at the end of the treatment ( 6 weeks post-infection ) through luciferase detection by bio-imaging ( IVIS Spectrum , Caliper Life Sciences , Inc . MA/USA ) as described [17] . Briefly , prior to imaging , mice received 75 mg/kg luciferin ( VivoGlo Luciferin , Promega ) intraperitoneally . Imaging was collected 20 min later , through high-resolution mode from a fixed-size region of interest . Results were quantified with Living Image software version 4 . 3 . 1 ( Caliper Life Sciences ) , and results were expressed as ph/sec/cm2/sr . Animal experiments were approved by the Ethics Committee for Animal Experimentation ( Protocol 033/42/02 ) . In vitro data were analysed for statistical significance by One-way ANOVA , followed by the Tukey post-test . Data on lesion progression and parasite burden were analysed for statistical significance by using the non-parametric Mann-Whitney test . Statistical analyses were performed using GraphPad Prism 5 software .
The activity of raloxifene was tested initially against promastigotes of different Leishmania species incubated in culture medium supplemented with FCS . Sensitivity was uniform across the genus , with EC50 values ranging from 30 . 2 to 38 . 0 µM after 24 h incubation ( Table 1 ) . Maximal effect was already observed after 2 h incubation with raloxifene in M199 medium , with an EC50 value of 30 . 6±1 . 0 µM against L . amazonensis promastigotes . The antileishmanial activity was more pronounced when the assay was carried out in HBSS+Glc , with an EC50 of 9 . 3±1 . 0 µM after 2 h incubation ( Table 1 ) . Cytotoxicity against the host cells in vitro was determined using cultures of J774 macrophages treated with raloxifene for 24 h with calculated CC50 of 28 . 6±0 . 5 µM . Drug activity was also tested ex vivo against L . amazonensis amastigotes obtained from mice infected tissue allowing the determination of an EC50 of 15 . 0±2 . 3 µM . Similar activity was demonstrated when raloxifene was used to treat cultures of BMDM infected with L . amazonensis or L . infantum chagasi ( Table 1 , Figure S1 ) . Raloxifene's Selectivity Index varied between 1 . 76 and 3 . 24 , for L . amazonensis and L . infantum chagasi intracellular amastigotes , respectively . The EC50 of miltefosine , a control standard drug against L . amazonensis promastigotes and intracellular amastigotes was calculated as 16 . 8±1 . 7 µM and 2 . 7±0 . 3 µM , respectively , in agreement with previously published data [25] . Transmission electron microscopy was used to investigate the effects of raloxifene on the parasite's ultrastructure and morphology . Untreated promastigotes showed typical ultrastructure ( Figure 1A ) while parasites treated with raloxifene displayed morphological alterations as early as 30 min after incubation with the drug , when vacuoles similar to autophagosomes were observed ( Figure 1B and C , arrows ) . After 2 h , severe mitochondrial damage was noted with marked swelling and loss of the matrix content ( Figure 1D and E , stars ) . In all cases the plasma membrane seemed to be continuous . The absence of membrane disruption was observed even in promastigotes treated with raloxifene for 14 h , which presented severe damage of the cytoplasm ( Figure 1F ) . In infected macrophages treated with raloxifene for 48 hours , parasitophorous vacuoles were filled with remnants of vacuolated cell bodies compatible with dead amastigotes ( Figure 1H , arrowheads ) . The ultrastructure of intracellular amastigotes in infected BMDM also displayed morphological alterations compatible with the formation of autophagosomes ( Figure 1I , arrow ) and mitochondrial swelling ( Figure 1I , star ) . MDC is considered a marker for the presence of autophagosomes [26] . Raloxifene-treated promastigotes were labeled with MDC and visualized by fluorescence microscopy . Spherical structures stained by the dye were observed in parasites treated with raloxifene for 30 min ( Figure 2C and D ) , in contrast to the weak and diffuse overall cytoplasmic staining in untreated parasites ( Figure 2A ) . PBS starvation was used as a known inducer of autophagy [27] . In these conditions promastigotes concentrated the label in round bodies ( Figure 2B ) with a pattern similar to the one observed after treatment with raloxifene . Membrane integrity can also be evaluated through permeation of vital dyes such as PI . Permeation of PI in promastigotes treated with raloxifene for 20 min or 2 h was measured by flow cytometry . While parasites treated with 25 µM digitonin showed early and pronounced increase in the uptake of PI , indicative of membrane damage , treatment with 30 or 60 µM raloxifene did not induce any significant changes in fluorescence after 20 min or 2 h ( Figure 3 ) . To verify whether raloxifene alters plasma membrane functions which could not be observed through transmission electron microscopy or PI staining , ΔΨp was monitored using DiSBAC2 ( 3 ) as a fluorescent probe . In cells loaded with the voltage sensitive fluorescent dye , raloxifene induced an early decrease in fluorescence indicating membrane hyperpolarization ( Figure 4 ) . This early effect was followed , after the first 5 min of treatment , by a dose-related fluorescence intensity increase to reach levels indicative of total depolarization after 30 min incubation . This was confirmed by treatment with the nonselective ionophore gramicidin , which only marginally increased the fluorescence in cells previously treated with raloxifene while a significant increase in fluorescence was noted in control untreated parasites . In order to confirm the mitochondrial damage observed under transmission electron microscopy , the mitochondrial function was evaluated using the fluorescent probe Rh123 by flow cytometry . Parasites treated with increasing concentrations of raloxifene for 20 min exhibited a gradual inability to concentrate the dye , indicating a progressive collapse of the ΔΨm ( Figure 5 ) . The EC50 calculated based on Rh123 fluorescence was 9 . 42±1 . 03 µM , in accordance with the EC50 based on mitochondrial activity measured in parasites incubated in HBSS+Glc for 2 h . The mitochondrial protonophore FCCP was used as positive control of mitochondrial membrane depolarization and induced a reduction of 63 . 5±0 . 9% in Rh123 accumulation . Treatment of L . amazonensis-infected BALB/c mice was initiated 3 weeks post-infection and the progression of lesion thickness in untreated and raloxifene-treated mice was recorded weekly . In a first series of experiments , treated mice received 40 mg/kg/day raloxifene orally for 10 doses on weekdays . Significant reduction in lesion thickness in treated groups was noticed ( Figure 6A ) . Five weeks after the end of treatment , the average size of lesions in treated groups was reduced by 41 . 7% as compared with control mice . After the interruption of treatment , lesion size in treated groups increased but did not reach sizes observed in control untreated animals ( Figure 6A ) . In a second series of experiments , mice were treated with 100 mg/kg/day orally in a total of 10 doses in alternate days . In this case , significant decrease in the lesion size was more pronounced with a 54 . 3% reduction in the average lesion size of treated mice five weeks after the end of treatment ( Figure 6B and C ) . In order to estimate the parasite burden in the site of infection , light emission of LaLUC parasites was recorded at the end of raloxifene treatment ( 6 weeks post-infection ) . Parasite burden was reduced by 89 . 7% in the group treated with raloxifene when compared with the untreated group , as indicated by bioluminescence quantification ( Figure 6D and E ) . Untreated and vehicle-treated groups were not significantly different ( data not shown ) .
The activity of the triphenylethilene tamoxifen against Leishmania , in vitro and in vivo was previously demonstrated [12] , [14] , [15] . Tamoxifen and raloxifene are classified as SERMs , a class of compounds that exhibit agonist and antagonist estrogenic effects depending on the target tissue . Here , we report that raloxifene – a benzothiophene belonging to a SERM class distinct from tamoxifen's – also presents in vitro broad-spectrum activity against different Leishmania species . This is the first report of the activity of raloxifene against protozoan parasites . Tamoxifen has been shown to be strongly incorporated into biomembranes , to disrupt membrane structure and to induce mitochondrial permeabilization [28] , characteristics that can be correlated with its high hydrophobicity ( LogP = 5 . 93 ) and are estrogen receptor-independent . Raloxifene is also highly lipophilic ( LogP = 5 . 69 ) and possesses estrogen receptor independent activities [29] . The observation of continuous plasma membrane ultrastructure as well as the lack of increased permeability to PI in raloxifene-treated promastigotes indicated absence of membrane disruption . On the other hand , more subtle membrane lesions can be monitored using potential-sensitive fluorescent probes such as DiSBAC2 ( 3 ) which enters depolarized cells where binding to intracellular proteins causes enhanced fluorescence . Interestingly , treatment with raloxifene induced an early hyperpolarization followed by dissipation of the membrane potential . The early event may be due to fast ionic currents , observed as drug interaction with the membrane takes place . This is followed by the collapse of ionic gradients across the membrane leading to depolarization . At this point , it is unclear whether the loss of membrane potential is due to a disarrayed membrane structure or to inhibition of ion channels . Interestingly , previous studies have shown that raloxifene inhibits voltage dependent calcium currents in mouse spermatogenic cells in an estrogen receptor independent way [30] . The presence of a voltage gated calcium channel sharing several characteristics with the human counterpart has been recently demonstrated in the plasma membrane of Leishmania [31] . Accordingly , calcium channel blockers used as anti-hypertensive drugs , have been shown to possess antileishmanial activity [32] , [33] , [34] . Thus , interference in calcium channels in raloxifene treated parasites cannot be ruled out . Raloxifene was previously shown to decrease ΔΨm in human endometrial carcinoma cells [35] . In raloxifene-treated parasites we observed severe mitochondrial damage by transmission electron microscopy and Rh123 labeling . As a result , the energy-coupling system in the mitochondria is most likely inactivated . The observation of autophagic vacuoles by transmission electron microscopy and the presence of MDC-labeling vacuoles in raloxifene-treated promastigotes suggest the participation of autophagy in raloxifene's mode of action . This is in accordance with previous findings indicating that SERMs induce autophagy in tumor cells [36] , [37] . Autophagosomes and MDC-labeled vacuoles have been previously described as effects of antileishmanial compounds [38] , [39] . Previous to raloxifene-induced cell death , loss of mitochondrial membrane potential and entrapment of cytoplasm content within autophagosomes were noted . The presence of autophagic vacuoles is most likely related to the degradation of damaged organelles induced by raloxifene treatment . Based on the obtained results , we propose the following sequence of events triggered by raloxifene: i ) partition of raloxifene in the plasma membrane; ii ) alteration in plasma membrane potential; iii ) mitochondrial membrane depolarization; iv ) mitochondrial dysfunction; v ) autophagy; vi ) parasite death . Raloxifene was developed as an anti-estrogen against breast cancer and was later found out to be a therapeutic agent for postmenopausal osteoporosis . Pharmacokinetic studies have shown good oral absorption; glucuronide conjugates are formed after extensive first-pass metabolism in the intestinal mucosa and liver . Acute toxicity is low with no mortality observed with doses up to 5000 mg/kg orally in mice and rats . Preclinical toxicology studies indicated that raloxifene was well tolerated in repeated dose assays in mice , rats , dogs and monkeys . In mice , daily doses of up to 120 mg/kg raloxifene for 3 months induced no serious toxic effects [40] . Based on this data , we chose the starting oral dose of 40 mg/kg/day ( 3 times the human dose for osteoporosis treatment and prevention , based on body surface area ) [41] for initial in vivo tests in a murine model of cutaneous leishmaniasis . Treatment of L . amazonensis infected BALB/c mice with raloxifene resulted in significant decrease in lesion size . As the lesion in treated animals did not heal , we then increased the dose to 100 mg/kg/day . As the drug is poorly soluble in water , for this second series of experiment raloxifene was prepared in Cremophor A25 , which is a vehicle capable of yielding stable emulsions of hydrophobic , pharmacologically active biomolecules . Treatment of L . amazonensis infected BALB/c mice with 100 mg/kg/day raloxifene resulted in significant decrease in lesion size and parasite burden . No toxic effects were observed during drug administration . Raloxifene-treated animals did not heal the lesions completely and , after the interruption of treatment , lesions worsened . However , these lesions did not reach sizes observed in untreated controls . Since BALB/c mice infected with L . amazonensis represents a model of extreme susceptibility , the significant reduction in the number of parasites in treated animals supports the proposal of further testing of this drug in other animal models of leishmaniasis . Furthermore , these data warrant the consideration of this molecule as a lead for further development . In fact , recent results from our laboratory indicate that the antileishmanial potency of synthetic benzothiophenes is increased 10-fold ( as compared to raloxifene ) by the presence of two basic side chains in the molecule [42] . Furthermore , structure-activity data showed that the most active antileishmanial benzothiophenes lack the pharmacophore for estrogen receptor activity confirming that the antileishmanial activity observed for benzothiophenes is independent of the interaction with the estrogen receptors [42] . In conclusion , the results of this work extend the investigation of SERMs as potential candidates for leishmaniasis treatment . Raloxifene's activity in vitro is mediated by functional damage to the plasma and mitochondrial membranes , which culminate in cell death . Further studies are necessary to ascertain whether other antileishmanial mechanisms are engaged in vivo .
|
Cutaneous and visceral leishmaniasis are part of the group we call neglected diseases . They are serious conditions that afflict millions in vast regions of the world . These diseases are very difficult to treat . This is due to the scanty choice of effective drugs together with their potentially severe side effects . One way of finding new treatments for these neglected conditions is to repurpose drugs that are already in use to treat other diseases . In this paper , we show that raloxifene , a drug that is used for the treatment of osteoporosis and also as an alternative in the treatment of breast cancer , is active against the causative agents of leishmaniasis and is effective in the treatment of cutaneous leishmaniasis in an experimental model . We also show that the antileishmanial mechanism of action of raloxifene is related to damage to the cell membrane and to the mitochondrion of the parasite .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases",
"veterinary",
"diseases",
"zoonoses",
"medicine",
"and",
"health",
"sciences",
"clinical",
"medicine",
"drug",
"research",
"and",
"development",
"leishmaniasis",
"neglected",
"tropical",
"diseases",
"biology",
"and",
"life",
"sciences",
"pharmacology",
"drug",
"discovery",
"tropical",
"diseases",
"veterinary",
"science"
] |
2014
|
Antileishmanial Activity of the Estrogen Receptor Modulator Raloxifene
|
Despite control efforts , human schistosomiasis remains prevalent throughout Africa , Asia , and South America . The global schistosomiasis burden has changed little since the new anthelmintic drug , praziquantel , promised widespread control . We evaluated large-scale schistosomiasis control attempts over the past century and across the globe by identifying factors that predict control program success: snail control ( e . g . , molluscicides or biological control ) , mass drug administrations ( MDA ) with praziquantel , or a combined strategy using both . For data , we compiled historical information on control tactics and their quantitative outcomes for all 83 countries and territories in which: ( i ) schistosomiasis was allegedly endemic during the 20th century , and ( ii ) schistosomiasis remains endemic , or ( iii ) schistosomiasis has been "eliminated , " or is "no longer endemic , " or transmission has been interrupted . Widespread snail control reduced prevalence by 92 ± 5% ( N = 19 ) vs . 37 ± 7% ( N = 29 ) for programs using little or no snail control . In addition , ecological , economic , and political factors contributed to schistosomiasis elimination . For instance , snail control was most common and widespread in wealthier countries and when control began earlier in the 20th century . Snail control has been the most effective way to reduce schistosomiasis prevalence . Despite evidence that snail control leads to long-term disease reduction and elimination , most current schistosomiasis control efforts emphasize MDA using praziquantel over snail control . Combining drug-based control programs with affordable snail control seems the best strategy for eliminating schistosomiasis .
Can we do better at controlling schistosomiasis ? Despite effective drug treatment options and large-scale drug distribution programs , most endemic areas have not yet achieved satisfactory schistosomiasis control . Today , schistosomiasis remains prevalent in Africa , Asia , and South America where trends over time forecast perpetual endemicity . Sometimes , endemicity has been because poverty constrains control efforts; otherwise , endemicity is due to failed or ineffective control attempts . With more than 250 million people still infected and elimination stalled [1 , 2] , the World Health Assembly ( WHA ) called for researching and applying complementary , non-pharmaceutical control strategies for eliminating schistosomiasis in its 2012 resolution 65 . 21 [3] . This resolution ignited debate over the best strategies for eliminating schistosomiasis [4–9] . To add quantitative data to this debate , we evaluated schistosomiasis control strategies over the past century and screened for factors associated with elimination or long-term prevalence reductions . Human schistosomiasis occurs where aquatic ( or amphibious ) intermediate host snails shed infective Schistosoma spp . cercariae that penetrate human skin upon contact . Infected humans suffer from anemia , stunted growth , cognitive impairment , fatigue , infertility , and sometimes , liver fibrosis or bladder cancer [10] . Most affected people live in poverty where there are few resources for research and control [11] . Schistosomiasis control efforts aim to disrupt the parasite’s complex life cycle ( Fig 1 ) : sanitation stops parasite eggs in urine or feces from moving into aquatic snail habitats; snail control reduces intermediate host density ( parasite larvae reproduce asexually in snails ) ; education ( or “information education and communication”; IEC ) helps people avoid high-risk water contacts and know when to seek treatment; and drugs—given as mass drug administrations ( MDA ) , targeted treatment campaigns ( “test-and-treat” or TAT ) , or through health services—kill the adult worm in the human host , with immediate and long-term health benefits for infected individuals [12] . Such efforts seem simple , but implementation often fails for economic or political reasons . Consensus on the “best” schistosomiasis control strategy has varied over the past century . Early Egyptian control efforts around the turn of the 20th century treated human infection , then shifted toward borehole latrines as a sanitary measure in the 1930s , but with little success [13] . In the 1940s the focus shifted again , this time toward snail control using copper sulfate [13] . Control strategy recommendations by the World Health Organization ( WHO ) then emphasized integrated control measures , including clean water access , sanitation , snail control , health education , and health services , in addition to drug treatments [12] . In the late 1970s and early 1980s –after praziquantel became the drug-of-choice for treating schistosomiasis [14]–the integrated approach was often supplanted by chemotherapy via MDA . Praziquantel is safe and effective against adult schistosomes [15] , but is ineffective against juvenile worms [16–18] , and drug treatment does not prevent reinfection [19–23] . Nonetheless , large-scale research projects and influential modeling results suggested that widespread drug treatment ( best when coupled with sanitation measures ) would reduce schistosomiasis more than other interventions [24–27] . MDA increased in the 1990s after generic ( inexpensive ) praziquantel became available [15 , 28] . Thus , in 2001 , the WHA endorsed preventive chemotherapy as the primary strategy to control schistosomiasis through reducing morbidity associated with high worm burdens [29] . Specifically , Engels et al . [30] summarized the modern , WHO-recommended schistosomiasis global control strategy as MDA in high-transmission areas to reduce morbidity and transmission reduction in low-transmission areas ( Fig 2a ) . Although MDA is now the most popular control strategy , some have argued that snail control is more effective [5 , 6 , 31–33] . However , a Center for Global Development working group focused on successful global health interventions points out: “We don’t know enough about what’s worked because scaled up programs are rarely evaluated systematically” [34] . Here , using objective criteria and a quantitative analysis to test for commonalities among successful control programs , we find snail control has been effective at reducing schistosomiasis .
We evaluated control programs for all areas around the world with active ( autochthonous ) human schistosomiasis transmission at some time in the 20th century . We considered countries with little to no control effort as having minimal control . However if minimal control corresponded to a loss of schistosomiasis , we defined the outcome as fortuitous elimination . We defined control as not ( yet ) successful where ( i ) control has been incomplete , ( ii ) transmission continues , or ( iii ) the disease has been almost , but not yet , eliminated . We defined control as successful for active programs that were reported to have stopped local transmission ( i . e . , elimination or becoming “non-endemic” ) in WHO reports or peer-reviewed assessments ( e . g . , [9 , 35] ) . “Elimination” implies reducing disease incidence to zero in a particular area [30 , 36] . There is some inconsistency in the literature on the term “eradication , ” which often refers to global disease extirpation [37] . With respect to schistosomiasis , this has been applied to the regional elimination achieved by Japan . Here , because elimination , “eradication , ” and non-endemicity all imply no local transmission , we treated these designations as successful . Our goal was to evaluate control success in all countries and territories with endemic schistosomiasis during the 20th century . We began with the nine countries often cited as “success stories” for schistosomiasis elimination: Iran , Japan , Lebanon , Malaysia , Martinique , Montserrat , Thailand , Tunisia , and Turkey [38] . Antigua , Jordan , and Morocco were three other potential “successes” [35 , 39] . We were also interested in countries that achieved great reductions in schistosomiasis prevalence including Brazil , China , the Philippines , and Egypt . Additional literature searches focused on characterizing disease and control history for all additional countries with: ( i ) historical disease data and ( ii ) recorded national- or territory-level schistosomiasis control programs . Although we found relevant data for most countries , data were contradictory for several Caribbean islands such as Guadeloupe and Dominican Republic , with some reports indicating elimination and others claiming ongoing risk . We considered these countries to be “not ( yet ) successful . ” For several in-conflict countries such as Chad and Syria , current schistosomiasis prevalence is “unknown” , with the potential for conflict and political unrest to hinder control [40] . We obtained country-specific data for several categories ( S1 Table ) by reading peer-reviewed published sources as well as non-peer-reviewed reports accessed through online ( or hard copy ) repositories , including PubMed , ISI Web of Science , Google Scholar , WHO , United Nations ( UN ) , World Bank , United States Agency for International Development , and the UN Food and Agriculture Organization ( see S1 Appendix for complete reference list ) . From these sources , we assessed 77 countries and six semi-autonomous territories ( including Western Sahara in northern Africa , Guadeloupe , Martinique , Montserrat , Puerto Rico , and Zanzibar ) . For each country , we collected information on schistosomiasis , control efforts , parasite life cycles , environmental factors , and economics . We focused on variables related to national schistosomiasis data ( country- or territory-wide prevalence , infected population size , at-risk population size ) and details about the control strategies implemented and their time-course . We also recorded snail and schistosome species present; island or mainland geography; and per-capita gross domestic product ( GDP ) in 2013 and in all years for which schistosomiasis disease data were available in each country . Further , we noted site-specific factors that might alter disease outcomes or resources for control activities ( S1 Table ) . We limited the prevalence information , in almost every case , to country-level ( or territory-level ) statistics . Only for Japan , where finer scale data were available over many years , did we use large-scale and long-term regional data to assess trends , and we included only the data from the largest endemic area ( the Kofu basin ) in the statistical analyses . We were careful to avoid small-scale , focal studies on prevalence that might not represent the whole country . We treated countries/territories as replicates in statistical analyses done in JMP Pro version 12 [41] and R version 3 . 1 . 2 [42] . To test the general hypothesis that control programs can eliminate schistosomiasis , we assessed whether schistosomiasis was eliminated/non-endemic using a logistic regression , with five predictors for 68 countries/territories ( excluding 15 that lacked enough data or were designated non-endemic to begin with ) : ( i ) the presence/absence of a national- or territory-level control program; ( ii ) status as a mainland or island ( because it should be easier to achieve elimination with more isolation ) ; ( iii ) the total human population infected with schistosomiasis at baseline , or before control began ( because it might be harder to eliminate schistosomiasis when the starting infected population is large ) ; ( iv ) the current fraction of people with access to improved water sources ( as a proxy for contemporary water , sanitation , and hygiene conditions , World Bank Development Indicators , 2012 [43] ) ; and ( v ) contemporary per-capita GDP ( as a proxy for “wealth” status , World Bank Development Indicators , 2013 [44]; S2 Table ) . Using 68 countries gave us considerable statistical power to evaluate when and where control has been effective in eliminating schistosomiasis . We next compared how well different control strategies reduced disease . The strategies used in historical schistosomiasis control efforts were categorized as: MDA , snail control , or engineering interventions ( e . g . , sanitation infrastructure , cement lined canals , drained wetlands ) . Each control category was further sorted according to our best estimate ( based on qualitative descriptions , or sometimes , quantitative reports ) : extensive/complete ( >70% of the population/area in need received treatment ) , intermediate ( >30% ) , or focal to none ( <30% ) . Although disease can be measured as intensity [45 , 46] ( as indicated by patient egg output ) , there was not enough published national-level data to assess intensity means and variances . Therefore , we compared disease prevalence on a continuous scale ( 0 to 100% based on the schistosomiasis national prevalence at each available time point for each country/territory ) . We included only countries with national control programs and enough longitudinal disease data . We excluded those countries/territories with no coordinated control effort ( “Minimal control” in Table 1 , Fig 3 ) . Further , among national control programs , we designated the category “low coverage” where snail control , or MDA , or both were included , but the program achieved low ( <30% ) coverage for either strategy . Similarly , engineering controls were considered present only where their coverage was high ( i . e . where more than 30% coverage was achieved ) . To test the hypothesis that control strategies differed in their ability to reduce prevalence , we used a quantitative generalized linear mixed model ( GLMM , function “glmer” from the R package “lme4” ( [42] , S2 Table ) . The statistical model assessed what factors best predicted relative change in prevalence over time for the 44 countries that applied concerted control and had quantitative , longitudinal data on prevalence , control strategies and covariates ( more details below and in S2 Table ) . This GLMM considered country as a random effect ( to account for the repeated measures over time within each country ) and the following fixed effects: ( i ) control program duration ( to test if longer efforts might be more successful ) ; ( ii ) a country’s status as an island or mainland; ( iii ) initial prevalence before control began ( to account for the control effort needed ) ; ( iv ) the percentage of the population with access to improved water sources ( World Bank Indicators , 2012 [43] ) ; and ( v ) the inflation-adjusted per capita GDP over time ( recorded at each time point with disease data from The Maddison-Project [47] ) . We were most interested in the interaction terms between the predictors and time [year] , which , if significant , would indicate an effect on prevalence reduction or increase over time . We first assembled a “full model” that contained all predictors and interaction terms , and then used a model selection procedure based on Akaike’s information criterion ( AIC ) to remove each interaction term in turn to find the best balance between parsimony and fit to the data ( [48] , S3 Table ) . After analyzing what control strategies were most successful , we became curious about the factors that might have determined which control strategies a country used . To that end , we assessed the correlations between the control strategies used and a country’s “wealth” status ( per-capita GDP for each country at each time-point ) as well as the control era ( the year each national- or territory-level control program began ) .
Counter to expectations , elimination/non-endemicity was not associated with having a control program . This unexpected result was due to two factors: several countries/territories achieved “fortuitous” elimination without any documented control effort ( Antigua , Djibouti , Malaysia , Montserrat , Thailand , and Turkey ) and several other countries failed to eliminate schistosomiasis , despite substantial prevalence reductions . Island/mainland did not predict elimination status ( Table 2 ) , however , our inclusion of population size , which is higher on continents and makes elimination harder , could have co-varied with a mainland-island effect . Possessing greater “wealth” ( indicated by a higher contemporary per capita GDP ) did not affect elimination . Elimination was , however , more likely where more people can access improved water sources . In summary , achieving elimination was idiosyncratic . It was easier with smaller infected populations and in countries with improved ( safer ) water sources . Although many programs have failed to eliminate schistosomiasis , sometimes elimination has occurred without a coordinated control program . Below , we discuss what factors in addition to control programs could affect schistosomiasis prevalence reductions and elimination success . Although fortuitous elimination in several countries confounded our ability to assess whether control programs eliminated schistosomiasis , many areas with control programs experienced durable prevalence reductions . A program’s effectiveness ( i . e . , the prevalence reduction rate ) depended strongly on strategy type and coverage and weakly on the intercept ( prevalence at baseline ) . Applying snail control , MDA , or both—with at least intermediate ( >30% ) coverage—worked better than any programs with low coverage . Snail control programs ( primarily mollusciciding and biological control using non-native , competitor snails ) showed the strongest prevalence reductions ( while accounting for other covariates , including: control duration [in years] , country “wealth” [as per capita GDP in each year with disease data] , and access to improved water sources; Table 3 , Figs 5 and 6 ) . In other words , all else being equal , prevalence reduction was highest with snail control at intermediate or better coverage . Although engineering controls ( e . g . , installing sanitation infrastructure , cementing canals , building bridges , or draining wetlands ) , were almost always accompanied by snail control , about half of the programs using snail control did not use engineering in their control programs . Programs that used MDA as a primary strategy ( without snail control ) also did not report using any large-scale engineering controls . The presence or absence and extent of engineering controls showed weak effects on prevalence , and including 3-way interactions with this variable—along with the other control strategies and time—in the quantitative statistical model did not improve model fit to the data ( based on AIC; S3 Table ) . Thus , engineering controls , although perhaps beneficial within some integrated programs , did not consistently reduce schistosomiasis prevalence . Population size affected control success . As expected , prevalence reductions were impaired where there were larger initial infected human populations , but this relationship differed among the control strategies . Snail control programs ( with or without MDA ) were less sensitive to initial infected human population size , than were other approaches ( Fig 7 ) . Control strategy depended on country wealth and the year in which control began . Richer countries ( measured by inflation-adjusted , per-capita GDP ) tended to begin their control programs earlier in the 20th century , with a stronger focus on snail control and greater success ( Fig 8 ) . Higher wealth was also correlated with greater access to improved ( safer ) water sources . Large-scale MDA programs were rare before praziquantel entered the global market in the 1980s . After this turning point , there was a new option ( using both MDA and snail control ) and this integrated strategy has been used since the 1980s in places like China , Egypt , and Brazil . Countries that began their control programs even more recently ( after the 1990s or 2000s ) were poorer and tended either to focus on MDA or achieved poor coverage ( designated as “low coverage” in Fig 8 ) .
Our results support recent suggestions that snail control is key to schistosomiasis reduction [5 , 6 , 49] . Such an effect has been anticipated . In 1985 , a lead researcher of the Caribbean “St . Lucia Project , ”–a Rockefeller-funded schistosomiasis control study—wrote , “chemotherapy is now assuming the major role in control programmes , but in most… a reservoir of infection inevitably remains . Transmission is thus likely to continue at a low , but probably increasing level unless a supplementary control strategy is present” [24] . Unfortunately , it took decades to assess this prediction . Chemotherapy has major benefits for infected humans , but , by itself , MDA has done little to curb re-infection . Although programs limited to MDA with praziquantel did not appear to do as well as the other strategies evaluated , the ( targeted ) chemotherapy for infection control remains an undeniable factor in improving health , especially when integrated with snail control . Countries whose programs focused on snail control often relied on distributing chemotherapy through means other than MDA , such as Morocco’s successful test-and-treat ( TAT ) campaigns using mobile teams [50] , Iraq’s early school-based TAT programs [51] , and Japan’s involvement as an early TAT site for praziquantel beginning in the late 1970s . This involvement might have carried Japan to country-wide elimination by 1996 [52] . One reason praziquantel seems less effective than expected is that it was applied later in history when control campaigns targeted more challenging countries . In other words , schistosomiasis elimination was more successful among programs started before praziquantel reached the global market than among those programs started after the drug’s introduction in the late 1970s . This might arise , in part , because wealthier countries tended to address the disease earlier in the 20th century , as they could afford molluscicides for widespread snail control ( although inexpensive biological control also sometimes succeeded ) . The “fortuitous elimination” of schistosomiasis from Antigua , Djibouti , Malaysia , Montserrat , Thailand , and Turkey without documented control efforts suggests cryptic factors have affected schistosomiasis , including: species invasions ( e . g . snail competitors or predators ) , sanitation or health care improvements outside control programs , and human-induced or natural ecosystem changes ( such as changes in dams , irrigated-agriculture , and urbanization ) . The least fortuitous of the fortuitous eliminations was the 1995 volcanic eruption that drove almost half of Montserrat’s population off the island and made the schistosomiasis transmission zones off limits to people [53 , 54] . ( See S2 Appendix and S1 Fig for more cryptic schistosomiasis control examples ) . These results suggest that programs have been most effective when snail control is coordinated soon thereafter—or simultaneously—with chemotherapy ( morbidity control ) via a rational progression from widespread , active drug distribution campaigns ( MDA or targeted treatment ( TAT ) ) to a focus on high-risk groups and finally passive distribution within health services coupled with surveillance ( e . g . “surveillance and response”[55 , 56] ) and health education ( e . g . IEC ) in the “end game” ( Fig 2b ) . As for how to control snails , the most common strategy has been to use expensive and toxic molluscicides; an effort that is neither feasible nor desirable for many poor countries where schistosomiasis is now endemic . Schistosomiasis has been hard to control without well-funded , national-level efforts , and the contemporary global health discussion has been focused on strategies that optimize efficiency and affordability . By recognizing the successful use of snail control for transmission reduction , and by fostering research directed toward the development of creative , safe and cheap tools to target the snail intermediate host , global schistosomiasis elimination might be attainable .
|
Schistosomiasis is a parasitic disease infecting more than 250 million people worldwide , with almost 800 million at risk . Over the past century , nations undertook schistosomiasis control programs , with outcomes varying from little effect to elimination . The biggest hope for elimination began about 40 years ago with the discovery of the antischistosomal drug praziquantel , after which snail control was seen as old fashioned . Here , we review control program outcomes over the past 100 years across all major schistosomiasis endemic zones , including Africa , Asia , and the Americas . We screened for differences in long-term schistosomiasis reductions among countries and found the most successful programs focused on transmission control ( most often snail control , with or without engineering interventions ) , sometimes in tandem with praziquantel . Although praziquantel has important human-health benefits , our results suggest old-fashioned snail control has been the key to schistosomiasis elimination .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"engineering",
"and",
"technology",
"cancer",
"treatment",
"clinical",
"oncology",
"tropical",
"diseases",
"parasitic",
"diseases",
"animals",
"health",
"care",
"oncology",
"control",
"engineering",
"clinical",
"medicine",
"gastropods",
"pharmaceutics",
"neglected",
"tropical",
"diseases",
"sanitation",
"infectious",
"disease",
"control",
"snails",
"public",
"and",
"occupational",
"health",
"infectious",
"diseases",
"molluscs",
"helminth",
"infections",
"schistosomiasis",
"environmental",
"health",
"chemotherapy",
"biology",
"and",
"life",
"sciences",
"drug",
"therapy",
"organisms"
] |
2016
|
Global Assessment of Schistosomiasis Control Over the Past Century Shows Targeting the Snail Intermediate Host Works Best
|
Type 1A topoisomerases ( topos ) are the only ubiquitous topos . E . coli has two type 1A topos , topo I ( topA ) and topo III ( topB ) . Topo I relaxes negative supercoiling in part to inhibit R-loop formation . To grow , topA mutants acquire compensatory mutations , base substitutions in gyrA or gyrB ( gyrase ) or amplifications of a DNA region including parC and parE ( topo IV ) . topB mutants grow normally and topo III binds tightly to single-stranded DNA . What functions topo I and III share in vivo and how cells lacking these important enzymes can survive is unclear . Previously , a gyrB ( Ts ) compensatory mutation was used to construct topA topB null mutants . These mutants form very long filaments and accumulate diffuse DNA , phenotypes that appears to be related to replication from R-loops . Here , next generation sequencing and qPCR for marker frequency analysis were used to further define the functions of type 1A topos . The results reveal the presence of a RNase HI-sensitive origin of replication in the terminus ( Ter ) region of the chromosome that is more active in topA topB cells than in topA and rnhA ( RNase HI ) null cells . The S9 . 6 antibodies specific to DNA:RNA hybrids were used in dot-blot experiments to show the accumulation of R-loops in rnhA , topA and topA topB null cells . Moreover topA topB gyrB ( Ts ) strains , but not a topA gyrB ( Ts ) strain , were found to carry a parC parE amplification . When a topA gyrB ( Ts ) mutant carried a plasmid producing topo IV , topB null transductants did not have parC parE amplifications . Altogether , the data indicate that in E . coli type 1A topos are required to inhibit R-loop formation/accumulation mostly to prevent unregulated replication in Ter , and that they are essential to prevent excess negative supercoiling and its detrimental effects on cell growth and survival .
DNA topoisomerases ( topos ) are nicking-closing enzymes that solve the topological problems inherent to the double-helical structure of the DNA [1] . Such problems arise during DNA transactions including replication , transcription and recombination and must be solved in order for these processes to be completed , and to maintain the integrity of the genome . Type I topos cut one DNA strand and either use a strand passage ( 1A ) or a rotation mechanism ( 1B ) to alter the topology of DNA . Type II enzymes cut two DNA strands and use a strand passage mechanism to change the topology of DNA . Type 1A topos are the sole ubiquitous topos , being present in the three domains of life [2 , 3] . They require single-stranded DNA ( ssDNA ) for binding , a substrate that is mostly provided by negatively supercoiled DNA for some of these enzymes , as is the case for bacterial topos I , the members of the first type 1A topo subfamily , encoded by topA [1] . The prototype enzyme of this group , also the first topo to be discovered , is topo I from E . coli [4] . To form visible colonies on solid media , E . coli topA null mutants need to acquire compensatory mutations in gyrA or gyrB , reducing the supercoiling activity of DNA gyrase [5 , 6] , or amplifications of a chromosomal DNA region carrying parC and parE [7 , 8] , the two subunits of topo IV ( a type II topo ) , increasing the cellular DNA relaxation activity . In fact , topo IV , at its wild-type cellular level , does play a role in the regulation of supercoiling in E . coli [9] . These observations lead to a model in which the chromosomal negative supercoiling in the cell is set by the opposing enzymatic activities of DNA gyrase , introducing negative supercoiling , and topo I and IV relaxing it [9] . In fact , the major function of topo I in supercoiling regulation , is the relaxation of transcription-induced negative supercoiling behind moving RNA polymerases ( RNAPs ) [10] . Topo I physically interacts with RNAP in E . coli [11 , 12] and this interaction is also found between M . tuberculosis topo I and RNAP , though in that case by a distinct mechanism [13] . In E . coli , one major consequence of transcription-induced negative supercoiling if not relaxed , e . g . in topA null mutants , is R-loop formation [14] . An R-loop forms when the nascent RNA re-anneals with the template strand behind the moving RNAP [14] . Overproducing RNase HI ( rnhA ) , the enzyme degrading the RNA of an R-loop , significantly rescue the growth defect of topA null mutants [15] , and double topA rnhA null mutants do not grow despite the presence of a gyrB compensatory mutation [15 , 16] . Evidence for R-loop formation inhibiting transcription , causing RNA degradation and hypernegative supercoiling in topA null mutants has been presented [17] . Furthermore , R-loop formation in yeast and human cells has been reported [18–21] . It is now clear that R-loops can have a wide range of positive and negative effects on cell physiology [22 , 23] . A second group of type 1A topos , the topo III subfamily , is widely distributed in the three domains of life [2 , 3] . Topo III is characterized by a stronger requirement for ssDNA as compared to topo I [1] . The first enzyme of this subfamily has been found in E . coli and is encoded by topB [24 , 25] . In vitro , E . coli topo III has a strong decatenase activity but a poor relaxation activity on DNA with a wild-type supercoiling density at 37°C [24] . In fact , topo III plays no role in global supercoiling regulation in E . coli [9 , 26] . As opposed to topA mutants , topB mutants grow normally , do not accumulate compensatory mutations and display no obvious phenotypes [25] . In vitro topo III can act as a decatenase to fully support replication of a circular dsDNA template [27] . Evidence for topo III acting as a replicative decatenase , as a back-up for the main cellular decatenase topo IV has been presented [28] . In vitro , an R-looped DNA template strongly stimulates the relaxation activity of topo III [29] . The initial studies performed with double topA topB null mutants of E . coli , lead to the conclusion that both topo I and III share a unique and essential function [30] . topA null cells carrying a gyrase compensatory mutation and depleted of topo III activity generated very long filaments fully packed with unsegregated and diffuse DNA , and eventually stopped growing [30] . These phenotypes were not observed for single type 1A topo mutants . Additionally , cells lacking both type 1A topos were said to be non-viable as topB null transductants of a topA null strain could not be obtained after an overnight incubation on minimal medium . As these phenotypes could be corrected by deleting recA , it was also concluded that they were related to homologous recombination [30] . In later studies , topA topB null transductants in strains carrying a gyrase compensatory mutation could be obtained after 48 hours of incubation [31 , 32] . These results demonstrated that E . coli cells can survive without type 1A topos but did not show if compensatory mutations were required for viability , and therefore how such mutants could survive . The fact that the typical phenotypes of cells lacking type 1A topos were observed , were exacerbated by decreasing the incubation temperature in strain carrying a gyrB ( Ts ) compensatory mutation , and were largely corrected by overproducing RNase HI , suggested that deleting topB in topA null mutants exacerbated the supercoiling and R-loop-dependent phenotypes of topA null cells [33] . More recently , replication from R-loops was shown to be activated in topA topB null cells [34] . Replication from R-loops was first described in E . coli cells lacking RNase HI ( rnhA ) [35 , 36] . This replication was named “constitutive stable DNA replication” ( cSDR ) because , as opposed to the normal one from the chromosomal origin oriC ( requiring de novo DnaA synthesis ) , it could go on for several hours after the full inhibition of protein synthesis . Replication from R-loops is initiated via the PriA-dependent primosome that also includes PriB and DnaT proteins [36] . The origins of replication from R-loops were tentatively mapped and were named oriKs [37] . The RecA recombinase has been shown to be involved in the initiation step of cSDR [38] , possibly at the step of R-loop formation since this protein is able to promote R-loop formation in vitro [39 , 40] . Since the strong phenotypes of topA topB null cells were significantly corrected by a dnaT mutation , it was concluded that cSDR was largely responsible for the sickness of cells lacking type 1A topos [33] . Furthermore , the observation that deleting recA corrected the strong phenotypes of topA topB null cells was explained , at least in part , in the context of the role of RecA in cSDR [33] . This is also supported by the observation that the positive effects of RNase H overproduction on the growth and phenotypes of topA topB null cells is not seen when recA is deleted [33] . Many things are still unclear regarding the phenotypes of topA topB null cells and the role of type 1A topos . For example , topo III is a protein of low abundance and there is no evidence that , at least at this level , it could act on R-loops to inhibit cSDR . In fact , no cSDR is detected in cells lacking topo III activity and deleting topB has no effect on growth , cell morphology and cSDR in rnhA mutant [34] . Furthermore , it is not obvious why a high level of cSDR could generate the severe phenotypes of topA topB null mutants , especially when considering the fact that these phenotypes are not observed in rnhA mutants that have an even higher level of cSDR [34] . In this work , we have used next-generation sequencing ( NGS ) and qPCR for marker frequency analysis ( MFA ) to further characterize the roles of type 1A topos . Moreover , we have used the DNA:RNA hybrids specific antibody S9 . 6 [41] to detect R-loops in rnhA , topA and topA topB null cells . Our results reveal three important findings: 1- Type 1A topos from both subfamilies can inhibit R-loop formation/accumulation mostly to prevent unregulated replication from R-loops . 2- High levels of replication from R-loops in the terminus region of the circular chromosome appear to be largely responsible for the strong phenotypes of topA topB null mutants . 3- topA topB null mutants are able to grow owing to the amplification of a chromosomal DNA region carrying the parC and parE genes that leads to topo IV overproduction .
We used MFA by NGS to map the putative oriKs on the chromosome of our isogenic strains . This approach was recently used for rnhA null strains , and five to ten peaks that could correspond to sites of replication initiation from R-loops were identified , including one prominent peak in the Ter region [42 , 43] . This approach is also very useful for the detection of various DNA rearrangements such as duplications/amplifications and inversions [44] . For our experiments , genomic DNA was extracted from log phase cells ( DO600 , 0 . 4 ) treated for two hours with spectinomycin ( 400 μg/ml ) , a protein synthesis inhibitor . With this treatment , we have previously shown that replication was fully completed in wild-type cells , whereas it could still be detected ( cSDR ) in rnhA , topA and topA topB null cells , unless RNase HI was overproduced for the type 1A topos mutants [34] . Fig 1 shows the results with the wild-type strain not treated ( log phase cells , top panel and stationary phase cells , middle panel ) or treated with spectinomycin ( log phase cells , bottom panel ) . For log phase cells ( top panel ) , a typical wild-type replication profile is seen where the highest and lowest copy numbers respectively are found in the oriC and Ter regions . This is the expected result for bidirectional replication initiated at oriC and terminated following the merging of the two replication forks in the Ter region . For stationary phase cells ( middle panel ) , a flattened profile with no significant peaks is generated , also as expected since oriC-dependent replication is inhibited in non-growing stationary phase cells . A uniform profile is also detected for wild-type cells treated with spectinomycin for two hours ( bottom panel ) . This is in agreement with our previous results showing that oriC-dependent replication was completed and cSDR not activated in wild-type cells treated with spectinomycin for two hours [34] . Thus MFA by NGS reveals that replication from oriC is active in wild-type log phase cells but not in stationary phase cells or in cells treated with spectinomycin for 2 hours . Fig 2 , top panel , shows that the treatment of rnhA null cells with spectinomycin generated a replication profile that is much more complex as compared to wild-type cells , with important fluctuations in copy numbers throughout the profile . In our cSDR study , we focused on a genomic region roughly delineated by the full and dashed black arrows ( e . g . Fig 2 top panel ) that respectively point to ydcM ( left; genomic position 1 . 505 ) and lepA ( right; genomic position 2 . 705 ) genes . These genes were used in our qPCR experiments and their position roughly correspond to the highest ( ydcM ) and lowest ( lepA ) copy numbers observed for topA topB null strains ( see below; excepted amp ) . Green arrows in Fig 2 , top panel , point to potential oriK sites ( peaks or bumps ) . The left one at position 1 . 52 is flanked by the TerA and TerB sites and the others ones are found at positions 1 . 83 , 2 . 23 and 2 . 54 . A prominent peak at position 1 . 52 was previously mapped for rnhA null cells treated with chloramphenicol [42] that also inhibits protein synthesis and allow cSDR to be detected . Bumps around 1 . 83 , 2 . 23 and 2 . 54 were also detected in previous studies [42 , 43] . The profile suggests that replication initiated in the Ter region ( peak at 1 . 52 ) is bidirectional and that the left- ( counterclockwise ) and right- ( clockwise ) moving forks are arrested respectively at TerA and TerB . In E . coli , replication forks are trapped within the Ter region by the Tus protein that binds to polar Ter sequences ( TerA to TerJ ) . Ter sequences within the left portion of the chromosome block clockwise-moving forks , whereas those in the right portion of the chromosome block counterclockwise-moving forks [45] . Note also that for the rnhA null cells ( Fig 2 , top panel ) , the profile is asymmetric is this area with the drop in copy numbers being much more important at TerA than TerB . This can be explained by the presence of counterclockwise-moving replication forks from putative oriKs at positions 1 . 83 , 2 . 23 and 2 . 54 that pass through the TerB site and are arrested at the TerA site . MFA by NGS can also reveal the occurrence of collisions between replication and transcription . Such collisions , especially head-on as compared to co-directional , can have a major impact on cell physiology , mostly when they involve the heavily transcribed rrn ( rRNA ) operons [43 , 46] . The impact of collisions with rrn operons are normally minimized because they are transcribed in the same orientation as the bi-directional replication forks coming from oriC . However , when cSDR is activated from oriKs , head-on collisions between replication forks and transcribed rrn operons are inevitable . As shown in Fig 2 top panel , such collisions can be revealed by the presence of “hollows” or “steps” at genomic positions corresponding to rrnG ( 2 . 73 ) , the rrnDBAC cluster ( 3 . 42 to 3 . 69 ) and rrnE ( 4 . 21 ) . Similar patterns were also observed for the replication profiles of topA topB null strains ( Fig 3 ) . To further support the notion that the identified peaks or bumps correspond to oriK sites , we used a dnaT18::aph mutation that was previously isolated as a suppressor of the growth defect of a topA rnhA null mutant [47] . This mutation was shown to abolish cSDR in rnhA null mutants [33 , 34] . This allele does not behave like a loss-of-function mutation because , as opposed to a dnaT null mutation that confers severe growth defects similar to priA null mutations [48] , our dnaT18::aph mutation does not affect cell growth [33] . In fact , the aph cassette is inserted within the dnaT promoter region [33] and the results of our qRT-PCR experiments reveal that this insertion caused a 5 to 6-fold reduction in dnaT mRNA levels ( S1 Fig ) . Importantly , we found that the expression of the dnaC gene that is believed to be co-transcribed with dnaT ( dnaTC operon ) [49] is not affected by the dnaT18::aph mutation ( S1 Fig ) . Thus , the dnaT18::aph mutation is well-suited for the study of cSDR . Fig 2 , middle panel , shows the replication profile of dnaT18::aph rnhA null cells treated with spectinomycin . It can be seen that the presence of this dnaT allele almost fully eliminated the peaks or bumps found in rnhA null cells , thus strongly suggesting that they indeed correspond to oriK sites . Fig 2 , bottom panel , shows the replication profile of an rnhA null strain ( treated with spectinomycin ) carrying a large DNA inversion ( inv; from 1 . 39 to 2 . 28 ) in the Ter region that includes the oriK site at 1 . 52 . We found that this strain has reduced cSDR activity as compared to the non-inverted rnhA null strain [47] . If the peak at 1 . 52 really corresponds to an origin of replication ( a fixed origin of replication ) its location should be modified accordingly in the strain carrying the inversion . This is what we found , i . e . , the peak is still located at position 1 . 52 according to the W3110 reference strain . Thus , this peak is not the result of replication initiation triggered by the collision of bidirectional forks from oriC , as suggested to explain over-replication in the Ter region of recG cells [50] . Altogether , our results strongly suggest that the peak at position 1 . 52 in the Ter region of rnhA null cells corresponds to a fixed origin of replication that is R-loop- and PriA-primosome-dependent ( oriK ) . Fig 3 top panel shows the replication profile of topA topB null cells ( all the topA topB null cells used in the present study also carry a gyrB ( Ts ) compensatory mutation ) not overproducing RNase HI ( strain VU425 , topA topB gyrB ( Ts ) /pSK762c ) and treated with spectinomycin . A prominent peak in the Ter region that corresponds to the oriK site mapped in rnhA null cells ( genomic position 1 . 52 ) is detected , together with much smaller bumps at genomic positions 1 . 88 , 2 . 23 and 2 . 56 ( green arrows ) . When compared to the replication profile of rnhA null cells ( Fig 2 top panel ) , it is obvious that the Ter-located oriK is significantly more active in the topA topB null mutant , whereas the oriKs lying outside Ter are much less active . As expected if these peaks and bumps correspond to oriKs , RNase HI overproduction ( strain VU422 , topA topB gyrB ( Ts ) /pSK760 ) considerably reduce their intensity; the small bumps outside Ter are no longer seen and the peak within the Ter region is much lower ( Fig 3 bottom panel ) . cSDR is also activated in single topA null mutants , though its level is lower as compared to topA topB null cells . Indeed , S2 Fig shows the replication profile of single topA null cells treated with spectinomycin that demonstrates the low level of cSDR activity in this strain , with only one small peak being found in the Ter region ( position 1 . 52 ) . Thus deleting the topB gene in a topA null strain significantly increases cSDR activity in the Ter region . This result shows that topo III also acts at the level of cSDR initiation . To further support our results of MFA by NGS we performed qPCR . As stated above we used probes that correspond to ydcM ( genomic position 1 . 505 ) and lepA ( genomic position 2 . 705 ) . Their position roughly correspond to the highest ( ydcM; excepted amp , see below ) and lowest ( lepA ) copy numbers observed for topA topB null strains ( Fig 3 , top panel , full and dashed black arrows respectively ) . The histogram in S3 Fig shows the result of qPCR experiments with samples of the genomic DNA preps that were used for MFA by NGS . Columns with an error bar correspond to the qPCR results ( at least two independent experiments ) represented as the ydcM/lepA ratio . The ratios were also calculated from the results of MFA by NGS ( columns with no error bars ) . It can be seen that very similar ydcM/lepA ratios are obtained with the two approaches , which confirms the results of MFA by NGS and validates the qPCR method . Fig 4 shows the ydcM/lepA ratios calculated from the results of qPCR experiments with genomic DNA samples from log phase cells not treated with spectinomycin . As it is the case for the DNA samples from cells treated with spectinomycin , the highest ratio is found for the topA topB null mutant not overproducing RNase HI ( VU425 , pSK762c no overproduction vs VU422 , pSK760 overproduction ) . As expected , the ratio is much lower when the dnaT18::aph mutation is present in the topA topB null mutant ( VU441 ) . Importantly , while the ratio is significantly higher in the topA null mutant not overproducing RNase HI ( VU296 ) as compared to the wild-type strain ( RFM443 ) , it is much higher when topB is also absent ( VU425 ) . This confirms the important regulatory role of topo III on cSDR initiation in the Ter region . However , this role is only observed when topA is also lacking as no increase in the ydcM/lepA ratio is observed for the isogenic topB null mutant ( VU403 ) as compared to the wild-type strain . This result is in agreement with our previous finding that cSDR was not activated in a single topB null mutant [34] . Moreover , deleting recA in both topA ( SB265 ) and topA topB ( VU243 ) null mutants , restored the ydcM/lepA ratio to the level seen in wild-type cells . This is the expected result as RecA is required for the initiation step of cSDR [38] . Overall , when the ratios in Fig 4 ( no spectinomycin ) are compared to those shown in S3 Fig ( spectinomycin ) the differences between the strains are very similar but the ratios are lower . This is expected as in log phase cells not treated with spectinomycin , replication from oriC is active . Because cSDR is strongly activated in cells lacking type 1A topos , it should be possible to detect R-loops in these cells . To test this , we performed dot blots with S9 . 6 antibody that recognizes DNA:RNA hybrids . S9 . 6 has been widely-used to detect and map R-loops in eukaryotic cells [41] . To validate this approach for E . coli cells , we considered two situations in which R-loop formation/accumulation has been supported by much experimental evidence and/or can be predicted to occur . The first situation is related to rnhA null mutants . Because of the absence of RNase HI and the occurrence of cSDR , R-loops are expected to accumulate in rnhA null mutants . To test that , genomic DNA was extracted from both the rnhA null mutant ( MM84 ) and the isogenic wild-type strain ( RFM443 ) . Fig 5 demonstrates the accumulation of R-loops in the rnhA null mutant but not in the wild-type strain as expected ( compare RFM443 and MM84 , —and + RNase HI ) . The second situation concerns topA null mutants . R-loop formation on plasmid , based on RNase HI-sensitive gel retardation and hypernegative supercoiling , has been shown to occur both in vitro and in vivo , in the absence of topo I but in the presence of gyrase [14 , 51] . Based on these assays and on the growth inhibition and RNA degradation phenotypes that can be corrected by RNase HI overproduction , R-loop formation can be predicted to occur in our topA null mutant following a temperature downshift from 37°C to 30°C and below [17] . To test that , genomic DNA was extracted from our topA null mutant ( RFM480: topA20:Tn10 , gyrB ( Ts ) ) 45 min after a temperature downshift from 37 to 30°C . Fig 5 shows the accumulation of R-loops in these cells as an RNase HI-sensitive spot could be detected in dot-blot experiments with S9 . 6 antibodies ( RFM480 downshift , compare–and + RNase HI ) . Thus , these results establish the validity of the S9 . 6 antibodies to detect R-loops in E . coli cells and directly demonstrate , for the first time , the accumulation of R-loops in topA null and rnhA null mutants . Next , we analyzed genomic DNA preps from three isogenic strains , RFM480 ( topA20:Tn10 , gyrB ( Ts ) ) , VU422 ( topA20:Tn10 , gyrB ( Ts ) , ΔtopB/pSK760 ) and VU425 ( topA20:Tn10 , gyrB ( Ts ) , ΔtopB/pSK762c ) grown under the same conditions used for NGS and qPCR , i . e . 30°C up to an OD600 of 0 . 4 . Fig 5 clearly shows the accumulation of R-loops in VU425 ( pSK762c: no RNase HI overproduced ) but not in VU422 ( pSK760: RNase HI overproduced; very weak signal ) or RFM480 . Thus , under these conditions , growth at 30°C and no temperature changes , R-loops are only detected significantly when both type 1A topos are absent . Thus , the high level of cSDR replication correlates with an accumulation of R-loops in topA topB null cells . Fig 3 also shows the presence of a DNA amplification in the topA topB mutant whether or not RNase HI was overproduced ( top and middle panels , amp: from genomic position 2 . 99 to 3 . 24 ) . As stated in the introduction , such an amplification that includes the parC ( 3 . 162 ) and parE ( 3 . 172 ) genes coding for the two subunits of Topo IV , is the most frequent compensatory mechanism for the absence of topo I [7] . DNA amplifications are very unstable as they are easily lost by a RecA-dependent mechanism [52] . There maintenance in a population of cells indicate that they confer a growth advantage . The other well-described compensatory mechanism is the occurrence of mutations reducing the supercoiling activity of gyrase ( gyrA and gyrB ) . Normally , only one mechanism is sufficient to allow the growth of topA null mutants . We have previously shown that upon prolonged incubation on plates , our topA null gyrB ( Ts ) mutants , mostly those carrying the topA20::Tn10 allele as compared to the Δ ( topA cysB ) allele , can generate larger colonies at a very high frequency [15] . These colonies are made of cells carrying a DNA amplification of the chromosomal region including parC and parE . This is likely because the gyrB ( Ts ) mutation is not a naturally selected one that arose to compensate for the absence of topA , and it is therefore probably not optimal for compensation . Thus , despite the fact that our topA null mutants can grow without the parC parE amplification ( S2 Fig ) , as also shown in this work , when it occurs it confers a growth advantage . However , according to our results the situation might be different for the topA topB null mutant as in addition to the gyrB ( Ts ) mutation , it also carries an amplification of the parC parE region ( Fig 3 ) . This may suggest that excess negative DNA supercoiling is more harmful for topA topB null cells than it is for topA null cells and/or that more relaxation activity is required in the double mutant . We used qPCR to look for a DNA amplification of the parC parE genomic region in various isogenic strains used in the present study . Cells were grown and DNA extracted as done for MFA by NGS , i . e . 30°C up to an OD600 of 0 . 4 , except that the spectinomycin treatment was omitted . For the qPCR , we used probes that correspond to qseC ( genomic position 3 . 169 ) , located between parC and parE , and lepA ( genomic position 2 . 705 ) . Fig 6A shows that the qseC/lepA ratio is close to one for both the wild-type ( RFM443; 0 . 95 ) and the Δ ( topA cysB ) gyrB ( Ts ) ( RFM475; 0 . 85 ) strains . In the absence of DNA amplification , the ratio is expected to be near one as qseC and lepA are close to each other , and therefore no significant copy number variations related to replication from oriC are expected to be seen . The result for the topA null strain confirms the absence of the parC parE amplification when topB is present as shown by NGS for the strain carrying the Δ ( topA cysB ) allele ( S2 Fig; RFM475 ) . Furthermore , the reverse is also true , i . e . no amplification is seen in the absence of topB when topA is present ( Fig 6A , VU403 ) . When the qPCR was performed with the topA null mutant carrying the topA20::Tn10 allele ( Fig 6A; RFM480 ) , a small but reproducible DNA amplification of the parC parE region ( qseC/lepA ratio of 1 . 7 ) was observed . So , the selective pressure to keep the parC parE amplification is stronger when the cells carry the topA20::Tn10 allele as compared to the Δ ( topA cysB ) allele . These results are in agreement with our previous observations showing that the phenotypes of topA null and topA topB null mutants are stronger when they carry the topA20::Tn10 allele instead of the Δ ( topA cysB ) one [33] . Next , we analyzed two pairs of topA topB null strains that have been constructed in our laboratory . Each pair includes strains carrying the gyrB ( Ts ) allele and either pSK760 ( RNase HI overproduction ) or pSK762c ( control , no RNase HI overproduction ) . The first pair , VU422 ( pSK760 ) and VU425 ( pSK762c ) , the one that has been used in MFA by NGS , includes strains with the topA20::Tn10 mutation and the ΔtopB allele from the Keio collection [53] . The second pair , VU306 ( pSK760 ) and VU333 ( pSK762c ) includes strains with the Δ ( topA cysB ) mutation and the ΔtopB::kan allele from ref . [30] . For the first pair , the topB null allele was introduced before the topA null allele , whereas for the second one , the topB null allele was introduced after the topA null allele . Fig 6A clearly shows the presence of a DNA amplification of the parC parE region in all the strains ( qseC/lepA ratios from 3 . 0 to 3 . 7; the ratio is 2 . 8 for strain VU421 , the topA topB null mutant from which VU422 and VU425 were obtained ) . So the parC parE amplification is observed whether RNase HI is overproduced or not , as shown for the NGS results ( Fig 3 ) and irrespective of the topA and topB null alleles present in the strains . We also looked for parC parE amplifications in topA topB null strains carrying the recA ( VU243 ) or dnaT ( VU441 ) mutation that were shown respectively to fully inhibit or considerably reduce cSDR ( Fig 4 ) . Fig 6A shows the presence of a DNA amplification in the topA topB null strain carrying the recA mutation , as the qseC/lepA ratio is close to 2 . 5 . Although this result may suggest that deleting recA did not eliminate the need for topo IV overproduction , it is difficult to interpret as the loss of duplications/amplifications is a RecA-dependent process [52] . Conceivably , introducing the recA mutation could have stabilized a parC parE amplification that was already present in the topA topB null mutant . Fig 6A shows the absence of a parC parE duplication/amplification in the topA topB null strain carrying the dnaT mutation , as the qseC/lepA ratio is close to one . This result shows that the dnaT mutation not only inhibited cSDR as shown above , but also allowed the topA topB null mutant to grow despite the lack of a parC parE amplification . This supports the hypothesis that the major problem of cells lacking type 1A topos activity is related to R-loop-dependent replication initiated from the PriA-dependent primosome that also includes DnaT . If the amplification of a genomic DNA region including parC and parE genes is indeed required to allow topo IV overproduction , the need for this amplification should be bypassed by introducing a plasmid from which topo IV can be produced into topA topB strains . To test this , we used the plasmid pET11-parEC producing a ParEC fusion protein that was shown to be active as a topo IV both in vitro and in vivo [54] , and that could complement the growth defect of the topA null strain RFM475 [32] . This plasmid was first introduced into the topA gyrB ( Ts ) strain ( RFM475 ) and then transduction with P1vir was performed to introduce the ΔtopB::kan allele into RFM475/pET11-parEC . The transduction was also performed in parallel in the RFM475 strain carrying no plasmid . The ΔtopB::kan transductants appeared after 24 and 48 hours of incubation respectively for the RFM475 strain with and without pET11-parEC . Upon re-streaking them at 37°C , the RFM475 transductants carrying pET11-parEC grew slightly better than the ones of strain RFM475 carrying no plasmid . At 30°C , no significant differences in term of colony number and size could be seen between these transductants . Fig 6B shows qseC/lepA ratios close to 1 ( 0 . 87 and 0 . 96 ) and above 2 ( 2 . 16 ) for ΔtopB::kan transductants of RFM475 respectively carrying ( 2 clones; JB37 , JB38 ) or not carrying ( 1 clone; JB40 ) pET11-parEC . This is the expected result if the selective pressure to maintain the amplification of a genomic region including parC and parE is related to topo IV overproduction . Thus , the amplification of a genomic DNA region including parC and parE allows topA topB mutants to grow because of topo IV overproduction . The parC parE amplification is likely maintained in topA topB null cells to provide a higher level of DNA relaxation activity , via topo IV , possibly to limit the accumulation of R-loops and its associated unregulated replication . We used plasmids pACYC184ΔEN and pACYC184ΔHE as markers for the global supercoiling level in isogenic topA and topA topB null strains . Various portions of the tet gene have been deleted in these plasmids so that they can be used to evaluate the global supercoiling level , on which topo IV would mostly act . Topo I acts on local transcription-induced supercoiling that is generated by tet transcription on pACYC184 [55 , 56] . S4 Fig clearly shows that more plasmid topoisomers migrated toward the relaxed state in the topA topB null strain ( CT170 ) as compared to the topA null strain ( RFM475 ) . This supports the hypothesis that indeed topo IV overproduction provides more DNA relaxation activity to topA topB null cells .
Previous results of in vitro experiments have suggested that E . coli type 1A topos can act at least at two levels to prevent the accumulation of R-loops . They can prevent their formation by relaxing transcription-induced supercoiling and they can likely destabilize R-loops by using them as hot-spots for DNA relaxation activity [14 , 29 , 60] . This activity may be reminiscent of a recently described reaction called D-loop dissolution [61] . In this reaction , yeast Top3 was able to dissolve Rad51 ( RecA ortholog ) -mediated D-loops on supercoiled templates , a reaction that was also accompanied by the simultaneous relaxation of the DNA template . In vitro , as predicted from the known biochemical properties of the two E . coli type 1A topos , an R-looped DNA template was shown to be a relatively better substrate for topo III than topo I , whereas the reverse situation was observed for a transcribed DNA template [29] . The results presented here showed that deleting topB from topA null mutants significantly stimulated both cSDR and the accumulation of R-loops detected by the S9 . 6 antibodies . Thus , by inhibiting the accumulation of R-loops , topo III can prevent unregulated replication . However , this function of topo III is only seen when topo I is absent . This suggests that topo III acts after topo I to prevent the accumulation of R-loops . Indeed , because topo III , unlike topo I , is a protein of low abundance , it is unlikely to act during transcription . In fact , a significant effect of topo III on transcription-induced negative supercoiling in a topA null mutant , was only observed when it was overproduced [29] . It is therefore more probable that topo III acts on the R-loop to destabilize it . This activity would be more compatible with its high specificity for ssDNA . Thus , in agreement with their biochemical properties and as shown in vitro , topo I , by interacting with RNAP , would act preferentially to relax transcription-induced supercoiling , whereas topo III would rather act on the R-loop to destabilize it . The results presented in this work confirm the previous findings that a limited number of chromosomal locations can be used as oriKs in rnhA null cells [37 , 42 , 43] . Moreover , despite the fact that deleting topA can lead to non-sequence-specific R-loop formation via hypernegative supercoiling [17] , oriK activity was observed almost exclusively at one location on the chromosome of our topA null mutant . Furthermore this location corresponded to an oriK originally identified in rnhA null mutants , the Ter located one . These results strongly suggest that the majority of R-loops do not lead to PriA-dependent replication . Factors such as the frequency of R-loop formation , the stability of the R-loops and some unknown additional properties of the R-loops and/or the surrounding nucleotide sequence are most likely very important for oriK activity . In this context , we can predict the occurrence of competing activities between proteins of low abundance such as PriA and topo III for the stable R-loops that have escaped degradation by RNase HI in the topA null mutant . Therefore , topo III would be specifically targeted to stable R-loops that can also be used by the PriA-dependent primosome , a function that would be compatible with its low copy number in the cell . In this context , it is important to mention that topo III , PriA and RNase HI have been shown to interact with SSB , the first protein that is predicted to bind to the ssDNA portion of an R-loop in vivo [62–65] . Despite the fact that the level of cSDR was shown to be higher in rnhA mutants that in topA topB null mutants , only in the later that the strong pathological state was observed [33 , 34] . Here , our results of MFA by NGS revealed that the Ter located peak that likely correspond to an oriK site , was much higher in topA topB null cells as compared to rnhA null cells . It is therefore reasonable to propose that the pathology of cells lacking type 1A topos is related , at least partially , to this strongly activated origin of replication . The question is then how the high level of replication initiation in the Ter region can contribute to this pathological state . We believe that it could be related to hyper-recombination . Indeed , the chromosomal Ter region is known to be hyper-recombinogenic [66–68] and DSBs ( double-strand breaks ) are found in this region [69] . DSBs are substrates for the binding of the RecBCD complex that degrade the DNA up to a chi site , from which RecA proteins are loaded on the DNA to initiate homologous recombination . It has been shown that a Ter-blocked replication fork ( Ter/Tus ) can lead to replication fork collapse , i . e . the formation of a double-stranded end when a second replication fork moving in the same direction run into the Ter-blocked fork [70] . A high level of RecA and RecBCD-dependent recombination that requires the presence of Tus protein and chi sites , thus likely related to forks collapse , has been shown to occur at TerA , B and C sites in rnhA null but not in wild-type cells [68] . This was proposed to be due to cSDR initiated in the Ter region of rnhA null cells [68] . Considering the very high level of cSDR in the Ter region of topA topB null cells with replication forks arrested at TerA and B sites , forks collapse is expected to be very frequent and may lead to hyper-recombination in these cells . Interestingly , topA null mutants deleted for recB were previously shown to be barely viable [33] . The high level of hyper-recombination in the Ter region of topA topB null mutants could impede chromosome segregation , either because of the accumulation of Holliday junctions or because of the additional PriA-dependent replication initiated from D-loops that have been assembled by RecA during DSB repair . We cannot exclude the possibility that a type 1A topo activity is required during cSDR to solve the topological problems of head-on collisions between replisomes or between a replisome and a heavily transcribed gene or operon , as we recently proposed [34 , 47] . Such conflicts may threaten cell viability [43 , 46] . Clearly more work are still required to fully understand the pathological state of cells lacking type 1A topos . The fact that topo IV needs to be overproduced for topA topB null mutants to survive despite the presence of a gyrB ( Ts ) compensatory mutation , likely indicates a major problem related to excess negative supercoiling in this strain . The observation that the dnaT mutation can both correct the cSDR phenotype and allow the requirement for a parC parE amplification to be bypassed in a topA topB null strain , may suggest that PriA-dependent replication from R-loops is the main problem related to excess supercoiling in strain lacking type 1A topos . We have previously shown that upon a temperature downshift , topA null mutants carrying the gyrB ( Ts ) allele accumulated both hyper-negatively supercoiled DNA and truncated RNAs , and stopped growing [17] . After less than two hours , the accumulation of full-length functional RNAs and growth resumption coincided with the relaxation of hyper-negatively supercoiled DNA by topo IV [16 , 17] . When RNase HI was overproduced hyper-negatively supercoiled DNA barely accumulated and was rapidly relaxed by topo IV and , as a result , growth was not inhibited . In agreement with the involvement of R-loops in this phenotype , we have shown here that during the transient growth arrest following a temperature downshift from 37 to 30°C , the topA null gyrB ( Ts ) mutant accumulated R-loops as detected by the S9 . 6 antibodies . As predicted , when this topA null mutant was growing at 30°C ( OD600 , 0 . 4; no temperature downshift ) , no R-loops were detected . Based on in vivo and in vitro results , we have previously proposed a self-promoting cycle of R-loop formation whereby negative supercoiling promotes R-loop formation , which , in turn increases negative supercoiling following gyrase activity before the action of RNase HI [71] . This further increase R-loop formation . Ultimately , hypernegative supercoiling leads to extensive non-sequence specific R-loop formation [17] . In our model , RNase HI overproduction would efficiently compete with gyrase to prevent hypernegative supercoiling whereas topo IV would act later to relax this hypernegative supercoiling before its transcription . The involvement of topo IV in the inhibition of R-loop formation is also supported by our observation that topA rnhA gyrB ( Ts ) mutants carry an amplification of the parC parE region ( qseC/lepA ratio slightly above 2 ) . Presumably , in the absence of both type 1A topos , R-loop formation/accumulation would be so efficient at least at some loci used to initiate cSDR that both RNase HI and topo IV would need to be overproduced to inhibit unregulated replication . It is also possible that the RNA of some R-loops , like the one used as a primer to initiate ColE1 replication , is resistant to RNase HI . In that case , the action of a topo could be required to prevent R-loop formation . Furthermore , at some specific sites R-loop formation may depend only on locally induced negative supercoiling during transcription . In this situation , only a type 1A topo would be able to inhibit their formation . Thus , our results suggest that in E . coli , topo I , III and IV all participate , to different extents , in the control of replication from R-loops to maintain the stability of the genome .
The bacterial strains used in this study are listed in S1 Table and are all derivatives of E . coli K12 . S1 Table also gives the details on their constructions as well as the list of plasmids used in this study . Transductions with phage P1vir were done as described previously [32] . PCR with appropriate oligonucleotides were performed to confirm the transfer of the expected alleles in the selected transductants . Cells were grown overnight in liquid LB medium supplemented with the appropriate antibiotics . Overnight cultures were diluted in LB medium to obtain an OD600 of 0 . 01 and grown at the indicated temperature to on OD600 of 0 . 4 . When indicated , spectinomycin ( 400μg/ml ) was added and the cultures were incubated for an additional two hours at the same temperature . Samples of 10 ml were transferred in tubes filled with ice and the cells were recovered by centrifugation . Genomic DNA was extracted by using either the GenElute bacterial genomic DNA kit ( Sigma Aldrich ) or the QIAamp DNA mini kit ( Qiagen ) . Similar results were obtained for both kits whether the genomic DNA was used for NGS , qPCR or dot-blots with S9 . 6 . For the stationary phase wild-type ( RFM443 ) cells , culture were incubated overnight and 1 . 5 ml samples were used for the genomic DNA extraction . The purity of the various DNA preps was evaluated by using the Nanodrop ( Thermofisher ) and the DNA concentrations were determined by using the Qubit dsDNA Assay kit ( Invitrogen ) with the Qubit Fluorometer ( Thermofisher ) . Shotgun libraries with PCR were prepared for Illumina sequencing . Sequencing was performed by using Illumina HiSeq 2500 v4 ( Génome Québec , Montréal , Canada ) to determine sequence copy number . Bioinformatics analysis was performed at the Canadian Centre for Computational Genomic ( C3G , McGill University , Montréal , Canada ) . For the read mapping ( 11 to 16 million sequencing reads per sample ) , the E . coli K12 W3310 genomic sequence AP009048 . 1 was used as the reference . To reduce miscalculation of depth of coverage due to reads mapping at multiple places in the genome , a minimum mapping quality of 10 ( phred scale based ) was used for a read to be kept during the calculation of depth of sequencing . In Figs 2 and 3 , the number of reads were normalized against a spectinomycin-treated wild-type control to take into account differences in read depth across the genome of spectinomycin treated cells . Enrichment in 500 bp windows ( on average ) across the genome ( 10 , 000 points ) was calculated and loess regression curves were generated with loess_span parameters set to 0 . 1 . Genomic DNA for qPCR was prepared as described above for NGS . The Quantinova SYBR Green PCR kit ( Qiagen ) was used with a Rotor-Gene 6000 ( Corbett ) apparatus . For each experiment and each set of primers two tubes were prepared that contained respectively 8 and 20 ng of genomic DNA . Experiments were repeated at least twice for each set of primers . The ratios were determined by using the 2-Δct formula and standard deviations were calculated from these values . The primers were designed by using the PrimerQuest tool ( IDT ) . Forward and reverse primer sequences ( 5’-3’ ) were GAGTACCGGGCAGACCTATAA and AGCCTACTTCGCCACATTTC for lepA , CGAGACTTCAGCGACAGTTAAG and CCTGCGGATATTTGCGATACA for ydcM and CTGGACTCACTGGATAACCTTC and TGCGCCGTGTGGTAAATA for qseC . Genomic DNA for the dot blots with S9 . 6 antibodies was prepared as described above for NGS except that the amount of RNase A added was reduced by half . For each genomic DNA prep , two tubes containing 300 ng of DNA in RNase III reaction buffer ( Ambion ) were prepared . In one tube RNase III ( 2U; Ambion/Invitrogen; a ribonuclease specific to dsRNAs ) was added , whereas in the second tube RNase III and RNase HI ( 0 . 8 μg; from Kefei Yu , Michigan State University ) were added . The tubes were incubated at 37°C for 3 hours and the DNA was purified by phenol/chloroform extraction and EtOH precipitation . The DNA was resuspended in 20 μl of TE . The RNase III treatment was found to be necessary as the S9 . 6 antibody can also recognize RNA:RNA hybrids ( dsRNAs ) , albeit with a 5 to 6 fold lower affinity than DNA:RNA hybrids [72] . Moreover , a strong RNase H-resistant signal in dot blots , especially for the genomic DNA from the topA topB null mutant not overproducing RNase HI , was found to be RNase III-sensitive . We believe that this signal might be due to the accumulation of short truncated RNAs , especially highly structured rRNA fragments , due to R-loop formation in topA null mutants as shown previously [73 , 74] . In fact , S9 . 6 was recently shown to be able to immuno-precipitate dsRNAs from yeast cells , and a RNase III treatment was found to be required to generate an accurate map of R-loops in the genome of yeast cells [75] . Dot blotting was performed essentially as described previously [76 , 77] . Ten μl of DNA were spotted on a Hybond-N+ membrane ( Amersham ) . The membrane was UV-crosslinked ( UV Stratalinker 1800 ) , blocked in 5% milk in TBST and incubated overnight at 4°C with 30 μg of S9 . 6 antibody ( obtained from Dr Michael Wilson , University of Toronto , Canada ) . The membrane was washed 3 times in TBST and the secondary antibody ( 1:500; Stabilized Goat Anti-Mouse IgG HRP , Thermo scientific ) was added for 1 hour at room temperature . The HRP signal was revealed by using the SuperSignal West Pico PLUS kit ( Thermo scientific ) and the membrane was exposed to an autoradiography film .
|
DNA topoisomerases are nicking closing enzymes with strand passage activity that solves the topological problems inherent to the double-helical structure of DNA . Topos of the type 1A family are the only ubiquitous topos . They are classified in two subfamilies , topo I and topo III respectively found in bacteria only and in organisms from the three domains of life . The prototype enzymes of these two subfamilies are topo I and topo III from Escherichia coli . Recent data suggest that duplications leading to topo I and III subfamilies occurred in the Last Common Universal Ancestor of the three domains of life . In this context , our finding reported here that both E . coli topo I and III control R-loop formation/accumulation , mostly to inhibit unregulated replication , may suggest that R-loops have been a problem early in the evolution of life . Furthermore , our data show that E . coli cells can survive in the absence of type 1A topos , owing to the surproduction of topo IV that can relax excess negative supercoiling and prevent R-loop formation . Thus , our results strongly suggest that a major function of type 1A topos is to control R-loop formation to preserve the integrity of the genome .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"blood",
"cells",
"sequencing",
"techniques",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"nucleases",
"enzymes",
"immunology",
"dna-binding",
"proteins",
"enzymology",
"next-generation",
"sequencing",
"mutation",
"dna",
"replication",
"genome",
"analysis",
"dna",
"molecular",
"biology",
"techniques",
"extraction",
"techniques",
"research",
"and",
"analysis",
"methods",
"white",
"blood",
"cells",
"genomics",
"animal",
"cells",
"proteins",
"ribonucleases",
"dna",
"amplification",
"molecular",
"biology",
"biochemistry",
"dna",
"extraction",
"hydrolases",
"cell",
"biology",
"nucleic",
"acids",
"k",
"cells",
"phenotypes",
"genetics",
"transcriptome",
"analysis",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"computational",
"biology",
"dna",
"sequencing",
"gene",
"amplification"
] |
2018
|
Topoisomerases I and III inhibit R-loop formation to prevent unregulated replication in the chromosomal Ter region of Escherichia coli
|
In multiple studies DNA methylation has proven to be an accurate biomarker of age . To develop these biomarkers , the methylation of multiple CpG sites is typically linearly combined to predict chronological age . By contrast , in this study we apply the Universal PaceMaker ( UPM ) model to investigate changes in DNA methylation during aging . The UPM was initially developed to study rate acceleration/deceleration in sequence evolution . Rather than identifying which linear combinations of sites predicts age , the UPM models the rates of change of multiple CpG sites , as well as their starting methylation levels , and estimates the age of each individual to optimize the model fit . We refer to the estimated age as the “epigenetic age” , which is in contrast to the known chronological age of each individual . We construct a statistical framework and devise an algorithm to determine whether a genomic pacemaker is in effect ( i . e rates of change vary with age ) . The decision is made by comparing two competing likelihood based models , the molecular clock ( MC ) and UPM . For the molecular clock model , we use the known chronological age of each individual and fit the methylation rates at multiple sites , and express the problem as a linear least squares and solve it in polynomial time . For the UPM case , the search space is larger as we are fitting both the epigenetic age of each individual as well as the rates for each site , yet we succeed to reduce the problem to the space of individuals and polynomial in the more significant space—the methylated sites . We first tested our algorithm on simulated data to elucidate the factors affecting the identification of the pacemaker model . We find that , provided with enough data , our algorithm is capable of identifying a pacemaker even when a weak signal is present in the data . Based on these results , we applied our method to DNA methylation data from human blood from individuals of various ages . Although the improvement in variance across sites between the UPM and MC was small , the results suggest that the existence of a pacemaker is highly significant . The PaceMaker results also suggest a decay in the rate of change in DNA methylation with age .
DNA methylation is an important component of the epigenetic code that defines and maintains the state of cells [1–3] . Mammalian cells contain three DNA methyltransferases that preferentially methylate CpG dinucleotides . These enzymes faithfully maintain cytosine methylation patterns during cell division . However , as cells undergo differentiation , from stem cells to mature cells , the patterns of DNA methylation change substantially , and help define the changing cellular states [4] . The genomic profiles of DNA methylation across multiple cell types have been defined during the past few years using techniques such as bisulfite sequencing and DNA methylation arrays , that allow one to measure the methylation state of many cytosines in the genome [5] . Consequently , it has been shown that DNA methylation also changes as organisms age [6–12] . The seminal work of Steve Horvath [13] has identified three hundred CpG dinucleotides , whose methylation state can be used to accurately predict the age of an individual . The epigenetic clock is now widely used in aging research and is far more accurate than alternative approaches that rely on the measurement of telomere lengths or gene expression . The Horvath epigenetic clock model uses a linear combination of the methylation status of several hundred sites to predict the age of an individual . It also uses a nonlinear transformation to modify the ages of young individuals ( less than 20 years ) , while leaving the ages of adults untransformed . Here we try to develop a more general formalism for modeling changes in DNA methylation during aging . To this end , we use the universal pacemaker ( UPM or simply pacemaker—PM ) of genome evolution [14 , 15] , which was devised in the setting of molecular evolution in order to relax the evolution rate constancy imposed by the molecular clock ( MC ) hypothesis [16] . Under UPM , the relative evolutionary rates of all genes remain nearly constant ( i . e constant pairwise ratio ) whereas the absolute rates can change arbitrarily ( See Fig 1 for illustration ) . It was shown on several taxa groups spanning the entire tree of life that the UPM model describes the evolutionary process better than the traditional molecular clock model [14 , 17 , 18] . The UPM model relies on a statistical framework encompassing simultaneously all evolving genes in genomes , and across the entire tree of life , therefore making it doubly universal . Here we propose to adapt the UPM to model changes in DNA methylation during aging , making no a priori assumption about the relationship between chronological and epigenetic time , i . e . linearity in time as asserted by the MC model . The UPM is one degree of freedom more relaxed than MC in the sense that it still requires rate uniformity of a site among all individuals , yet it allows the individual’s aging rate to play a role . By relaxing the constraint that epigenetic age is linear with chronological age , we can explore a rich parameter landscape , and identify complex nonlinearities using the UPM formalism . Our goal is not only to develop site specific models of changes in DNA methylation as a population ages , but also to discover the nonlinearities in the rates of change . This richness has its cost in terms of computational intensity . In general , statistical analysis and in particular the approches we pursue here—maximum likelihood ( ML ) solutions—are computationally intensive [19] . However , although the current setting , methylation modeling , is more complex than the evolutionary model considered in [14] due to an additional array of variables to be optimized , under the MC model we were able to formalize it as a linear least squares , allowing us to obtain a closed form solution in polynomial time . Under the PM model , we show that no closed form solution is achievable . However , through a series of observations , we could reduce the search space significantly to the degree that the heuristic search , done by a fast optimization method , is performed only in the , relatively small , space of individuals . The rest of the search is polynomial is the space of methylation sites , hence enabling us to analyze problems of non-negligible size . Although the focus of this work is on the description of the algorithm , such as the model formulation and the statistics involved , we also demonstrate its performance in a real dataset . We first applied this formalism in a simulation study to discover the effect of the parameters involved and their interplay . Among other things , we show that the scheme is capable of identifying a pacemaker , i . e . a deviation from linearity in time , even when the pacemaker signal is relatively faint , if enough data is provided . Next we analyze a dataset of DNA methylation collected from the blood of humans of different ages . The signal in these data is indeed fairly small , however , the size of the data allows us to confidently infer coordinated , nonlinear changes in methylation . Further analysis shows that the changes in the rates resemble the empirical transformations used in the Horvath model .
Our basic objects are a set of m individuals and n methylation sites in a genome ( or simply sites ) . Each individual has an age , forming the set T of time periods {tj} corresponding to each individual j’s age . Henceforth we will interchangeably refer to individuals with their age . Each individual has a set of sites si undergoing methylation changes at some characteristic rate ri . Each site si starts at some methylation start level s i 0 . All individuals have all the sites si . As ri and s i 0 are characteristic of the site si , by the model , they are the same in all individuals . The latter fact , links between same sites but across different individuals , but also between different sites within and across individuals by the fact that sites generally maintain the same characteristic rates across the whole population . Henceforth , we will index sites with i and individuals with j . Now , let si , j measure the methylation level at site si in individual j after time tj . Hence , under the molecular clock model , we expect: s i j = s i 0 + r i t j . However , in reality we have a noise effect εi , j that is added and therefore the observed value s ^ i j is s ^ i j = s i 0 + r i t j + ε i , j . ( 1 ) Our goal is to find , given the input matrix S ^ = [ s ^ i , j ] , the maximum likelihood ( ML ) values for the variables ri and s i 0 for 1 ≤ i ≤ n . For this purpose , we assume a statistical model for εi , j by assuming that it is normally distributed , εi , j ∼ N ( 0 , σ2 ) . In contrast to the MC , in the UPM model we do not just use the given chronological age but estimate the age of each individual . Therefore under the UPM we must find the optimal values of s i 0 , ri , and tj . The solution to this optimization is described in detail below . We note that the deviation between the chronological age and the estimate epigenetic age under the UPM results is an age difference which , when positive , we denote as age acceleration , and when negative as age deceleration . Our first result is a maximum likelihood ( ML ) scheme to detect a coordinated , or rather genome wide change in methylation rate under UPM . We note that such a change is distinct from a single , uncoordinated , site change . We start with an overview of the approach . Two competitive explanations ( i . e . likelihood functions ) are developed , in which one ( MC ) is restricted to linearity with time by estimating a constant rate of methylation at each site , and using the given chronological age of each individual . The competing , relaxed , model ( UPM ) has no such restriction , and we estimate an “epigenetic” age for each individual . By definition , the ML solution under the relaxed model cannot be worse than the constrained model . Therefore , in order to compare the approaches , we use the likelihood ratio test that penalizes the UPM model proportionally to the loss of parameters in the MC model . In the Methods section we prove that under our model , the ML solution is equivalent to minimizing a quantity denoted as the residual sum of squares , RSS . The computational question of how we solve the problem , i . e . minimizing the RSS , under the two models is unique to this framework and hence we describe it here in the Results section below . In order to test our method we first conducted a simulation study as we now describe . The goal was to examine the effect of the various parameters on the performance of the method , i . e . its capability to distinguish between a PM and the MC . Performance was measured by means of the p-value of the likelihood ratio test ( LRT ) . We now describe the study’s parameters . Our model is comprised of an m-dimensional vector times T where tj corresponds to the jth individual’s age that we draw randomly to obtain variation in individuals’ ages . Next we have two n-dimensional vectors , rates r and methylation starting position s0 , where ri and s i 0 correspond to the ith site’s methylation rate and methylation starting position respectively . Both vectors were drawn randomly . These are the base parameters used to generate the input matrix S ^ . However recall that our goal was to test the sensitivity of our algorithm to distinguish between a PM and a MC . Also recall that by Lemma 0 . 5 , a PM is simply another linear correlation to time periods t j ′ only that these correspond to the PM ticks and each such PM ticks at an arbitrary rate . Therefore , to simulate the PM perturbation of the astronomical clock , we perturbed each tj by some εj ( i . e . multiplied by 1 + εj ) where ε j ∼ N ( 0 , σ t 2 ) . Hence , the constant parameters of the PM model are the ( perturbed ) times t j ′ and the original ri and s i 0 values . So by our model we have s i , j = s i 0 + r i t j ′ . Finally , to simulate biological noise , we sampled s ^ i , j ∼ N ( 0 , σ s 2 ) . Given the matrix S ^ and the time vector T , we ran both algorithms on that input and compared the results . The MC model fit the site rates and methylation start levels while adhering to the times in T while the PM model considered only the matrix S ^ and disregarded the times in T . Both models returned their RSS’s . Since under PM the times T′ are also inferred , we used LRT to compare between the models with m degrees of freedom which is the size of the vector T′ . The score of a single run is the p-value of the χ2 test . Since that setting is non trivial , we now discuss the parameters and their interpretation . Obviously , the signal to the method comes only if there is any variation in the pacemaker ticks with respect to the chronological clock , since otherwise both the PM procedure and the MC procedure will converge to the same values and will produce the same error ( RSS ) . Therefore our first parameter , the PM variance σ t 2 , that determines the size of the deviation of the PM from chronological time , is distinct from other parameters . Indeed we divided the study into two parts in which different values were used and the differences are significant . The second parameter is the variance at each site , or simply the amount of pure noise in the signal . Our experiments show that this is a major factor inhibiting the identification of the PM . The last two parameters are the number of sites that are included and the number of individuals . The results of our simulations are presented in Figs 3 and 4 . In all figures , the y axis represents the success rate in terms of the p-value returned from the LRT . The x axis represents the noise σ s 2 , the site variance . We now explain the results . The graphs in Fig 3 correspond to experiments with weaker PM signals , σ t 2 = 0 . 1 . Fig 3 ( a ) corresponds to 50 individuals . The graph contains three curves that correspond to individuals with {50 , 70 , 100} sites ( colors blue , red , and green respectively ) . That is , each experiment is done over a population of 50 individuals , each with 50 ( alternatively 70 or 100 ) methylation sites . Additionally , each individual is associated with a PM that modifies the methylation rate of that individual . That PM rate distributes , IID at each individual , normally with variance σ t 2 = 0 . 1 . The x-value of a point represents the background noise we apply to each site , that also distributes normally and IId at each individual and site , with variance σ s 2 . The y-value of a point represents the relative number of times ( or success frequency ) our scheme described in the Results section , was able to identify the PM ( a PM always exists but its signal may disappear due to confounding signals ) . Let us focus on the curve in Fig 3 ( a ) that corresponds to 50 sites ( blue curve ) . It is shown that for a small amount of noise , σ s 2 ≤ 2 , reconstruction quality is high but then it starts to diminish with success rate less than 1/2 for σ s 2 ≥ 7 . We can also see that this trend is generally true for each curve in the experimental study . We also see that there is an obvious benefit for the inclusion of additional sites ( red and green curves in Fig 3 ( a ) ) or individuals ( Fig 3 ( b ) ) . Fig 4 depicts a situation in which a stronger PM signal σ t 2 = 0 . 15 is embedded and the two graphs represent experiments with 50 and 100 individuals as in Fig 3 . Here we can observe that the clear trend of a weak PM and small number of individuals , as depicted in Fig 3 ( a ) , is not always maintained due to the high success rate and the stochastic nature of the process . However , that general behavior is still maintained . As can be seen , under this PM signal , the PM is identified with a high rate , ( ≥ 85% ) , even with only 50 individuals ( Fig 4 ( a ) ) and 50 sites for all levels of noise . With 100 individuals ( Fig 4 ( b ) ) , 100 sites suffice for almost perfect identification . We conclude this part by noting that for a fairly weak signal of PM and even under quite high levels of noise , our procedure is capable of identifying the deviation of methylation rate from linearity in time . This observation is critical when analyzing real data where we expect that the signal is stronger and noise is weaker . We remark that due to the fairly involved setting with many confounding parameters such as the amount of information ( sites , individuals ) , stochastic processes ( PMs , sites ) , the same behavior as we observed in Figs 3 and 4 , can be observed for many other combinations of parameters . Based on our simulation results , we next tested our approach on DNA methylation data previously reported in [22] . The data was collected using the Illumina 450K DNA methylation array platform . The resulting data matrix contains about 450 , 000 CpG sites measured across 657 human individuals . In order to limit ourselves to a manageable size for parameter estimation of our model we had to apply a selection criterion over the sites . We took the 300 sites with the maximum variance where the highest variance was 0 . 105 and the lowest around 0 . 0079 . These sites are more likely to be relevant for our model , as they have methylation levels that vary across the population . We ran both algorithms on this reduced data . The following results were obtained . The average error per entry in S ^ under MC was 0 . 138 . The UPM search algorithm started from 10 random stating points all of them converged to the same ML point—0 . 135 . This is a mild improvement of about 2% indicating that sites are correlated and also there are shifts from linear correlations to chronological time . The χ2 for these values under LRT is 3517 . 468 . Since we had measurements across 300 individuals and under PM their values were optimized , we had an additional 300 free variables ( the “epigenetic” age ) in the PM model with respect to MC . Under the χ2 distribution with degree of freedom 300 , in order to achieve a p-value 0 . 01 , a χ2 of 360 is required . Therefore the null hypothesis ( MC ) is rejected outright . As illustrated , the PM model guarantees an optimal ranking between the rates of sites such that the model likelihood is optimized . However there is one degree of freedom here , allowing us to assign an arbitrary value to one of the rates . This value in turn determines the values of the rest of the variables . By picking one of our ML points we obtain an ML assignment to rates . In order to compare how MC and PM rates behave under the different sites , we did the following . For each of the sites , we calculated the ratio between its MC and PM rates . We sorted the sites according to that value . After removing a few Eq ( 8 ) outliers at each side , we plotted this result . Fig 5 ( a ) depicts this result . We note a few facts about this ratio . The majority of the sites ( 5/6 ) maintain the same sign ( i . e . increasing or decreasing methylation ) , about half ( 55% ) of these sites decelerate ( i . e . ratio ≤ 1 ) . Fig 5 ( b ) shows an even more interesting phenomenon that corroborates certain conjectures . The figure depicts the ratio between the chronological times ( ages ) , taken as parameters ( i . e . fixed , unoptimized ) under the MC model , versus ML times inferred under PM . The x axis is the chronological time of the individual , meaning that ratios are presented from the youngest individual at the left to the oldest at the right . The y axis is the MC/PM age ratio . A conspicuous phenomenon emerging from this figure is the diminishing ratios between times ( or equivalently aging ) as individual becomes older . Another property arising from that comparison , is that the variance of this measure ( MC/PM age ratio ) in young ages is substantially larger than in more advanced ages . We comment that this data set of [22] does not contain individuals of very young ages . Therefore we expect even more extreme contrasts in data that does include young individuals , however this is beyond the scope of the current work and is left for further research .
In this work we developed an approach to model changes in DNA methylation with age and measure acceleration/deceleration of methylation rates with age . This approach is based on a novel , probabilistic framework where two competing explanations are compared , where one of the explanations is a special , restricted case of the other , and the comparison is made by the likelihood ratio test . The underlying mechanism in the novel framework is the universal pacemaker that was devised to find correlations among evolving genes in a genome , while relaxing the rate constancy imposed by the traditional molecular clock model . The methylation setting is typically more complex than the genomic evolution setting as it involves more variables , making the procedure and the analysis more computationally demanding . Therefore , we believe we have made here only the first step in this direction . Nevertheless , the results we present , first in the simulation analysis , but especially in the analysis of a human blood dataset with individuals of different ages , mark this approach as promising These results on the human methylation data , although based only on a sample of CpG sites , indicate that the rate of methylation changes tend to diminish with age , suggesting that the use of the PM framework is appropriate in this setting . We remark that the emphasis in this work is on the mathematical and computational aspects of this approach . These properties , as illustrated also in our simulation study , but also in the algorithmic part of the Method section , are far from being trivial and we believe further investigation will follow . The same also holds for the biological findings we indicate in our real data study . These result are significant , but should be verified on larger data sets . In particular , the finding of diminishing ratios PM/MC should be tested in a population that contains young individuals . Finally , we expect that the model may also be of use when investigating epigenetic aging in other species , and in the future intend to apply this formalism to datasets across species .
We now show that , under our formulation , the RSS is minimized at the Maximum Likelihood ( ML ) solution . Let the residual sum of squares , RSS be defined as follows: R S S = ∑ 1 ≤ i ≤ n ∑ 1 ≤ j ≤ m ε i , j 2 . ( 10 ) The formulation in Eq ( 10 ) is called least squares ( LS ) and is a very common criterion in optimization [23] . Although the fact that RSS is minimised under least squares under a normal distribution , since our formulation is somehow unique , we now show the the following lemma ( see detailed proof in S1 Text ) : Lemma 0 . 6 Minimizing RSS is equivalent to finding the maximum likelihood solution to our formulation . The likelihood ratio test ( LRT ) is a statistical test used to compare the goodness of fit of two competing models , one of which ( the null model ) is a special case of the other , more general , one . The log of the ratio of the two likelihood scores distributes as a χ2 statistic and therefore can be used to calculate a p-value . This p-value is used to reject the null model in the conventional manner . Specifically , let Λ = L0/L1 where L0 and L1 are the ML values under the restricted and the more general models respectively . Then asymptotically , −2log ( Λ ) will distribute as χ2 with degrees of freedom equal the number of parameters that are lost ( or fixed ) under the restricted model . In our case , ( see Eq ( 6 ) in the S1 Text for a detailed explanation ) , it is easy to see that log Λ ) = - n m 2 log R S S ^ M C R S S ^ P M ( 11 ) where R S S ^ M C and R S S ^ P M are the ML values for RSS under MC and PM respectively . Hence we set our χ2 statistic as χ 2 = n m log R S S ^ M C R S S ^ P M . ( 12 )
|
DNA methylation is an important component of the epigenetic code that defines and maintains the state of cells . Recently , it has been found that certain sites in the genome undergo methylation changes at different rates during aging . The seminal work of Steve Horvath found that the methylation of a couple hundred CpG sites could be linearly combined to accurately predict the age of an individual in a number of tissues . Such a pattern resembles the Molecular Clock ( MC ) concept prevailing in molecular evolution , which suggests that there are sites in the genome that change linearly with age . In this work , we adapt the Universal PaceMaker ( UPM ) model to the setting of DNA methylation changes during aging . UPM relaxes the rate constancy of MC and was found to provide a better statistical explanation for genome evolution across the entire tree of life . This adaptation requires the solution of a complex optimization problem . Nevertheless , in a series of observations we show that the problem can be solved efficiently under the MC model and slightly less efficiently under the UPM model . This allows us to solve problems of non-trivial size . We chose as a proof of concept to analyze DNA methylation data collected from the blood of humans of different ages . Our results show that , similarly to genome evolution , the UPM provided an improvement of about 2% in the fit to the data . The statistical significance of this improvement is very high . Although tested on a small data set , this improvement demonstrates that the UPM more accurately captures age related DNA methylation changes than the MC model .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] |
[
"biotechnology",
"medicine",
"and",
"health",
"sciences",
"genome",
"evolution",
"simulation",
"and",
"modeling",
"optimization",
"methylation",
"mathematics",
"algebra",
"pacemakers",
"epigenetics",
"dna",
"dna",
"methylation",
"chromatin",
"polynomials",
"research",
"and",
"analysis",
"methods",
"genomics",
"chromosome",
"biology",
"gene",
"expression",
"molecular",
"evolution",
"chromatin",
"modification",
"chemistry",
"dna",
"modification",
"biochemistry",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"chemical",
"reactions",
"physical",
"sciences",
"computational",
"biology",
"evolutionary",
"biology",
"medical",
"devices",
"and",
"equipment"
] |
2016
|
A Statistical Framework to Identify Deviation from Time Linearity in Epigenetic Aging
|
Crimean-Congo hemorrhagic fever ( CCHF ) is a serious disease with a high fatality rate reported in many countries . The first case of CCHF in Oman was detected in 1995 and serosurveys have suggested widespread infection of humans and livestock throughout the country . Cases of CCHF reported to the Ministry of Health ( MoH ) of Oman between 1995 and 2017 were retrospectively reviewed . Diagnosis was confirmed by serology and/or molecular tests in Oman . Stored RNA from recent cases was studied by sequencing the complete open reading frame ( ORF ) of the viral S segment at Public Health England , enabling phylogenetic comparisons to be made with other S segments of strains obtained from the region . Of 88 cases of CCHF , 4 were sporadic in 1995 and 1996 , then none were detected until 2011 . From 2011–2017 , incidence has steadily increased and 19 ( 23 . 8% ) of 80 cases clustered around Eid Al Adha . The median ( range ) age was 33 ( 15–68 ) years and 79 ( 90% ) were male . The major risk for infection was contact with animals and/or butchering in 73/88 ( 83% ) and only one case was related to tick bites alone . Severe cases were over-represented: 64 ( 72 . 7% ) had a platelet count < 50 x 109/L and 32 ( 36 . 4% ) died . There was no intrafamilial spread or healthcare-associated infection . The viral S segments from 11 patients presenting in 2013 and 2014 were all grouped in Asia 1 ( IV ) lineage . CCHF is well-established throughout Oman , with a single strain of virus present for at least 20 years . Most patients are men involved in animal husbandry and butchery . The high mortality suggests that there is substantial under-diagnosis of milder cases . Preventive measures have been introduced to reduce risks of transmission to animal handlers and butchers and to maintain safety in healthcare settings .
Crimean-Congo hemorrhagic fever ( CCHF ) is a serious and often fatal infection caused by the CCHF virus ( CCHFV ) . Ixodid ticks , especially Hyalomma spp , act as both reservoirs and vectors . This virus has the greatest geographic range of any tick-borne virus and there are reports of viral isolation and/or disease from more than 30 countries in Africa , Asia , Eastern and Southern Europe , and the Middle East [1–3] . Numerous domestic and wild animals , such as cattle , goats , sheep and small mammals , such as hares and rodents , serve as asymptomatic amplifying hosts for the virus [4] . CCHFV can be transmitted between animals and humans by Hyalomma ticks . It can also be transmitted by direct contact with blood and other body fluids of viremic humans and animals and has the potential to cause population-based outbreaks [5–6] . Clinical features commonly include fever of abrupt onset , myalgia , headache and thrombocytopenia , and can progress to hemorrhage , multiorgan failure and death . The levels of liver enzymes , creatinine phosphokinase , and lactate dehydrogenase are raised , and bleeding markers are prolonged [7–8] . The crude mortality rate of CCHF differs from country to country , ranging from 2–80% [1] . Early diagnosis and supportive management are essential for a favorable outcome . CCHFV is a negative-sense single-stranded RNA virus classified within the Orthonairovirus genus of the Nairoviridae family . The CCHFV genome is comprised of single-stranded negative-sense RNA divided into 3 distinct segments designated small ( S ) , medium ( M ) , and large ( L ) . Comparisons of full S segment sequences have shown that CCHFV forms 7 distinct clades , each with strong geographical associations [1 , 9–11] . Subtle links between distant geographic locations , shown by phylogenetic analysis , may have originated from the international livestock trade or from long-distance carriage of CCHFV by infected ticks via bird migration [5 , 10 , 12–13] . Oman is situated in the southeastern corner of the Arabian Peninsula , bordering the Kingdom of Saudi Arabia , United Arab Emirates , and Yemen . The summer is hot and humid with temperatures reaching as high as 49°C and the winter relatively cooler and with some rain . The total population is 4 , 615 , 269 individuals , of whom 54 . 6% are Omanis and the remainder are expatriates [14] . Cases of CCHF were first detected in Oman in 1995 when there were 3 unrelated sporadic cases , followed by a further case in 1996 [15 , 16] . Cases related to animal movement and slaughter were also reported the following year from Western Saudi Arabia [17–18] and from the UAE [15 , 19 , 20] , where an imported case had previously resulted in fatal infections of health care workers in 1979 [21] . A survey conducted in 1996 in Oman revealed asymptomatic seropositivity for CCHFV exposure in 1/41 ( 2 . 4% ) of Omanis compared to 73 ( 30 . 3% ) of 241 non-Omani citizens with occupational animal contact [22] . However , no further human cases of CCHF were reported in Oman until 2011 [23] , and since then there has been a steady increase [24] . A recent survey has shown infection in a variety of animals and ticks in Oman [25] . Limited data are available on the prevalent clade ( s ) , or group ( s ) of organisms from a single ancestor , of CCHFV in the Arabian Peninsula . Sequencing of the S , M , and L segments of CCHFV isolated from the 1996 patient in Oman ( recorded as Oman 1997 in GenBank ) showed that it belonged to Asia lineage 1 ( clade IV ) [10 , 26] , as was the virus isolated from a patient who returned to India with CCHF acquired in Oman in 2016 [27] . Virus isolates from 4 patients in the UAE in 1994 and 1995 also align with the Asia 1 ( clade IV ) lineage , as did contemporaneous isolates from Hyalomma ticks obtained from livestock imported into the UAE from Somalia [9 , 19 , 28] . A further human isolate in the UAE in 1994/95 aligned with lineage Africa 1 ( clade III ) [9 , 19 , 28] . The aims of this study are to describe the clinical and epidemiological features and outcomes of cases of CCHF diagnosed in Oman between 1995 and 2017 . We also investigated the local molecular epidemiology of CCHFV by partial and complete S segment sequencing of stored CCHFV isolates from patients recently diagnosed in Oman .
A retrospective descriptive record-based review and analysis of CCHF cases was conducted over the period 1995 through 2017 . CCHF has been listed as a notifiable disease in Oman since 1995 and surveillance forms from suspected cases are submitted by all healthcare providers to the Communicable Diseases Department at Ministry of Health headquarters . Blood samples obtained from suspected cases were submitted at the same time to the Central Public Health Laboratory ( CPHL ) at the MoH in Muscat , Oman . All CCHF cases reviewed and included in this study were detected by this routine communicable disease surveillance combined with the CPHL results during the study period . A generic national form is used for initial notification of a suspected case of CCHF; once the diagnosis is confirmed , a more detailed form is submitted that includes patient identifiers , demographic and geographic variables , relevant exposure history , key clinical features , and some clinical laboratory test results . The form is not unique to CCHF , so prompts for some specific CCHFV-related exposures and laboratory variables are missing . The report format remained similar until 2017 when the paper form was replaced by an electronic version . Data were systematically extracted from the surveillance forms of all laboratory confirmed cases of CCHF . Demographic variables included age , sex , nationality , location , and date of notification . Risk factors included history of tick bite , occupational exposure and contact with tissues , blood or other biological fluids from an infected animal , or contact with a case within 14 days prior to the onset of symptoms . Clinical data included presence/absence and duration of fever , headache , myalgia , nausea , vomiting , diarrhea , petechial rash , and bleeding from sites including gums , nose , lung , gastrointestinal tract , or skin . Key laboratory variables include platelet counts , hemoglobin , urea and electrolytes , and liver function tests . The main clinical outcomes were death or survival . The case definition for a suspected case in Oman is: an illness with sudden acute onset with the following clinical findings: a fever ≥ 38 . 5°C ( > 72 hours to < 10 days ) associated with severe headache , myalgia , nausea , vomiting , and/or diarrhea; thrombocytopenia < 50 x 109/L; hemorrhagic manifestations which develop later and may include petechial rashes , bleeding from the gums , nose , lungs , gastrointestinal tract , etc . ; history of tick bite , occupational exposure , contact with fresh tissues , blood , or other biological fluids from an infected animal [24] . Good laboratory practice and a high level of effective biosafety precautions are required by laboratory staff handling materials from suspected CCHF cases due to the potential for sample-to-person , or indirect , transmission [6 , 29] . Blood samples collected from suspected cases of CCHF admitted to all MoH and non-MoH health care institutions in Oman are sent to CPHL in triple pack containers , using the most direct and timely route available . These samples are considered urgent and results are provided within 24 hours of their arrival at CPHL . National guidelines are in place to instruct local laboratories where suspected cases are admitted on safe handling of all material collected for any diagnostic purpose [24 , 30] . Both serum and plasma samples are requested for CCHFV testing . Plasma is preferred for molecular testing using a commercial CCHFV real time reverse transcription polymerase chain reaction ( rRT-PCR ) kit ( In vitro Diagnostics , Liferiver Shanghai ZJ Bio-Tech Co . , Ltd . Shanghai , China ) . At CPHL , the plasma extraction takes place inside a gloved box using a manual extraction system , QIAamp Viral RNA Kit ( QIAGEN , Hilden , Germany ) . The samples are first treated with AVL buffer ( QIAGEN ) to inactivate infectious viruses and RNases . Intact viral RNA is then purified by selective binding and washing steps . The screening RT-PCR reaction is based on a one step real-time RT-PCR . Briefly , CCHFV RNA is converted into cDNA and a thermostable DNA polymerase is used to amplify specific CCHFV S segment sequence targets by standard thermocycling in a PCR as per manufacturer instruction . The kit contains an internal control to identify possible PCR inhibition . A positive result from a RT-PCR screening for CCHFV RNA is used to confirm infection . In such cases , the serum sample is not tested further . If the RT-PCR is negative , heat inactivated serum ( 56°C water bath for 30 minutes ) is tested for CCHFV antigen and IgM and IgG antibodies using a commercial kit ( Vector-Best , Novosibirsk , Russia ) . For samples that are negative for all parameters , a convalescent serum is requested for CCHF IgG testing . The CPHL takes part in regular internal and external quality assurance reviews in association with WHO EMRO and WHO Quality Management Standards . All available stored serum samples , collected from 21 CCHF patients in 2013 and to 2014 , were inactivated with AVL buffer and sent to PHE Porton Down , England , UK . At PHE , AVL samples were processed with a standard QIAamp Viral RNA Kit . Eluted RNA was evaluated for the presence of CCHFV RNA using an in-house RT-PCR assay [31] . Sequencing was performed using standard CCHFV S segment sequencing primers as described previously [32] . Assembled sequence data for the S segment of each sample were manipulated and analyzed using the Lasergene suite of programs ( DNAStar , Maddison , WI , USA ) . For phylogenetic analysis , sequences were aligned using the Clustal W computer program ( The European Bioinformatics Institute , Wellcome , UK ) [33] and output in PHYLIP Format ( scikit-bio ) . To construct maximum-likelihood phylogenetic trees , quartet puzzling was applied using the program , Tree-Puzzle , at the Institut Pasteur [34 , 35] . The Tamura-Nei model of substitution was adopted , as has been performed in other phylogenetic studies demonstrating reassortment [36] . Phylogenetic trees were drawn using the program TreeView ( JAM Software GmbH , Trier , Germany ) [37] . The values at the tree branches represent the puzzle support values . S segment sequences were submitted to GenBank . The data analysis was conducted at the MoH Department of Surveillance in Muscat . A descriptive analysis compared age , sex , nationality , location , and date of cases . Risk factors and clinical and laboratory parameters were also tabulated . Missing data items ( positive or negative ) were omitted from analysis . Statistical comparisons were performed using SPSS 11 . 0 package program ( SPSS Inc , Chicago , IL , USA ) . Ethical approval was sought from the MoH , Oman . The study is considered free from ethical constraints as it is a secondary analysis of the data collected routinely for the purpose of public health surveillance and reporting . No personal identifying information accompanied the samples sent to PHE .
A total of 88 cases were reported between 1995 and 2017 . Of these , 82 ( 93 . 2% ) were confirmed by RT-PCR and 4 by CCHFV IgM alone . Two further probable cases ( both fatal ) in 2011 and 2016 were included on the basis of typical clinical and laboratory features as per electronic records . There were 3 isolated cases in January , May , and June 1995 with a further case in 1996 , and then no cases were reported until 2011 . Since then , there has been a steady increase in numbers , peaking at 20 cases in 2015 ( Fig 1 ) . Annual notifications of suspected cases were not systematically recorded until 2011 and data about notifications of suspected cases and possible missed cases are incomplete . In the years 2001 to 2011 inclusive , there were 35 notifications of possible VHF cases , of which 2 were proven CCHF ( in 2011 ) and at least 23 were confirmed to be cases of dengue ( 2 fatal ) . The patients had a median ( range ) age of 33 ( 15–68 ) years and 79 ( 90% ) were male . The most common nationality affected was Omani 51 ( 59% ) followed by Bangladeshi 18 ( 21% ) , Pakistani 7 ( 8% ) , Yemeni 3 ( 4% ) , Indian 4 ( 5% ) , Somali 2 , and Sri Lankan 1 . Cases occurred in all governorates ( wilayats ) except Musandam and Al Wustah ( Fig 2 ) . There was no geographic or source-related clustering of cases; however , several cases followed Eid Al Adha , a festival associated with animal sacrifice . In the years 2013–2017 , 19/80 ( 23 . 8% ) of all cumulative cases had their onset within 3 weeks after Eid Al Adha ( Fig 3 ) . There was also a smaller peak of cases in the spring weeks 6–19 ( Fig 3 ) . The main exposure risk identified was animal/fresh tissue exposure in 73/88 ( 83% ) , with only 1 case attributed to tick bite alone . Exposure risk was not identified in 14 ( 15 . 9% ) ( Table 1 ) . Clinical features in 88 patients included fever in 80 ( 90 . 9% ) , hemorrhagic features 41 ( 46 . 6% ) , vomiting 32 ( 36 . 4% ) , myalgia 30 ( 34 . 1% ) , diarrhea 20 ( 22 . 7% ) , respiratory symptoms 17 ( 19% ) , abdominal pain 11 ( 12 . 5% ) , other symptoms in 29 ( 33% ) . Severe thrombocytopenia ( platelet count < 50 x 109/L ) was present in 64 ( 72 . 7% ) . There were 32 deaths , resulting in a cumulative case fatality rate of 36 . 4% . The case fatality rate in Omanis was 16/53 ( 30 . 2% ) and in Bangladeshis was 10/18 ( 55 . 6% ) ( P>0 . 05 ) . Of the 21 serum samples that were sent to PHE , 20 were RT-PCR positive using an in-house assay . However , of these , only 12 samples provided suitable cycle threshold values ( the cycle threshold being 28 or under ) to warrant further sequencing of CCHFV S segments and only 12 samples provided sequencing data which spanned the entire ORF of the S segment . Sequence data have been submitted to GenBank and sequences have been assigned the following accession numbers: MH037279 ( Oman 2012-40S ) , MH037280 ( Oman 2013-116S ) , MH037281 ( Oman 2014-828P ) , MH037282 ( Oman 2014-979P ) , MH037283 ( Oman 2014-602P ) , MH037284 ( Oman 2013-825P ) , MH037285 ( Oman 2013-92S ) , MH037286 ( Oman 2013-108S ) , MH037287 ( Oman 2013-179P ) , MH037288 ( Oman 2014-860P ) , MH037289 ( Oman 2014-624S ) , and MH037290 ( Oman 2014-747P ) . Sequences were compiled with a range of other CCHFV S segment ORF sequences and used to make the maximum likelihood phylogenetic tree shown in Fig 4 .
This report summarizes the clinical , epidemiological , and virological findings in 88 people with symptomatic CCHF throughout the Sultanate of Oman in the past 2 decades . Cases were detected by passive surveillance , starting with a few sporadic reports in 1995 and 1996 , followed by no cases until 2011 . Since then there has been a sustained increase in yearly reports , of which 19/80 ( 23 . 8% ) have clustered around the Eid Al Adha festival , occurring in summer months in the years 2013 to 2017 , with a possible smaller peak in the spring months . Ninety percent of all patients were male with a median age of 33 years . Both Omanis and citizens of other nationalities were affected , the predominant risk factors being exposure to animals and meat products , especially involvement in butchering or slaughtering . Diagnosis was confirmed by RT-PCR in 82 ( 93 . 2% ) cases and by serology alone in 4 ( 4 . 5% ) . Stored viral RNA from 12 patients presenting in 2013 and 2014 was sequenced for the entire S segment ORF of each of the 12 samples , all grouped in the Asia 1 ( IV ) clade . The cumulative mortality was 36 . 4% , and no cases of healthcare related or intrafamilial spread of infection were reported . This is the largest series of cases of CCHF reported from a GCC country , and provides the first data about locally prevalent strains of CCHFV in almost 20 years . The findings raise a number of questions about the origin and distribution of CCHFV in Oman and neighboring countries , the reasons for the high observed mortality , and the appropriate human and veterinary public health responses in Oman and other GCC states . The clinical features of the cases were similar to those reported in other countries [13 , 38–40] . The most common symptoms reported were fever , fatigue , headache , loss of appetite , myalgia , and abdominal pain . Hemorrhagic manifestations were described in 34/67 ( 50 . 8% ) and severe thrombocytopenia ( < 50 x109/L ) was present in 64/88 ( 72 . 7% ) at presentation . There is no internationally agreed case definition for CCHF , but at least 3 scoring systems to assess severity of illness have been proposed [41–43] . Mortality is known to be associated with older age , presence of underlying illness , and CCHFV viral load at presentation [8] . Case numbers were too small to show a link with mortality in our series and details about the latter 2 risks were not recorded . Representative case fatality rates elsewhere include 5% in Turkey , 17 . 6% in Iran , and 15% in Pakistan [1 , 44–45] . However , the CFR of 15% in Pakistan was reported from a center with substantial experience , whereas overall mortality rates of up to 41% have been reported more recently in Pakistan , especially during outbreaks [46] . The lower mortality in Turkey and Iran could be explained by improved surveillance and early diagnosis of CCHF in patients with fever and thrombocytopenia , following prolonged campaigns to raise awareness in both healthcare personnel and the general public in those countries . A serosurveillance study conducted in Oman in 1996 showed that none of the 74 antibody-positive individuals identified recalled ever being hospitalized for an illness resembling CCHF with associated fever and bleeding , suggesting that there is a substantial incidence of subclinical CCHF human infections in Oman [3 , 12–13 , 22] . Serosurveillance studies in other countries have shown seroprevalence rates of approximately 10–13% in high risk human populations [5 , 47] . Based on mortality data from Turkey we believe that under-diagnosis of mild cases has skewed the mortality data in Oman . This is the first study to describe the complete sequence of the S segment ORFs of a series of CCHFV isolates from the region . The results largely confirm findings from partial sequencing of sporadic isolates from the UAE [19 , 28] and Oman [10 , 27] since the mid 1990’s . The phylogenetic relationship of these sequences with other published sequences from the region is depicted in Fig 4 . The similarity of all these sequences to the human cases in Oman and the UAE over 2 decades is striking , and these sequences align with those from Pakistan . In the future , with the advance of cheaper sequencing technologies , it will be valuable to compare full length genomes from multiple locations with clinical data . This may help address hypotheses about alternative strain pathogenicity , including the relative contribution of segment reassortment in CCHF disease [10 , 48] . The average age of patients in Oman was 33 years , and 90% were men . Most infections were acquired while butchering or slaughtering animals or from other close animal tissue and blood exposure as in earlier and more recent cases in Dubai [20 , 49] and the Kingdom of Saudi Arabia [17] . This contrasts with the situation in Turkey , Kazakhstan , and Iran , where tick bites are the most commonly reported risk factor [11 , 39–40 , 44] . Slaughtering animals during Eid Al Adha is known to pose a particularly high risk of infection [5] . Sporadic unregulated slaughtering without using appropriate personal protective equipment still occurs in Oman during Eid Al Adha . It is also common for non-professional individuals to become involved and for butchers to freelance , going from house to house to sacrifice animals , as people find it more convenient to have the sacrifice performed at farms and backyards . During skinning and subsequent tanning of the hides , ticks can bite humans . We examined the effect of Eid as a possible cause for the apparent increase in cases over the past 7 years as the festival has moved back into the summer months when tick activity is most prominent . However , this is not the only factor . Fig 3 demonstrates that there is the expected clustering of cases after Eid Al Adha , but this accounts for only 23 . 6% of the total cases . Similar findings have been reported in Pakistan [50] and the data suggest that climatic factors affecting tick activity are most important in promoting seasonal variation in human infection risk together with extra added risk at the time of Eid . These data change our perceptions about the duration and origin of CCHFV activity in Oman and neighboring countries . Previously , it had been postulated that sporadic cases were related to the importation of infected livestock from other countries . A large amount of livestock is imported to Oman every year: in 2016 , over 1 . 5 million farm animals were imported , including sheep ( 85 . 2% ) , cattle ( 7% ) , and goats ( 5 . 6% ) . The origins of these included Armenia , Australia , Djibouti , India , Iran , Jordan , Pakistan , Somalia , Sudan , Turkey and the UAE . A 21-day quarantine procedure is in effect for animals arriving from other countries by sea or land . Once in Oman , animals are distributed to sales centers , feedlots , and distribution points throughout the country . Livestock are held in large holding pens and not segregated according to country of origin or time from entry into Oman . Spread of infection could result from unrestricted entry of tick-infested and potentially viremic domestic animals during religious holidays; the abundance of virus-infected ticks within stockyards and holding pens; the uncontrolled movement of livestock animals infested with CCHFV-carrying Hyalomma ticks to ranches , farms , and markets throughout the country; and the indiscriminate mixing and crowding of tick-infested and potentially viremic animals with uninfected and tick-free animals [3 , 13 , 22] . There was partial support for the possibility of intermittent importation of CCHFV with livestock into the UAE in the 1990s , where ticks were found on animals with different clades of CCHFV S segment corresponding to African as well as Asia 1 clades [9 , 19 , 28] . However , human serosurveys in Oman in 1996 [22] and the finding of the virus in ticks and animals throughout Oman in 2013–2014 [25] suggest that all areas of the Sultanate have had a substantial burden of CCHFV infection for at least 2 decades , probably related to all the risks mentioned above . Moreover , all virus isolates from humans in Oman and the UAE have had remarkably similar S Segments , apart from the nosocomial outbreak in Dubai in 1979 from an Indian index case [21] . In contrast , several different S segments are circulating in Somalia , Iran , and Turkey [9 , 10 , 51] . This suggests that the Asia 1 S segment of CCHFV has been circulating in Oman for more than 20 years . It will be of interest to fully sequence the complete genomes of the recent isolates from ticks in Oman ( and elsewhere ) to explore this hypothesis further [25] . Reports of CCHFV antibody positivity in earlier human serosurveys in Kuwait [52] and intermittent occupational-related outbreaks in the UAE [19 , 20 , 28] , and KSA [17 , 18] , since then suggest that this is also the case throughout GCC countries . The Oman MoH has undertaken a number of activities and initiatives to educate and inform the public about the risks of CCHF infection associated with slaughtering . A joint strategic initiative was developed in collaboration with the Ministry of Agriculture and Fisheries and Ministry of Municipalities and Water Resources . Education and information on prevention of CCHF in different languages has been targeted at those involved in slaughtering and handling animals . This includes placing advertisements on social media platforms , TV , radio , billboards , magazines , and newspapers before and during Eid Al Adha . Knowledge about CCHF is increasing in Oman with hospitals now following guidelines for the management of suspected cases of CCHF [53] . In addition , guidelines have been produced for culturally acceptable safe burials [24 , 54] . It is reassuring that no healthcare related infections were detected in this series . The data suffer from the limitations of a retrospective study that spans over 20 years , based on notifications of suspected illness and laboratory reports . In particular , the completeness of notification has been highly variable and is likely to have underestimated the incidence of symptomatic infections . Data on notifications of suspected cases that later turned out to be negative for CCHFV have not been systematically recorded and it is likely that the gap in notified cases between 1997 and 2011 is due to missed diagnoses and underreporting , rather than absence of cases . However , the records of confirmed cases at CPHL are thought to be complete . Conversely , the increase in notifications since 2011 may be due to a genuine increase in cases and/or be due to increased physician awareness and hence case recognition and reporting . This is the largest reported series of CCHF from any of the GCC countries to date and brings together all published viral sequences in this region . The implication is that CCHF is endemic and under-recognized in Oman and surrounding countries and that prospective studies are needed to determine how often less severe cases of fever and thrombocytopenia are presenting in Oman . Proven and suspected cases have been reported in expatriate travelers returning from Oman to India [27] and Pakistan [55] and the possibility of CCHF should be considered in febrile travelers arriving from GCC countries , especially if they have been involved in animal slaughtering [56] . Oman has responded by improving its notification systems and laboratory support . Active local and regional programs of health promotion and human illness prevention need to be maintained together with surveillance and control of infection in animals and local tick vectors .
|
Crimean-Congo hemorrhagic fever , an often fatal tick-borne viral disease , has made an impact in the Sultanate of Oman—affecting nationals and expatriates alike—for the past 20 years . In this retrospective review of the epidemiology and outcomes of cases in Oman from 1995 to 2017 , we identified 4 sporadic cases in 1995 and 1996 , then none until 2011 , followed by a steady increase until 2017 . The mortality rate of 32 of 88 cases ( 36 . 4% ) is high in comparison to studies from other countries and this could be explained by under-diagnoses of milder cases in the Sultanate . Transmission is commonly associated with animal husbandry and butchering and 88% cases were infected by contact with animals , whereas transmission by tick bite is more commonly recorded in some countries . A proportion of cases ( 23 . 8% ) were clustered around the Eid-Al-Ahda festival which has , from 2011–2017 , occurred in the summer months , which have a higher risk of transmission . This additional risk has been noted and preventive measures have been introduced to reduce the risk of transmission to animal handlers and butchers .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"death",
"rates",
"reverse",
"transcriptase-polymerase",
"chain",
"reaction",
"taxonomy",
"invertebrates",
"livestock",
"sequencing",
"techniques",
"medicine",
"and",
"health",
"sciences",
"ixodes",
"crimean-congo",
"hemorrhagic",
"fever",
"tropical",
"diseases",
"geographical",
"locations",
"animals",
"organisms",
"phylogenetics",
"data",
"management",
"phylogenetic",
"analysis",
"neglected",
"tropical",
"diseases",
"molecular",
"biology",
"techniques",
"population",
"biology",
"ticks",
"rna",
"sequencing",
"research",
"and",
"analysis",
"methods",
"viral",
"hemorrhagic",
"fevers",
"infectious",
"diseases",
"computer",
"and",
"information",
"sciences",
"artificial",
"gene",
"amplification",
"and",
"extension",
"oman",
"evolutionary",
"systematics",
"molecular",
"biology",
"disease",
"vectors",
"agriculture",
"arthropoda",
"people",
"and",
"places",
"population",
"metrics",
"arachnida",
"eukaryota",
"asia",
"biology",
"and",
"life",
"sciences",
"viral",
"diseases",
"species",
"interactions",
"evolutionary",
"biology",
"polymerase",
"chain",
"reaction"
] |
2019
|
Clinical and molecular epidemiology of Crimean-Congo hemorrhagic fever in Oman
|
Blastomyces dermatitidis belongs to a group of human pathogenic fungi that exhibit thermal dimorphism . At 22°C , these fungi grow as mold that produce conidia or infectious particles , whereas at 37°C they convert to budding yeast . The ability to switch between these forms is essential for virulence in mammals and may enable these organisms to survive in the soil . To identify genes that regulate this phase transition , we used Agrobacterium tumefaciens to mutagenize B . dermatitidis conidia and screened transformants for defects in morphogenesis . We found that the GATA transcription factor SREB governs multiple fates in B . dermatitidis: phase transition from yeast to mold , cell growth at 22°C , and biosynthesis of siderophores under iron-replete conditions . Insertional and null mutants fail to convert to mold , do not accumulate significant biomass at 22°C , and are unable to suppress siderophore biosynthesis under iron-replete conditions . The defect in morphogenesis in the SREB mutant was independent of exogenous iron concentration , suggesting that SREB promotes the phase transition by altering the expression of genes that are unrelated to siderophore biosynthesis . Using bioinformatic and gene expression analyses , we identified candidate genes with upstream GATA sites whose expression is altered in the null mutant that may be direct or indirect targets of SREB and promote the phase transition . We conclude that SREB functions as a transcription factor that promotes morphogenesis and regulates siderophore biosynthesis . To our knowledge , this is the first gene identified that promotes the conversion from yeast to mold in the dimorphic fungi , and may shed light on environmental persistence of these pathogens .
The endemic dimorphic fungi are comprised of seven ascomycetes that include Blastomyces dermatitidis , Histoplasma capsulatum , Coccidioides immitis , Coccidioides posadasii , Paracoccidioides brasiliensis , Sporothrix schenckii , and Penicillium marneffei . These fungi possess the unique ability to switch between two different morphologies , yeast and mold , in response to external stimuli [1] . In nature , they grow as mycelia that produce conidia , which are the infectious particles; when aerosolized spores are inhaled into the warmer lungs of a mammalian host , they convert into pathogenic yeast and cause necrotizing infection [1] . The dimorphic fungi collectively are the most common cause of invasive fungal disease worldwide and account for several million infections each year [2] . Unlike opportunistic fungi , such as Cryptococcus or Aspergillus , the dimorphic fungi can infect both immunocompetent and immunocompromised hosts [3]–[5] . The size of the inhaled inoculum and the integrity of the cell-mediated immune system influence the extent and severity of infection [1] , [3] . Clinical manifestations range from asymptomatic infection to symptomatic disease and include pneumonia , acute respiratory distress syndrome , and disseminated disease involving multiple organ systems [1] , [3] . The ability of the dimorphic fungi to switch between the two different morphologies is crucial for pathogenesis . Although temperature is postulated to be the major stimulus that induces phase transition , other stimuli , including carbon dioxide tension , steroid hormones , and oxidative stress influence this morphologic switch [1] , [6]–[9] . Phase transition is a complex process that involves the coordinated expression and repression of many genes in response to external stimuli , which alters cell wall composition , metabolism , intracellular signaling , and morphology [10]–[13] . The identification of DRK1 ( dimorphism-regulating kinase-1 ) in B . dermatitidis and H . capsulatum offered strong genetic evidence that phase transition is required for pathogenicity [10] . DRK1 functions as a global regulator and has pleotropic effects on the cell , controlling morphogenesis , cell wall composition , sporulation , expression of yeast-phase specific genes , and virulence . DRK1 null mutants remain locked in the mycelial phase , fail to sporulate or express the essential virulence factors BAD1 ( Blastomyces adhesin-1 in B . dermatitidis ) and CBP1 ( Calcium binding protein-1 in H . capsulatum ) , and are avirulent in a murine model of infection [10] . Three additional genes , RYP1 , RYP2 , and RYP3 , have been described that regulate morphogenesis in H . capsulatum . Silencing the expression of RYP1 , 2 or 3 results in hyphal growth at 37°C and inappropriate sporulation [12] , [13] . The goal of this study was to identify and characterize additional genes that regulate the phase transition in dimorphic fungi , using B . dermatitidis as a model system . While progress has been made in identifying genes that regulate the morphological transition from mold to yeast , to our knowledge , no genes have been identified that regulate the switch in the other direction in the dimorphic fungi – that is , from the yeast to mold form . The mold form is believed to be required for the growth and survival of the dimorphic fungi in the environment by enabling propagation in soil and transmission to humans through the generation of conidia . Herein , we describe a gene , SREB , identified through insertional mutagenesis , which impacts multiple disparate fates in B . dermatitidis , including the phase transition of yeast to mold , cell growth at 22°C , and the biosynthesis of siderophores .
Agrobacterium tumefaciens-mediated DNA transfer was used to mutagenize haploid , uninucleate conidia of B . dermatitidis strain T53-19 . Following selection with hygromycin , 22 , 000 transformants were visually screened by light microscopy for morphologic alterations including growth as hyphae or pseudohyphae at 37°C or as yeast at 22°C . In this study , one of the mutants identified by the screen , 3-15-1 , was characterized in detail . This mutant , unlike the parent strain , was pigmented yellow and failed to complete the conversion from yeast to mold ( Figure 1A , 1B ) . Southern blot hybridization demonstrated a single site of insertion ( Figure S1 ) . The genomic DNA flanking the insert in 3-15-1 was amplified using adapter PCR , sequenced , and analyzed using a BLASTn search against the genome sequence of B . dermatitidis strain 26199 . No rearrangements or deletions were identified in the DNA flanking the insert . Additional BLAST analyses indicated that the insert interrupted a region 692 base-pairs ( bp ) upstream of a predicted open reading frame with nucleotide homology to Pencillium chrysogenum SREP , which encodes a GATA transcription factor that regulates the biosynthesis of siderophores [14] . We named this homolog SREB ( siderophore biosynthesis repressor in Blastomyces ) in B . dermatitidis . FGENESH analysis of the nucleotide sequence predicted that SREB contained a 1909 nucleotide ( nt ) coding region interrupted by two short introns ( 88 and 74 nt ) . Each intron was located in a zinc-finger coding region and contained the expected 5′-splice donor ( GTNNGT ) and 3′-splice acceptor ( pyrimidine-AG ) sequences [15] . The length , location , and number of introns interrupting the open reading frame were conserved among B . dermatitidis SREB , H . capsulatum SRE1 , A . nidulans SREA , and N . crassa SRE [16]–[18] . The SREB coding region was predicted to encode a 636 amino acid protein . The predicted amino acid sequence of SREB had homology to siderophore biosynthesis repressors in other fungi including Aspergillus nidulans SREA , Penicillum chrysogenum SREP , Neurospora crassa SRE , Ustilago maydis URBS1 , Schizosaccharomyces pombe FEP1 , Candida albicans SFU1 , Cryptococcus neoformans CIR1 , and Histoplasma capsulatum SRE1 ( Figure 1C ) [14] , [16]–[22] . SREB contained several conserved domains characteristic of GATA transcription factors that regulate iron assimilation , including two zinc finger motifs separated by a cysteine-rich region ( CRR ) and a C-terminus predicted to have a coiled-coil domain ( Figure 1C ) [17] , [23] . With the exception of C . neoformans CIR1 , fungal GATA transcription factors that regulate the acquisition of iron contain two zinc fingers [22] . This zinc finger arrangement is unique because most GATA transcription factors in fungi contain only one zinc finger [17] . The cysteine residues in each zinc finger of SREB were arranged in a conserved class IV motif , Cys-X2-Cys-X17-Cys-X2-Cys [24] . The cysteine-rich region contained four conserved cysteine residues , which have been demonstrated to coordinate the binding of iron in H . capsulatum [16] . Mutant 3-15-1 failed to convert from yeast to mycelia or produce conidia following a shift in incubation temperature from 37°C to 22°C ( Figure 1A ) . In contrast , the parent strain T53-19 converted to mycelia when grown at 22°C and produced conidia . Mutant 3-15-1 accumulated little biomass at 22°C , but remained viable ( as measured by the exclusion of 0 . 2% eosin stain ) , and converted to normal yeast morphology when the incubation temperature was shifted back to 37°C ( data not shown ) . The yellow-orange pigmentation of mutant 3-15-1 and the predicted amino acid sequence suggested that SREB functioned as a repressor of siderophore biosynthesis . Deletions of SREB homologs in P . chrysogenum ( SREP ) , A . nidulans ( SREA ) , and N . crassa ( SRE ) produce similar discoloration [14] , [17] , [18] . To assess for the dysregulation of siderophore biosynthesis in the insertion mutant , we used a colorimetric assay to detect the production of hydroxymate-type sideophores in culture supernatants [25] . Under iron-poor conditions , both T53-19 and 3-15-1 produced an abundance of siderophores as measured by this assay ( data not shown ) . Under iron-replete conditions , mutant 3-15-1 continued to produce siderophores , whereas parent strain T53-19 repressed siderophore biosynthesis ( Figure 2A ) . To determine if the mutant phenotype was from altered expression of SREB , and not due to another mutation incurred during insertional mutagenesis , we set out to complement the mutant phenotype . Insertional mutant 3-15-1 was re-transformed via A . tumefaciens to provide an intact gene copy of SREB and its endogenous promoter . Complemented strains A5 and D5 grew as white colonies that did not discolor the medium , suppressed siderophore production under iron-replete conditions ( 10 µM FeSO4 ) , and converted fully to mycelia when grown at a temperature of 22°C ( Figure 2A-C ) . Retransformation of 3-15-1 with a vector lacking SREB did not complement the mutant phenotype ( empty vector strain ) ( Figure 2A-C ) . Whereas Northern analysis demonstrated a reduction in the abundance of SREB transcript in mutant 3-15-1 compared to the parental strain , message levels were overexpressed in both complemented strains ( Figure 2D ) . Thus , complementation reversed the mutant's phenotypic defects , supporting the idea that the insert was responsible for the dysregulation of siderophore biosynthesis and the alteration in morphogenesis . To confirm that SREB represses the biosynthesis of siderophores and affects morphogenesis in B . dermatitidis , we disrupted this gene in wild-type isolate 26199 using homologous recombination . To minimize the probability that the phenotype observed in mutant 3-15-1 was unique to strain T53-19 , we used a different B . dermatitidis strain , 26199 , to generate a null mutant . The rate of allelic replacement was 0 . 04% ( 1/2670 ) . The null mutant , SREBΔ , grew as yellow-pigmented colonies that discolored the surrounding medium and failed to properly repress siderophore biosynthesis when iron was abundant ( Figure 3A , 5B ) . The intensity of pigmentation was dependant on exogenous iron and independent of temperature ( 37°C vs . 22°C ) ( data not shown ) . In contrast , the parent strain grew as white-colored yeast and repressed the production of siderophores under iron-replete conditions as measured by the ferric perchlorate assay ( Figure 3A , 5B ) . SREBΔ failed to complete the yeast-to-mold phase transition following a shift in temperature from 37°C to 22°C , did not exhibit radial growth , and accumulated little biomass at 22°C ( Figure 3A , 3B ) . The defect in phase transition persisted during prolonged incubation ( >14 days ) at 22°C; however , a few hyphal strands would develop and could only be observed by light microscopy . Similar to insertional mutant 3-15-1 , SREBΔ remained viable at 22°C ( as measured by 0 . 2% eosin exclusion ) and converted back to yeast following a shift in temperature from 22°C to 37°C ( data not shown ) . In the yeast form , the SREBΔ mutant grew at the same rate as the parent strain ( Figure 3C ) . The morphologic defect at 22°C was independent of exogenous iron concentrations ( data not shown ) . Analysis of the null mutant by PCR indicated disruption of SREB and the absence of any deletion or rearrangment of the genomic DNA flanking the transgene ( data not shown ) . Southern blot analyses demonstrated replacement of SREB with a hygromycin resistance cassette and the absence of additional deletions in the genomic DNA flanking the transgene in SREBΔ ( Figure 4A-E ) . Northern analysis demonstrated the loss of SREB transcript in SREBΔ ( Figure 4F ) . To confirm the phenotype in SREBΔ was due to disruption of the siderophore biosynthesis repressor gene , we re-transformed the null mutant using A . tumefaciens to insert a copy of SREB . Complemented strains grew as white-colored colonies and properly suppressed the biosynthesis of siderophores when iron was abundant ( Figure 5A , 5B ) . Following a temperature shift from 37°C to 22°C , complemented yeast strains converted to mold ( Figure 5C ) . This conversion was slower in the complemented strains ( 14–17 days ) when compared to the wild-type isolate ( <7 days ) ( data not shown ) . The complemented strains underwent radial growth at 22°C; however , colony expansion was less than the wild-type isolate ( data not shown ) . Prolonged incubation did not result in catch-up growth . Analysis of transcript abundance demonstrated restoration of message levels in C#25 and overexpression in C#6 when compared to wild-type and SREBΔ strains ( Figure 5D ) . To test if the expression of SREB was influenced by the concentration of exogeneous iron , we grew wild-type B . dermatitidis strain 26199 under iron-poor and –replete conditions . Northern blot analysis demonstrated that the expression of SREB was increased during conditions of iron abundance and repressed when iron was limited ( Figure 4F ) . In fungi , the expression of genes that encode proteins involved with iron assimilation are often co-expressed or -repressed when iron is limited or abundant , respectively . To investigate whether this was also true in B . dermatitidis , we analyzed the expression of several genes in response to exogenous iron . Under iron-poor conditions , B . dermatitidis wild-type strain 26199 induced the expression of genes involved in the biosynthesis of siderophores ( SIDA ) , transport of ornithine from the mitochondria into the cytosol ( AMCA ) , uptake of siderophores ( MIRB , MIRC ) , and a bZIP transcription factor ( HAPX ) ( Figure 6 ) . Conversely , these genes were repressed when iron was abundant ( Figure 6 ) . The disruption of SREB de-repressed the expression of each of these genes . Thus , SREB regulates genes involved in siderophore biosynthesis and uptake in B . dermatitidis ( Figure 6 ) . To further characterize the regulatory role of SREB on siderophore biosynthesis , we used LC/MS and reverse-phase HPLC to identify the specific type ( s ) of siderophores secreted by B . dermatitidis wild-type and null mutant yeast cells . Starting with wild-type cells grown under iron-limited conditions , siderophores from culture supernatant were isolated using column chromatography . Mass spectroscopy of the eluate showed two large peaks at 4 . 16 and 7 . 26 minutes with molecular weights of 538 . 2 and 822 . 2 that correspond to dimerum acid and coprogen , respectively ( Figure 7A-C ) . Reverse-phase HPLC of the eluate and comparison of retention times to siderophore standards confirmed the identities of these siderophores ( Figure 7D ) . Under iron-replete conditions , wild-type B . dermatitidis repressed the biosynthesis of dimerum acid and coprogen ( Figure 7D ) . In contrast , the null mutant continued to produce both siderophores ( Figure 7D ) . To identify candidate genes regulated by SREB that may promote the phase transition , we first used MAST analysis to search the Blastomyces genome for GATA transcription factor-binding motifs in intergenic regions located ≤2000 bp upstream of predicted genes . Our initial search for the classic GATA transcription factor-binding motif , HGATAR , revealed the presence of this motif upstream of nearly all B . dermatitidis genes . This finding is similar to Schrettl et al . , who found widespread distribution of this motif in Aspergillus fumigatus [26] . An extended version of the HGATAR motif , ATC-w-gAta-a , has been recently described and was demonstrated to occur at a 5 . 4-fold higher frequency in the promoter of genes regulated by A . fumigatus SREA , an SREB homolog , when compared to the entire A . fumigatus genome [26] . We revised our strategy and searched for this extended motif in the promoter of genes in the B . dermatitidis genome . We identified a total of 1 , 213 genes with at least one of the following motifs located ≤2 kb upstream of the start codon: ATC- ( A/T ) -GATA- ( A/G ) , ATC- ( A/T ) -GATA- ( T/C ) , ATC- ( A/T ) -GATT-A , ATC- ( A/T ) -GATC-A , ATC-A-GATG-A , ATC-C-GATA-A , and ATC-A-AATA-A . This gene-set included genes involved in siderophore biosynthesis and uptake ( i . e . SIDA , MIRB , AMCA ) . Two or more upstream GATA motifs were present in 232 ( 19 . 1% ) in the gene-set . Hwang and colleagues identified the motif ( G/A ) -ATC- ( A/T ) -GATA-A upstream of siderophore biosynthesis and transport genes regulated by SID1 in H . capsulatum [27] . We found this longer motif upstream of 271 ( 22 . 3% ) of our 1 , 213 MAST-identified genes; however , MIRB and MIRC , both involved in siderophore uptake , lacked the motif . To classify the 1 , 213 candidate genes into functional categories and facilitate further analysis , we annotated the predicted protein products of these genes as well as the complete B . dermatitidis predicted proteome against the eukaryotic orthologous groups ( KOG ) database . The results , shown in Table 1 , indicate that the KOG-annotated GATA-containing genes fall into many categories of gene function ( i . e . transcription , RNA metabolism , signal transduction , cell remodeling and metabolism ) . The frequency of KOG-annotated genes with upstream GATA motifs within a particular KOG category was compared to the frequency of genes in the same KOG category within all KOG-annotated genes in the B . dermatitidis genome . Three KOG categories were significantly over-represented in the candidate gene-set harboring GATA sites: amino acid transport and metabolism ( KOG code E ) , secondary metabolites biosynthesis , transport and catabolism ( KOG code Q ) , and lipid transport and metabolism ( KOG code I ) ( Table 1 and Table S1 ) . This suggests that these cellular process pathways may be important for SREB regulation , although it does not exclude a role for the GATA-containing genes in other KOG groupings . In a complimentary approach to identify genes that may be regulated by SREB , we performed a preliminary microarray analysis . Using an expression array with 70-mer oligonucleotides representing the 10 , 567 open reading frames of B . dermatitidis strain 26199 , we used two-color spotted analysis to compare isogenic wild-type vs . SREBΔ at 37°C and at 22°C 48 hours after the temperature shift downward ( data not shown ) . At least 38 of the genes identified by MAST analysis were differentially expressed ( increased or decreased by ≥2-fold ) , including seven genes classified by KOG to be involved in lipid transport and metabolism . To validate the microarray results , we performed quantitative RT-PCR on a subset of four genes found to be altered in expression; three from the lipid transport and metabolism KOG category , and one from the carbohydrate metabolism category . At 22°C , the null mutant strain failed to upregulate the expression of a lipid transfer protein and acetoacetyl-CoA synthase ( Figure 8 ) . Conversely , the expression of a peroxisomal dehydratase was over-expressed at 37°C and 22°C , when compared to the wild-type isolate ( Figure 8 ) . We also confirmed the altered expression of a glycosyl hydrolase postulated to be involved in cell-wall remodeling . In the null mutant , this gene is over-expressed at 37°C and 22°C , when compared to the wild-type isolate ( Figure 8 ) . Thus , we have begun to identify candidate genes and processes that may be direct or indirect targets of SREB and contribute to the phase transition from yeast to mold .
The use of A . tumefaciens-mediated DNA transfer for insertional mutagenesis has advanced our understanding of the endemic dimorphic fungi at the molecular level [10] , [12] , [13] , [28] . We used this technology to mutate B . dermatitidis conidia and screen for transformants with altered morphology during growth at 22°C and 37°C . Analysis of mutant 3-15-1 uncovered a GATA transcription factor , SREB , which regulates siderophore biosynthesis and affects morphology in B . dermatitidis . GATA transcription factors are zinc-finger proteins that bind conserved motifs to induce or repress gene expression [16] , [22] , [26] , [27] , [29] . These genes are found widely in eukaryotes , but they function differently in fungi , plants , and animals [30] , [31] . In fungi , GATA transcription factors regulate diverse functions including the response to blue light , switching of mating-type , uptake of nitrogen , pseudohyphal growth during nitrogen starvation , biosynthesis of siderophores , and iron assimilation [17] , [22] , [29] , [32] , [33] . Our analysis indicates that SREB has pleotropic effects in B . dermatitidis - it promotes the transition from yeast to mold at environmental temperature and represses the biosynthesis of siderophores . Following a shift of incubation temperature from 37°C to 22°C , the insertional and null mutants were unable to complete the phase transition or accumulate significant biomass when compared to the parent strain . To our knowledge , B . dermatitidis SREB is the first gene identified in the dimorphic fungi that promotes the conversion of yeast to mold . Much of the field's attention has been focused on genes that regulate the phase transition from mold to yeast; only a few genes have been identified that regulate growth or morphology in the dimorphic fungi at environmental temperature ( i . e . 22–25°C ) . In H . capsulatum , the mold-specific gene MS8 regulates mycelial morphology and growth , but not the phase transition [34] . In P . marneffei TupA is required for maintenance of mycelial morphology at 25°C; null mutants convert to mycelia following a temperature shift from 37°C to 25°C , but revert to yeast morphology with prolonged incubation [35] . We hypothesize that B . dermatitidis SREB binds DNA to regulate many genes that , in turn , control such disparate functions as phase transition and the response to abiotic stress , including iron availability . Using MAST analysis we identified a large number of genes with putative GATA transcription factor binding sites . When compared to the entire B . dermatitidis genome , candidate genes involved with the biosynthesis of secondary metabolites as well as amino acid and lipid metabolism were found to be over-represented . Some of these candidate genes were indeed altered in expression in SREBΔ , as detected in preliminary microarray analysis and validated by RT-PCR . The enrichment of genes involved in secondary metabolism and amino acid metabolism were not unexpected , in part , because SREB regulates siderophore biosynthesis , a process that requires the transport and metabolism of amino acids . The abundance of genes containing GATA binding sites involved in lipid transport and metabolism was surprising . To our knowledge , regulation of lipid metabolism and transport in fungi by GATA transcription factors has not been described . Changes in fatty acid metabolism in the dimorphic fungi are associated with the phase transition and are postulated to impact morphogenesis [36]–[42] . Exposure of H . capsulatum mycelia to unsaturated fatty acids prolongs the mold-to-yeast conversion following a shift in temperature from 25 to 37°C [36] . In contrast , treatment with saturated fatty acids accelerates the phase transition [36] . In C . immitis , exposure to exogenous fatty acids alters the conversion of spherules to mycelia [37] . Reduced expression of the Δ9-desaturase gene , OLE1 , in C . albicans , impairs hyphal formation [38] . Differences in the concentration of unsaturated fatty acids ( oleic and linoleic acids ) and unsaturated sphingolipids ( N-2′-hydroxy- ( E ) -Δ3-octadecenoate ) have been described in the yeast and mold forms of H . capsulatum and P . brasiliensis [39]–[42] . In P . brasiliensis , several genes involved in lipid metabolism , have been demonstrated to be phase-regulated [43] . Thus , further investigation of genes involved in fatty acid metabolism may clarify the mechanism by which SREB promotes the phase transition from yeast to mold . B . dermatitidis insertional and null mutants have multiple alterations in the regulation of iron assimilation , as indicated by their yellow-orange appearance , constitutive production of siderophores , and derepression of iron-regulated genes during conditions of iron abundance . Iron acquisition must be tightly regulated for proper cellular function and to avoid toxicity due to iron overload [17] , [44] . Under iron-replete conditions , SREB represses genes involved in the production ( SIDA , AMCA ) and uptake ( MIRB , MIRC ) of siderophores . AMCA encodes a transferase that shuttles ornithine from the mitochondria to the cytosol [44] . The first step in siderophore biosynthesis involves the conversion of ornithine into N5-hydroxy-L-ornithine , which is catalyzed by an L-ornithine-N5-monooxygenase encoded by SIDA [45] . Siderophores secreted into the environment bind iron and then can be taken up by the cell through permeases such as MIRB and MIRC [46] . Analysis of the B . dermatitidis genome did not reveal an ortholog to A . nidulans MIRA , which facilitates the uptake of xenosiderophores , specifically enterobactin [46] . Deletion of SREB resulted in de-repression of SIDA , AMCA , MIRB , and MIRC expression under iron-replete conditions . Similar to P . chrysogenum SREP , N . crassa SRE , and A . nidulans SREA null mutants , disruption of SREB in B . dermatitidis resulted in discoloration of the fungus [14] , [18] , [45] . In addition , we identified two extracellular siderophores , dimerum acid and coprogen , produced by B . dermatitidis when grown under iron-poor conditions . When iron is abundant , SREB represses the biosynthesis of both these siderophores . Similar to A . nidulans and H . capsulatum , the expression of B . dermatitidis SREB is upregulated when iron is abundant , and repressed when iron is limited [16] , [17] . Repressors of siderophore biosynthesis are not uniformly regulated at the transcriptional level in other fungi , as orthologs of SREB including SRE , URBS1 , FEP1 , and SFU1 are constitutively expressed regardless of exogenous iron concentrations [18]–[21] . SREB is expressed as a single transcript , similar to SRE and URBS1 [18] , [19] . In contrast , SREP , SREA , and FEP1 are expressed as two separate transcripts due to the presence of two transcriptional start sites [14] , [17] , [20] . B . dermatitidis SREB may participate in a regulatory circuit with the bZIP ( basic leucine zipper ) transcription factor , HAPX . Computational analysis of the promoter region of HAPX in B . dermatitidis revealed putative GATA binding sites . Moreover , iron-poor conditions induced HAPX expression in wild-type B . dermatitidis , whereas iron abundance reduced its expression . In A . nidulans , HAPX represses SREA as well as genes that encode iron-dependent proteins such as CYCA ( cytochrome C ) , ACOA ( aconitase ) , LYSF ( homoaconitase ) when iron availability is limited [44] . We found that deletion of SREB resulted in the expression of HAPX under iron-poor and iron-replete conditions . Our findings support the idea that B . dermatitidis SREB functions as a transcription factor that regulates the biosynthesis of siderophores and promotes the conversion from yeast to mold . We propose that SREB inhibits genes involved with the biosynthesis and uptake of siderophores under conditions of iron abundance . Our findings also suggest that SREB affects phase transition independently of iron assimilation , perhaps , by altering the expression of genes involved with lipid metabolism or cell wall remodeling . The iron-related defects do not explain the failure to convert from yeast to mold since growth under iron-poor conditions had no effect on the defect in morphogenesis . GATA transcription factors in other fungi have been demonstrated to regulate morphogenesis as well as the response to temperature . S . cerevisiae ASH1 encodes a GATA transcription factor that inhibits mating-type switching and induces filamentous growth under conditions of nitrogen limitation [29] . C . neoformans CIR1 , an ortholog of B . dermatitidis SREB , regulates genes involved in reductive iron assimilation and siderophore transport , but also genes critical for virulence including those required for thermotolerance , capsule production , and melanin biosynthesis [22] . In summary , we identified and characterized a GATA transcription factor that represses the biosynthesis of siderophores and promotes the phase transition from yeast to mold . To our knowledge , B . dermatitidis SREB is the first gene identified in dimorphic fungi that promotes the conversion of yeast to mycelia . By using bioinformatic and expression analyses we identified several genes whose expression may be directly or indirectly regulated by SREB . We investigated a sample of these genes , including ones in KOG categories for lipid and carbohydrate metabolism , and found that their expression is affected by the deletion of SREB . Future work will strive for a more complete description of how SREB promotes the yeast to mold phase transition . Because growth in the mold form is thought to be essential for the survival of dimorphic fungi in nature and the generation of infectious particles , SREB may be needed for the evolutionary maintenance of this species . The generation of an SREB null mutant provides a unique opportunity to elucidate the SREB regulon and identify genes that govern growth in the mold form , as well as other traits in this human fungal pathogen .
Blastomyces dermatitidis strains used in this study included T53-19 and American Type Culture Collection ( ATCC ) 26199 . T53-19 sporulates , but is weakly virulent in a murine model of infection , and ATCC strain 26199 is highly virulent , but does not sporulate [10] , [47] . The genome of strain 26199 has been sequenced by the Genome Sequencing Center at Washington University ( http://genome . wustl . edu ) . B . dermatitidis yeast and mold were grown on Histoplasma macrophage medium ( HMM ) , 3M medium ( 3M ) , Potato dextrose agar ( PDA ) , or Middlebrook 7H10 agar medium containing oleic acid-albumin complex ( 7H10; Becton Dickinson and Company , Franklin Lakes , NJ ) [48]–[50] . Agrobacterium tumefaciens strain LBA1100 harboring the Ti helper plasmid pAL1100 ( gift from C . van den Hondel; Leiden University , The Netherlands ) was maintained on Luria-Bertani ( LB ) medium supplemented with 0 . 1% glucose , spectinomycin 100 µg/ml , and kanamycin 100 µg/ml once transformed with a binary vector [28] . Conidia from B . dermatitidis strain T53-19 were mutagenized using A . tumefaciens containing pBTS165 [10] , [28] , [51] . This binary vector contains a resistance cassette , hygromycin phosphotransferase ( hph ) , integrated into the T-DNA that is driven by a glyceraldehyde-3-phosphate dehydrogenase ( gpdA ) promoter derived from Aspergillus nidulans [10] . Conidia harvested from mycelial cultures by manual disruption were counted using a hemocytometer , suspended in phosphate buffered saline ( PBS ) to a final concentration of 2×107/ml , and co-cultivated with A . tumefaciens ( 6×108 cells/ml ) on a Biodyne A nylon membrane ( Pall Gelman , Ann Arbor , MI ) on induction medium containing 200 µM acetosyringone ( IMAS medium ) [28] . After 72 hours of incubation at 22°C , the biodyne membranes were transferred to 3M medium supplemented with hygromycin 100 µg/ml ( AG Scientific Inc . , San Diego , CA ) and cefotaxime 200 µM ( Sigma-Aldrich ) , and incubated at 37°C or 22°C . Individual transformants were visually screened by light microscopy for altered morphology: growth as hyphae or pseudohyphae at 37°C or yeast at 22°C . Replica plates were used to identify transformants that lost viability upon shifting the incubation temperature from 22°C to 37°C . Adaptor PCR was used to amplify DNA flanking the pBTS165 insert from insertional mutant 3-15-1 [52] . Following the digestion of genomic DNA by restriction enzymes StuI , HpaI , and XmnI , which do not cut in pBTS165 , adaptors were ligated to the restriction fragments using T4 DNA ligase ( New England Biolabs , Ipswich , MA ) . PCR was performed using primers specific for the adaptors and pBTS165 . The PCR products were separated by agarose gel electrophoresis and purified using the QIAquick gel extraction kit ( Qiagen , Valencia , CA ) and sequenced by the DNA Sequencing Laboratory at the University of Wisconsin Biotechnology Center . Sequence flanking the insert was analyzed using GSC ( Genome Sequencing Center ) BLAST ( http://genome . wustl . edu/tools/blast ) and National Center for Biotechnology Information ( NCBI ) tBLASTx ( http://blast . ncbi . nlm . nih . gov/Blast . cgi ) . FGENESH was used to identify predicted exons and introns in the SREB gene ( www . softberry . com ) . Two vectors , pBTS4-KO1 and pBTS4-KO2 , were used to delete SREB in B . dermatitidis strain 26199 by homologous recombination and resulted in two null mutants , T1#23 and T12#16 , respectively . Although both null mutants had similar phenotypes , T1#23 contained an additional 2 , 214 bp deletion in the 5′ untranslated region that was upstream of the disrupted SREB gene . Herein , T12#16 , which has no additional deletions , is referred to as SREBΔ . Plasmid pBTS4-KO2 contained 1611 bp of 5′ upstream sequence and 1747 bp of coding and 3′ downstream sequence flanking hph . The 1611 bp and 1747 bp products were amplified from B . dermatitidis 26199 genomic DNA using F and R primers containing SacI , BbsI , SbfI , or ClaI restriction sites ( F-1611-SacI 5′-TTTGAGCTCACTTTACTCTTCGGACGGGTTTT; R-1611-BbsI 5′-TTTTCGATTGTCTTCAGCCAAAAGCCCCGTCATTCCTGT; F-1747-SbfI 5′-TT-TCCTGCAGGTTGCAGCGTGAGGCGGAAGA; R-1747-ClaI 5′-TTTATCGATTGACAGGGCAG-GCTACATA ) . PCR products were separated by agarose gel electrophoresis , purified using QIAquick PCR purification kit ( Qiagen , Valencia , CA ) , sequenced , and ligated into pBTS4 in sequential fashion following restriction digest to flank the hph-resistance cassette [53] . After sequence and restriction digest analyses confirmed integration of the ligated PCR fragments , pBTS4-KO2 was electroporated into A . tumefaciens strain LBA1100 [28] . B . dermatitidis strain 26199 ( 2×107 yeast/ml ) was transformed with A . tumefaciens containing pBTS4-KO2 ( 6×108 bacteria/ml ) on Biodyne A membranes on IMAS medium . After 72 hours of incubation at 22°C , the Biodyne membranes were transferred to HMM medium supplemented with 10–20 µM FeSO4 , hygromycin 25 µg/ml , cefotaxime 200 µM , and incubated at 37°C . Transformants were visually screened for yellow pigmentation . The null mutant was cloned to obtain individual colonies and establish a line of cells . SREB gene deletion was confirmed by PCR , and Southern and Northern blot analyses ( see below ) . Insertional mutant 3-15-1 was re-transformed with pBTS47-11+13 using A . tumefaciens-mediated DNA transfer . This plasmid contained the SREB coding region , 1990 bp of 5′ sequence upstream of the start codon , 603 bp of 3′ sequence downstream of the stop codon , and a nourseothricin resistant cassette . Genomic DNA was amplified using primers ggp11-XbaI ( 5′-TTTCTAGAACAACTACCTCTACATGACACT-GC ) and ggp13-SbfI ( 5′-TTTCCTGCAGGGAGCCTTTTCTTTCTGTCAA ) . The PCR products were separated by agarose gel electrophoresis , purified using QIAquick PCR gel extraction kit ( Qiagen , Valencia , CA ) , sequenced , and ligated into pBTS47 to generate pBTS47-11+13 . The null mutant , SREBΔ , was re-transformed by A . tumefaciens with pBTS47-5331 , which contains the SREB coding region , 2655 bp of 5′ sequence upstream of the start codon , 603 bp of 3′ sequence downstream of the stop codon , and a nourseothricin resistant cassette . The protocol for A . tumefaciens-mediated DNA transfer was similar to that described in the previous section . Transformants were screened for white colony pigmentation on HMM medium supplemented with 20 µM FeSO4 , nourseothricin 25 ug/ml ( Werner Bioagents , Germany ) , and cefotaxime 200 µM at 37°C incubation . B . dermatitidis was grown to late log phase in liquid HMM at 37°C incubation . Genomic DNA was extracted using the method described by Hogan and Klein [54] . Southern blot hybridization was performed as described [28] , [55] . The fate of the transforming DNA in the insertional mutant was determined using probes specific for T-DNA and non-T-DNA sequences . An 822 bp amplicon constructed using primers 5′-CGATG-TAGGAGGGCGTGGATA and 5′-GCTTCTGCGGGCGATTTGTGT was used to probe hph within the T-DNA . An 8 kb BglII restriction fragment generated from pBTS4 was used to probe the non-T-DNA sequence . Deletion of SREB in the null mutant was analyzed using PCR-generated probes specific for the SREB coding region ( 1303 bp; 5′-CCCGCTCTTTGCTTAACC-CGTATG and 5′-CTGGTGATAAAGAAGGGCTGAA ) , hph ( 822 bp; 5′-CGATGTAGGAGGGCG-TGGATA and 5′-GCTTCTGCGGGCGATTTGTGT ) , 5′ region flanking SREB ( 1663 bp; 5′-ACTT-TACTCTTCGGACGGGTTTTC and TATCTGCGCTTTTGGTAGTAGGAG ) , and the 3′ region flanking SREB ( 1747 bp; 5′-TTGCAGCGTGAGGCGGAAGA and 5′-ACAAATCGTAGCACCAG-TC ) . All probes were radiolabeled with α-32P dCTP using a Prime-a-Gene labeling system ( Promega , Madison , WI ) . Unincorporated radionucleotides were removed using ProbeQuant G50-micro columns ( GE Healthcare , Buckinghamshire , UK ) . Following hybridization , the blot was washed sequentially with low stringency ( 0 . 25 M NaPO4 , 2% SDS , 1 mM EDTA ) and high stringency ( 0 . 04 M NaPO4 , 1% SDS , 1 mM EDTA ) solutions , exposed to a storage phosphor screen ( Molecular Dynamics , Sunnyvale , CA ) and scanned using a Storm 660 imaging system ( Molecular Dynamics , Sunnyvale , CA ) . Ferric perchlorate was used to measure siderophore production semi-quantitatively [25] . B . dermatitidis was grown at 37°C in liquid 3M or HMM under iron-poor or replete ( 10 µM FeSO4 ) conditions . Iron-poor media consisted of HMM or 3M prepared with F-12 Ham's nutrient mixture lacking FeSO4 , or trace elements lacking FeSO4 , respectively . In addition , exogenous iron was not added to these media . As the yeast entered stationary growth ( A600 = 3 . 5−4 . 0 ) , culture supernatants were collected , filtered ( 0 . 2 µM ) , and added to a ferric perchlorate solution ( 5 mM Fe ( ClO4 ) 3 in 0 . 1 N HCl ) . Absorbance was measured at 425 or 495 nm . Plasticware was used whenever possible . Glassware was treated with 2N HCl to remove residual traces of iron [56] . Analysis of variance ( ANOVA ) was used to analyze the results from the ferric perchlorate assay . Tukey's Honest Significant Difference method was used to adjust the p-values for multiple comparisons . B . dermatitidis was grown to mid-log phase at 37°C in liquid HMM with no added iron ( iron-poor medium ) , 10 µM FeSO4 , or 50 µM FeSO4 . Total RNA was extracted using the phenol-guanidinium thiocyanate-1-bromo-3-chloropropane extraction method [55] . In brief , yeast were washed with PBS , beaten with beads , and treated with TRI Reagent followed by 1-bromo-3-chloropropane ( Molecular Research Center Inc . , Cincinnati , OH ) . RNA was precipitated using a 1∶1 concentration of isopropanol and a high salt solution ( Molecular Research Center Inc . , Cincinnati , OH ) , washed with 75% ethanol , and resuspended in water that was pre-treated with diethyl pyrocarbonate ( DEPC; Calbiochem , San Diego , CA ) . Total RNA was further purified using RNeasy kit ( Qiagen , Valencia , CA ) and enriched for mRNA using oligo ( dT ) -polystrene chromatography ( Sigma-Aldrich ) . Northern hybridization was performed as described using 2 . 0-2 . 3 µg poly ( A ) +-enriched mRNA per sample [55] . Gene expression was analyzed using probes constructed by PCR against SREB ( SreF 5′-CCCGCTCTTTGCTTAACCCGTATG; SreR 5′-CTGGTGATAAAGAAGGGCTGAA ) SIDA ( SidA-F1 5′-AGACAGTACTCAAGAACGACAA; SidA-R1 5′-GCTGTCATCGCTGGGCTTTAGTGC ) , MIRB ( MirB-F 5′-CTCCTCCTCGTCGCTTTCGCACTA; MirB-R 5′-CCCTGAGGTCCCCGT-AGATGAG ) , MIRC ( MirC-F 5′-TGATGGCATTCTCAACCTCCC; MirC-R 5′-AACCTGCGGTGAT-GAAACCAC ) , AMCA ( AmcA-F 5′-GTCCGCATTACTCATCTG; AmcA-R 5′-CGCCTCATAAATC-GTAA ) , HAPX ( HapX-F 5′-CCGGTACCCCTCAAGCCCACAACT; HapX-R 5′-AAATACTTCAAC-ACGCCCATAACG ) , and actin ( Actin-F 5′-TCGGCCGTCCTCGCCATC and Actin-R 5′-TCCAG-ACTCGTCGTAGTCCTGC ) . Total RNA was extracted from B . dermatitidis wild-type and SREB null mutant strains grown in HMM at 37°C and 22°C in a similar fashion as described above; modifications included grinding cells frozen in liquid nitrogen in a mortar and pestle . Wild-type and SREB null mutant cells were grown for 48 hours at 22°C prior to RNA extraction . RNA , at 10 ug/sample , was treated with Turbo DNase ( Applied Biosystems/Ambion , Austin , Tx ) and further purified using RNeasy kit ( Qiagen , Valencia , CA ) . cDNA was generated from 1 ug of DNase-treated RNA using iScript cDNA synthesis kit ( Bio-Rad , Inc . , Hercules , CA ) . Real-time PCR reactions were comprised of 1x SSoFast EvagGreen supermix ( Bio-Rad ) , 0 . 5 mM of each primer , and 1 ul of 10-fold diluted cDNA template in a total volume of 10 ul . All reactions were performed in triplicate for two biological replicates . Real-time PCR was performed using a Bio-Rad iCycler MyiQ . Cycling conditions were 1 cycle at 95°C for 30 seconds followed by 40 cycles of 95°C for 5 seconds and 60°C for 10 seconds . Melting curve analysis was performed following the completion of the PCR . Gene expression was normalized relative to the expression of alpha-tubulin based on R ( relative expression ) = 2−ΔCt , ΔCt = Cttarget gene–Cttubulin [57] . Primers used to amplify transcripts from the following genes were: Lipid transfer protein ( BDBG_03618-1F 5′- CCATCAATGCTGCCATCAAC; BDBG_03618-1R 5′-GGTCTCACCCTTGTCGTTTG ) , glycosyl hydrolase ( BDBG 03183-1F 5′-GCTCTCCCAAGACATACATCAG , BDBG_03183-1R 5′-CCAT-AGCAAACTTCCCAAAAG ) , peroxisomal dehydratase ( BDBG_00052-1F 5′-CCCATTGTGCTA-ACCTTCAAG , BDBG_00052-1R 5′-AACTCCATCCGTCGCCTC ) , acetoacetyl-CoA synthase ( BDBG_09522-1F 5′-GCTCTCGGCACGCTCATAC , BDBG_09522-1R 5′-GGTGGTGACGG-GAGAAATG ) and alpha-tubulin ( BDBG_00020-2F 5′-GGTCACTACACCATCGGAAAG-3′ , BDBG_00020 2R 5′-CTGGAGGGACGAACAGTTG ) . The annotated genome and predicted proteome of B . dermatitidis strain SLH14081 was used for MAST analysis and KOG annotation . The genome of this strain ( 75 . 35 Mb; 9 , 555 genes ) has been sequenced and annotated by the Broad Institute ( www . broadinstitute . org/annotation/genome/blastomyces_dermatitidis/MultiHome . html and ACBT01000000 ) . The absence of annotation in the sequenced genome of 26199 precluded its use for computational analysis . MAST/MEME ( multiple em for motif elicitation ) software in unix ( version 4 . 2 . 0 ) was used to identify GATA transcription factor binding motifs in the genome of B . dermatitidis SLH14081 [58] . A fifth-order Markov background model was built for SLH14081 using the MEME utility fasta-get-markov . To find the location of previously identified motifs , MAST was run with a given motif frequency table , the Markov background model ( -bfile ) and options to produce text output as a ‘hit list’ ( –text –hit_list ) . For a search with the ATCwgAtaa motif [26] , a p-value of 0 . 0005 was used ( –mt 0 . 0005 ) . MAST output and Broad gene coordinates ( http://www . broadinstitute . org/annotation/genome/blastomyces_dermatitidis/MultiHome . html ) were parsed using a custom perl script to find intergenic motifs <2kb upstream of predicted genes . A total of 84 , 965 motifs were found in the genome assembly , of which 79 , 458 were in intergenic regions . Of these , 3 , 372 copies were found <2 kb upstream of 2 , 468 genes . Genes with the following motifs were retained: ATC- ( A/T ) -GATA- ( A/G ) , ATC- ( A/T ) -GATA- ( T/C ) , ATC- ( A/T ) -GATT-A , ATC- ( A/T ) -GATC-A , ATC-A-GATG-A , ATC-C-GATA-A , and ATC-A-AATA-A . These motifs are found upstream of genes regulated by A . fumigatus SREA , an SREB homolog [26] . To discover new motifs using MEME , we identified orthologs in SLH14081 of the iron-upregulated genes from A . fumigatus ( BDBG_00046 , BDBG_00047 , BDBG_00048 , BDBG_00050 , BDBG_00053 , BDBG_00054 , BDBG_00055 , BDBG_01314 , BDBG_02226 , BDBG_06775 , BDBG_06965 , BDBG_08034 , BDBG_08208 , BDBG_09322 ) and searched the 1 kb upstream for common motifs; MEME options were set for any number of motifs per region ( -mod anr ) , the above described Markov background model , and a minimum width of 6 ( -minw 6 ) . This identified a motif of vATCwGATAA , which is similar to the motif described by Hwang and colleagues [27] . For KOG annotation and analysis , the predicted proteome from B . dermatitidis strain SLH14081 was retrieved from the Broad institute ( http://www . broadinstitute . org/annotation/genome/blastomyces_dermatitidis/MultiDownloads . html , accessed: 11/09/2009 ) and compared against the NCBI KOG database ( ftp://ftp . ncbi . nih . gov/pub/mmdb/cdd/ , accessed: 11/09/2009 ) ) using RPSBLAST ( e-value 1e-05 ) [59] , [60] . Two data sets were generated with the first containing all B . dermatitidis genes encoding proteins that registered a KOG annotation . The second set included B . dermatitidis proteins encoded by the candidate genes with upstream GATA sites . The KOGs for both sets were correlated to their associated categories , and the total number of proteins within each category was tabulated . A two-tailed Fisher's exact test was used to determine if the number of proteins in each category were over- or under-represented when compared to all KOG-annotated proteins in the B . dermatitidis proteome . Categories were considered over-represented if the p-value of the right of the Fisher's exact test was less than 0 . 05 and over-represented if the left tail was less than 0 . 05 . To isolate and identify siderophores produced by B . dermatitidis 26199 wild-type and null mutants , we used column chromatography , liquid chromatography/mass spectroscopy ( LC/MS ) , and reverse-phase high-pressure liquid chromatography ( HPLC ) . Supernatants were harvested from B . dermatitidis grown in liquid HMM at 37°C under iron-poor ( no added iron ) and iron-replete ( 10 µM FeSO4 ) conditions when the cultures entered stationary growth ( A600 = 3 . 5−4 . 0 ) . Culture supernatants were filtered ( 0 . 2 µM ) , treated with 2% ferric chloride and applied to a column ( K 9/30 , GE Healthcare ) packed with Amberlite XAD-2 resin ( Supelco , Bellefonte , PA ) . The resin and column were prepared according to the manufacturer's recommendations . Following a water wash ( 7 bed volumes; flow rate of 0 . 2 ml/min ) , siderophores were eluted from the resin using methanol ( 1 . 7 bed volume; flow rate of 0 . 1 ml/min ) , reduced to dryness , and re-suspended in water ( 100 µl ) . Colorless supernatants that contained siderophores developed an orange color when treated with ferric chloride . This allowed for visual assessment of binding and elution of siderophores from the resin [61] , [62] . The Mass Spectroscopy Facility at the University of Wisconsin Biotechnology Center performed LC/MS analysis of concentrated eluate collected from wild-type B . dermatitidis grown under iron-poor conditions following XAD-2 column chromatography . For HPLC , siderophores were separated on a C18 column ( Agilent Eclipse XDB-C18 column; 4 . 6×150 mm ) using a water-acetonitrile gradient containing 0 . 1% trifluoroacetic acid ( Sigma-Aldrich ) . The gradient of acetonitrile was increased from 5% to 15% over 15 minutes , and 15% to 25% over 35 minutes . The flow rate was 0 . 5 ml/min and the absorbance was measured at 465 nm . Retention times were compared to siderophore standards ( HPLC calibration kit – coprogens and fusarinines; EMC microcollections , Tubingen , Germany ) . The nucleotide sequences for SREB , SIDA , AMCA , MIRB , MIRC , and HAPX from B . dermatitidis strain 26199 were obtained from the Genome Sequencing Center , Washington University , Saint Louis , MO ( http://genome . wustl . edu/tools/blast ) . Although this genome is publically available , it is not annotated . Allelic sequences can be found at the Broad Institute ( http://www . broadinstitute . org/annotation/genome/blastomyces_dermatitidis/MultiHome . html ) and have the following gene locus identification numbers: SREB ( BDBG_01059 ) , SIDA ( BDBG_00053 ) , AMCA ( BDBG_00128 ) , MIRB ( BDBG_05798 ) , MIRC ( BDBG_08034 ) , HAPX ( BDBG_01314 ) . Additional gene locus numbers include: lipid transfer protein ( BDBG_03618 ) , glycosyl hydrolase ( BDBG_03183 ) , peroxisomal dehydratase ( BDBG_00052 ) , acetoacetyl-CoA synthase ( BDBG_09522 ) , and alpha-tubulin ( BDBG_00020 ) .
|
The dimorphic fungi are the most common cause of invasive fungal disease worldwide . In the soil , these fungi grow as mold that produce infectious spores; when inhaled into the warmer lungs of a mammalian host , the spores convert into yeast , which cause infection . The change in shape between mold and yeast is a crucial event in the lifecycle of these fungi . The molecular regulation of this morphologic switch , or phase transition , is poorly understood . The goal of our research was to identify and characterize novel gene ( s ) that govern the phase transition in dimorphic fungi using Blastomyces dermatitidis as a model organism . Using insertional mutagenesis , we identified a gene , SREB , which encodes a transcription factor that affects phase transition and regulates the production of iron-gathering molecules or siderophores . When SREB is deleted , B . dermatitidis fails to complete the conversion from yeast to mold , grows poorly at environmental temperature , has yellow-orange colony pigmentation , and cannot properly repress the biosynthesis of siderophores . We also identified two types of siderophores produced by B . dermatitidis . To our knowledge , SREB is the first gene identified that promotes the conversion from yeast to mold , a process important for survival in the environment and generation of infectious spores .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"infectious",
"diseases/fungal",
"infections",
"microbiology/medical",
"microbiology"
] |
2010
|
SREB, a GATA Transcription Factor That Directs Disparate Fates in Blastomyces dermatitidis Including Morphogenesis and Siderophore Biosynthesis
|
A biophysical model that captures molecular homeostatic control of ions at the perisynaptic cradle ( PsC ) is of fundamental importance for understanding the interplay between astroglial and neuronal compartments . In this paper , we develop a multi-compartmental mathematical model which proposes a novel mechanism whereby the flow of cations in thin processes is restricted due to negatively charged membrane lipids which result in the formation of deep potential wells near the dipole heads . These wells restrict the flow of cations to “hopping” between adjacent wells as they transverse the process , and this surface retention of cations will be shown to give rise to the formation of potassium ( K+ ) and sodium ( Na+ ) microdomains at the PsC . We further propose that a K+ microdomain formed at the PsC , provides the driving force for the return of K+ to the extracellular space for uptake by the neurone , thereby preventing K+ undershoot . A slow decay of Na+ was also observed in our simulation after a period of glutamate stimulation which is in strong agreement with experimental observations . The pathological implications of microdomain formation during neuronal excitation are also discussed .
Astroglia determine the architecture of neural tissue and maintain central nervous system ( CNS ) homeostasis [1–3] . Astrocytes are organised into functional syncytia that show anatomical specialisation [4 , 5] , which allow intercellular diffusion of ions , second messengers and metabolites . Astroglial membranes closely enwrap the majority of excitatory synapses in the CNS , forming astroglial cradles [6 , 7]; a structure which facilitates synaptogenesis , synaptic maturation , synaptic transmission and synaptic extinction . Astroglial membranes are densely packed with transporters and ion pumps that maintain molecular homeostasis in the synaptic cleft and in the brain interstitium [8–11] . Furthermore , astrocytes maintain the homeostasis of many neurotransmitters and neuromodulators and supply neurones with glutamine , an essential precursor for the synthesis of glutamate and gamma-Aminobutric acid ( GABA ) , the main excitatory and inhibitory neurotransmitters respectively [12–15] . K+ homeostasis is a canonical function of astroglia proposed in the mid-1960s; both energy dependent Na+/K+ATPase ( NKA ) and passive ( inward rectifier K+ channels ) pathways were considered as molecular mechanisms [9 , 16 , 17] . Subsequently the Na+/K+/Cl- transporter NKCC1 was suggested to participate in K+ buffering , especially at high ( pathological ) K+ concentrations [10 , 18 , 19] . The local K+ uptake is supposedly supported by spatial K+ buffering ( K+ diffusion through gap junctions from regions of elevated [K+] to regions of lower [K+] ) . Under physiological conditions , however , the main pathway for K+ influx is associated with NKA , whereas Kir4 . 1 inward rectifying channels mediate K+ efflux which is needed to restore K+ gradients in neuronal compartments [10 , 18 , 19] . These observations are consistent with astrocytic K+ being re-released via Kir4 . 1 channels at distal synapses after distribution in the astrocytic functional syncytium via gap junctions [18] . However , in our paper we are dealing with K+ microdomains at the PsC , due to a low conductance pathway between the PsC and astrocyte soma , causing a significant increase in [K+]PsC which returns to baseline level via K+ leak and Kir4 . 1 channels post neural excitation . Astroglial homeostatic function is , to a large extent , controlled by transmembrane Na+ gradients and is regulated by cytoplasmic Na+ signals [20 , 21] . Dynamic fluctuations of [Na+]i affect Na+-dependent transporters and associated molecular cascades that link Na+ dynamics to homeostasis of K+ and neurotransmitters [20 , 21] . Astroglial Na+ signals are spatially heterogeneous with the existence of Na+ microdomains; these Na+ signals may also propagate through astroglial syncytia via gap junctions [22] . The main pathway for astroglial Na+ entry occurs through the excitatory amino acid transporters 1 and 2 ( EAAT 1/2 ) [15] . Glutamate transport is powered by transmembrane ion gradients where 3 Na+ and 1 H+ ions are exchanged for one K+ ion , hence the Na+/glutamate transporter generates a net inward Na+ current [23 , 24] . In addition , glutamate opens astroglial ionotropic receptors ( AMPA and NMDA receptors ) and activates ( indirectly through store-operated Ca+ influx pathway ) TRPC channels which further contribute to stimulus-dependent Na+ entry [25–28] . Glutamate transporters co-localise with sodium-calcium exchanger ( NCX ) that couple Na+ and Ca2+ signalling [29 , 30] . The existence of astroglial ionic microdomains [21 , 31 , 32] at the PsC indicates that there must be mechanisms that slow ionic diffusion along the processes . A candidate for this mechanism is associated with fixed negative charges existing in cell membranes . Previous experiments [33] had found localised fixed negative charges in the membranes of neurones and glia . Biological membranes consist of a continuous bilayer of lipid molecules in which membrane proteins are embedded . These bilayers are filled with polar and non-polar portions in their structure ( amphipathic ) [34] . All eukaryotic membranes are also asymmetric such that biophysical properties differ between the intracellular and extracellular surfaces; this asymmetry is necessary for many key cellular processes including cell fusion and cell clearance [35] . As astrocyte processes are very thin we hypothesise that this effect will have a dominant role in restricting cation conduction along the membrane . Specifically , we propose that cation retention in the potential wells requires that cations must hop from well to well as they move along the thin astrocyte process and therefore , this hopping effect serves to semi-isolate the astrocytic perisynaptic cradle ( PsC ) from the astrocytic main body . In this paper we show , for the first time , that this slow leakage of cations could explain the formation of K+ and Na+ microdomains at the PsC which points to a new theory for K+ clearance . For example , it is widely reported [9 , 36] that excitatory presynaptic neurones release K+ into the extracellular space ( ECS ) which is subsequently cleared at the PsC and buffered away through diffusion to the main astrocyte body . However , we propose that the flow of K+ away from the PsC is not volume diffusion limited , but rather is restricted due to the hopping effect along the process , and therefore a K+ microdomain forms at the PsC during sustained presynaptic neuronal excitation . Furthermore , we will show through mathematical modelling , that the formation of a K+ microdomain at the PsC may very well be advantageous as it facilitates a low energy return pathway for K+ to the ECS , after neuronal excitation has ceased . Additionally , cation retention along thin processes may also affect homeostasis for Na+ ions as they also carry a positive charge . It has been shown in other work [23] that the decay rate of Na+ following a sustained level of glutamate uptake ( in the absence of extra-cellular K+ change ) through EAAT1/2 is in the order of seconds . We will also show in this paper that this could potentially result from the restricted flow rate of Na+ ions along thin processes which results in the formation of a Na+ microdomain at the PsC that can only be removed by the NKA , reversal of NCX and other transporters .
Both sides of the bilayer phospholipid membrane surface are negatively charged , however , it has been shown that the distribution of charge is non-uniform at the atomic scale [49] . This non-uniformity gives rise to a potential profile including deep potential wells that can trap ions close to the membrane . Basic quantum mechanics show that regardless of the depth of a well it will have at least one state that can trap the travelling ions . The main effect of the depth of the well is the energy needed for the ion to continue its motion . This means that the ions cannot move easily along the membrane and must “hop” from well to well [49 , 50] . Charge hopping transport has been extensively studied in dielectric and semiconductor materials . The present case is particularly complex because of the presence of mobile cations in the cytoplasm , causing the formation of an electrical double layer near the interface between the cytoplasm and the membrane . The negative fixed charge present on the membrane is due to phosphatidylserine ( PS ) which is composed of phosphatidic acid ( PA ) , with the negatively charged phosphate group attached to the amino acid serine at the hydroxyl end . The associated negative charge will cause cations to move towards the membrane creating a dynamic space charge distribution in the cytoplasm as cations move from one well to an adjacent one; that is , there will be a net movement of cation ions due to the electric field along the length of the process . This leads to a complex and dynamic 3D potential distribution , which is beyond the scope of this paper . However , in order to investigate the interplay between the retention of cations along the process length and the formations of microdomain at the PsC we draw on a well-established charge hopping model for electrons in dielectrics containing Coulombic trapping centres [51] . In our case , the membrane fixed charge and associated potential wells are deemed analogous to Coulombic trapping centres . The current flow IiPF through the thin process is then represented as: IiPF=KiVA−Vm−Vrlexp[−Qi ( φw−Qi ( VA−Vm−Vr ) lπϵ ) kBT]CSAP ( 18 ) where K is a constant which represents mobility and concentration of mobile ions , Vm is the resting membrane potential of the astrocyte , φw is the well activation energy or potential barrier to ion flow . It has been shown that φw lies typically in the range ( 1–20 ) × kBT [51] . and is initially taken as 10 kBT but the effect of changing φw will be considered later; l is the length of the process , Q is the charge on a single ion taken as the charge on an electron , T is the absolute temperature , CSAP is the cross-sectional area of the process , ϵ is the dynamic permittivity and is given by ϵ = ϵ0 ϵr , where ϵ0 is the absolute permittivity and ϵr is the relative permittivity of the cytoplasm , and kB is the Boltzmann constant . Note that the potential across the length of the process is assumed linear in this formulation; that is the lateral electric field is constant . The square root term in the argument of the exponential term of Eq 16 represents the field-dependent lowering of the activation energy , φw . The ‘trapping centres’ are assumed to be spaced relatively widely such that their potential distributions do not overlap . The concentrations of K+ and Na+ in the astrocyte soma are held constant but will be continuously changing at the PsC thus establishing a dynamic concentration gradient associated with these cations . Consequently , we formulate a Nernst-like reversal potential for Na+ and K+ between the astrocyte soma ( AS ) and the PsC as: Vr=RTFln ( [i]AS[i]PsC ) ( 19 ) where i is the ion under consideration . A schematic of the hopping concept is shown in Fig 3 . The diffusion of K+ between the PsECS and the GECS is modelled as a simple gradient controlled channel and is given by: IKECSL=gECSEECSSAECSL ( 20 ) where gECS is the conductance of the channel , SAECSL is the surface area between the PsECS and the GECS , and EECS is the Nernst like potential of the channel given by: EECS=RTFln ( [K+]PsECS[K+]GECS ) ( 21 ) The neuronal model utilised in the work consists of the biophysical Hodgkin and Huxley ( HH ) type model ( described in supplementary material S1 Text ) [55] with the addition of NKA . All parameter values for the NKA can be found in Table 4 . Neurone potassium channel ( KNeu ) . The HH model simulates current flow of K+ through a voltage gated channel , therefore the current flow of K+ from the neurone can be modelled as: IKNeu=−gKNeun4 ( VNeu−EKNeu ) SASyn ( 22 ) where gKNeu is the maximum K+ channel conductance , EKNeu is the reversal potential of the potassium channel , VNeu is the membrane voltage of the neurone and SAsyn is the surface area of the synapse ( neurone parameters described in supplementary material S1 Table ) . Neurone sodium potassium pump ( NKANeu ) . Similar to the astrocytic NKA , the neuronal NKA exchanges intracellular Na+ for extracellular K+ against the gradient of both ions and has a stoichiometry of 3:2 . The K+ current component of the pump is given by [46]: IKNKANeu=2FPNKAmaxNeu[Na+]Syn1 . 5[Na+]Syn1 . 5+KNaiNeu1 . 5[K+]PsECS[K+]ECS+KKENeuSASyn ( 23 ) where PNKAmaxNeu is the NKA maximum pump rate , KNaiNeu is the Na+ threshold of the pump , and KKENeu is the K+ threshold of the pump . Since the neural model does not contain all ion channels necessary for homeostasis , Na+ changes due to the neuronal NKA pump both inside and outside the synapse are not taken into account . The value of PNKAmaxNeu was chosen in such a way that IKNKANeu = 0 at steady state .
To explore how K+ retention in the astrocyte process gives rise to a K+ microdomain at the PsC and eliminates K+ undershoot , several simulations were carried out with the presynaptic neurone stimulated using external currents to produce firing rates of 20Hz , 40Hz , 60Hz and 80Hz . These firing rates are all within physiological frequencies of most cortical pyramidal neurones and fast spiking neurones . The neural stimulus has a duration of ~1 minute where the first 0 . 1 minute allows the model to reach a steady state condition and the stimulus ceases after 1min . Although this is a long period of time , it allowed an investigation into how extracellular and intracellular ionic concentrations would be affected during a sustained period of neural activity . For each simulation , PsECS [Glu] was held constant at the background level . Fig 4 describes [K+] and [Na+] dynamics for each of the 4 different stimuli where it can be seen that neuronal release of K+ into the PsECS leads to an increase in the astrocyte membrane voltage ( VA in Fig 4A ) because of the change in ionic currents through the PsC membrane . It can also be seen that K+ steadily increases within the PsECS ( [K+]PsECS in Fig 4B ) and after a period of ~0 . 8 minutes it approaches steady state at higher frequencies where the release rate of K+ by the presynaptic neurone equates to the clearance rate by NKA and KB on both the PsC and the presynaptic terminal , and also K+ lost into the GECS . It is also worth noting that as the concentration of K+ increase in the PsC ( [K+]PsC in Fig 4C ) , the Na+ concentration with the PsC decrease due to efflux by NKA at the PsC ( [Na+]PsC in Fig 4D ) . Note: the astrocyte membrane voltage VA , [K+]PsECS and [K+]PsC all increase with the presynaptic neurone firing rate while [Na+]PsC decreases . During neural activity , the NKA and Kir channel currents are responsible for K+ uptake while the background K+ and KPF currents release K+ from the PsC . These currents can be seen in Fig 5 where Fig 5B shows that , contrary to the current thinking [56] , the NKA is the dominant driving force for K+ uptake while Kir channel ( Fig 5A ) is much less so for K+ clearance: furthermore , clearance by Kir diminishes over time because the changes in the associated reversal potential due to the [K+]PsC microdomain . Fig 5C shows that IKPF is several orders of magnitude lower than IKir and therefore this slow leakage of K+ away from the PsC appears to be a plausible explanation for the emergence of a K+ microdomain . Note the saturation and subsequent fall off of IKB at higher frequencies is a direct result of the K+ background reversal potential approaching VA . This is caused by the rapid build-up of K+ in the PsECS and cradle . The high frequency oscillatory behaviour which appears as a thickening of Fig 5A–5D is due to the astrocytic response to the pulsed nature of presynaptic neuronal K+ release . As the potassium in the PsECS fluctuates so does the astrocyte NKA pump and to a lesser extent the astrocyte membrane voltage . These fluctuations in the NKA and membrane voltage are also reflected in Na+ and K+ currents . Inserts in Fig 5A–5D , column 1 , are used to show detail of astrocyte K+ current dynamics in response to neurone K+ release . Note: for clarity only the first column shows this detail as the dynamics for each current is similar for each of the stimulus frequencies . When the neurone stops releasing K+ ( ~ 1min ) it quickly flows from the PsECS into the ECS which reduces the K+ gradient between the PsECS and PsC thereby reducing NKA pump rate , after which a net efflux of K+ takes place from the stored K+ in the associated microdomain . This points to a new theory whereby K+ microdomain formation during neuronal excitation ( due to ion retention in the astrocyte process ) provides the driver for the return of K+ to the PsECS , via background K+ leak and Kir4 . 1 channels , for uptake by the neurone . Fig 6A shows the net transfer of K+ across the perisynaptic membrane while Fig 6B shows the net current flow along the process ( out of the perisynaptic cradle ) . During stimulation ( 0 . 1 min to 1min ) it can be seen that there is a net transfer of K+ into the perisynaptic cradle across the membrane ( Fig 6A ) . Since the current flowing along the process to the soma ( Fig 6B ) is 3 orders of magnitude smaller than the currents entering the cradle , there is a net build-up of K+: essentially a K+ microdomain forms because of the low conductance pathway from the cradle to the astrocyte soma . Furthermore , this microdomain allows the efflux of K+ from the PsC into the PsECS after neurone stimulation ceases . This can be seen as a spike like current in Fig 6A after 1min and is more pronounced in the 80Hz simulation . Fig 7 shows the Na+ currents for the four different stimulus frequencies . All Na+ channels , except the NKA ( Fig 7B ) Na+ current , result in Na+ influx to the PsC . When the neurone stops firing there is a net influx of Na+ into the PsC . The decrease in INaB ( Fig 7A ) can be explained as follows: Since INaB is dependent on the astrocyte membrane potential as well as Na+ gradient there is a sharp decrease in the current due to the astrocyte membrane potential depolarising . As well as K+ buffering , astrocytes also provide a critical role in glutamate uptake and recycling via the glutamate-glutamine cycle ( GGC ) [57] . In this simulation , the role of glutamate transport via EAAT1/2 is investigated and results show that the slow leakage of Na+ ions in the astrocyte process causes Na+ to increase in the PsC before being returned to the PsECS via the NKA . These results support previously published experimental work [23]: there is no neuronal excitation and therefore the concentration of K+ in the PsECS is held constant . The concentration of glutamate in the PsECS was modulated using a Gaussian function as shown in Fig 8A . Fig 8B–8D presents the results of the PsC ionic [K+]PsC and [Na+]PsC concentrations and membrane voltage , VA , for this simulation . From Fig 8C and 8D we clearly see that the [K+]PsC decreases while [Na+]PsC increases , this is the opposite dynamics to that observed in Fig 4C and 4D . This is because K+ in the PsECS is now held constant at 3 mM and therefore all K+ channels except the NKA and slow leakage through the astrocyte process remove K+ from the PsC ( Fig 9 ) resulting in a net K+ efflux . The main driving force behind Na+ uptake by the PsC is the EAAT1/2 transporter which is also responsible for the removal of glutamate from the PsECS ( Fig 10 ) . During [Glu]PsECS injection , the EAAT1/2 and Kir release K+ at an accelerated rate . This is opposed by NKA and the transport of K+ from the astrocyte soma to the PsC . When glutamate falls to baseline levels , the EAAT1/2 and Kir channels quickly revert to their initial rates . NKA and transport of K+ from the astrocyte soma is then able to establish baseline ionic concentrations at the PsC . As in the previous simulation , retention of Na+ ions as they flow within the astrocyte process substantially limits the transport rate of these ions away from the PsC . In this case , Na+ is restricted and therefore a Na+ microdomain forms at the PsC . Note: similar to the results presented in [23] there is a long decay ( ~80s ) transient of Na+ which far outlasts the glutamate signal decrease ( Fig 8D ) and we propose that this is due to the slow removal of Na+ by the NKA . These observations could explain previously observed experimental results [23] . The previous two simulations have shown that K+ or Na+ microdomains form in the PsC when the system is stimulated with PsECS changes in K+ or Glu respectively . However , while these simulations show that our hypothesis could potentially explain experimental observations , we now wish to use our model to predict ionic dynamics at the PsC under physiological conditions where both K+ and Glu are released at the presynaptic terminal . In this case K+ is released by the neurone as before and a 100 μM puff of Glu is released into the PsECS , with each spike event . Presynaptic neurone firing rates are 20Hz , 40Hz , 60Hz and 80Hz , for a period of 0 . 1min to 1min . The results presented in Fig 11 show that the overall behaviour of the model , i . e . microdomain formation of K+ in the PsC , occurs . However , the astrocyte membrane voltage VA oscillates ( ~7mV amplitude ) ( Fig 11A ) caused by the periodic reversal of the Kir channel ( See Fig 12A ) . This reversal is caused by the efflux of K+ via the EAAT1/2 ( Fig 12E ) channel . Moreover , the dynamic behaviour of the reversal potential of the Kir and VA continuously cause reversal of the overall polarity ( Fig 13 ) , thus causing the Kir channel to periodically reverse direction resulting in an efflux of K+ into the ECS; this can be seen as oscillations in [K+]PsECS . It can be observed in Fig 11C and 11D that a K+ microdomain is formed in the PsC and its magnitude increases with frequency while the magnitude of Na+ reduces . This is due to the behaviour of the K+ uptake by NKA dominating over the K+ efflux pathways ( See Fig 12 ) . Fig 14 shows the Na+ currents for the four different stimulus frequencies . As expected all Na+ channels on the PsC membrane , except the NKA ( Fig 14B ) result in Na+ influx to the PsC . INaEAAT has a large peak amplitude for a short duration ( few milliseconds ) due to the EAAT channel slowing down after removal of Glu from PsECS . Having analysed the formation of microdomains and model behaviour in the previous three simulations we now explore the sensitivity of the model to model parameters . These parameters are PsC surface area , the maximum NKA pump rate , Pmax , and the potential barrier to ion flow along the process , φw . In these simulations a neuronal firing rate of 40Hz was chosen . Microdomain Sensitivity to PsC Surface Area ( SA ) . Three different values of PsC SA were chosen for this simulation; PsC SA × 0 . 75 , PsC SA × 1 and PsC SA × 1 . 25 . The results of these simulations are shown in Fig 15 where it can clearly be seen that the amplitude of the K+ microdomain increased with PsC SA with a corresponding drop in the concentration of Na+ . Also , the K+ and Na+ currents efflux/influx also increased with PsC SA ( See Supplementary S1 Fig for the changes in K+ currents ) . Microdomain Sensitivity to Pmax . Four different values of Pmax were chosen for this simulation; Pmax × 0 . 2 , Pmax × 0 . 5 , Pmax × 1 and Pmax × 5 . The results of these simulations are shown in Fig 16 where it can clearly be seen that [K+]PsC and [Na+]PsC is strongly dependent on Pmax . Using the Pmax x 0 . 2 value causes [K+]PsC to decrease and [Na+]PsC to increases and as Pmax increases , [K+]PsC begins to form a microdomain with [Na+]PsC steadily decreasing . From these simulations we can conclude that when the NKA pump rate is low it is no longer the dominant co-transporter and both the EAAT co-transporter and Kir channel dictate [K+]PsC and [Na+]PsC dynamics . The opposite is true when the pump rate is large . Microdomain Sensitivity to φw . In this simulation φw was varied from 4 kBT to 15 kBT . Fig 17 shows the peak K+ current along the process for the different values of φw . As φw is decreased , the peak current along the process increases exponentially . Therefore , with decreasing φw the formation of a microdomain becomes less likely as IKPF , max is increasing and eventually IKPF , max approaches an electro-diffusion limited model with no likelihood of a microdomain forming at the PsC . From these simulations it is clear that the mechanism responsible for the formation of microdomains is the well formation along the process which effectively semi-isolate the PsC from the astrocyte soma when φw is 10 kBT or greater . It is also clear that the PsC SA can limit the maximum amplitude of the microdomain concentration . This is due to the increase/decrease of ion channel densities on the membrane of the PsC . Moreover , the NKA maximum pump rate also has an important role in the formation of microdomains whereby if the pump rate is low then K+ clearance by NKA weakens; effectively these to ion transporters compete to move K+ and Na+ ion across the membrane but in opposite directions .
Homeostatic control over synaptic cleft , of which K+ buffering and glutamate uptake is of the most fundamental importance , represent the quintessential function of astroglial cradle formed by the perisynaptic process [7 , 8] . Physiological K+ buffering is essentially K+ recycling between astroglial and neuronal compartments: NKA-dependent astroglial uptake limits the peak of extracellular K+ rise , whereas K+ efflux is imperative for restoration of [K+]i in neuronal terminal [3 , 10] . Glutamate uptake , as well as glutamate conversion into glutamine and glutamine shuttling to neuronal terminals are regulated by Na+ concentration in the cytosol of astroglia; astroglial [Na+]i in addition controls a multitude of SLC ( Solute Carrier ) transporters responsible for various homeostatic pathways [20 , 21] . Both mechanisms require localisation of ionic signalling within the confines of astroglial synaptic cradle , and indeed local and long-lasting [Na+]i transients are routinely recorded from astroglial processes [23 , 58] . Molecular machinery responsible for localisation of [K+]i and [Na+]i increases remains unknown . Here we table a novel mechanism of localisation of ionic signals in astroglial cells . We propose that ion retention within thin astrocyte processes can give rise to the formation of K+ and Na+ microdomains at the PsC . This localisation of astroglial ionic microdomains arises because in thin processes , surface conduction dominates over volume conduction , and because membrane lipids are negatively charged , deep potential wells form near the dipole heads restricting the flow of cations along the process . Therefore , cations must hop from well to well which restricts ion conduction along the membrane . This hopping effectively semi-isolates the PsC from the astrocytic main body allowing the formation of K+ and Na+ microdomains at the PsC under different conditions . We modelled ionic responses of the PsC to the neuronal excitation that results in an increase in K+ concentration in the synaptic cleft . For the simulation , glutamate in the cleft was held constant at the background level and , during neuronal excitation , K+ was released into the PsECS leading to a depolarisation in the PsC membrane voltage due to ionic currents flowing through the PsC membrane . The simulations demonstrate that a K+ microdomain formed at the PsC due to the restricted flow of these ions along the astrocyte process . We further contemplate that the K+ microdomain provides the driving force for the return of K+ to the PsECS via background K+ channels for uptake by the neurone via its NKA . Essentially , K+ is transiently “stored” at the PsC during neuronal excitation where it decreases the electrochemical gradient of K+ , so reducing inward flow of potassium through Kir; this “stored Potassium” is then available to replenish neuronal K+ levels when the excitation ceases , thereby preventing K+ undershoot in the extracellular space . These observations are consistent with in vivo experimental data [59] , and partly explain why inward rectifying K+ channels may play a prominent role for K+ uptake at large volume glial processes ( e . g . terminal endfeet of retinal Muller cells [60] but not at low volume perisynaptic cradles . These results will also necessitate a reappraisal of the mechanisms and role of astrocytes in potassium accumulation during seizure activity [61] , especially given the observation that loss of function mutations in the gene encoding Kir4 . 1 are associated with a human epilepsy syndrome [62] and astrocyte Kir4 . 1 expression is decreased in acquired epilepsy models [63] . Moreover , other conditions that have been proposed to be due to abnormalities of potassium homeostasis such as familial hemiplegic migraine are associated with mutations of the gene encoding the alpha2 subunit of NKA , which is predominantly expressed in astrocytes [64] . Our model also shows that the influx of Na+ ions into the astrocyte process causes a Na+ microdomain to form at the PsC where the decay rate of Na+ is governed by the NKA . In this simulation , there is no neuronal excitation and the concentration of glutamate in the cleft was modulated using a Gaussian function and was taken up at the PsC by EAAT1/2 . A slow decay of Na+ was observed after the glutamate uptake ceased which is in strong agreement with experimental observations [21 , 23] . In summary , we accept our model for ion retention in thin astrocyte processes requires much more refinement . For example , the main challenge would be to account for the dynamic interaction between the charge present in membrane proteins and the charged ions in the astrocyte medium . Additionally , the dimensions of the astrocyte are such that the membrane proteins are unlikely to be represented by point charges and a more atomistic view of the proteins would need to be found to create a map of the charge distribution at the atomic scale . Also , any simulations would require a large number of atoms to be taken into account to obtain the electrostatic potential profile at the membrane-cytoplasm interface . Once the electrostatic potential in thin process is found and combined with the cation distribution , then the movement of ions can be modelled . Moreover , we have only considered K+ and Na+ ions in our model and therefore a more biophysical model would need to consider Ca2+ microdomains and Cl- ion dynamics with the inclusion of the associated membrane transporters such as Sodium/Calcium exchanger and Sodium/Potassium/Chloride cotransporter . We have assumed throughout the model an infinite GECS but this in reality would not be the case . However more biological data about the shape and size of extracellular spaces and morphology of the perisynaptic cradle would be required before modelling in such a way . Despite this our model does however indicate that the morphological and biophysical properties of the astroglial perisynaptic processes facilitate emergence of Na+ and K+ microdomains that are essential for astroglial homeostatic support of synaptic transmission in the central nervous system , and point to new and important implications for potassium homeostasis during pathological activity such as seizures . For example , the long-held view that spatial potassium buffering plays an important role in seizure activity has been challenged by a number of experimental observations including only a small effect on potassium buffering of knocking out astrocytic gap junction proteins [65] , and an antiepileptic effect of gap junction blockers [66] . Indeed , it has been proposed that gap junctions are not necessary for potassium buffering but instead are important for maintaining neuronal metabolism during seizure activity [67] . Our model supports the experimental data and indicates that potassium buffering is a locally restricted phenomenon , and therefore our model challenges the orthodox view of the role of glial spatial potassium buffering during pathological activity . Indeed , our model supports the existence of a mechanism that prevents local potassium depletion during excessive neuronal firing and indicates a novel mechanism by which astrocytes maintain neuronal excitability during pathological activity . Finally , we would like to point out that ion hopping is a current transport mechanism that involves ions surmounting potential barriers . It is therefore strongly temperature dependent , following Arrhenius behaviour , and has a distinctive temperature and voltage dependence or 'signature' that enables identification against other possible current mechanisms; this may provide a means to test our hypothesis as the kinetics of ionic microdomains formed by Na+ , K+ or even Cl- in perisynaptic processes could be quantified in experimental brain slices using respective intracellular probes .
|
During periods of neuronal activity , ionic homeostasis in the surrounding extracellular space ( ECS ) is disturbed . To provide a healthy environment for continued neuronal function , excess ions such as potassium must be buffered away from the ECS; a vital supportive role provided by astrocyte cells . It has long been thought that astrocytes not only removed ions from the ECS but also transport them to other areas of the brain where their concentrations are lower . However , while our computational model simulations agree that astrocytes do remove these ions from the ECS they also show that these ions are mainly stored locally at the PsC to be returned to the ECS , thus restoring ionic homeostasis . Furthermore , we detail in this paper that this happens because of a previously overlooked biophysical phenomenon that is only dominant in thin astrocyte processes . The flow of these cations within thin processes is primarily by surface conduction where they experience the attraction of fixed negative charge at the membrane inner surface . This negative charge constrains cation movement along the surface and so their flow rate is restricted . Consequently , ions such as potassium that are released during neuronal excitation enter the PsC and are stored locally due to the low conductance pathway between the PsC and the astrocyte soma . Our simulations also show that this local build-up of K+ is returned to the ECS after the neuronal activity dies off which could potentially explain why K+ undershoot has not been observed; this result agrees with experimental observations . Moreover , the same mechanism can also explain the transient behaviour of Na+ ions whereby in thin processes a slow decay time constant is experimentally observed . These findings have important implications for the role of astrocytes in regulating neuronal excitability under physiological and pathological conditions , and therefore highlight the significance of the work presented in this paper .
|
[
"Abstract",
"Introduction",
"Models",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"neurochemistry",
"nervous",
"system",
"astrocytes",
"membrane",
"potential",
"ions",
"electrophysiology",
"neuroscience",
"macroglial",
"cells",
"sodium",
"neurotransmitters",
"cations",
"physical",
"chemistry",
"animal",
"cells",
"chemistry",
"glial",
"cells",
"glutamate",
"biochemistry",
"cellular",
"neuroscience",
"cell",
"biology",
"anatomy",
"synapses",
"physiology",
"neurons",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"physical",
"sciences",
"neurophysiology",
"chemical",
"elements"
] |
2018
|
Potassium and sodium microdomains in thin astroglial processes: A computational model study
|
Slow-cycling subpopulations exist in bacteria , yeast , and mammalian systems . In the case of cancer , slow-cycling subpopulations have been proposed to give rise to drug resistance . However , the origin of slow-cycling human cells is poorly studied , in large part due to lack of markers to identify these rare cells . Slow-cycling cells pass through a noncycling period marked by low CDK2 activity and high p21 levels . Here , we use this knowledge to isolate these naturally slow-cycling cells from a heterogeneous population and perform RNA sequencing to delineate the transcriptome underlying the slow-cycling state . We show that cellular stress responses—the p53 transcriptional response and the integrated stress response ( ISR ) —are the most salient causes of spontaneous entry into the slow-cycling state . Finally , we show that cells’ ability to enter the slow-cycling state enhances their survival in stressful conditions . Thus , the slow-cycling state is hardwired to stress responses to promote cellular survival in unpredictable environments .
From an evolutionary perspective , individuals that give rise to the highest number of progeny are considered to be the fittest . However , proliferation rate is often highly variable , even in a genetically identical population in optimal growth conditions . This heterogeneity is marked by the presence of a small population of slow-cycling cells observed in bacteria [1] , yeast [2] , and human cells [3–5] . In unicellular organisms such as bacteria and yeast , this heterogeneity in proliferation rate has been proposed to serve as a bet-hedging mechanism , in which the slow-cycling subpopulation can be better suited to tolerate harsh conditions , giving rise to increased fitness in a variable environment in the long term [6] . The long-term benefit allows the heterogeneity itself to be selected as a conserved trait . In slow-cycling cells , the relatively long time between two cell division events is often attributed to a prolonged noncycling state that precedes commitment to the cell cycle [7–9] . In multiple primary , immortalized but non-transformed , and cancerous human cells , we previously reported that a majority of cells commits to another cell cycle soon after mitosis , marked by increasing CDK2 activity and hyperphosphorylated Rb ( CDK2inc cells ) , while a separate subpopulation enters a transient G0/quiescence marked by low CDK2 activity , hypophosphorylated Rb , and declining Ki67 levels ( CDK2low cells ) [10–12] . These cells can later re-enter the cell cycle by increasing CDK2 activity , indicating the reversibility of this state [10 , 13] . We refer to this unprompted entry into the CDK2low state as “spontaneous G0” or “spontaneous quiescence , ” to contrast with canonical quiescent states , in which cells are forced into quiescence by serum starvation or contact inhibition [14] . We and others have recently shown that most CDK2low cells express high levels of a cyclin-dependent kinase ( CDK ) inhibitor , p21 , which is the dominant cause of entry into the spontaneous CDK2low state [15–17] . About 50% of these CDK2low cells harbor low levels of endogenous DNA damage , marked by the presence of 53BP1 nuclear bodies and γH2AX foci , which triggers p53-mediated up-regulation of p21 [15 , 16] . However , the trigger for entry into the CDK2low slow-cycling state in the other 50% of CDK2low cells remains unknown . In this study , we isolate spontaneous CDK2low cells from normally cycling cells and characterize the transcriptome of this subpopulation of slow-cycling cells . We first show that the slow-cycling state is a long-lived but reversible state . Transcriptomic analysis reveals a stress-response signature in the spontaneous CDK2low subpopulation , which is not present in cells forced into quiescence using four other methods . More specifically , we detect a strong signature of a p53 transcriptional program as well as activation of the integrated stress response ( ISR ) . Knockout of p53 or its transcriptional target p21 eliminates the spontaneous CDK2low subpopulation , giving rise to a more homogenous fast-cycling population . However , such cells that are unable to enter the spontaneous CDK2low state are quite vulnerable to exogenous stress , manifesting as a dramatic drop in fitness in stressful environments . Our data suggest that entering a CDK2low quiescence is an important mechanism to protect cells from stress .
To investigate the source of variation in cell-cycle length , we used an immortalized non-transformed human epithelial cell line , MCF10A , which has intact cell-cycle checkpoints . The intermitotic time ( IMT ) of individual MCF10A cells follows a distribution that peaks at 13 h with a long right tail ( Fig 1A ) , confirming the existence of a slow-cycling subpopulation . We further confirmed that the slow-cycling subpopulation cannot be explained by experimental settings such as expressing tagged histones or phototoxicity from imaging ( S1A–S1C Fig ) . The long IMT could either be due to entry into a long noncycling phase , such as G0 , or due to a lengthened proliferative cell cycle consisting of G1–S–G2–M . To distinguish between these two possibilities , we used MCF10A cells expressing DHB-mVenus , a live-cell sensor of CDK2 activity , to mark the time at which CDK2 activity first begins to rise , a molecular event that we have previously shown to coincide with Rb hyperphosphorylation and cell-cycle commitment at the Restriction Point ( R-point ) ( Fig 1B ) [10] . As previously reported , the majority of cells increase their CDK2 activity shortly after anaphase and immediately commit to another cell cycle ( Fig 1B , first cell cycle ) [10] . As a consequence , the IMTs of this CDK2inc subpopulation correlate well with time spent in G1–S–G2–M ( Fig 1C–1E and S1A Fig ) . This CDK2inc subpopulation also tends to have short IMTs and makes up the dominant mode in the IMT distribution in Fig 1A ( Fig 1C ) . The cells in the right tail of the distribution , however , tend to spend long periods in the CDK2low state ( Fig 1C and 1D ) . Time spent in the CDK2low state is the main contributor to the long IMT and explains 45% of the variation in IMT , in contrast to 3% explained by the length of G1–S–G2–M ( Fig 1C and 1D and S1D Fig ) . Taken together , our data suggest that slow-cycling cells ( those with long IMTs ) result from prolonged transits through the CDK2low state . To isolate and profile the CDK2low population , we took advantage of the previous finding that CDK2 activity strongly anticorrelates with p21 protein level ( Fig 1F ) [10 , 12 , 13] , which enables us to separate the CDK2low population by fluorescence-activated cell sorting ( FACS ) based on high p21 levels . We used an MCF10A cell line in which an mCitrine-p21 fusion protein is expressed from the endogenous p21 locus [12] . The distribution shape of mCitrine-p21 intensity resembles that of IMT with a long right tail: the majority of cells have low to zero levels of mCitrine-p21 , similar to wild-type cells that do not express the mCitrine fusion protein , while a small subpopulation of cells have elevated p21 levels ( Fig 1G ) . We used FACS to collect two subpopulations , p21low and p21high , corresponding to CDK2inc and CDK2low cells , respectively ( Fig 1G ) . We validated that the sorted p21high subpopulation indeed expresses higher levels of p21 protein , and lower level of phospho-Rb ( S1E Fig ) . We then followed the two sorted subpopulations by live-cell imaging for three days , starting from one day after the sorting , and confirmed that the p21high subpopulation indeed proliferates more slowly than the p21low subpopulation ( Fig 1H ) , indicating a fitness disadvantage in optimal growth conditions . How stable is the p21high slow-cycling state ? We previously showed that cells that pass through the CDK2low state and then re-enter the cell cycle and divide are more likely to generate daughters that enter the CDK2low state after mitosis [10] . We have also shown that daughter cells entering the CDK2low state have mothers with longer IMTs [15] . Together , these results suggest that the slow-cycling state is heritable beyond a single generation . To determine the longevity of the CDK2low/p21high state , we sorted the p21high and p21low subpopulations and remeasured the distribution of p21 intensity of the two subpopulations every few days . Strikingly , the two subpopulations take about 3 wk to re-establish the steady-state p21 distribution ( Fig 1I ) . To test whether both interconversion of the two subpopulations and a higher proliferation rate of p21low cells contribute to re-establishing the steady-state p21 distribution , we labelled the mCitrine-p21 cell line with either H2B-mTurquoise or H2B-mCherry , and sorted them into p21high and p21low subpopulations , respectively ( Fig 1J , left ) . The labelling of the two subpopulations allowed us to coculture them and monitor the population dynamics . If no outgrowth of the faster cycling p21low cells occurs , the two colors should remain at the initial mixing fraction . However , we found that the fraction of H2B-mCherry , initially p21low cells , gradually increased , consistent with fast-cycling p21low cells outcompeting slow-cycling p21high cells ( Fig 1J , right ) . The fractions of two colors reached a steady state after 3 wk , a similar time frame to that of re-establishing the steady-state p21 distribution ( Fig 1J , right ) . If no interconversion between CDK2low/p21high and CDK2inc/p21low cells occurs , we would expect that the H2B-mTurquoise , initially p21high cells , will eventually be excluded from the population . In contrast to this scenario , at an initial mixing ratio of 1:1 , the H2B-mTurquoise , initially p21high cells , reached a steady-state fraction of 20% of the population ( Fig 1J , right ) , suggesting that interconversion occurred . Moreover , this 20% of the population cannot be explained by a small fraction of p21low cells in the collected p21high subpopulation , given the sorting purity estimated by remeasuring the mCitrine-p21 intensity right after the initial sort ( 99 . 8% and 99 . 1% for p21low and p21high subpopulations , respectively ) . Therefore , convergence to the steady-state p21 distribution is partially driven by interconversion of p21high cells and p21low cells , and partially by outgrowth of faster-cycling p21low cells over slower-cycling p21high cells . These data suggest that although the p21high state is reversible , it is a relatively long-lived state at the population level . We sought to characterize the CDK2low/p21high slow-cycling state by performing RNA sequencing ( RNA-seq ) on the sorted subpopulations described above . To minimize the bias from cell-cycle phase differences , we sorted only G0/G1 cells using the geminin sensor of the fluorescence ubiquitination cell cycle indicator ( FUCCI ) [18] . The geminin sensor is absent in G0/G1 cells and accumulates from S phase to mitosis [18] . We confirmed that the intensity of the geminin sensor shows a bimodal distribution , and that the geminin-negative cells almost exclusively have 2N DNA content ( S1F and S1G Fig ) . We collected only the geminin-negative cells , and sorted these into p21high and p21low subpopulations ( S1G Fig ) . To reduce clonal effects , we carried out the sorting followed by RNA-seq in two different clones of mCitrine-p21 cells , one clone with one wild-type CDKN1A allele and one mCitrine knock-in allele ( clone 2e2 ) , and the other clone with one mCitrine knock-in allele and one knock-out allele ( clone 3b6 ) [12] . Despite the copy number difference of the gene , the two clones express similar levels of p21 at the single-cell level ( S1H Fig ) . The genes that are up-regulated in the CDK2low/p21high population are most highly enriched in transcriptional targets of p53 ( Table 1 , S1 and S2 Tables ) , consistent with the fact that p53 is a transcription factor for p21 [19 , 20] . The up-regulation of p53 target genes ( and p53 protein itself , S2A and S2B Fig ) is also consistent with the observation that about half of the CDK2low subpopulation has 53BP1 nuclear bodies and γH2AX foci , markers of DNA lesions ( S2C Fig ) [15 , 16] . Additionally , the mitophagy and autophagy-lysosome pathways are up-regulated in the CDK2low/p21high cells ( Table 1 ) , a result we validated with the observation of up-regulated PINK1 and LAMP1 proteins by western blot ( S2A and S2B Fig ) . This result is consistent with previous findings that these pathways protect quiescent cells from oxidative stress and mitochondrial dysfunction , and have been suggested to be important for the maintenance of quiescence [14 , 21] . We also observed that CDK2low/p21high cells show increased expression of extracellular matrix ( ECM ) receptors , such as E-cadherin and integrin , and ECM components such as collagen , fibronectin , and laminin ( Table 1 , S1 and S2 Tables ) . This observation suggests that CDK2low/p21high cells remodel and interact with their extracellular environment more extensively than CDK2inc/p21low cells . Similarly , many receptors on the plasma membrane and secreted signaling ligands are up-regulated in CDK2low/p21high cells , including the NOTCH family , WNT5A , TGF-β receptors , and BMP family ( Table 1 , S1 Table ) , suggesting increased cell–cell communication in these cells . Interestingly , a few markers of cancer and normal tissue stem cells are up-regulated in the CDK2low/p21high cells ( S1 Table ) , which often function to protect stem cells from environmental insults . Examples include elevated levels of aldehyde dehydrogenase ADLH1A3 to detoxify drugs [22 , 23]; increased expression of FOXO1 , a transcription factor implicated in the regulation of an anti-reactive oxygen species gene expression program [24]; and up-regulation of ATP-binding cassette ( ABC ) transporters that are capable of drug efflux [25] . These shared molecular features between slow-cycling cells and cancer stem cells suggest that slow cycling may be an intrinsic property of cancer stem cells that protects them from drug treatment . The pathways that are most significantly down-regulated in CDK2low/p21high cells relative to CDK2inc/p21low cells are DNA replication and cell cycle ( for example , the minichromosome maintenance protein complex [MCM] complex and the DNA replisome complex ) ( Table 2 , S1 and S2 Tables ) . Although both sorted subpopulations were geminin negative and therefore in G0/G1 at the time of sorting and sequencing , the data suggest that the p21low subpopulation was preparing for DNA replication and cell-cycle progression , whereas the p21high subpopulation appears less ready to complete the cell cycle . Consistent with our previous finding that the CDK2low/p21high cells are not senescent [13 , 15] , the cellular senescence pathway is down-regulated in these cells . Four major DNA repair pathways ( mismatch repair , nucleotide excision repair , base excision repair , and homologous recombination ) are down-regulated in CDK2low/p21high cells , an initially surprising finding given that these cells have been shown to have increased levels of DNA damage markers relative to CDK2inc/p21low cells ( S2C Fig ) [15 , 16] . A closer inspection of the gene lists reveals that the down-regulated genes mostly function not in DNA damage detection ( e . g . , p53 target genes ) but in actual repair of DNA damage , and often require S-phase entry for their expression ( e . g . , DNA polymerases and ligases ) ( S1 Table ) [26] . The down-regulation of these pathways in CDK2low/p21high cells therefore likely results from a lack of preparation for S phase in CDK2low/p21high cells relative to CDK2inc/p21low cells . To compare the transcriptome of the spontaneously quiescent CDK2low cells with more canonical quiescent states , we also performed RNA-seq on cells forced into quiescence for 48 h using traditional ( serum starvation , contact inhibition ) [14] and somewhat less traditional methods ( Mek inhibition , CDK4/6 inhibition ) ( S3A Fig ) . We confirmed that these perturbations forced cells out of the cycling state , marked by reduced mRNA levels of proliferation markers , such as Ki67 and Cyclin B1 ( S1 Table ) , and absence of phospho-Rb and 5-ethynyl deoxyuridine ( EdU ) incorporation ( a marker for DNA synthesis ) ( S3B Fig ) . We note that under contact inhibition and CDK4/6 inhibition , a small fraction of cells are phospho-Rbhigh and EdU+ , likely explaining the lack of significance of some cell-cycle gene down-regulation , especially genes that are lowly expressed in MCF10A cells , such as CCNE1 and E2F1 . We also verified that these perturbations did not force cells into senescence , because the cells can readily revert back to proliferation within 24 h of restoring full-growth conditions , with a similar fraction of phospho-Rbhigh and EdU+ as control cells ( S3B Fig ) . Because these quiescent states are induced by different pathways , we reasoned that genes that are differentially regulated in these various forms of quiescence will likely reveal the causes of quiescence , whilst genes that are commonly regulated in all forms of quiescence may indicate consequences of quiescence . For example , we recently showed that the bifurcation of Ki67 levels follows , and therefore is a consequence of , the proliferation-quiescence decision [11] , which is consistent with Ki67 being down-regulated in all five forms of quiescence examined here ( S1 Table ) . The distribution of expression fold-change immediately shows that different quiescence-induction methods perturb cells to dramatically different degrees , with serum starvation causing the widest range of differential expression , followed by Mek inhibition , contact inhibition , and CDK4/6 inhibition , with spontaneous quiescence showing the least differential expression relative to cycling cells ( Fig 2A ) . This is consistent with the fact that at one extreme , serum starvation will lead to loss of mitogenic signaling from many receptor signaling pathways , whereas CDK4/6 inhibition is a highly specific perturbation , and spontaneously quiescent CDK2low cells are not even artificially perturbed . Principal component analysis ( PCA ) of transcriptome data from the spontaneously quiescent CDK2low state and the four forced-quiescence states revealed five distinct clusters that correlated with the quiescence-induction method ( Fig 2B ) , although we note there is clonal variation in the positioning of the clusters ( S4A Fig ) . The PCA analysis therefore suggests that the five different triggers of quiescence lead to five transcriptomically distinct states of quiescence , at least at the 48 h time point examined here . Consistent with a previous study [14] , these data suggest that quiescence is not a single state . The first principle component separates serum starvation and Mek inhibition from other samples ( Fig 2B and S4A Fig ) , indicating that the dominant feature separating the five types of quiescence is the mitogen-activated protein kinase ( MAPK ) pathway , which accounts for 35%–49% of the variance ( Fig 2B and S4A Fig ) . We also note that p21high and p21low cells are not far separated in the PCA , consistent with their relatively small transcriptional differences compared with that between control and forced-quiescence samples ( Fig 2A and 2B ) . Seventy genes are transcriptionally up-regulated in all five forms of quiescence ( Fig 2C , S4B Fig and S3 Table ) . However , no particular signaling pathway is enriched in this group of genes , and the only enriched biological process in Gene Ontology ( GO ) is “epithelial cell differentiation” . Given that MCF10A cells originate from epithelium , this result suggests that genes universally turned on in all quiescence conditions may be constrained by the cell identity . Consistent with this interpretation , this list of genes overlaps poorly with a previously published “fibroblast quiescence program” , which includes 116 genes consistently up-regulated in three quiescence conditions in fibroblast cells [14] , with only one overlap between the two lists ( S4 Table ) . Among the 128 genes that are universally downregulated in all forms of quiescence , the enriched pathways are DNA Replication , Spliceosome and Cell Cycle , part of the “fibroblast quiescence program” in a previous study ( Fig 2C , S4B Fig and S3 Table ) [14] . Highly enriched in these genes are targets of the dimerization partner , RB-like , E2F4 , and multi-vulval class B ( DREAM ) complex , which includes transcriptional targets of E2F1 , 2 , 3 in the early cell cycle and targets of the MuvB complex with B-Myb or FOXM1 ( MMB-FOXM1 ) in the late cell cycle [27] ( S5 Table ) . In quiescence , the DREAM complex binds to the promoter of its targets and suppresses their expression . Once cells commit to the cell cycle , the DREAM complex disassembles and liberates the promoter for E2F- and , later , MMB-dependent transcription [28] . Notably , expression of DREAM targets is repressed in all forms of quiescence examined here ( Fig 2D ) , suggesting that a DREAM-dependent suppression program is active and that it is likely a consequence of quiescence . To test this notion experimentally , we considered that a logical requirement of differential gene expression being a consequence of quiescence is that the differential expression occurs after the cell’s decision to enter quiescence . We therefore sought to validate our RNA-seq results using single-cell time-lapse microscopy , followed by RNA fluorescence in situ hybridization ( RNA FISH ) or immunofluorescence ( IF ) , as in Gookin and colleagues [13] . By tracking hundreds of asynchronously cycling cells , we can populate the trajectory of the mRNA or protein of interest and reproduce the dynamics of that gene’s expression in single cells throughout the cell cycle [10 , 13] . We considered four DREAM targets , E2F1 , CCNE1 , CCNA2 , and CCNB1 , as well as PGK1 as a negative control , because it is not a DREAM target [27] ( Fig 3 and S5 Fig ) . To determine the relative timing of the proliferation-quiescence decision and the bifurcation of DREAM-target mRNAs , we first confirmed previous reports describing two subpopulations of cells with distinct levels of phosphorylated Rb that are visible immediately after anaphase [10 , 12 , 13] ( Fig 3A ) . The majority of cells ( about 80% ) have high levels of phospho-Rb , corresponding to the fraction that will become CDK2inc cells ( Fig 3G ) . A smaller subset of cells ( about 20% ) have low levels of phospho-Rb , corresponding to the fraction that will become CDK2low cells . These two populations are already discernible in late anaphase ( Fig 3A ) [12] . By contrast , the mRNA levels of E2F targets , E2F1 and CCNE1 , are low in the vast majority of cells at anaphase ( Fig 3B , 3C and 3G ) . From there , they begin to rise in CDK2inc cells , peaking at about 5 h after anaphase , whereas they remain low in CDK2low cells ( Fig 3B and 3C ) . Thus , CDK2inc cells have already committed to proliferation at anaphase ( as marked by high levels of phosphorylated Rb ) , but only up-regulate E2F1 and CCNE1 a few hours later . To further demonstrate this , we stained for phospho-Rb and E2F1 or CCNE1 mRNA in the same cells and identify a subpopulation of phospho-Rbhigh/E2F1low/CCNE1lowcells , indicating that Rb hyperphosphorylation occurs prior to E2F and CCNE1 transcription ( Fig 3H ) . mRNA levels of CCNA2 and CCNB1 , DREAM targets that function later in the cell cycle , diverge even later in the CDK2inc versus CDK2low subpopulations , at 5–7 h after anaphase ( Fig 3D and 3E ) . Together , these results confirm that differential expression of DREAM targets between the CDK2inc and CDK2low subpopulations occurs after the proliferation-quiescence decision and therefore is a consequence of the proliferation-quiescence decision . We next examined the RNA-seq data to find causes of spontaneous entry into the CDK2low state from genes uniquely up- or down-regulated in cells in this state . There are 287 genes up-regulated and 168 genes down-regulated in the spontaneous CDK2low cells that are not differentially expressed relative to other forms of quiescence ( Fig 2C and S4B Fig ) . The only enriched pathway within this set of genes is the p53 signaling pathway ( Fig 2C and S4B Fig ) , suggesting that up-regulation of p53 signaling is a cause of spontaneous quiescence entry . We first confirmed that p53 protein levels are higher in p21high/CDK2low cells versus p21low/CDK2inc cells and that p53 transcriptional targets are strongly induced in p21high cells but not in other forms of quiescence ( Figs 2D and 4A , S2A and S2B Fig ) . As previously reported , knockout of p53 or p21 eliminates the spontaneous CDK2low subpopulation ( Fig 4B and 4C ) [10 , 15–17] . Conversely , increasing p53 protein levels in G2 by adding Nutlin ( an inhibitor of MDM2 , a ubiquitin ligase responsible for p53 degradation ) can force wild-type cells into a CDK2low state after mitosis ( Fig 4B and 4C ) [16 , 17] . Therefore , activation of p53 is necessary and sufficient for entering the spontaneous CDK2low state . Activation of the p53 pathway is consistent with the finding that CDK2low cells harbor increased DNA damage relative to CDK2inc cells [15 , 16] . However , only approximately half of the cells in the CDK2low state show signs of DNA damage [15 , 16] , raising the question of whether other stresses may be involved in promoting the remainder of the transits through the CDK2low state . We reasoned that the specific stress may vary from cell to cell , but that stress signaling can converge at stress hubs such as p53 and emerge as a subpopulation phenotype . We therefore examined a second stress-response pathway that integrates multiple stresses , the ISR-eIF2α pathway . Four upstream kinases , GCN2 , PERK , HRI , and PKR , which sense amino acid deprivation , endoplasmic reticulum ( ER ) stress , heme deprivation , and viral infection , respectively , phosphorylate eIF2α , a critical component of translation initiation . Phosphorylated eIF2α impairs general cap-dependent translation , but enhances translation of specific downstream effectors of the ISR , such as ATF4 , a transcription factor for a group of stress-responsive genes [30] . We first attempted to examine the overall expression level of ATF4 target genes in our dataset . However , unlike p53 , for which chromatin immunoprecipitation sequencing ( ChIP-Seq ) and global run-on sequencing ( GRO-Seq ) studies have systematically discovered direct transcriptional targets [27 , 31] , no such dataset is available for ATF4 . We therefore constructed from the literature a list of six ATF4 transcriptional targets that are expressed in MCF10A cells [32 , 33] . We found that five of these six genes are up-regulated in CDK2low/p21high cells , whereas they are regulated in a less consistent way in other forms of quiescence , suggesting activation of ATF4 in CDK2low/p21high cells but not in other forms of quiescence ( S6A Fig ) . ATF4 protein was below the detection limit of our antibody in unperturbed conditions ( S6B Fig ) . We therefore examined phosphorylation of eIF2α , the upstream inducer of ATF4 . Consistent with the RNA-seq result , we detected much stronger phosphorylation of eIF2α in CDK2low/p21high cells relative to CDK2inc/p21low cells ( Fig 4D ) . Taken together , these data indicate that the ISR-eIF2α pathway is activated in CDK2low cells . We next examined whether inhibition of eIF2α dephosphorylation would lead to an increase in the fraction of cells entering the CDK2low state . We imaged cells expressing the CDK2 activity sensor in unperturbed conditions for 16 h , followed by treatment with salubrinal , an inhibitor of the eIF2α phosphatase Gadd34/PP1 [34] , and further imaging for another 24 h . Treatment with salubrinal caused an increased fraction of cells to enter into a CDK2low state after mitosis ( Fig 4E ) , suggesting that activation of the ISR-eIF2α pathway can force cells into a CDK2low quiescence . What is the relationship between ISR-activated and DNA-damaged CDK2low/p21high cells ? Can ISR activation account for the cells transiting through the CDK2low state that do not show signs of DNA damage ? Answering these questions requires a single-cell readout of the ISR; however , the ATF4 and phospho-eIF2α antibody signals are not detectable by IF in unperturbed cells . Instead , we measured the global translation rate in single cells using the O-propargyl-puromycin ( OPP ) assay , in which an alkyne analog of puromycin , OPP , is incorporated into nascent peptides and can later be visualized by fluorescence microscopy using a copper-catalyzed alkyne-azide cycloaddition ( CuAAC ) reaction with a fluorescent azide [35] . To compare OPP incorporation between CDK2low and CDK2inc cells , we imaged cells for 24 h by time-lapse microscopy to identify each cell’s CDK2 activity state , pulsed cells for 24 min with OPP , and immediately fixed the cells with paraformaldehyde . We then matched each fixed cell back to its live-cell trace , thereby linking its CDK2 activity state to its rate of protein translation . We found that CDK2low cells have a significantly lower translation rate than CDK2inc cells ( Fig 4F , top ) , consistent with their activation of the ISR . In contrast to translation rate , we detect no difference in transcription rate between CDK2low cells and CDK2inc cells ( Fig 4F , bottom , and S6C Fig ) , using a 5-ethynyl uridine ( EU ) incorporation assay similar to the OPP incorporation assay [36] . This suggests that the reduced translation rate in CDK2low cells is likely a specific signaling event , consistent with activation of the ISR in these cells . We therefore used OPP incorporation as single-cell readout of the ISR to assess overlap in the activation of DNA damage and ISR stress pathways . We co-stained cells with proliferation/quiescence markers phospho-Rb/p21 , DNA damage marker 53BP1 , and ISR marker OPP . Consistent with our observation in CDK2low cells ( Fig 4F , top ) , phospho-Rblow/p21high cells show reduced levels of OPP incorporation compared with phospho-Rbhigh/p21low cells . This reduction of translation is apparent in phospho-Rblow/p21high subpopulations with and without 53BP1 nuclear bodies ( Fig 4G ) , indicating that activation of the ISR is independent of DNA damage . Therefore , activation of the ISR could account for some of the cells entering the CDK2low state without DNA damage . However , we have not yet found a way to eliminate the ISR-associated slow-cycling state ( including via small interfering RNA [siRNA] knockdown of the four eIF2α kinases , S6D and S6E Fig ) , which would be necessary to show that endogenous activation of the ISR indeed causes entry into the CDK2low state . The essence of a stress response is to protect cells from stress . We hypothesized that cells in the spontaneous CDK2low state would be more resistant to stress than their proliferating CDK2inc counterparts . We initially sought to sort p21high/CDK2low cells from p21low/CDK2inc cells and test their resistance to exogenous stresses . However , applying exogenous stress rapidly converts many ( 17%–92% , depending on the stress ) CDK2inc cells into CDK2low cells that may activate the same stress response as the spontaneous CDK2low cells and hence obscure the result . We therefore used live-cell imaging to follow the fate of cells after applying exogenous stresses and computationally sorted the population into three categories: ( 1 ) cells that are CDK2low at the time of stress addition; ( 2 ) those that are CDK2inc at the time of stress addition but convert to CDK2low in response to the stress , either by dividing into CDK2low daughters or by dropping their CDK2 activity without a mitosis; and ( 3 ) those that are CDK2inc at the time of the stress and remain CDK2inc until their death or the end of the imaging period . Remarkably , in all three stress conditions , genotoxic stress , oxidative stress , or unfolded protein stress , the cells that persist in the CDK2inc state have significantly worse survival than cells existing in or converting to the CDK2low state ( Fig 5A–5C ) , suggesting that the CDK2low state indeed enables cell survival under stressful conditions . Thus , while CDK2low cells have a fitness disadvantage under optimal growth conditions ( Fig 1H ) , they have a fitness advantage under stressful conditions ( Fig 5A–5C ) . We further hypothesized that if entering quiescence is an essential part of the stress response , loss of the ability to enter quiescence should render cells particularly susceptible to exogenous stresses . We tested this idea using p21−/− MCF10A cells , which are not able to enter the spontaneous CDK2low state but should remain capable of activating other stress response pathways [10] . In unperturbed conditions , cell death occasionally occurred at a rate of 1 . 5% ± 0 . 2% per 50 h in p21−/− cells , but never occurred in wild-type cells during the 50-h imaging period ( Fig 5D , S1 and S2 Movies ) . Upon challenge by genotoxic stress , oxidative stress , or unfolded protein stress , p21−/− cells displayed substantially worse survival than wild-type cells ( Fig 5E–5G , S3 and S4 Movies ) . This result indicates that p21−/− cells are less tolerant to a wide range of stresses and that the ability to enter a quiescent CDK2low state in response to stress represents a survival advantage in stressful conditions .
Slow-cycling subpopulations have been implicated in antibiotic resistance in bacteria and drug resistance in cancer [3–5] . In bacteria , slow cycling has been proposed as a bet-hedging strategy , in which phenotypic variation improves long-term fitness by increasing survival through unpredictable environments , a beneficial trait that would have been selected by evolution . In the case of bet hedging , slow or fast cycling is a stochastic decision for a cell . In this work , we have investigated the origins of a slow-cycling subpopulation in a noncancerous human cell line and found that it largely results from a prolonged transit through a CDK2low quiescent state . We further show that the CDK2low state originates from activation of stress-response pathways , a deterministic molecular cause . In addition to DNA damage-induced up-regulation of p21 , p21high cells also show activation of the ISR . Although it has been reported that p21 can be activated by the ISR via transcription- and translation-based mechanisms under exogenous stresses [37–39] , precise determination of the connection between p21 and the ISR under spontaneous/endogenous stress conditions will require development of more sensitive detection methods for ATF4 and phospho-eIF2α in unperturbed single cells . While ISR activation appears to be independent of DNA damage as marked by 53BP1 nuclear bodies , it is not independent of p53 , because p53 is required for entry into the spontaneous CDK2low state ( Fig 4B and 4C ) [16 , 17] . The discrepancy between 53BP1 nuclear bodies ( DNA damage ) and p53 implies that p53 senses stresses beyond DNA damage . This is supported by the findings that p53 can be activated by stresses independent of DNA damage , such as nucleolar stress , hypoxia , and oncogene activation [40] . Thus , our current results implicate p53 and p21 as downstream effectors of stress signaling beyond their canonical role in response to DNA damage . While genetic knockout of p53 or p21 eliminates this CDK2low subpopulation , we have not found a way to reduce entry into the spontaneous CDK2low state by reducing potential stresses . Therefore , it remains unproven whether cellular stress explains all transits through the spontaneous CDK2low state . It is possible that each type of stress only accounts for a small fraction of cells entering the spontaneous CDK2low state . Alternatively , cell stress may be intrinsically associated with the stochastic nature of biochemical reactions in cells ( e . g . , DNA replication , transcription , and metabolism ) [41 , 42] and therefore may be impossible to eliminate . A third interpretation is that spontaneous CDK2low cells hijack stress-response pathways to enter a noncycling state in the absence of stress . Both the p53 pathway and the eIF2α pathway have been implicated in maintaining stem cell quiescence in mice as a physiological function [43 , 44] . In this case , cells may simply activate these pathways without stress in response to extrinsic cues , or in a stochastic manner , as a means to trigger quiescence . The bet-hedging theory relies on the fact that slow-cycling micro-organisms are more resistant to harsh environments [1] . In the case of cancer cells , an intuitive interpretation is that dormant cancer cells escape treatment by avoiding S phase , which is often the target of chemotherapy and radiotherapy [45] . Our findings show that the ability to enter spontaneous quiescence is beneficial not only in genotoxic stress conditions but also in diverse other stress conditions , suggesting that quiescence actively protects cells from exogenous stress . Together with our finding that activation of stress responses causes entry into spontaneous quiescence , these results indicate that the nature of spontaneous quiescence is a stress response to various endogenous stresses , given that these cells are in optimal growth conditions without exogenously added stress . Our observation that blocking entry into spontaneous quiescence ( by p21 knockout ) compromises the protective function of the stress response suggests that the slow-cycling state is not a passive consequence of stress , but an essential and inseparable part of the stress response that fortifies cells against various stresses . The stress response is perhaps one of the most conserved biological functions across all kingdoms of life . Our data suggest that reduced proliferation is hardwired into the stress response in human cells . With similar findings in bacteria and yeast [1 , 8] , we propose that this hardwiring of a slow-cycling state is as conserved as the stress response itself , and that reduced proliferation is a core feature of cellular adaptation for survival through changing environments and on a longer timescale , through evolution .
MCF10A cells were maintained in DMEM/F12 ( Gibco , Waltham , MA ) , supplemented with 5% horse serum ( Gibco ) , 20 ng/mL epidermal growth factor ( [EGF] Sigma-Aldrich ) , 0 . 5 mg/mL hydrocortisone ( Sigma-Aldrich , St . Louis , MO ) , 100 ng/mL cholera toxin ( Sigma-Aldrich ) , 10 μg/mL insulin ( Thermo Fisher , Waltham , MA ) , and 1× penicillin/streptomycin . For serum starvation media , the horse serum , EGF , and insulin were removed , and 0 . 3% BSA was added . For live-cell time-lapse imaging , phenol red–free DMEM/F12 ( Gibco ) was used . To prepare forced quiescent samples for RNA-seq or validation experiments , cells were seeded at a density of 200 , 000 cells per well in a six-well dish , or 2 , 000 cells per well in a 96-well dish . Cell number was increased 2-fold for serum starvation , Meki , and CDK4/6i conditions and by 10-fold for the contact inhibition condition . All forced quiescence treatments lasted 48 h . To test the reversibility of the quiescence , cells were released from the treatment into full-growth media for 24 h . Wild-type and p21−/− MCF10A H2B-mTurquoise DHB-Venus cells were described by Spencer and colleagues [10] . MCF10A mCitrine-p21 knock-in cells were described by Moser and colleagues [12] . MCF10A p53−/− cells ( clone H2PC13 ) were described by Weiss and colleagues [46] , and H2B-mTurquoise and DHB-mVenus were expressed in these cells by means of lentiviral transduction . Lentivirus was used to express Geminin1–110-mCherry , H2B-mCherry , or H2B-mTurquoise in MCF10A-mCitrine-p21 cells . Geminin is used as a marker for the start of S phase [18] . FACS was used to sort positive cells . The following inhibitors and chemicals were used: CDK4/6 inhibitor , palbociclib ( S1116 , Selleckchem , Houston , TX ) at 1 μM; Mek inhibitor , PD-0325091 ( S1036 , Selleckchem ) at 100 nM; bortezomib ( 10008822 , Cayman Chemical , Ann Arbor , MI ) at 1 μM; nutlin-3 ( 10004372 , Cayman Chemical ) ; salubrinal ( 14735 , Cayman Chemical ) ; ganetespib ( STA-9090 , Selleckchem ) ; gemcitabine ( G6423 , Sigma ) ; and hydrogen peroxide ( 5240–05 , Maron ) , at indicated concentrations . Five biological replicates of the Gemininlow/p21high and Gemininlow/p21low subpopulations were collected for each clone by FACS . Two biological replicates of the forced-quiescence cells were collected for each clone after treatment for 48 h . In each sample , a cell pellet of 100 , 000 cells was snap-frozen and stored at −80 °C before library construction . Total RNA was extracted from each sample using Quick-RNA MicroPrep kit from Zymo ( Irvine , CA ) . The Lexogen RiboCop rRNA Depletion Kit was used for ribosomal RNA removal and the Lexogen SENSE total RNA kit was used to prepare paired-end , stranded libraries . The libraries were pooled at equal amounts using the concentrations measured on a Qubit Fluorometer ( Thermo Fisher ) and sequenced on an Illumina NextSeq sequencer with 1% PhiX spike-in . Overall sequence quality was examined using FastQC version 0 . 11 . 2 . Adaptor sequences were clipped using Trimmomatic version 0 . 32 [47] . Reads were mapped to the human genome ( GRCh38 ) using HISAT2 version 2 . 0 . 3-beta [48] . Mapped fragments were counted using GenomicAlignments package in R [49] . Differential expression analysis was carried out using DESeq2 version 1 . 10 . 1 [50] , with a negative binomial-generalized linear model including batch and cell clone factor in the design . Genes with an adjusted p-value less than 0 . 05 were considered to be differentially expressed . GO analysis was carried out using g:Profiler [51] , with significance threshold at p < 0 . 01 . Pathway analysis was carried out using GAGE R package [52] , with significance threshold at p < 0 . 05 . The UpSetR plot was generated using UpSetR R package [53] . The following antibodies and siRNA were used: anti–phospho-Rb Ser807/811 ( CST , Danvers , MA , production number 8516 ) at 1:1 , 000 for western blots and at 1:250 for IF; anti-p21 ( CST , 2947 ) at 1:1 , 000 for western blots and at 1:250 for IF; anti-eIF2α phospho-Ser51 ( CST , 3398 ) at 1:1 , 000; anti-eIF2α ( CST , 5324 ) at 1:1 , 000; anti-p53 ( Santa Cruz Biotechnology , Dallas , TX , sc-126 ) at 1:1 , 000; anti-Lamp1 ( CST , 9091 ) at 1:1 , 000; anti-Pink1 ( CST , 6946 ) at 1:1 , 000; anti-GAPDH ( abcam , Cambridge , MA , ab9485 ) at 1:2 , 000; anti-tubulin ( CST , 86298 ) at 1:2 , 000; anti-PKR ( CST , 12297 ) at 1:2 , 000; anti-PERK ( CST , 3192 ) at 1:1 , 000; anti-GCN2 ( CST , 3302 ) at 1:1 , 000; anti-rabbit IgG , HRP-linked ( CST , 7074 ) at 1:2 , 000; anti-mouse IgG , HRP-linked ( CST , 7076 ) at 1:2 , 000; goat anti-rabbit and goat anti-mouse 800 ( Azure Biosystems , Dublin , CA ) at 1:5 , 000; and Alexa Fluor-647 secondary antibody ( Thermo Fisher ) at 1:500 . siRNA oligos were synthesized by IDT ( San Jose , CA ) : EIF2AK1 ( hs . Ri . EIF2AK1 . 13 . 3 ) ; EIF2AK2 ( hs . Ri . EIF2AK2 . 13 . 2 ) ; EIF2AK3 ( hs . Ri . EIF2AK3 . 13 . 3 ) ; EIF2AK4 ( hs . Ri . EIF2AK4 . 13 . 2 ) , and negative control DsiRNA ( 51-01-14-04 ) . siRNA transfection was carried out using DharmaFECT 1 ( Dharmacon , Chicago , IL ) following the manufacturer’s instruction . Cells were fixed 24 or 48 h after transfection . Cells were plated on a 96-well plate ( Cellvis , Mountain View , CA , P96-1 . 5H-N ) coated with collagen ( Advanced BioMatrix , Carlsbad , CA , #5015 ) 24 h prior to the start of imaging , at a density such that cells were sub-confluent throughout the imaging period . Cells were imaged on either an ImageXpress Micro XLS wide-field microscope ( Molecular Devices , San Jose , CA ) with a 10× 0 . 45 NA objective or on a Nikon Inverted Microscope Eclipse Ti-E PFS ( Nikon , Japan ) with a 10× 0 . 45 NA objective with appropriate filter sets . Images were taken by a Zyla 5 . 5 sCMOS camera ( Andor Technology , UK ) or an ORCA-Flash 4 . 0 CMOS camera ( Hamamatsu , Japan ) at the frequency of 1 frame per 12 min . During the imaging , cells were kept in a humidified , 37 °C chamber at 5% CO2 . Total light exposure time for each time point was kept under 350 ms . For experiments involving drug treatments , cells were first imaged without drug for 16–24 h; the movie was then paused for 1–5 min and drugs were added by exchanging 50% of the media in each well with media containing a 2× drug concentration . Cells were then imaged for an additional 24 to 48 h . In experiments in which live-cell imaging was followed by IF or RNA FISH , cells were fixed immediately after live-cell imaging by incubation in 4% paraformaldehyde for 15 min . In EdU , OPP , or EU incorporation experiments , cells were incubated in media containing 10 μM EdU for 15 min , 20 μM OPP for 24 min , or 1 mM EU for 24 min and then fixed and processed according to the manufacturer’s instructions ( ThermoFisher C10340 , C10458 , and C10330 ) . For IF , cells were incubated with a blocking/permeabilization buffer ( 3% BSA , 0 . 1% Triton X-100 ) for 1 h at room temperature . Primary antibody staining was carried out overnight at 4 °C in the blocking buffer and visualized using secondary antibodies conjugated to Alexa Fluor 647 . RNA FISH was carried out using the ViewRNA ISH Cell Assay kit ( Thermo Fisher ) following the manufacturer’s instruction . The FISH probes used were as follows: PGK1 ( VA1-12352 ) ; CCNE1 ( VA6-3167995 ) ; E2F1 ( VA6-3168356-VC ) ; CCNA2 ( VA6-15304 ) ; and CCNB1 ( VA6-16942 ) . Image processing and cell tracking were performed using a published pipeline [54] available at https://github . com/scappell/Cell_tracking . In brief , camera dark noise was subtracted from the raw images , which were then divided by the illumination bias . Dark noise was measured by a blank image taken with light power off . The illumination bias of each fluorescent channel was estimated by the averaged cell-free contour of all images in that channel . Log-transformed H2B-mTurquoise images were then convolved with a rotationally symmetric Laplacian of Gaussian filter , and objects were defined as contiguous pixels exceeding a threshold filter score . Segmented cell nuclei were tracked by screening the nearest future neighbor . The background of each image was subtracted using top-hat filtering . Mean nuclear intensities were measured by averaging the background-subtracted pixel intensities in each nucleus as defined by a segmented nuclear mask . CDK2 activity was calculated as the ratio of cytoplasmic to nuclear median sensor fluorescence , with the cytoplasmic component measured in a four-pixel-wide cytoplasmic ring outside of the nuclear mask . In experiments in which IF or FISH signals were matched back to the live-cell imaging , the IF/FISH images were mapped to the last frame of the corresponding live-cell image using nearest neighbor screening after jitter correction . FISH intensity was quantified as median pixel value in a four-pixel-wide cytoplasmic ring outside of the nuclear mask . To segment 53BP1 nuclear bodies and γH2AX foci , images of 53BP1 and γH2AX staining were top-hat filtered to remove nuclear background before thresholding . A segmented focus was assigned to a nucleus if they shared at least one pixel . Cells with at least one 53BP1 nuclear body were classified as 53BP1 n . b . + . Each daughter cell was classified as CDK2inc , CDK2low , or CDK2emerge as follows: CDK2inc cells have CDK2 activity greater than 0 . 5 at 3 h after anaphase; CDK2low cells have CDK2 activity less than or equal to 0 . 5 at 3 h after anaphase and stay below 0 . 5 for the rest of the movie; and CDK2emerge cells have CDK2 activity less than 0 . 5 at 3 h after anaphase and rise above 0 . 5 later in the movie . The R-point was determined as the time CDK2 activity first begins to rise . Computationally , this involves calculating slopes of CDK2 activity using windows of 6–10 time points , and then maximizing a linear function for time since mitosis , CDK2 activity , and CDK2 slope ( long times since mitosis , low CDK2 activity , and high CDK2 slope ) . The rise point of CDK2 activity was manually verified for each cell . To determine low/high thresholds for staining intensity of phospho-Rb , p21 protein , and E2F1/CCNE1 mRNA , the Otsu method was used to compute a global threshold on intensity distributions of all cells in the same experiment [29] . Lineage survival is defined as follows: each cell at the frame of drug addition is considered as a lineage; the end of a lineage is defined as the time of cell death of the last cell in the lineage . For example , if a cell divides once after drug addition , the end of this lineage is the time that both daughter cells have died . The times of cell death were manually determined by condensation of the nuclear marker H2B without cell division .
|
Even within a genetically identical population , some cells proliferate more slowly than others . Slow-cycling cells have been implicated in resistance to antibiotics , antifungals , and cancer therapies , yet the origin of the slow-cycling state remains poorly understood . Here , we isolate a naturally slow-cycling subpopulation of human cells and find that the slow-cycling state is induced by moderate activation of stress responses . We further show that the ability to enter this slow-cycling state protects cells from further stress , consistent with its association with drug resistance . We propose that the existence of the slow-cycling state thereby promotes long-term survival of populations that occasionally experience mildly stressful environments .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"sequencing",
"techniques",
"cellular",
"stress",
"responses",
"anaphase",
"gene",
"regulation",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"cloning",
"dna",
"damage",
"mitosis",
"molecular",
"biology",
"techniques",
"dna",
"rna",
"sequencing",
"research",
"and",
"analysis",
"methods",
"small",
"interfering",
"rnas",
"chromosome",
"biology",
"gene",
"expression",
"molecular",
"biology",
"biochemistry",
"rna",
"cell",
"biology",
"nucleic",
"acids",
"genetics",
"biology",
"and",
"life",
"sciences",
"non-coding",
"rna"
] |
2019
|
Spontaneously slow-cycling subpopulations of human cells originate from activation of stress-response pathways
|
Treatment with nifurtimox ( NF ) for Chagas disease is discouraged during breast-feeding because no information on NF transfer into breast milk is available . NF is safe and effective for paediatric and adult Chagas disease . We evaluated the degree of NF transfer into breast milk in lactating women with Chagas disease . Prospective study of a cohort of lactating women with Chagas disease . Patients were treated with NF for 1 month . NF was measured in plasma and milk by high performance liquid chromatography ( HPLC ) . Breastfed infants were evaluated at admission , 7th and 30th day of treatment ( and monthly thereafter , for 6 months ) . Lactating women with chronic Chagas disease ( N = 10 ) were enrolled ( median age 28 years , range 17–36 ) . Median NF dose was 9 . 75 mg/kg/day three times a day ( TID ) . Six mothers had mild adverse drug reactions ( ADRs ) , but no ADRs were observed in any of the breastfed infants . No interruption of breastfeeding was observed . Median NF concentrations were 2 . 15 mg/L ( Inter quartil range ( IQR ) 1 . 32–4 . 55 ) in milk and 0 . 30 mg/L ( IQR 0 . 20–0 . 95 ) in plasma . Median NF milk/plasma ratio was 16 ( range 8 . 75–30 . 25 ) . Median relative infant NF dose ( assuming a daily breastmilk intake of 150 mL/kg/day ) was 6 . 7% of the maternal dose/kg/day ( IQR 2 . 35–7 . 19% ) . The low concentrations of NF in breast milk and the normal clinical evaluation of the breastfed babies imply that maternal NF treatment for Chagas disease during breastfeeding is unlikely to lead to clinically relevant exposures in the breastfed infants . Clinical trial registry name and registration number: ClinicalTrials . gov NCT01744405 .
Chagas disease ( CD ) or American trypanosomiasis , is a parasitic zoonosis caused by infection with Trypanosome cruzi of worldwide distribution , endemic to the Americas , predominantly affecting the poor and medically underserved [1 , 2 , 3] . Most CD patients are asymptomatic during the acute phase , but progress to a chronic phase that if the disease is left untreated can lead to cardiac and/or digestive complications in almost 30% of the patients [4 , 5] . More than 10 million people are infected in South and Central America . Most people acquire CD during childhood . If girls are not treated , they can transmit the infection to their babies . Recently , CD has become a global health problem expanding to virtually all regions of the world via immigration , with many cases reported in Europe and North America [2 , 6] . The only two drugs available for the treatment of CD are nifurtimox ( NF ) and benznidazole ( BNZ ) , Both drugs have similar effectiveness and limitations [3 , 7]; Their mechanisms of action , pharmacokinetics or toxicokinetics are still unclear , but they have been used since decades nonetheless , in spite of a high risk of toxicity in adults , especially dermatological reactions [3 , 7] . However , the reported incidence of adverse drug reactions ( ADR ) is much lower in infants and children [8 , 9 , 10 , 11] . Most physicians rarely give women treatment for CD during lactation due to a perceived risk of infant exposure to these drugs through breastmilk . On the other hand , discontinuation of breastfeeding to allow for maternal treatment is not advisable , given that breast milk is the ideal food for newborns , as well as a source of multiple benefits [12] . However , in areas with high birth rates and limited access to health care , the postpartum breastfeeding period may be the only , and brief , period of time when a woman has consistent contact with health services , and may be amenable for CD treatment . The aim of this study was to prospectively study , for the first time , NF transfer into breast milk in a cohort of lactating women with CD in order to clarify the safety of this practise . The secondary objective is to provide support for evidence-based recommendations for the management of CD during lactation .
The study protocol was approved by the Research and Teaching Committee , and Bioethics Committee of the Buenos Aires Children’s Hospital “Dr . Ricardo Gutierrez” . The approval number is 2011-06-22 . 1LAC . Written informed consent was obtained from all participants . All parents provided informed consent on behalf of all minor participants evaluated in this trial . The protocol was registered in ClinicalTrials . gov ( #NCT01744405 ) . Women with chronic CD who were breastfeeding and their infants were enrolled in this prospective cohort study at the Parasitology and Chagas Service , Buenos Aires Children´s Hospital “Ricardo Gutierrez” , between October 2011 and February 2012 . CD diagnosis we performed with at least two independent serological tests for T . cruzi antibodies , as per routine clinical care . Exclusion criteria included: None of the included patients were taking any co-medications . Patients with known medical conditions that could affect result interpretation , a positive pregnancy test , history of NF hypersensitivity or previous NF treatment were excluded from the study . Treatment: Lactating women with CD received 8 to 12 mg/kg/day , TID NF ( 120 mg NF tablets ) p . o . ( Lampit , Bayer , El Salvador ) , for 30 days [13 , 10] . A detailed clinical history , physical examination and routine laboratory tests were obtained at diagnosis , at the end of the first week and 30 days into NF treatment . Patients were then followed as per CD treatment and follow up guidelines . Treatment response was evaluated by T . cruzi specific real-time Polymerase Chain Reaction ( PCR ) performed at diagnosis and at the end of treatment [14 , 15] . Patients were instructed to use contraception during treatment; a pregnancy test was performed before enrolment . Diagnosis of congenital Chagas disease: All infants under 8 months old were monitored for CD using microhematocrit test . Infants with negative parasitemia were later tested by serology at 8 months of age [10] . Children older than 8 months of age were evaluated with two serological tests for T . cruzi antibodies [10] . Growth and psychomotor development was assessed in children by experienced paediatricians . Pediatric evaluations were performed at days 0 , 7 and 30 of maternal treatment , and monthly thereafter for at least 6 months . Breast milk samples ( approximately 30 mL ) were collected before the start of NF treatment , and on the 7th ( +/- 3 days ) and 30th ( +/- 3 days ) day of treatment . Each milk sample was mixed , total volume recorded and an aliquot stored at -20 C° until analysis . Breastmilk lipid content was not measured . Venous blood was sampled in heparinized tubes , centrifuged at 3 , 000 g for 10 min and plasma stored at -20 C° and lyophilized prior to analysis . A high performance liquid chromatography ( HPLC ) method was used to determine NF concentration in plasma and milk , as described previously [16] . Briefly , plasma samples were deproteinized with 100 μL tricloroacetic acid ( 30% w/v ) , vortexed 20 seconds , sonicated for five minutes and then centrifuged at 8 , 000 rpm for 5 minutes . Supernatants were mixed with 500 μL of ethyl acetate , precipitated with 100 mg of anhydrous sodium sulfate ( to a concentration near saturation ) and vortexed for one minute . The mixture was centrifuged at 8 , 000 rpm for 5 minutes and the organic phase of three consecutive liquid/liquid extraction procedures were recovered together and rotoevaporated to dryness at 40°C and 40–80 bars . The residue was resuspended in 250 μL of methanol , vortexed for 20 seconds and centrifuged 2 minutes before injection in the HPLC . Breast milk samples ( 1000 μL ) were deproteinized by adding 100 μL of trichloroaceticacid ( 30% w/v ) , vortexed for 1 minute and sonicated for 10 minutes , after which the samples were filtered through a 0 . 45 micron membrane by centrifugation at 8 , 000 rpm for 20 minutes to obtain an ultrafiltrate of breast milk . The ultrafiltrate was directly injected into the HPLC [17 , 18] . The limit of detection ( LOD ) and limit of quantitation ( LOQ ) for plasma and breastmilk were 0 . 01 mg/L and 0 . 1 mg/L , respectively . Milk-to-plasma ( MP ) ratios were calculated from single milk and plasma concentration measurements . In those patients that had plasma concentrations below the LOQ ( but above LOD ) , a value equal to half the LOQ ( i . e . 0 . 05 mg/L ) was imputed in order to provide a realistic estimate of plasma concentrations that would overall avoid under- or overestimating MP ratios in these patients . Single-point maximum observed milk concentration for each individual was multiplied by 0 . 15 L/kg/day ( i . e . estimated median milk intake for an infant ) to yield the absolute infant daily NF dose ( in μg/kg/day ) that the infant would ingest per day through breastfeeding . The absolute infant daily NF dose was then divided by the weight-normalized maternal NF dose ( in μg/kg/day ) and multiplied by 100 to estimate the percent Relative Infant Dose ( RID ) [19 , 20 , 21] . In cases where more than one RID estimate was available for the same patient , the highest RID was chosen for the statistical calculations . The RID represents the percentage of the therapeutic dose ( usually taken from the maternal dose ) that a baby would be exposed during breastfeeding . The NF dose used for calculations ( i . e . 10–15 mg/kg/d ) is the actual pediatric dose used in clinical practice [10 , 22] .
Ten women and their 10 babies were enrolled in the study . All mothers were in the chronic CD stage; six of them had acquired the infection in Bolivia , 3 in Argentina and 1 in Paraguay . Median age and weight of the mothers were 28 years ( range 17–36 years ) and 58 , 5 kg ( range 52–73 kg ) , respectively . Median infant age at the start of maternal treatment was 6 . 8 months ( range 1 month-11 months ) , and median weight 7 . 6 kg ( range 5–9 . 5 kg ) . All infants were healthy , within 25th to 95th percentiles for weight and height for their respective ages . Three babies were exclusively breastfed and seven also received solid foods . Median maternal daily dose of NF was 9 . 82 mg/kg/day ( range 8 . 3–12 mg/kg/day ) . [Table 1] Six mothers ( 60% ) had adverse drug reactions ( ADR ) to NF: 4 were mild ( 1 vomiting and fever , 1 headache and dizziness , 1 eosinophilia and 1 mild leukopenia ) and were able to continue treatment , and 2 were moderate ( psychomotor agitation and headache ) and led to medication discontinuation by patient decision after 9 and 19 days of treatment , respectively . There were no serious ADRs and no infant had to stop breastfeeding . No ADRs were observed in the breastfed infants , nor any changes in their behaviour , weight progress or other effects potentially attributable to NF . All infants were healthy during and after the study , as assessed by paediatricians skilled in the evaluation of paediatric patients with CD . Breast milk samples , a total of 17 , were taken at a median 9 . 4 days ( range 4–21 ) after start of NF treatment , so that all patients are assumed to have been at steady state for NF plasma concentrations at the time of sampling . Post-treatment breast milk samples were taken within 24 hours after the last dose . Median plasma NF concentration was 0 . 30 mg/L ( 9 samples were LOQ ) ( IQR of samples that were not LOQ , 0 . 20–0 . 95 mg/L ) . Median milk concentration was 2 . 15 mg/L ( IQR 1 . 32–4 . 55 ) . Median milk/plasma NF concentration ratio ( MPR ) was 16 ( IQR 8 . 75–30 . 25 ) . Assuming a 150 mL/kg daily milk intake , the estimated median NF daily infant dose was 0 . 50 mg/kg/day ( IQR 0 . 20–0 . 69 ) , representing a median RID of 6 . 70% of the maternal weight-corrected daily dose ( IQR 2 . 35–7 . 19% ) . Among the 10 infants enrolled in the study , 8 turned out not to have congenital CD , as confirmed by serology at 9 months of age; the remaining 2 were diagnosed with congenital CD and were treated accordingly; Both had a serological response and negative conversion of PCR; None of these infants had any medication related ADRs . Only one mother showed positive qPCR at the end of treatment . The measured NF concentrations for this mother in blood and milk were below LOD . After re-evaluation , this patient admitted to not taking the drug correctly ( and therefore her data were left out of the analysis ) . A new 60 days NF treatment course was started and the qPCR was negative at the end of treatment and during posttreatment follow-up .
CD transmission can take place by contact with the vector ( i . e . known as “kissing bugs” ) , congenitally , and via transfusions or organ transplantation . Every year an estimated 1 , 300 children are born with congenital CD in Argentina , but less than half are offered access to treatment . Recently , small outbreaks have also been linked to ingestions of parasite-contaminated food [3 , 23] . T . cruzi has rarely been detected in human milk , only in mothers with bleeding nipples during acute CD infection . In a previous study of 21 lactating women , our group found no presence of T . cruzi in human milk using qPCR [24] . Even though risks for parasite exposure from breastmilk are unclear , they are unlikely to be significant and CD in the mother is not considered a reason to avoid breastfeeding [25 , 26 , 27] . In a previous study by our group , we observed limited transfer of benznidazole ( the other drug available for CD ) into breastmilk , and no significant risks to the infants [24] , and Vela et al later confirmed that benznidazole used during postpartum in women with CD had no negative impacts on the breastfed child , suggesting that there is no need to interrupt breastfeeding [28] . Unfortunately , benznidazole is not consistently available in all endemic countries , which led us to study NF during breastfeeding , encouraged by a theoretical pharmacokinetic model that suggested that the transfer of NF into breastmilk was likely to be very limited [29] . In rural Latin America young women may only sporadically interact with the health system except during pregnancy , delivery and the early postpartum period . Also , short inter-pregnancy intervals may leave few opportunities for CD treatment beyond breastfeeding periods . The heretofore lack of data supporting safety of NF during breastfeeding put health care professionals in the uncomfortable position of deciding between supporting breastfeeding or CD treatment for the mother , thus forgoing widespread recommendations to support exclusive breastfeeding , and to treat CD [12 , 30] . However , this choice between Chagas disease treatment and breastfeeding implies risks such as losing the opportunity to treat the mother and hopefully prevent congenital infections in future babies , as well as preventing long term cardiac complications in the mother , or , if treatment is chosen over breastfeeding , increased risks of infant diarrhea , infections and other formula-associated problems . This study describes the first prospective study of NF transfer to breastmilk in CD patients , suggesting that infants’ exposure to NF via breastmilk would amount to less than 5% of the usual infant weight-corrected NF dose ( i . e . 10–15 mg/kg/day ) . This exposure is below the 10% cut-off commonly used as threshold evaluate risk for exposure to maternal drugs during breastfeeding [31 , 32 , 33] . Taking into account the known safety of NF in children , observed NF milk concentrations ( i . e . ~10 times lower than therapeutic doses ) would not be expected to produce exposures associated to infant ADRs or any other risks . Furthermore , treatment with NF is better tolerated in infants and children with CD than in adults [10 , 34] . No ADRs were observed in the breastfed infants in our study , and careful evaluation by experienced paediatricians found no behavioural , growth or weight impacts potentially attributable to NF . The potential difficulties of detecting adverse events in children and infants ( especially central nervous system events in small infants ) have not escaped our attention . However , even if specific ADRs may be hard to pinpoint ( e . g . headache ) , these events do have detectable manifestations that trained pediatricians can detect . Our group also participated in a multidisciplinary study in children using NF to treat Chagas disease and an incidence of 19% of NF related ADRs were observed , the most common being weight decrease , decreased appetite , headache and rash . All ADRs were readily identified by the pediatricians evaluating these children , many of which participated in this study . The overall observed incidence of ADRs in adults in our cohort ( 60% ) is in agreement with the rate previously described in adults [37 , 13] . Transfer of drugs into breastmilk is a function of molecular weight ( MW ) and maternal plasma level [31 , 32 , 35] . NF is a small molecule ( MW = 287 ) with high oral bioavailability and moderate plasma protein binding ( 50% ) [31] . Our results show clear evidence that the milk concentrations are a function of the plasma concentrations ( Table 1 ) . These results follow the general rule stating that drug concentration in human milk are usually low and will seldom lead to levels that could produce a pharmacological response in the nursing infant [31 , 36] . The MP ratio estimates in our patients was hampered by the fact that many plasma concentrations were below LOD ( i . e . detectable but not measurable ) , thus forcing us to estimate a concentration in order to calculate MP ratios . We chose the value of 50% LOQ ( i . e . the median plasma level that is detectable but not measurable ) as a good overall estimate for the observed but non-measurable NF concentrations . MP ratios are intended to provide a general ( over ) estimate of drug transfer into breastmilk for medications taken by the mother , but contain limited information to judge potential exposure of the baby through breastmilk . MP ratios are , in fact , not generally the preferred estimator of potential for infant drug exposure if other , better; indicators of degree of exposure risk are available such as RID . There is an abundance of examples in the literature of drugs that have high MP ratios but negligible infant exposures due to very low milk concentrations [37] . In the case of NF , the median MP ratio of 16 suggests a significant accumulation of NF in breastmilk . Many potential explanations can account for this , but the main possible reason is that NF is a substrate of breast cancer resistance protein ( BCRP ) , which may be responsible for actively transferring it into the breast ( and other tissues ) 35 . One patient ( P1 , Table 1 ) had an estimated MP ratio of 190 . This large MP ratio may be related to BCRP polymorphisms , or other factors . Unfortunately , we do not have enough data to explore this interesting observation further [32] . Given the nature of the design of this study ( e . g . in many cases , mothers expressed milk at home and brought it to the clinic the next day ) , we cannot ascertain whether fore or hind milk was obtained in most occasions , as the main objective was to obtain leftover milk and in no way interfere with infants’ breastfeeding . NF concentrations do not vary significantly depending on fat content , and therefore we did not expect to see much variation between hind and fore milk . A limitation of this study is the small number of infants enrolled , which makes it impossible to rule out uncommon ADRs . However , relatively large numbers of paediatric CD patients , including infants and neonates , have been treated with NF at therapeutic doses ( approximately 8 to 10 times higher than the expected exposure through breast milk based on our data ) for the past few decades in many centres in Latin America , and no significant developmental problems or other significant ADRs have been identified to date [10 , 38 , 39 , 29] . We have no reason to believe that a significantly lower exposure would lead to ADRs not observed at therapeutic doses .
The results of this study , the first of its kind in CD , suggest that NF may be compatible with breastfeeding due to limited drug transfer into breast milk , and low overall infant exposure . The currently perceived contraindication to NF treatment during lactation , so far unsubstantiated by any evidence , may lead to lost opportunities to treat lactating women . This conclusion is further supported by the complete absence of ADRs attributable to NF in the breastfed infants . Our study provides , for the first time , support for continuation of breastfeeding during maternal CD treatment with NF , a practice that can potentially benefit many women and their breastfed infants in settings where maternal treatment during breastfeeding may be advantageous .
|
It is not known whether Nifurtimox , a drug for Chagas disease , is significantly transferred into breast milk , and no clinical trials were conducted to evaluate this topic . Treatment with nifurtimox is safe and effective in children and newborns with Chagas disease . Treatment of young women before pregnancy prevents congenital transmission of Chagas disease . This is the first study to measure nifurtimox concentrations in breast milk . We found that presence of nifurtimox into breast milk is limited , and that breastfed babies had normal clinical evaluations with no observable adverse events . None of the mothers had to discontinue breastfeeding due to adverse events . The exposure of nifurtimox through breast milk during the treatment of mothers with Chagas disease does not seem to pose significant risks for the breastfed infants .
|
[
"Abstract",
"Introduction",
"Population",
"and",
"methods",
"Results",
"Discussion",
"Conclusion"
] |
[
"neonatology",
"children",
"medicine",
"and",
"health",
"sciences",
"milk",
"breast",
"milk",
"body",
"fluids",
"maternal",
"health",
"liquid",
"chromatography",
"tropical",
"diseases",
"diet",
"parasitic",
"diseases",
"pediatrics",
"age",
"groups",
"nutrition",
"women's",
"health",
"infants",
"neglected",
"tropical",
"diseases",
"pharmacology",
"high",
"performance",
"liquid",
"chromatography",
"families",
"research",
"and",
"analysis",
"methods",
"adverse",
"reactions",
"protozoan",
"infections",
"blood",
"plasma",
"chromatographic",
"techniques",
"people",
"and",
"places",
"breast",
"feeding",
"chagas",
"disease",
"blood",
"anatomy",
"beverages",
"physiology",
"population",
"groupings",
"biology",
"and",
"life",
"sciences"
] |
2019
|
Negligible exposure to nifurtimox through breast milk during maternal treatment for Chagas Disease
|
Many broadly neutralizing antibodies ( bNAbs ) against human immunodeficiency virus type 1 ( HIV-1 ) were shown effective in animal models , and are currently evaluated in clinical trials . However , use of these antibodies in humans is hampered by the rapid emergence of resistant viruses . Here we show that soft-randomization can be used to accelerate the parallel identification of viral escape pathways . As a proof of principle , we soft-randomized the epitope regions of VRC01-class bNAbs in replication-competent HIV-1 and selected for resistant variants . After only a few passages , a surprisingly diverse population of antibody-resistant viruses emerged , bearing both novel and previously described escape mutations . We observed that the escape variants resistant to some VRC01-class bNAbs are resistant to most other bNAbs in the same class , and that a subset of variants was completely resistant to every well characterized VRC01-class bNAB , including VRC01 , NIH45-46 , 3BNC117 , VRC07 , N6 , VRC-CH31 , and VRC-PG04 . Thus , our data demonstrate that soft randomization is a suitable approach for accelerated detection of viral escape , and highlight the challenges inherent in administering or attempting to elicit VRC01-class antibodies .
A large number of potent broadly neutralizing antibodies ( bNAbs ) have been generated from HIV-1-infected individuals ( reviewed in [1–5] ) . Many of these bNAbs , including 2F5 , 4E10 , PGT121 , VRC01 , 3BNC117 and 10–1074 , have been or will be evaluated in clinical trials [6–12] . A number of animal studies have shown that administration of bNAbs can prevent infection , and reduce viral loads in an established infection [13–22] . Previous human studies also show that bNAbs can reduce viral loads or delay viral rebound upon treatment interruption [6–8 , 11 , 12 , 23] . However , HIV-1 generally escapes these bNAbs when they are administered to infected animals or humans [6–8 , 11–15 , 17 , 18 , 20 , 21 , 23–26] . Escape pathways of HIV-1 from inhibitors against protease and reverse transcriptase have been comprehensively documented through large-scale studies of patients treated with these inhibitors . In contrast , escape from neutralizing antibodies has not been similarly described , in part because no antibody has yet been tested in a sufficiently large cohort of HIV-1 positive individuals . Most reported in vivo escape variants emerged from a limited number of animals or volunteers who participated in clinical trials [6–8 , 11 , 13 , 20 , 24–27] . In addition , although most bNAbs have been characterized with large panels of HIV-1 isolates [28–30] , these studies do not provide direct insight into the likelihood of viral escape , because most HIV-1 isolates have never been exposed to or selected by these exceptional and rare antibodies . In some cases , the number of avenues to escape may be small because of structural and functional constraints on conserved epitopes of the envelope glycoproteins ( Env ) . Nevertheless , to use bNAbs in humans or to design vaccines based on their epitopes , it is essential to understand how viruses can escape them . Although such information is now available for some antibodies from clinical trials , and other mutations have been observed in animal models or selected in cell culture , only a limited number of escape variants have been described for any given antibody or combination of antibodies . This small number is a consequence of inefficient process by which resistant variants emerge . This process relies on errors naturally generated by the viral reverse transcriptase , which introduces approximately one mutation per newly integrated provirus . Because these mutations are distributed throughout the viral genome , fewer than 1 in 10 of such mutations affect the env gene , the target of all bNAbs . Typically , the emergence of resistant variants is much faster in vivo than in vitro—escape variants usually emerge within a few weeks after antibody infusion [6–8 , 11 , 13 , 23–26]—likely owing to much higher number and diversity of viruses found in each individual compared to that used in in vitro experiments . Nonetheless , only a limited number of escape pathways were detected in vivo , possibly because once a resistant virus emerges , it soon dominates the viral swarm , eliminating the selection pressure that would otherwise lead to additional resistant mutations . Therefore , to accumulate a meaningful number of escape mutations , this process needs to be repeated great many times or in many individuals or animals . Several mutagenesis methods have also been adopted to accelerate viral evolution . Of these , the mutation rate of error-prone PCR can be difficult to control , and alanine-scanning mutagenesis is limited to mutating residues only to alanine , precluding other substitutions or combinations of substitutions available to the virus . A library constructed by codon mutagenesis , Exceedingly Meticulous and Parallel Investigation of Randomized Individual Codons ( EMPIRIC ) and Single-site saturation mutagenesis ( SSM ) were used to accelerate the evolution of HIV-1 or influenza A virus [31–33] . These methods are comprehensive , but labor intensive and only one mutation is introduced per copy of genome . On the other hand , random mutagenesis would generate high number of mutations , but in an overwhelming number of cases , the resulting Envs would not fold or function properly . Another mutagenesis technique , soft randomization , originally developed for phage display applications , allows generation of libraries whose members have a small , desired number of mutations distributed throughout a target region [34 , 35] . This technique permits it by using primers synthesized with nucleotides mixed at a non-equimolar ratio , favoring original nucleotides [36] . The optimal composition and use of a hand mix ratio depend on the number of target codons to be altered , the library size , and how wobble codons are addressed . Here , we developed a method that can help anticipate a range of viral escape pathways from antibodies with a well-defined epitope . To assess its usefulness , as a proof of concept , we tested this method against the CD4-binding site ( CD4bs ) bNAbs . We generated a replication-competent HIV-1 library expressing Envs diversified by soft randomization in the CD4bs , and selected the escape variants that were resistant to neutralization by VRC01-class CD4bs antibodies . Soft randomization introduced a controlled number of mutations into the majority of library members , and a large number of antibody-resistant isolates could be identified after only a few passages . Most randomly chosen clones of escape variants were resistant to all well-characterized VRC01-class antibodies , including N6 , an exceptionally broad antibody of this class [37] , in addition to the antibodies used in the selection . Because the epitope regions of CD4bs antibodies overlap with the CD4 binding sites , the resistant variants exhibit growth kinetics slower than that of the parental virus , as previously shown , but nonetheless they readily outgrew the parental virus at antibody concentrations commonly targeted in clinical trial [8 , 10 , 38] . Collectively , our data demonstrate that soft randomization can be usefully applied to identifying viral escape pathways in the face of a specific selection pressure .
To determine if soft randomization could generate a useful library of HIV-1 proviruses , we soft randomized the epitope regions of VRC01-class antibodies in the env gene , and selected the resulting virus library with the VRC01-class bNAb NIH45-46 . We chose NIH45-46 because it was isolated from the same patient from which VRC01 was derived , but it is more potent than VRC01 [39] , and because it has been evaluated in a clinical trial . As shown in Fig 1A , the binding epitopes of NIH45-46 are primarily located in the four regions of Env ( Los Alamos National Laboratory ( LANL ) database CATNAP ) : Loop D , CD4 binding loop , bridging sheet and variable region 5 ( V5 ) . To minimize interference with CD4 binding , only the residues in Loop D , V5 and in their vicinity were randomized , whereas the CD4-binding loop and the bridging sheet were left unmodified . The compositions of the PCR primers used to build our library are shown in Fig 1B , and based on the sequence of ADA , a well-characterized clade B R5 isolate . Soft-randomizing primers were synthesized by handmixing original nucleotide and the other three nucleotides at a selected ratio for the first two positions of each codon . To incorporate a small number ( 1 to 3 ) of mutations per region , we use a handmixing ratio of 88:4:4:4 for the 11 amino acids of Loop D , and 91:3:3:3 for the longer V5 region ( 17 residues ) . An equimolar mix of G and T for the wobble positions of all amino acids encoded by 4 or 6 codons in Loop D , and an equimolar mix of C and A was used for V5 regions , because the soft-randomizing primer for V5 is anti-sense . As shown in S1 Fig , Loop D and V5 libraries were separately generated by PCR amplification of the entire pBR322 plasmid containing ADA env gene , pBR322-Env ( ADA ) , as a template , and ligated using 5’ phosphate in the soft- randomizing primers . V5 library was subcloned into the plasmid containing Loop D library using an enzyme site engineered between Loop D and V5 regions . The resulting env library containing Loop D and V5 regions was then subcloned into the pNL4-3 proviral plasmid carrying ADA env gene , pNL-ADA , yielding a proviral library encoding diverse Env proteins , pNL-ADA-Lib . An NL-ADA-Lib virus population ( described here as a “swarm” ) was then produced by transfecting HEK-293T cells with the pNL-ADA-Lib plasmid . To assess the composition of the library , a fragment encompassing both Loop D and V5 was amplified by RT-PCR and sequenced by paired-end Illumina MiSeq , and the total number of mutations in Loop D and V5 regions of each library member was analyzed . Only 21% of the total reads contained either stop codons or unknown amino acids , fewer than what is typically observed with less controlled mutagenesis approaches . These reads were excluded before analysis . The most common number of amino-acid substitutions in Loop D and V5 were 1 to 2 and 2 to 4 , respectively , typically yielding 3 to 6 substitutions combined ( Fig 2A ) , consistent with our predicted values . The proportion of non-mutated members was relatively low: 15% for Loop D and 4% for V5 . More than 74% of the library members contain 1–3 mutations in Loop D , and 69% contain 1–4 mutations in V5 , indicating that the majority of NL-ADA-Lib members contain any number of substitutions between 2 and 7 . The substitution profile of each soft-randomized residue is shown in Fig 2B . Substitutions of each residue were relatively evenly distributed among 19 amino acids , with substitution biases reflecting one or two shared nucleotides between mutated and parental codons . Thus , soft-randomization can introduce a high degree of controlled diversity in selected regions of a viral genome . To validate that escape variants can be easily detected from the soft-randomized library of viruses , we passaged the library swarm in the presence of indicated concentration of NIH45-46 ( S2A Fig ) . Virus-containing culture supernatants were harvested two days later ( NIH45-46 passage 1 swarm ) , and this process was repeated four additional times at the indicated antibody concentrations ( S2A Fig ) . The virus swarm from each passage was then assessed for their resistance against NIH45-46 in TZM-bl luciferase reporter cells ( Fig 3A ) . Both the parental NL-ADA-Lib swarm and that passaged 5 times in GHOST cells in the absence of any antibody ( control passage 5 ) were used as controls . We observed that a single passage already conferred high level of resistance , and five passages , taking only two days per passage , were sufficient to select near complete resistance . We next assessed the NIH45-46 passage 5 swarm for their resistance to other VRC01-class CD4bs antibodies ( Fig 3B ) . VRC01 and VRC07 were isolated from the same individual from whom NIH45-46 was derived , but 3BNC117 , VRC-CH31 , VRC-PGV04 , and VRC-PG20 were isolated from different donors . Average IC50 and IC80 values of these antibodies for various HIV-1 isolates are shown in S2B Fig . As shown in Fig 3B , NIH45-46 passage 5 swarm was resistant to all VRC01-class CD4bs antibodies assessed , with the exception of 3BNC117 and VRC07 to a limited degree . To determine whether complete resistance against all VRC01-class CD4bs antibodies can be obtained , we grew NIH45-46 passage 5 swarm in the presence of 3BNC117 . The resulting swarm was then passaged four additional times at the concentrations indicated in S2A Fig . To prevent de-selection of NIH45-46 resistant viruses , NIH45-46 was included in all passages . 3BNC117 passage 5 swarm was then assessed for its resistance to 3BNC117 and VRC07 ( Fig 4A ) . Interestingly , although the virus has become completely resistant to 3BNC117 even at 30 times of its IC80 , its sensitivity to VRC07 was not changed . We therefore passaged this swarm in the presence of VRC07 . The resulting VRC07 passage 5 swarm was resistant to all the VRC01-class antibodies we tested ( Fig 4B ) , including the antibody , N6 , a recently isolated CD4bs bNAb with the highest breadth thus far described in the class . We further assessed the same swarm against non-VRC01-class CD4bs antibody , b12 , and non-CD4bs antibodies 10–1074 , 10E8 and PGDM1400 ( Fig 4C ) . VRC07 passage 5 swarm was sensitive to all of these antibodies , while their vulnerability to PGDM1400 was modestly reduced , suggesting its epitopes might partially overlap with those of CD4bs antibodies . These data show that observed resistance was not due to a generalized resistance mechanism or any non-specific growth advantages , and indicate that escape variants can readily be detected from a soft-randomized library swarm . To identify the sequence variation that contributes to escape phenotype , VRC07 passage 5 swarm was then deep sequenced in the Loop D and V5 regions and compared to the sequences of the parental library swarm or that passaged 15 times in GHOST cells in the absence of any antibody ( control passage 15 ) . The total number of variants containing a substitution at the indicated soft-randomized residue are shown in Fig 5A , and their specific substitutions in Fig 5B . Although the original library swarm had fewer mutations in Loop D than in V5 region ( Fig 2 ) , higher number of Loop D mutations was detected in the escape variants ( Fig 5A ) . Some substitutions are enriched in the control swarm , as expected , likely because they provided growth advantages in this experimental setting . The residues found most frequently substituted in the escape virus in this study are N276 , D279 and A281 in the Loop D , and N461 and S465 in the V5 regions , similar to previous observations [7 , 11 , 25 , 27 , 40] . These residues are indicated in the crystal structure of an Env trimer ( Fig 5C ) . To rank enriched sequences , copy number of each sequence found in VRC07 passage 5 swarm was normalized by the copy number of the same sequence detected in the control passage 15 swarm . To include in normalization the sequences that were detected in VRC07 passage 5 swarm but not in the control passage 15 swarm , a control copy number of 1 . 0 was added to all sequences . The sequences of the top 150 escape variants after normalization are listed in S3A Fig . The same sequences are presented in S3B Fig , with each residue marked with a different color to highlight their similarities and substitution patterns . To confirm the resistance of the escape variants to CD4bs antibodies , we first chose six clones ( 1 , 5 , 10 , 15 , 20 and 25 ) from the top 25 . The sequences of these clones—771 bp fragments containing only the mutations found in the Loop D and V5—were synthesized and cloned first into the pBR322-env ( ADA ) using engineered NcoI and MluI sites ( S1 Fig ) and then into the pNL-ADA plasmid . Replication-competent clonal viruses were produced from 293T cells by transfecting corresponding plasmids . When assessed with the three antibodies they were selected against , we observed that all these virus clones were completely resistant to them ( Fig 6A , lower panel ) . We also characterized several additional clones from the top 150 ( S3 Fig ) , which contained a small number ( 2 or 3 ) of substitutions ( Fig 6B , top panel ) . As shown in the lower panel of Fig 6B , most of these clones were also resistant to all three antibodies they were selected against . These results suggest that many , if not most , of the top 150 escape variants are resistant to all three antibodies . These clones were then tested against additional VRC01-class antibodies: N6 , VRC-CH31 , VRC-PG04 and VRC01 . All clones except clone 1 were resistant to all of these antibodies ( Fig 6C ) . Clone 1 was sensitive to N6 but resistant to all other antibodies . Of note , clone 142 , bearing only two substitutions ( D279G , N280Y ) , was resistant to all VRC01-class antibodies tested , including N6 . Although D279G was not sufficient by itself to confer resistance to VRC07 ( S4A Fig ) , more than half of known HIV-1 isolates contain an amino acid other than D at the residue 279 ( Fig 7A and LANL sequence database ) , indicating high flexibility at this position . Whereas none of the top 150 clones gained an N-linked glycan , a majority ( 136 ) has lost the well-characterized N-linked glycan at N276 in Loop D through a substitution either at N276 and/or at T278 , and 66 clones have substitution either at N461 and/or at S463 in V5 region . The loss of a N276 glycan is consistent with the reports that this glycan is required for virus neutralization by VRC01-class antibodies [22 , 29 , 41] . While it was previously detected in in vivo escape variants [25 , 27] , most natural isolates maintain this glycan ( Fig 7A and LANL sequence database ) . When examined , however , loss of the N-glycan at residue 276 alone did not confer resistance to VRC07 ( S4B Fig ) . This mutation was also not necessary for resistance to other VRC01-class antibodies; indeed , three clones ( 25 , 14 and 142 ) we tested retained the N-glycan at N276 , but they were nonetheless resistant to all VRC01-class antibodies ( Fig 6 ) . In addition to the aforementioned mutations at the residues 276 , 279 and 461 , several other mutations identified in our top 150 escape variants are found in naturally-occurring isolates ( LANL sequence database ) , including L277 , S278 , I283 , S460 , T/S/K at reside 461 , and N/T at 462 ( Fig 7A ) . Although the sequence variation among different parental swarms in different patients complicates analysis , a subset of the mutations found in our escape variants has also been previously reported in the clinical trials evaluating CD4bs bNAbs , VRC01 and 3BNC117 [6 , 7 , 11 , 12] . These include K276 , L277 , S278 , R282 , S460 , S/D at residue 461 , N/T at residue 462 , and R463 ( Fig 7B ) . Thus , some of the mutations identified through our approach are found in infected humans . Collectively , our data demonstrate that there are abundant and diverse pathways through which HIV-1 can escape CD4bs antibodies . Many of CD4bs bNAb escape variants were shown to exhibit fitness cost [42 , 43] , because changes in the antibody epitopes are likely alter virus binding to CD4 . Typically compensatory mutation emerge that restore fitness while maintaining resistance . To assess fitness , we first measured the ability of WT and escape variants to use CD4 by measuring their neutralization sensitivity to CD4-Ig ( Fig 8A–8C ) . Infection of TZM-bl cells by WT virus and clone 1 was similarly inhibited by CD4-Ig . However , CD4-Ig less potently neutralized other escape clones , indicating that they have reduced affinity for cellular CD4 . Next , we assessed their fitness in the CD4+ T cells prepared by activating peripheral blood mononuclear cells with phytohemagglutinin-L ( Fig 8D ) . Progeny virus production was measured by RT-qPCR in the culture supernatants of the infected cells every two days until day 8 post infection . The growth difference between WT and escape clones was substantial , except clone 1 , especially in early time points , indicating the fitness of most escape variants was compromised . However , progeny virus production by clone 1 reached similar level as that of WT virus by day 4 . In addition , as was shown in Fig 6 , whereas replication of escape variants was not affected by the presence of VRC07 , more than 99% of WT replication was inhibited . These data are consistent with the observations repeatedly made in the clinical trials evaluating CD4bs bNAbs [6 , 7 , 10–12 , 23] that although generally less fit than WT virus , once selected , the escape variants outcompete WT virus in the presence of bNAbs .
Most bNAbs are able to neutralize majority of known HIV-1 isolates in vitro and prevent a new infection or control an established infection in animal studies . It is well established , however , that resistant viruses easily emerge in vitro and in vivo in the presence of these antibodies ( reviewed in [1 , 4] ) . While much effort is focused on using passively administered bNAbs and eliciting bNAbs through vaccination , less effort has been dedicated to understanding how viral escape will impact the utility of those approaches . Comprehensive insight into the ways HIV-1 can escape bNAbs , and methods by which this escape potential could be rapidly assessed , are critical to the use of bNAbs in humans . Without such insight , it is difficult to determine whether an antibody will be therapeutically useful , how it might be improved , whether it would work best in concert with other antibodies or antiviral drugs , or whether its epitope would be a useful target for a therapeutic or prophylactic vaccine . Most importantly , such information is necessary to determine whether the use of bNAbs in humans will easily promote emergence and spread of resistant variants . Although there are a number of reports describing in vivo bNAb escape mutations , such studies have been limited to only a few antibodies , and by no means comprehensive . Whereas these in vivo escape variants emerge rapidly , within a few weeks after antibody infusion [6–8 , 11 , 13 , 24–26] , conventional in vitro escape studies typically take much longer . These studies are slow because viral reverse transcriptase introduces only approximately one nucleotide mutation per viral genome per replication cycle . Because a typical antibody epitope is encoded by 1–2% of the genome , it takes at least 50 replication cycles to introduce a single relevant mutation into a virus population . To accelerate this process , Dingens et al . recently adopted codon mutagenesis , a scanning mutagenesis-based approach [44] . This method generates a library of the env gene encoding every possible single amino-acid change in the Env ectodomain . Although comprehensive , this approach has several disadvantages relative to the soft-randomization approach we used here . First , the costs and labor of library generation are markedly greater; library generation with codon mutagenesis requires hundreds of primer pairs and corresponding numbers of PCR reactions , whereas soft randomization uses only a single pair of primers and one PCR reaction per target region . More critically , with codon mutagenesis , one cannot identify escape variants bearing more than one mutation . It therefore can only identify short pathways of escape unique to the specific Env under study . In contrast , as we show here , soft randomization can identify many escape variants with multiple changes , better reflecting the pathways available to a huge number of highly diverse viruses in circulation . In fact , most previously characterized bNAb escape variants identified in vivo bear multiple mutation [6–8 , 11 , 13 , 24–26] . Nonetheless , our approach has one key disadvantage relative to codon mutagenesis , namely it requires structural knowledge of the epitopes of the bNAbs under investigation , and excludes escape pathways involving changes outside of those epitopes . In the case of HIV-1 Env and bNAbs , however , the amount of available structural and functional data largely compensates for this limitation . As both approaches and their limitations are complementary , in the future , they may be combined to provide maximum insight into viral escape . Here we extended soft-randomization to HIV-1 and diversified the regions encoding a key antibody epitope , and showed that this approach could create a library of functional replication-competent HIV-1 proviruses with a controlled numbers of amino acid substitutions in each Env . Using well-characterized VRC01-class antibodies as an example , we further showed that a disturbingly high number of such substitutions facilitated viral escape from those antibodies . In addition , a subset of the escape mutations identified in this study was also detected in in vivo studies and in natural isolates . For example , among the five residues found most frequently substituted in this study ( N276 , D279 , A281 , N461 and N465 ) , loss of glycosylation at N276 was also observed in escape variants derived from humans and humanized mice [25 , 27 , 45] . This glycan is highly conserved ( in 92% of known HIV-1 isolates , HIV Sequence Compendium 2017 [46] ) , and its loss further exposes the CD4-binding site to humoral immunity . However , loss of this glycan among escape variants is not unexpected because it participates in the binding of VRC01 antibodies [29 , 41 , 45] . On the other hand , glycosylation at N461 is present only in 27% of known isolates , indicating that mutations can be readily accommodated at this position . Over 60% of HIV-1 isolates listed in the HIV Sequence Compendium 2017 [46] has residues other than D at the position 279 , indicating that variations at this position would also not be difficult , as indicated by its rapid evolution in an infected individual [27] . The similarities between the escape variants from our study and those derived from humans or humanized mice validate the utility of a soft-randomized library approach for rapidly assessing viral escape from a bNAb . We observed a very high number of escape pathways from VRC01-class antibodies in vitro . This large number of escape pathways is important and problematic for the use of these antibodies in humans , for several reasons . First , it suggests that any effort to “checkmate” the virus by using a cocktail of antibodies , in which viruses escaped from one antibody are designed neutralized by others of the same class , is unlikely to be successful . There are simply too many possible escape pathways to cover them all . Second , our data make clear that the current panels of Env used to assess antibody breadth and potency do not in any way encompass the possible ways through which HIV-1 responds to bNAbs . Occasionally breadth is discussed as a surrogate for difficulty of escape , but our data indicate that these concepts are dissociable . For example , N6 is among the broadest CD4bs bNAbs thus far described , but the virus appears to have many available pathways to escape it ( Figs 4B and 6C ) . In fact , in some cases an antibody may be broad because it is rare , and thus its epitope has not been under pressure to diversify . Third , our data indicate that there is considerable overlap in the ways viruses can become resistant to different VRC01-class antibodies . For example , resistance to N6 readily emerged when viruses were selected against other VRC01-class members . Thus population-level escape from VRC01 may easily promote escape from 3BNC117 or N6 . Finally , because many different sets of substitutions can resist all VRC01-class antibodies , engineered vaccines designed to mainly elicit VRC01-class antibodies [47 , 48] may not sufficiently suppress population-level escape , and therefore may need to be supplemented by constructs designed to elicit bNAbs targeting complementary epitopes [49 , 50] . Although , as a proof of principle , we focused here only on VRC01-class antibodies and ADA isolate , HIV-1 likely has a similarly wide range of pathways of escaping other bNAbs . Indeed , escape might pose greater challenges with other bNAbs such as V2-loop/apex antibodies and 332-glycan antibodies , whose epitopes are less conserved and less functionally important than the CD4 binding site . One possible exception may be the antibodies recognizing the highly-conserved gp41 MPER epitope . A similar study with an MPER-region library may reveal greater constraints on escape than observed here for VRC01-class antibodies . In summary , soft-randomization of the key epitopes of HIV-1 Env is a useful approach for rapidly and extensively identifying antibody-resistant viruses , and thereby provides important insights into the propensity of a bNAb to promote viral escape and the potential pathways of escape . Its first application to the CD4bs bNAbs in this study extends previous observations and shows that there are a large number of discrete pathways by which HIV-1 can escape all VRC01-class antibodies .
Human embryonic kidney ( HEK ) -293T cells were obtained from the American Type Culture Collection ( ATCC , CRL-3216 ) and used to generate library and clonal viruses by transfection . The TZM-bl cells were obtained from NIH AIDS Reagent Program and used as an indicator cell line to measure the infectivity of various viruses [51] . Both cell lines were maintained in high-glucose Dulbecco’s minimal essential medium ( DMEM ) containing 10% fetal bovine serum ( FBS ) . GHOST cells are derived from the human osteosarcoma cells line , HOS . GHOST ( 3 ) CCR3+CXCR4+CCR5+ cells were obtained from NIH AIDS Reagent Program and used to passage virus library in the presence of antibodies , and were maintained in DMEM containing 10% FBS , 500 μg/ml G418 , 100 μg/ml hygromycin , and 1 μg/ml puromycin [52] . In later text , this media is referred as GHOST-cell complete media , and GHOST ( 3 ) CCR3+CXCR4+CCR5+ cells are referred as GHOST-R3/X4/R5 . All cells were grown at 37°C under 5% CO2 . The gene for the envelope glycoprotein ( Env ) of HIV-1 ADA ( GenBank AY426119 . 1 ) was cloned into pBR322 using SalI and BamHI sites , and used as a template to generate soft-randomized libraries . An AscI site was engineered between Loop D and V5 regions at amino acid positions 309–311 in order to combine independently generated Loop D and V5 libraries . The Env library from pBR322 was cloned into pNL4-3 proviral plasmid containing the env gene from ADA isolate ( pNL-ADA ) , using SalI and BamHI sites , to generate a replication-competent virus library . Loop D and V5 fragments containing the escape mutations identified by deep sequencing were synthesized by Integrated DNA Technologies ( IDT ) and cloned into pBR322 , and then into pNL-ADA . Plasmids encoding antibodies and CD4-Ig are described in Antibodies section . The broadly-neutralizing antibodies ( bNAbs ) used in this study are VRC01 , 3BNC117 , NIH45-46 , VRC07 , VRC-PG04 , VRC-PG20 , VRC-CH31 , b12 , N6 , 10E8 , 10–1074 , PGDM1400 . Of these , VRC01 , 3BNC117 , NIH45-46 , VRC07 , VRC-PG04 , VRC-PG20 , VRC-CH31 , b12 , N6 are CD4 binding-site ( CD4bs ) antibodies , 10E8 binds to the gp41 MPER epitope , 10–1074 binds a glycan on the V3 loop , and PGDM1400 recognizes an Env oligomer . IC80 and IC50 of these antibodies for ADA were obtained from the LANL database CATNAP ( http://hiv . lanl . gov/catnap ) [53] , and are provided in S2B Fig . The expressor plasmids for antibodies NIH45-46 , 3BNC117 and 10–1074 were kindly provided by Michael Nussenzweig ( The Rockfeller University ) . Those for VRC01 and 10E8 were obtained from AIDS Reagent Program , and the rest were constructed by cloning the genes for the heavy- and light-chain variable regions synthesized by IDT into the plasmids encoding the constant regions of the light chain and human heavy chain of IgG1 , as previously described [54] . GenBank numbers for the synthesized heavy and light chains are: VRC07 ( H , KT365998 . 1; L , KM408147 . 1 ) ; VR-PG04 ( H , JN159464 . 1; L , JN159466 . 1 ) ; VRC-PG20 ( H , KF515514 . 1; L , KF515513 . 1 ) ; VRC-CH31 ( H , JN159435 . 1; L , JN159438 . 1 ) ; N6 ( H , KX595108; L , KX595112 ) ; b12 ( H , AAB26315 . 1; L , AAB26306 . 1 ) ; PGDM1400 ( H , KP006370 . 1; L , KP006383 . 1 ) . All antibodies were produced in Expi293 Expression Medium ( Life Technologies ) by transfecting Expi293 cells with the corresponding expressor plasmids . Cells were grown for three days at 37°C with humidified air containing 8% CO2 on an orbital shaker platform rotating at 125 rpm . Proteins were purified from the culture supernatants using Protein A-Sepharose beads . The protocol for soft randomization was previously described [36] . In brief , primers were designed for Loop D ( amino acids 274–284 by HXB2 numbering ) and V5 ( amino acids 455–471 ) regions of ADA env gene . The soft-randomizing primer for Loop D was synthesized by hand-mixing 88% of the original nucleotide and 4% each of the other three nucleotides ( 88:4:4:4 ) for the first two positions of each codon of the 11 residues . A ratio of 91:3:3:3 was used for the longer ( 17 residues ) V5 region . For the wobble positions of the Loop D primer , an equimolar mix ( 50:50 ) of G and T was used for amino acids encoded by 4 or 6 codons . For the wobble positions of other codons , the ratios of 88:4:4:4 ( for Loop D ) or 91:3:3:3 ( for V5 ) was used . Both the 5’ and 3’ ends of the primers were extended outside the soft-randomized regions to match the melting temperature of pairing primers . Soft-randomizing primers were phosphorylated at the 5’ ends . The sequences of the soft-randomizing primers and pairing primers are shown in Fig 1B . The Loop D and V5 soft randomization was performed independently via whole-plasmid PCR of pBR322 carrying SalI-BamHI fragment of ADA env gene , using Q5 HotStart DNA Polymerase ( New England Biolabs ) . 1 μg of DpnI-digested and purified PCR product was ligated in 40 μl with 0 . 1 μl of concentrated T4 ligase ( New England Biolabs ) at 16°C overnight . 1 μg of ethanol-precipitated ligate was electroporated at 1700V into 25 μl of Electrocompetent NEB 10-beta ( New England Biolabs ) , and the culture was grown in S . O . C . media with shaking at room temperature for 1 . 5 hours . The culture was then added to 300 mL Lennox broth containing 50 μg /ml ampicillin and grown with shaking until the culture reached an OD600 of 1 . 7 . Chloramphenicol was added to final 170 μg/mL and the culture was grown for additional 36 h . To combine Loop D and V5 libraries , V5 library was subcloned into pBR322-Env ( ADA ) containing Loop D library , using AscI , engineered between Loop D and V5 regions , and BamHI sites . Ligation , DNA precipitation , electroporation , and plasmid preparation were performed as described above . pNL-ADA library was then constructed by subcloning the SalI-BamHI fragment of pBR322-Env ( ADA ) containing soft-randomized Loop D and V5 regions . Ligation , DNA precipitation , electroporation and plasmid preparation were again performed as described above . A total of 24 x 300 ml cultures , each 300 ml of which was derived from 1 ug ligate electroporated into 25 μl of NEB 10-beta , were grown to an OD600 0 . 8–1 . 0 . DNA was prepared as described above , pooled and used as the final pNL-ADA-Lib . WT pNL-ADA and its library , and all clonal viruses ( sequences are shown in S3 Fig ) were produced by calcium phosphate transfection of HEK-293T cells with corresponding plasmids , and cells were grown in DMEM containing 10% FBS . The virus-containing supernatants were harvested 48 hours post infection ( hpi ) , aliquoted , and stored at -80°C . All experiments involving replication-competent viruses were performed in biosafety level 3 laboratory following the protocols approved by the Institutional Biosafety Committee of The Scripps Research Institute . Library swarms , produced by transient transfection of HEK-293T described above , or obtained from a previous passage were incubated in GHOST-cell complete media for 30 minutes with indicated concentrations of NIH45-46 , 3BNC117 or VRC07 antibody . Antibody concentrations used for selection are provided in S2A Fig for each antibody and passage . For this study , we chose virus dilution that yielded 60–70% infection of GHOST cells at 48–72 hpi . The virus-antibody mixture was then added to GHOST-R3/X4/R5 cells pre-seeded at ~40% confluence in 10 x T75 flasks . After 6–8 h , cells were washed twice with Phosphate-Buffered Saline ( PBS ) and further incubated in GHOST-cell complete media containing the same concentration of an antibody used for selection . The Tat-regulated GFP expression in GHOST cells was used to assess virus infection levels . When infection level reaches 60–70% , the virus-containing culture supernatants were harvested , spun to remove cell debris , aliquoted , stored in the -80°C , and used for subsequent viral passages or RNA extraction for deep sequencing . RNA was extracted from 250 ul of cell-free culture supernatants of library or antibody-escaped swarms , using RNAqueous Total RNA Isolation kit ( Ambion , ThermoFisher Scientific ) . cDNA was synthesized using High Capacity cDNA Reverse Transcription Kit ( Applied Biosystems ) using a gene-specific primer ( 5’ GAACCCAAGGAACATAGCTCCTATC 3’ ) . An env fragment encompassing Loop D and V5 regions was amplified via 18 cycles of PCR reaction using Takara Taq Hot-Start DNA polymerase ( Takara Bio Inc . ) and ADA-Env-F3 ( 5’ GGCAGTCTAGCAGAAGAAGAGGTAGTAATTAG 3’ ) and ADA-Env-B3 ( 5’ CACTTCTCCAATTGTCCCTCATATCTCCTC 3’ ) primers . The PCR products ( amplicons ) were purified by AMPure XP beads ( Beckman Coulter ) at a bead to sample ratio of 3:2 , quantified using Qubit dsDNA HS ( high sensitivity ) assay kit ( Life Technologies ) and run on the DNA high sensitivity chip in the Bioanalyzer 2100 ( Agilent ) to confirm their length . 100ng of these amplicons were then end-repaired , 5’ phosphorylated , and A-tailed at the 3’ ends using TruSeq Nano kit ( FC-121-4001 , Illumina ) . Next , these amplicons were ligated with illumina barcoded-partial adapters , size selected using AMPure XP beads , and further amplified for 4 PCR cycles . The final products were validated on the Bioanalyzer 2100 and quantified using Qubit dsDNA HS assay , pooled at equimolar and sequenced by paired-end 150bp reads on the Illumina MiSeq at the Scripps Genomics Core Facility in La Jolla , California . The reads were trimmed of adapters , and cropped to isolate Loop D and V5 fragments . These sequences were translated , filtered for open-reading frames by eliminating those with stop codons and unidentifiable amino acids , and analyzed using in-house python scripts . To avoid removal of rare but intended mutations , reads were not corrected for the low level of errors that are potentially introduced during reverse transcription , PCR and sequencing . Enrichment was determined by normalizing the copy number of each sequence found in VRC07 passage 5 swarm by the copy number of the same sequence detected in the control passage 15 swarm . To include the sequences in normalization , which were detected in VRC07 passage 5 swarm but not in the control swarm , a control copy number of 1 is added to all sequences . Neutralization assays were performed in TZM-bl cells , as described previously [51] . To use TZM-bl cells , antibiotics were removed from virus stocks by growing them in GHOST cells in the absence of antibiotics . This process also removed antibodies present in the culture supernatants . Virus-containing supernatants were harvested at 48–72 hpi , spun to remove debris , aliquoted , and used for neutralization assays . Viruses were incubated with 0–10 ug/ml of antibodies at room temperature for 30 minutes and then added to TZM-bl cells plated at 10 , 000 cells per 96 well one day before the assay . At 48 hpi , infection levels were measured by luciferase assays using Luc-Pair Firefly Luciferase HS assay kit ( GeneCopoeia ) . The relative light unit was read at 575 nm using a Victor X3 plate reader ( PerkinElmer ) . To assess virus fitness , the growth of WT virus or virus clones that are resistant to bNAbs was assessed in CD4+ T cells . Cryopreserved human peripheral blood mononuclear cells ( PBMC , StemCell Technologies ) , were thawed and incubated in RPMI containing 15% FBS and 20 U/ml human IL-2 ( Roche ) for overnight . Next day , cells were enriched with CD4+ T cells population by negative selection using Human CD4+ T Cell Isolation kit ( Biolegend ) and activated with 1 μg/ml PHA-L ( Sigma ) for 48 hours in RPMI supplemented with 15% FBS , 20 U/ml IL-2 at 1 x 106 cells/ml . Viruses ( 5 x 108 genome copy number , quantified by RT-qPCR ) were pre-incubated for 20 min at room temperature with or without 10 μg/ml VRC07 in 100 μl RPMI supplemented with 15% FBS and 20 U/ml IL-2 , and added to 1 . 5 x 105 PHA-L-activated CD4+ T cells in 150 μl . After 6 h incubation at 37°C , cells were washed with PBS and resuspended in 500 μl fresh RPMI containing 15% FBS with or without 10 μg/ml VRC07 . Every 2 days , 180 μl of supernatant was harvested and replaced with same amount of fresh media . RNA was extracted from 150 μl these supernatants using TRIzol LS ( Ambion ) , and cDNA synthesized using High Capacity cDNA Reverse Transcription kit ( Applied Biosystems ) and an ADA Env-specific primer ( 5’-GAACCCAAGGAACATAGCTCCTATC-3’ ) . Probe qPCR was performed using iTaq Universal Probes Super mix ( Bio-Rad ) , ADA-Env-qPCR-sense ( 5’-CAAAGCCTAAAGCCATGTGTAAA-3’ ) and ADA-Env-qPCR-antisense ( 5’-CTCCTCTCATTCCCTCACTACTA-3’ ) , primers , and ADA-Env-qPCR probe ( 5’-/56-FAM/CCCATCCTG/ZEN/TGTTACTTTAAATTGCACTGA/3IABkFQ/-3’ ) in CFX96 Touch Real-Time PCR Detection System ( Bio-Rad ) . The frequency of mutations in the Loop D and V5 regions is analyzed using the AnalizeAlign tool available at LANL ( www . hiv . lanl . gov ) . The sequences included in the analyses are: 5471 natural isolates available at LANL database , 298 Loop D and 250 V5 sequences of in vivo escape variants previously identified in the clinical trials for CD4bs antibodies , VRC01 and 3BNC117 [6 , 7 , 11 , 12] , and the top 150 escape variants identified in this study . Because of the large number of parental sequences in in vivo studies , parental residues are excluded from the analyses of in vivo escape mutations . Insertion mutations found in some of in vivo variants are also excluded . Statistical analysis of the data was performed using GraphPad Prism software . The difference between groups for all neutralizing assays and virus growth curves in CD4+ T cells were tested using a two-way ANOVA . The null hypothesis was rejected when p<0 . 05 in all cases .
|
Several potent antibodies against human immunodeficiency virus type 1 ( HIV-1 ) have been evaluated in clinical trials . Use of these antibodies in humans , however , is problematic , because easy viral escape remains a major concern . To gain greater insights , we sought to develop an approach to rapidly assess the likelihood of viral escape from such antibodies . We show here that soft-randomization mutagenesis is a suitable approach to introduce a controlled number of changes into defined target regions . As a proof of concept , we used this approach to detect the HIV-1 variants fully resistant to VRC01-class of antibodies . We observed that within a few passages of the soft-randomized library of viruses in the presence of potent HIV-1 antibodies , a remarkably wide array of variants emerged , including variants resistant to every VRC01-class antibody . This study provides insights into a wide range of escape pathways , and describes a method for rapidly assessing the likelihood of viral escape from antibodies or small molecules targeting the HIV-1 envelope glycoprotein .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"medicine",
"and",
"health",
"sciences",
"microbial",
"mutation",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"immunology",
"microbiology",
"cloning",
"retroviruses",
"viruses",
"immunodeficiency",
"viruses",
"mutation",
"substitution",
"mutation",
"rna",
"viruses",
"molecular",
"biology",
"techniques",
"antibodies",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"immune",
"system",
"proteins",
"artificial",
"gene",
"amplification",
"and",
"extension",
"bioinformatics",
"proteins",
"medical",
"microbiology",
"hiv",
"microbial",
"pathogens",
"biological",
"databases",
"hiv-1",
"viral",
"replication",
"molecular",
"biology",
"biochemistry",
"sequence",
"databases",
"polymerase",
"chain",
"reaction",
"virology",
"physiology",
"viral",
"pathogens",
"database",
"and",
"informatics",
"methods",
"genetics",
"biology",
"and",
"life",
"sciences",
"lentivirus",
"organisms"
] |
2018
|
Diverse pathways of escape from all well-characterized VRC01-class broadly neutralizing HIV-1 antibodies
|
A dengue early warning system aims to prevent a dengue outbreak by providing an accurate prediction of a rise in dengue cases and sufficient time to allow timely decisions and preventive measures to be taken by local authorities . This study seeks to identify the optimal lead time for warning of dengue cases in Singapore given the duration required by a local authority to curb an outbreak . We developed a Poisson regression model to analyze relative risks of dengue cases as functions of weekly mean temperature and cumulative rainfall with lag times of 1–5 months using spline functions . We examined the duration of vector control and cluster management in dengue clusters > = 10 cases from 2000 to 2010 and used the information as an indicative window of the time required to mitigate an outbreak . Finally , we assessed the gap between forecast and successful control to determine the optimal timing for issuing an early warning in the study area . Our findings show that increasing weekly mean temperature and cumulative rainfall precede risks of increasing dengue cases by 4–20 and 8–20 weeks , respectively . These lag times provided a forecast window of 1–5 months based on the observed weather data . Based on previous vector control operations , the time needed to curb dengue outbreaks ranged from 1–3 months with a median duration of 2 months . Thus , a dengue early warning forecast given 3 months ahead of the onset of a probable epidemic would give local authorities sufficient time to mitigate an outbreak . Optimal timing of a dengue forecast increases the functional value of an early warning system and enhances cost-effectiveness of vector control operations in response to forecasted risks . We emphasize the importance of considering the forecast-mitigation gaps in respective study areas when developing a dengue forecasting model .
This study aims to identify the optimal timing for issuing a dengue early warning that will allow sufficient time for the local authority in Singapore to execute preventive measures to mitigate the potential risks . Our objectives are to 1 ) establish links between the estimated forecast lead time ( lag time ) and the time frame required by the local authorities for successful mitigation , 2 ) analyze time gaps between dengue forecast and successful mitigation , 3 ) suggest an optimal dengue forecast lead time that provides sufficient time for successful mitigation , and 4 ) identify possible factors influencing the gap between dengue early warning and mitigation .
Singapore is an island state nation with an area of about 700 km2 and population density of approximately 7000 per km2 . The nation experiences tropical climate , with temperature between 25°C and 30°C whole year round . The average annual rainfall is 2 , 200 mm , with the cooler monsoon season from November to January contributing 37% of the precipitations . We identified the optimal time for dengue early warning by analyzing 1 ) lead time based on the risk of increasing dengue cases in each lag time between dengue and weather predictors and 2 ) the time frame required by local authorities to mitigate the risk of dengue outbreak using retrospective data on duration of vector control in dengue clusters . First , we developed a Poisson regression model to analyze the relative risks of dengue cases as functions of mean temperature and cumulative rainfall at lag times of 4–20 weeks . We determined the lag times between weather predictors and dengue cases based on cross correlation function ( CCF ) and literature review on the effects of weather on vectors and dengue transmission [17] , [18] , [19] , [20] , [21] . Current number of dengue cases could be influenced by the number of cases in the past . We analyzed the period of this influence ( autoregressive terms ) using autocorrelation function ( ACF ) , partial autocorrelation function ( PACF ) , and literatures on dengue disease transmission [1] , [35] . We applied cubic spline function on temperature and rainfall to allow a non-linear exposure and response association between weather predictors and dengue cases . We included smoothing spline of time trend from week 1 of 2000 to week 52 of 2010 with 11 degrees of freedom ( df ) to account for other influences such as circulating dominating dengue virus serotypes and vector control measures that could potentially confound the relationship between weather predictors and weekly dengue cases . Sensitivity of degrees of freedom on the effects of the smoothing function of time trend was tested with df range from 7–16 . Furthermore , we offset midyear population to consider annual population movement and used Quasi-Poisson regression to allow for over-dispersion of data . We also estimated incidence rate ratio ( IRR ) between weather predictors and dengue cases using piecewise linear spline function . A best fit model was selected and validated using Akaike's Information Criterion ( AIC ) , Generalized Cross-Validation ( GCV ) score , post estimation PACF , residuals diagnosis including normality and sequence plots . Data inferences were based on 95% confidence interval using R [36] and STATA 11 ( StataCorp . , Texas , USA ) . Model:μt represents predicted mean dengue cases in week; ar denotes autoregressive terms ( 2–6 weeks ) of dengue cases; s ( temp ) denotes cubic spline of mean temperature with 3 df; s ( rain ) represents cubic spline of cumulative rainfall with 3 df; t = week; i = lag times of 4 , 8 , 12 , 16 , and 20 weeks; s ( trend ) means smoothing spline of time trend in weeks with 11 degrees of freedom; pop = midyear population Next , we examined the distribution of cases in identified dengue clusters , percentage of cluster-associated dengue cases , and the duration of dengue cluster management or vector control operations corresponding to the number of dengue cases in each cluster . We used the findings as indicators of estimated time or duration required for a successful mitigation . Finally , we evaluated the time gaps between possible forecast windows ( lag times ) and estimated time required for mitigation . Forecast-mitigation deficit or surplus was defined as the negative or positive time difference between forecast window and duration for mitigation .
Our findings show that increasing weekly mean temperatures and cumulative rainfall precede risks of increasing dengue cases by 4–20 and 8–20 weeks , respectively ( Figure 1 ) . Each degree increase of mean temperature from 25 . 5°C–28 . 2°C elevated risk of dengue almost linearly at lag week 4–16 with peak at lag week 12; while an inverse relationship was observed between mean temperature and dengue cases at lag week 20 . Figure 1 also shows mean temperature above 28 . 2°C raised the risk of dengue cases at lag weeks 4–20 with highest risk at lag week 16 . Overall , the highest risk of dengue as a function of mean temperature occurred at lag week 12 followed by week 16 . Simultaneously , each unit increase of weekly cumulative rainfall below 60 mm and above 150 mm elevated the risk of dengue cases at lag weeks 8–16 and weeks 12–20 , respectively . Likewise , Table 2 shows each unit increase of weekly mean temperature raises higher incidence rate ratios for dengue cases at lag week 12 ( IRR = 1 . 46 ) and week 16 ( IRR = 1 . 39 ) . Overall rate ratios for dengue cases in response to one mm rise of weekly cumulative rainfall peaked at lag week 16 ( IRR = 1 . 011 ) and every unit increase of cumulative rainfall below 60 mm and above 150 mm elevated risks of dengue cases by 0 . 6% and 0 . 8% , respectively . Our model explained about 91% of the variance in dengue cases using weather predictors , time trend , and past cases . Residuals diagnoses and PACF indicated that the model was fit for analysis with predicted cases against observed cases as shown in Figure 2 . Sensitivity tests using various degrees of freedom on the spline function of trend showed little change in risk functions . In the past two decades , the total number of cluster-associated dengue cases contributed an average 27% of overall reported dengue cases for 1990–1999 and 30% for 2000–2010 ( Figure 3 ) . Since 2000 , the proportion of cluster-associated dengue cases had been on an upward trend with a peak of 47% in year 2007 . Whereas , dengue clusters that reported a minimum of 10 cases represented approximately a third of the total cluster-associated cases . From 2000–2010 , the mean and median numbers of cases per cluster ( > = 10 cases ) were 22 and 17; mean and median numbers of cases were 21 and 16 for non-epidemic years and 23 and 19 for epidemic years ( 2004 , 2005 , and 2007 ) , respectively ( Figure 4 ) . As shown in Figure 4 , all the dengue clusters of 30 or less cases fell in the 75th percentile , except in years 2002 and 2005 when only 70% of clusters had fewer than this number . The differences in time required for dengue cluster control between non-epidemic and epidemic years is minor . Figure 5 indicates most of the dengue clusters have fewer than 30 cases and take up to 2 months to control dengue outbreaks in both non-epidemic and epidemic years . During the study period , approximately 23% ( non-epidemic = 28% , epidemic = 16% ) of the dengue clusters were managed within 1 month , 64% ( non-epidemic = 60% , epidemic = 71% ) was managed within 2 months , and 13% ( non-epidemic = 13% , epidemic = 13% ) required maximum 3 months of vector control and cluster management to curb outbreaks ( Figure 5 ) . Longer duration ( 2–3 months ) for cluster management was required as the number of dengue cases in each cluster exceeded 36 and 30 for non-epidemic and epidemic years , respectively . Overall , the mean and median cluster duration was about 2 months for both non-epidemic and epidemic years . The usefulness of a dengue early warning may be reflected in the time between a forecast window and the time required by local authority to mitigate an outbreak . A forecast-mitigation deficit occurs when the duration for successful mitigation exceeds the forecast window . In view of the maximum period required to curb transmission , a 3 month forecast is deemed appropriate . Incidentally , the highest risk of dengue as a function of mean temperature and rainfall occurs at lag week 12–16 . Forecasting on the basis of the observed mean temperature and cumulative rainfall 12 to 16 weeks previously could therefore provide accurate warnings while allowing sufficient time for local authorities to mitigate , or even avoid , an impending outbreak .
An early dengue warning system that allows time for successful mitigation will enhance effectiveness of preventive measures . Our findings show that a rise in weekly mean temperature and rainfall precede risks of increasing dengue cases by 1 to 5 months with higher risks being evident at 3–4 months . The lag times could partly be explained by high desiccation resistance of Aedes mosquito's eggs which could survive several months without water [37] . Our results are consistent with studies in Singapore that analyzed relationship between weekly temperature and dengue cases up to 20 weeks [4] , [38] . A study in Bangkok shows that temperature and rainfall precede dengue cases up to 6 months and 3 months , respectively [39] . Using the average duration of a dengue cluster as an indicator of the period needed for successful mitigation of transmission in a localized area , our analysis has shown that the local authority typically required an average of 2 months with maximum 3 months for effective mitigation in both non-epidemic and epidemic years . As cluster management could reflect the national situation at a local level , we suggest that a similar period is required for vector source reduction to prevent an outbreak island wide . The limitations of this assumption are: 1 ) the time required for vector control in respond to an early warning might be shorter compare to the time needed to control vectors during an outbreak as was measured here; 2 ) the manpower resources allocated per unit area for an island wide source reduction effort may be smaller than those committed to local outbreak control . Nevertheless , considering that the two limitations could marginally reduce and increase the period needed for island wide preventive measures respectively , an estimation of 3 months is considered a reasonable period required for planning and implementing preventive measures . Moreover , the time lag between onset of dengue fever among cases and the identification of a dengue cluster could be a crucial but as yet unmeasured factor in outbreak prevention . Therefore , if a dengue early warning was in place , it is possible that the maximum mitigation duration could be even less than 3 months . A dengue early warning at the optimal time boosts the success of vector control operations and cost-effectiveness of intervention . A study by Oki et al . ( 2011 ) has suggested that optimal timing of a vector control such as insecticide fogging increases the impacts of intervention on reduction of dengue cases [40] . Likewise , optimal timing of an early warning of dengue outbreak inevitably increases the impact of the early warning on the effectiveness of preventive measures . Dengue control programs are an economic burden to government and communities . During 2000 to 2009 , dengue control in Singapore cost approximately US $500 million to the nation [41] . Effective vector controls could help to minimize the economic burden possibly by reducing the number of dengue cases , preventing loss of working days and income due to disease , increasing saving on disability-adjusted life years ( DALY ) and boosting effectiveness of each dollar spent on intervention [42] . The duration of dengue cluster management could be influenced by a complex spatiotemporal interplay of risk factors unique to respective dengue clusters: 1 ) Demographic characteristics , density , and herd immunity among host population or residents in clusters could have an influence on the number of dengue cases . A community with lower herd immunity could possibly experience dengue epidemic with a low mosquito-population density [3] , [43] , 2 ) Environmental factors such as conditions , types and ages of structural buildings , construction activities , public drainage systems , and presence of parks possibly increase the challenges and prolong the duration of vector control operations in certain clusters , 3 ) Differential vector populations due to inaccessibility of larval breeding sites , larval indices , density of adult Aedes aegypti , and numbers of potential breeding sites complicate the dynamics of dengue transmission and require greater effort for outbreak control in respective clusters , 4 ) With the co-circulation of all four serotypes of dengue virus in Singapore , the circulating serotypes and herd immunity to the specific serotype of concern also influence the number of cases in clusters , 5 ) Community commitment to prevent dengue has a direct impact on the duration of vector control . Residents not granting permission to dengue officers to enter their premises to conduct mosquito larvae inspection and elimination could prolong the duration as well as reduce effectiveness of vector control measures . Since 2005 , Singapore has monitored circulating dengue virus to detect switches in predominant serotype which could signal an impending outbreak . More recently , virus genetic data has shown that clade replacement without a switch in serotype , could also lead to increase in cases [44] , [45] . Here we determine that 3 months lead time could be optimal for a warning to be issued; and that temperature and rainfall data could provide a forecast in that timeframe , to allow for preparation of control measures . Intervention measures include public and stakeholders' engagement , gaining political support and systematic source reduction exercises . The findings in this study were geographically based because of the heterogeneity of the environment including local weather , circulating viral serotypes , and herd immunity in the respective study areas . Owing to the fact that a model-based dengue early warning system has not previously been adopted for dengue surveillance in the study area , we utilized existing available vector control and cluster duration data as indicative references . Further studies could be undertaken to evaluate the forecast-mitigation gaps more accurately when such data are available in the future . Our study identified a short term forecast window of 3–4 months . In an environment where long duration of response or mitigation is anticipated , the effectiveness of a dengue early warning can be improved by re-considering the forecast-mitigation gap . One approach is to forecast dengue outbreaks with a longer lead time using other predictors of weather . Several studies in various geographical areas have revealed the feasibility of forecasting dengue cases several months in advance using weekly or monthly weather predictors and up to 10 months ahead using Southern Oscillation Indices ( SOI ) , El Niño Index , or El Niño Southern Oscillation indices ( ENSO ) [23] , [46] , [47] , [48] . Although long term forecast could possibly be compromised by lower forecast precision , the long lead time may be useful for longer term planning such as allocation of resources and acquisition of control tools such as insecticides . For several decades , dengue control has been a challenge for regions where dengue is endemic . Optimal timing of an early warning could help to bridge the forecast-mitigation gap between theoretical research predictions and practical control operations . Identifying the optimal lead time for dengue forecast and duration of local vector control could help to improve the functional aspects of a dengue forecasting model , reduce risk of dengue epidemic , increase cost-effectiveness of control strategies , and encourage local authorities to adopt a model-based dengue early warning . The lead time needed for mitigation varies according to different influencing factors in respective study areas . We emphasize the importance of considering the forecast-mitigation gaps in respective study areas when developing a dengue forecasting model .
|
A dengue early warning system that would provide an accurate forecast could enhance the effectiveness of dengue control , but only if it is given in sufficient time for local authorities to implement those control operations . In this study , we have suggested the optimal timing for issuing a warning of a dengue outbreak in Singapore that will allow authorities adequate time to respond . We first analyzed the relationship between the risk of dengue cases and weather predictors at 1–5 month lag times to gauge the possible lead time for providing an accurate dengue forecast . We then determined the average time needed for local authorities to curb the outbreak of clusters of 10 dengue cases or more using vector control and cluster duration records for the period 2000–2010 . Increasing weekly mean temperature and cumulative rainfall preceded a rise in dengue cases up to 5 months with higher risks evident at a lag time of 3–4 months . Local authorities required an average of 2 months with a maximum of 3 months for effective control . Therefore , a dengue early warning given at least 3 months ahead of time would provide sufficient time for local authorities to moderate an outbreak .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"infectious",
"diseases",
"public",
"health",
"and",
"epidemiology",
"epidemiology",
"global",
"health",
"neglected",
"tropical",
"diseases",
"infectious",
"disease",
"control",
"infectious",
"disease",
"modeling",
"public",
"health"
] |
2012
|
Optimal Lead Time for Dengue Forecast
|
Genome-scale metabolic network reconstructions ( GENREs ) are repositories of knowledge about the metabolic processes that occur in an organism . GENREs have been used to discover and interpret metabolic functions , and to engineer novel network structures . A major barrier preventing more widespread use of GENREs , particularly to study non-model organisms , is the extensive time required to produce a high-quality GENRE . Many automated approaches have been developed which reduce this time requirement , but automatically-reconstructed draft GENREs still require curation before useful predictions can be made . We present a novel approach to the analysis of GENREs which improves the predictive capabilities of draft GENREs by representing many alternative network structures , all equally consistent with available data , and generating predictions from this ensemble . This ensemble approach is compatible with many reconstruction methods . We refer to this new approach as Ensemble Flux Balance Analysis ( EnsembleFBA ) . We validate EnsembleFBA by predicting growth and gene essentiality in the model organism Pseudomonas aeruginosa UCBPP-PA14 . We demonstrate how EnsembleFBA can be included in a systems biology workflow by predicting essential genes in six Streptococcus species and mapping the essential genes to small molecule ligands from DrugBank . We found that some metabolic subsystems contributed disproportionately to the set of predicted essential reactions in a way that was unique to each Streptococcus species , leading to species-specific outcomes from small molecule interactions . Through our analyses of P . aeruginosa and six Streptococci , we show that ensembles increase the quality of predictions without drastically increasing reconstruction time , thus making GENRE approaches more practical for applications which require predictions for many non-model organisms . All of our functions and accompanying example code are available in an open online repository .
Metabolism is the driving force behind the wondrous flurry of biological activity carpeting our planet . An organism’s metabolism is determined by the metabolic enzymes encoded in its genome , the chemical reactions catalyzed by those enzymes , and whether or not those enzymes are actively expressed [1] . The simplest bacteria have hundreds of metabolic enzymes , while the most complex eukaryotes have thousands . The products of these enzymatic reactions serve as substrates for other reactions , such that the chemical transformations carried out in a cell can be represented as a vast network [2] . Mass and energy flow through such networks , transforming environmental inputs into the building blocks of life . Every species has a unique metabolic network driving its growth and interaction with the environment . Genome-scale metabolic network reconstructions ( GENREs ) are formal representations of metabolic networks [3] . GENREs serve as comprehensive collections of metabolic knowledge about particular organisms and are amenable to mathematical analysis [4] . The process of reconstructing a GENRE takes months to years , but the reconstruction process often leads to new discoveries [5] . Mathematical analysis of GENREs gives insight into how particular metabolic pathways are used by an organism , what substrates it can utilize , which of its genes are essential in a given environment , how a metabolic network can be engineered to produce more of a desired product , or which enzymes within the network should be targeted in order to halt growth of an organism [6–8] . The reconstruction and analysis of GENREs for single species has greatly contributed to the understanding of microbes and the ability to engineer them . Recently , analyses have been developed which predict metabolic interactions between microbes [9 , 10] . However , the application of these recent analyses has been greatly limited by the large investment in time required to reconstruct a useful GENRE . Many microbial communities of interest consist of hundreds of species [11 , 12] . It is decidedly impractical to spend decades manually curating hundreds of GENREs . Many automated methods have been developed for rapidly reconstructing GENREs [13–16] . We present a novel ensemble method that is complimentary to these existing automated methods which we refer to as Ensemble Flux Balance Analysis ( EnsembleFBA ) . EnsembleFBA pools predictions from many draft GENREs in order to more reliably predict properties that arise from metabolic network structure , such as nutrient utilization and gene essentiality ( Fig 1 ) . The primary benefits of this new method are that it relies on automatically-generated GENREs ( which can be generated in a matter of minutes to hours ) and yet produces more reliable predictions than individual GENREs within the ensemble . We implement and discuss one possible way of generating useful ensembles , but emphasize that other automated methods could be modified to generate useful ensembles . We begin by discussing a common GENRE curation procedure known as gap filling . We demonstrate that a global gap filling procedure ( bringing the network into agreement with all positive growth conditions simultaneously ) does not perform any better than a sequential one . Instead , we introduce an ensemble approach to pool the many possible network structures resulting from different sequences of the input media conditions ( Fig 1 ) . We demonstrate that an ensemble reliably outperforms most of its constituent GENREs in terms of predicting growth and gene essentiality . By tuning the stringency of the voting threshold ( e . g . requiring a majority of GENREs to agree vs . complete consensus ) it is possible to achieve greater precision or recall than any of the constituent GENREs . We show how additional steps to increase the diversity among GENREs within the ensemble ( e . g . reconstructing each member GENRE using subsets of the available data ) can further improve recall . Furthermore , we found that incorporating negative growth information into our GENREs improved overall accuracy of the ensemble . We present proof of concept of the use of ensembles by predicting carbon source utilization and gene essentiality in Pseudomonas aeruginosa , a well-studied , clinically-relevant pathogen . We provide an example workflow using EnsembleFBA by predicting gene essentiality in six Streptococcus species and mapping the predicted essential genes to small molecule ligands in DrugBank . All of our data and code are available in an online repository , including example scripts to simplify adoption of EnsembleFBA . Our ability to make mechanistic predictions about complex cellular communities requires advances in the way we leverage the data available to us , and the way we handle uncertainty . Ensemble FBA is a novel tool that maintains the speed of automated reconstruction methods while improving predictions by intentionally managing uncertainty in network structures .
Gap filling is the process of identifying mismatches between computational predictions and experimental results , and identifying changes to the network structure which will bring the computational predictions into agreement with the experimental data . “Gaps” are missing reactions and can be filled by drawing from a database of possible metabolic reactions . Given that there are usually many mismatches between computational and experimental results , we demonstrate that simply changing the order in which computational results are brought into agreement with experimental data can result in different network structures . For example , suppose that it is experimentally determined that a microbe can grow on glucose minimal media and sucrose minimal media , but the computational predictions do not match . Gap filling the GENRE against a representation of glucose minimal media first and sucrose minimal media second , may result in a different network in the end than if gap filled against a representation of sucrose minimal media first . In practice , the order of gap filling is often arbitrary . We implemented a custom gap fill algorithm based on the algorithms FASTGAPFILL and FastGapFilling ( see Materials and Methods ) [17 , 18] . We used the Model SEED biochemistry database as our “universal” reaction database from which to draw reactions for gap filling [13] . We used the Model SEED web interface to automatically generate a draft GENRE for Pseudomonas aeruginosa UCBPP-PA14 ( without using the Model SEED gap filling feature ) . We gap filled this draft GENRE using 2 , 5 , 10 , 15 , 20 , 25 and 30 media conditions that experimentally support growth . For each number of media conditions , the gap filling procedure was repeated 30 times . Each replicate consisted of drawing a random set of media conditions of the appropriate size , generating two random permutations of the drawn media conditions , gap filling in the order specified by those two permutations , and finally , comparing the resulting gap filled networks to each other . For example , for a single replicate of 5 media conditions , we selected five media conditions at random and selected two random permutations of these conditions ( in this case , there are 120 possible permutations ) . We gap filled twice , in the order prescribed by the two permutations , and compared the resulting networks . We repeated the process 30 times , each time drawing a new set of five random media conditions and gap filling using two random permutations of those five media conditions . We found that even with as few as two media conditions , gap filling in a different order resulted in an average of 25 unique reactions per GENRE ( Fig 2 ) . As the number of media conditions increased , so did the average difference between the resulting GENRES . We hypothesized that rather than gap filling sequentially , perhaps a “global” gap fill approach would result in more parsimonious , biologically-relevant solutions without the ambiguity associated with changing the gap fill order . We extended our custom gap fill algorithm to identify a minimal set of reactions which could be added to a GENRE to permit growth in multiple media conditions simultaneously ( see Materials and Methods ) . We started with the P . aeruginosa UCBPP-PA14 draft GENRE from the Model SEED and repeated the 30 replicates from 2 to 30 media conditions as above , but using the global gap fill approach that we developed ( see Materials and Methods; Fig 3 ) . For each iteration , a subset of media conditions of the appropriate size was selected , the order randomized , and one network created using sequential gap filling while the other was created using global gap filling . We found that this global approach did not identify solutions that were any more parsimonious ( Fig 3A ) , and lead to dramatic increases in solve times with increasing media conditions ( Fig 2B ) . The decrease in parsimony with the global approach is counterintuitive , and arises due to the use of a linear rather than an integer optimization approach [17] . However , a linear approach is necessary to solve the global problem for anything more than a few media conditions . In order to determine whether the global solution was any more “biologically-relevant” , we also compared the ability of the global and sequential approaches to reconstruct a GENRE for P . aeruginosa UCBPP-PA14 . These draft reconstructions were then compared to a well-curated GENRE called iPAU1129 ( Bartell et al . In press ) . For each iteration ( 30 total ) , we randomly removed 20% of reactions from iPAU1129 and used the sequential and global approaches to gap fill from the universal database using a random selection of five media conditions . We chose to remove 20% of reactions to leave enough network context to constrain gap filling , as has been done previously [17] . The resulting networks were compared to the original iPAU1129 , under the assumption that the most biologically-relevant approach would most faithfully reconstruct the curated GENRE , iPAU1129 . We found no statistically significant difference between the two approaches ( Fig 3C ) ( p-value = 0 . 63 by two-sided , paired Wilcoxon signed rank test ) . Because the sequential gap filling approach produces different results depending on the order of gap filling , we chose to maintain many possible structures resulting from random permutations of the input media conditions rather than select a single GENRE structure for downstream analysis ( given that all candidate network solutions are equally consistent with the available data ) . Not knowing the “true” network structure , we considered each different structure to be a “hypothesis” and analyzed them collectively . For each of 2 to 30 training media conditions we produced 21 GENREs by randomizing the gap fill order ( S1 Fig ) . We then evaluated each GENRE individually by predicting growth or no growth on 34 test media conditions ( 17 media conditions which experimentally supported growth and 17 which did not ) using flux balance analysis ( FBA ) . We found that each GENRE produced slightly different growth predictions , resulting in some GENREs being more accurate than others ( Fig 4 ) . In order to generate predictions using the ensemble , we treated each GENRE’s prediction as a single vote , and pooled the votes using a threshold ( Fig 1B ) . We tested three qualitatively different thresholds; “any” , “majority” , and “consensus” . The “any” threshold simply requires that at least one GENRE predict growth in a particular media condition . The “majority” threshold requires greater than half to predict growth , and the “consensus” threshold requires all GENREs to predict growth . In general , we chose to create ensembles from an odd number of GENREs , in order to prevent ties and make decisions using the “majority” threshold unambiguous . We evaluated the growth predictions in terms of accuracy , precision , and recall ( see Materials and Methods ) . We found that the “majority” threshold led the overall ensemble to achieve average accuracy with respect to the individual GENREs , consistently outperforming the least accurate of the individual GENREs ( Fig 4 “Order Only” ) . The “any” threshold decreased overall accuracy and precision to be worse than any individual GENRE , but increased the recall to match the best individual GENREs . At the other extreme , we found that the “consensus” threshold led to accuracy and precision that matched the very best individual GENREs but diminished recall . While different gap fill order does result in different GENRE structures , the differences are relatively small ( tens of differences relative to hundreds of reactions overall ) . In order to span a greater range of potential GENRE structures , we added random weights ( drawn from a uniform distribution ) to the reactions in the gap filling step ( see Materials and Methods ) . The rationale is that given two pathways of slightly different length but the same biological function , random weights will occasionally favor the longer pathway , thus exploring alternatives that would otherwise be unobserved given a strictly parsimonious procedure . Additionally , each GENRE was reconstructed using a random subset of only 80% of the reactions from the draft GENRE from the Model SEED . Using this new procedure , we reconstructed ensembles of 21 GENREs using 2 through 30 training media conditions ( S2 Fig ) . We evaluated the accuracy by predicting growth on the same 34 test media conditions as before . The resulting accuracy , precision and recall of the individual GENREs were essentially the same on average ( Fig 4 “Diverse” ) , but the distribution spanned a much greater range , both positively and negatively . In this case , the “majority” threshold again achieved average behavior with respect to the individual GENREs , outperforming the least accurate individual GENREs ( Fig 4 “Diverse” ) . The “any” threshold tended to achieve the best accuracy and precision , although not quite as good as the best individual GENREs . However , the “any” threshold achieved the best recall , better than the best individual GENREs and better than the recall achieved with a less diverse ensemble . Our experimental growth data for P . aeruginosa UCBPP-PA14 included both positive ( media conditions which supported growth ) and negative results ( media conditions which did not support growth ) . We formulated an iterative , optimization-based procedure which allowed us to incorporate information inherent in negative growth conditions into our automated curation ( see Materials and Methods ) . In brief , the optimization problem identifies a minimal number of reactions to “trim” from a GENRE in order to prevent growth on negative media conditions while maintaining growth on positive media conditions . As before , we generated ensembles of 21 GENREs for 2 through 30 positive media conditions ( S3 Fig ) . We used random reaction weights , random subsets of 80% of the reactions from the draft GENRE from Model SEED , and this time we selected 10 negative media conditions for each GENRE ( distinct from the negative conditions used to assess accuracy ) and incorporated them using our trimming procedure . We found that incorporation of the negative media conditions increased the accuracy and precision of both the individual GENREs and the ensembles by ~15% ( Fig 4 “Negative Growth Data” ) . The “majority” threshold once again approximately tracked the average GENRE accuracy , precision and recall . The “any” threshold achieved accuracy and recall that were often better than the top individual GENREs , with recall exceeding that achieved previously . Additionally , we explored the effect of generating GENREs within an ensemble using random subsets of the available growth data ( S4 Fig ) . We found that the results are threshold-dependent . The “majority” threshold achieves greater accuracy and precision—but slightly lower recall—when GENREs are generated using large fractions of the available data . Generating GENREs with larger fractions of the data , the performance of the “consensus” threshold is unaffected except for a slight increase in recall . The “any” threshold achieves greater accuracy and precision with lower fractions of data used to train GENREs , but achieves higher recall when higher fractions are used . We evaluated the ability of ensembles to predict gene essentiality . We generated an ensemble of 51 GENREs , each created by gap filling with a random subset of 25 of the total 47 positive media conditions ( 53% ) , 10 randomly-selected negative media conditions from the total of 40 ( 25% ) , and 1 , 210 randomly-selected reactions from a total of 1 , 512 in the draft GENRE generated by Model SEED ( 80% ) . Genes were associated with reactions based on the assigned gene-protein-reaction ( GPR ) relationships from the draft GENRE . We used an in silico representation of CF sputum medium and predicted gene essentiality by removing reactions associated with each gene in turn ( according to the GPR logic ) and running FBA . We compared the resulting gene essentiality predictions with experimental results [19] . We found that the “majority” threshold resulted in better accuracy and recall than the average of individual GENREs , and drastic improvement over the worst GENREs ( Fig 5 ) . The “consensus” threshold resulted in a ~20% increase in precision over the best individual GENRE and greater than 100% increase over the worst individual GENRE . Unsurprisingly , the increased precision of the “consensus” threshold comes at the cost of reduced recall . The “any” threshold resulted in lower precision but a ~40% increase in recall over the best individual GENRE and a ~170% increase over the worst individual GENREs . Using the same ensemble of 51 GENREs from above , we examined the effect of ensemble size on predicting essential genes . We sampled with replacement 10 , 000 small ensembles from among the 51 GENREs for ensemble sizes of 2 through 51 . We evaluated the accuracy , precision and recall against the same gene essentiality data set using the “any” , “majority” , and “consensus” thresholds . We found that for the “majority” threshold , smaller ensembles were less accurate , less precise , and more variable than larger ensembles ( Fig 6 ) . Increasing size improved predictions but with diminishing benefits as the ensemble grew larger . Interestingly , with this “majority” threshold , average recall is essentially unchanged as the ensemble grows larger ( Fig 6 ) . In contrast , for the “any” threshold , we found that increasing ensemble size diminished accuracy and precision , but greatly increased recall . The opposite trend was observed for the “consensus” threshold , where increasing ensemble size greatly increased precision but diminished recall . The “majority” and “consensus” thresholds achieved higher accuracy on average than the “any” threshold . In order to characterize the way gap filling distributes reactions throughout the ensemble , we generated an ensemble of 100 GENREs ( Fig 7A ) . Each GENRE was reconstructed using a randomly-selected 80% of the reactions in iPAU1129 , and then sequentially gap filled from the independent , universal reaction database using 25 random positive growth conditions . We refer to reactions from iPAU1129 as “correct” , acknowledging that a manually-curated GENRE is only a surrogate for a “true” metabolic network . We found that before gap filling , the “correct” reactions from iPAU1129 were initially distributed in a bell-shaped curve throughout the ensemble ( Fig 7B ) . The vast majority of “correct” reactions were found in 50 or more of the GENREs , and in 80 GENREs on average . In contrast , the “incorrect” reactions ( those added by gap filling but which were not in the original iPAU1129 ) were distributed sporadically , with the majority being found in 10 or fewer GENREs . After the gap filling step , 65 “correct” reactions were found to have been added to every GENRE , suggesting a core set of “correct” reactions that were required for biomass production in any condition . We observed that the most common reactions ( found in 50 or more GENREs ) were overwhelmingly “correct” reactions from iPAU1129 ( Fig 7C ) . All of these most common reactions ( both “correct” and “incorrect” ) were involved in the production of biomass components , particularly amino acids . We demonstrate how EnsembleFBA can be implemented in a systems biology workflow . We selected six species from the genus Streptococcus which all have growth phenotype data available through a previous study [20] . We reconstructed an ensemble for each species: Streptococcus mitis , Streptococcus gallolyticus , Streptococcus oralis , Streptococcus equinus , Streptococcus pneumoniae and Streptococcus vestibularis ( Fig 8A ) . For each species , we generated a draft GENRE using the Model SEED online interface . We generated an ensemble of 21 GENREs from each Model SEED draft , and gap filled each member GENRE using the “diverse” methodology and 25 random growth conditions specific to that species . We mapped all genes ( translated to protein sequences ) from each Streptococcus species to small molecule protein binding sequences from DrugBank using NCBI standalone BLASTP and an e-value threshold of 0 . 001 [21 , 22] . For all potential gene targets , we used the ensembles to predict gene essentiality using a “majority” threshold in rich media . We found 261 small molecules in DrugBank that potentially bind to the products of 169 essential genes ( evenly distributed throughout the six species ) . Many of these small molecules ( 113 ) interact with an essential gene in only one of the species , while 44 were predicted to target conserved essential genes in all six species ( Fig 8B ) . S . equinus was predicted to have the most essential genes interact with unique small molecules while S . pneumoniae was not predicted to have essential genes interact with any unique small molecules ( Fig 8C ) . As an example of a conserved small molecule interaction , DB04083 ( N'-Pyridoxyl-Lysine-5'-Monophosphate ) is predicted to interact with essential aspartate aminotransferases in all six species . Alternatively , DB03222 ( 2'-Deoxyadenosine 5'-Triphosphate ) is only predicted to interact with an essential ribonucleotide reductase in S . gallolyticus . To better understand the differences between the metabolic networks which underpin these small molecule screen results , we predicted reaction essentiality in rich media for all six species using a “majority” threshold . We found that several metabolic subsystems were enriched among essential reactions beyond what would be expected from random chance ( Fig 8D ) . Some subsystems were enriched in all six species , such as peptidoglycan biosynthesis , indicating that these reactions related to cell wall biosynthesis are disproportionately essential in all six species . Other subsystems were enriched among essential reactions in a unique species . For example , S . mitis is predicted to have a greater proportion of essential reactions related to amino acid metabolism than other species , perhaps indicating that S . mitis has less redundancy in those pathways than the other six species . Essential reactions related to butanoate metabolism were most enriched in S . pneumoniae , while essential reactions in lysine degradation were most enriched in S . equinus . Interestingly , reactions associated with core metabolic functions ( e . g . amino acid biosynthesis , valine and leucine biosynthesis , phenylalanine biosynthesis ) were not equally enriched among essential reactions for all species .
Genome-scale metabolic network reconstructions ( GENREs ) have been used for decades to assemble information about an organism’s metabolism , to formally analyze that information , and in so doing , to make predictions about that organism’s behavior in unobserved or unobservable contexts . A major barrier preventing more widespread use of GENREs , particularly in non-model organisms , is the extensive time and effort required to produce a high-quality GENRE . Many automated approaches have been developed which reduce this time requirement ( e . g . Model SEED , GLOBUS , CoReCo , RAVEN ) [13–15] . We demonstrate that gap filling—although our results apply to many automated curation approaches—can lead to many potential GENRE structures depending on the ordering of the input data . Rather than arbitrarily selecting a single GENRE from among many possible networks ( which are all reasonably consistent with the available data ) , we found that collecting many GENREs into an ensemble improved the predictions that could be made . We call this approach “EnsembleFBA” and emphasize that ensembles are a useful tool for dealing with uncertainty in network structure . We demonstrated how ensemble diversity impacts predictions . We show that EnsembleFBA correctly identifies many more essential genes in the model organism P . aeruginosa UCBPP-PA14 than the best individual GENREs . We showcase how EnsembleFBA can be utilized in a systems biology workflow by predicting how small molecules interact with different essential genes in six Streptococcus species . Ensembles increase the quality of predictions without incurring months of manual curation effort , thus making GENRE approaches more practical for applications which require predictions for many non-model organisms . We have provided code to facilitate the creation and analysis of ensembles of GENREs . Gap filling is a common step during the GENRE curation process , both for manually- and automatically-curated GENREs [5] . We used a linear ( rather than binary ) gap filling algorithm to expand GENREs so that they are capable of producing biomass in silico on growth media which supports growth of the organism in vitro . Gap filling algorithms suggest parsimonious reaction sets from some “universal” biochemical database which , if added to a GENRE , will allow growth in the new environment [6 , 17 , 18 , 23] . Often , multiple reaction sets can enable growth , so some heuristics are needed to select a final solution . Sometimes gene homology metrics are used to select a solution , such that genes which catalyze the suggested reactions are compared to the current genome , and the reaction set with the best matches in the current genome are selected as the final solution . When validated , these solutions can lead to re-annotation of the genome [6] . During automated curation , there is less opportunity for extensive validation , and so the first or the most parsimonious solution is selected . As we demonstrated , the order of gap filling can change the final outcome , thus producing GENREs with different structures from the exact same input data ( Fig 2 ) . Under these circumstances , it is difficult to know which solution is most correct without additional data . A possible way around this issue of gap fill order is to remove the sequential nature of gap filling entirely and use a global gap filling approach . We demonstrate that not only is such a global approach much slower ( quadratic increases in solution time as growth media conditions are added ) , but the solutions are no more parsimonious or biologically relevant ( Fig 3 ) . The reduction in parsimony is attributable to the use of a linear rather than an integer optimization approach [17] . The linear approach is unavoidable , however , as the scale of the global problem makes integer optimization prohibitively slow in practical applications . Alternatively , we found that the collection of multiple GENREs into ensembles improved the predictions that could be obtained from automatically-generated GENREs . Ensembles have been used for many years in the machine learning community to leverage the strengths of many different models to improve predictions [24 , 25] . Ensembles have been used previously to analyze GENREs from a kinetic standpoint [26] . Because kinetic parameters are usually unknown for an entire genome-scale network , ensembles of kinetic parameters are generated such that all parameter sets lead to the same steady state [26] . In this way , ensembles can represent the space of allowable kinetic parameters . Our approach to generating ensembles is different in that we attempt to represent the space of allowable GENRE structures rather than kinetic parameters . Ensembles provide a significant advantage over individual GENREs by tuning for specific results with defined decision thresholds ( Figs 4 and 5 ) . Consistently , by using the “any” threshold , recall can be made to equal or exceed the best individual GENREs . This result makes sense , considering that different network structures will result in different growth or gene essentiality predictions . By accepting any essential gene prediction from among the constituent GENREs , we cast a wider net and capture many more of the true essential genes and growth conditions . The fact that many individual GENREs contribute unique but true predictions suggests that each GENRE recapitulates elements of the “true” network structure ( Fig 7 ) . Similarly , by using the “majority” threshold , the ensemble predictions perform like the average GENRE ( Figs 4 and 5 ) . By requiring a majority of GENREs to agree , the ensemble guards against poor predictions and , in most cases , outperforms the worst individual GENREs . Finally , if precision is the overall goal , a “consensus” threshold provides confidence that the majority of positive predictions are true positives ( Fig 5 ) . We observed that ensemble performance is limited by the quality of the GENREs which form the ensemble . The choice of decision threshold ( “any” , “majority” , or” consensus” ) did not drastically improve overall accuracy of the ensemble . However , by improving the individual GENREs using negative growth information , the overall ensemble accuracy improved dramatically ( Fig 4 ) . Also , it should be noted that the computational burden required by ensembles will always be greater than the burden of a single GENRE . For all the examples in this study , computational burden scales linearly with the number of GENREs in the ensemble ( ensemble of size N GENREs will require N times longer to calculate FBA solutions ) which is a modest expectation in practice . Other applications , like predicting species interactions , would not scale linearly if all possible pairs of GENREs between two ensembles were simulated . Several previous groups have developed methods for incorporating negative growth conditions into GENRE curation . The GrowMatch algorithm seeks to remove reactions in order to prevent growth in specific conditions [23] . In that case , the reactions are not removed from the GENRE , but rather , prevented from carrying flux under particular conditions . This approach was supported by a biological justification that certain enzymes may not be functional under certain conditions [23] . Alternatively , the Model SEED and the CROP algorithm both remove reactions from the network entirely [27 , 28] . The CROP algorithm also allows for the consideration of multiple negative growth conditions simultaneously in order to select the reactions to trim [28] . Our approach is different in that it considers the positive growth conditions simultaneously with the negative growth conditions while selecting reactions to trim ( see Materials and Methods ) . Additionally , our iterative approach to integrating all growth conditions is an innovation that automates the search for a globally-consistent GENRE structure . By applying this new approach , we found that average GENRE accuracy ( with respect to predicting growth under specific media conditions ) increased by ~15% ( Fig 4 ) . Increasing ensemble diversity impacted ensemble recall , but did not have an obvious effect on overall accuracy . Some degree of diversity is required in order to gain any advantage through an ensemble representation . In the “Order Only” ensemble ( generated simply by changing the order of gap filling; Fig 4 ) there were only small differences between any of the GENREs so it was difficult to improve on the best GENRE . By injecting greater diversity through random weights and random subsets of the data , we observed much greater variation in individual GENRE performance ( both positively and negatively ) , but the average accuracy was the same as the low diversity ensemble ( Fig 4 ) . The advantage of diversity is in casting a wide net and thus improving ensemble recall , particularly when combined with an “any” decision threshold . In practice , the choice to increase diversity or not will depend on the goals of the analysis . If the goal is to generate many candidate essential genes or media conditions , then more diversity will be advantageous . If the goal is to generate fewer , more confident predictions , then minimizing diversity will be most effective . EnsembleFBA is easily integrated into systems biology workflows . As an example , a current challenge in systems biology is to identify species-specific drug targets so that therapies will not disrupt the healthy microbiome structure [29 , 30] . We reconstructed ensembles for six Streptococcus species by gap filling with growth phenotype data , we predicted essential genes and mapped those genes to potential small molecule binding partners within a matter of hours , and can have more confidence in the quality of the gene essentiality predictions than if we were to work with single GENREs for each species ( Fig 8 ) . The process scales well with the number of species , such that 12 or 100 species would not take significantly longer than six , and the quality of the predictions is maintained with scale . It is interesting to note that among Streptococcus species , there are generally small molecules which can be selected to uniquely interact with essential genes in a single species , and other small molecules which interact with conserved essential genes ( Fig 8C ) . The observed interactions between essential genes and small molecule ligands are species-specific because of differences in network structure which lead to some metabolic subsystems being disproportionately represented among essential reactions ( Fig 8D ) . In the search for species-specific drug targets , it is important to consider , not only the presence or absence of a particular gene , but also the role of that gene in the broader network context , and improved systems biology tools such as EnsembleFBA can help to elucidate that context with greater confidence . Gap filling is not the only GENRE reconstruction approach that produces many possible solutions . Likelihood-based gap filling produces a distribution of possible annotations for each gene in a genome , assigning a probability to each [14 , 16 , 31] . Network structure is then based on maximizing the likelihood over all possible solutions . Ensembles could be generated easily using this type of framework by sampling many alternative solutions around the maximum likelihood . Indeed , it may be beneficial to create an ensemble using GENREs reconstructed using several different methods or software tools . Additionally , during a manual curation process it is often necessary to make decisions using uncertain information , and ensembles offer a way to capture and represent the alternative model structures that can result . We suggest that there are many possible ways to generate ensembles such that they will allow researchers to generate better predictions about under-studied organisms . Finally , we foresee ensembles playing an important role beyond improving predictions . For example , ensembles of GENREs may improve the processes of experimental design and model reconciliation . Within a diverse ensemble , many possible network structures are represented , and it is expected that some structures will be closer to the truth than others . We suggest that ensembles can be leveraged to design an optimal series of experiments to weed out the most incorrect network structures . For instance , such an approach could select the most differentiating carbon sources to experimentally test , or the most differentiating essential genes . This sort of ensemble-guided experimental design could save time and experimental resources . Model reconciliation is another field that could benefit from ensembles [8 , 32] . Given GENREs for two different species , reconciliation is the process of removing systematic differences from the two GENREs so that any differences which remain are due to biology alone . Systematic differences often result from arbitrary choices during the process of reconstruction . Ensembles could be used to automate the reconciliation process by representing the space of possible GENREs for each species and the reconciled versions would be the two models from the two spaces that are most similar to each other . Thus , ensembles have potential to improve other tasks than prediction , including experimental design and mapping the space of GENRE structures for tasks like reconciliation .
All data , Matlab ( Natick , MA , USA ) implementations of algorithms , Matlab simulation scripts , results files and figure generation scripts are publically available in our online repository: https://github . com/mbi2gs/ensembleFBA All biochemical reference data was obtained from the Model SEED database ( https://github . com/ModelSEED/ModelSEEDDatabase ) . The metabolic reaction and compound databases were parsed and formatted for use in Matlab using a custom Python script available in our repository ( “format_SEED_data . py” ) . A draft network for P . aeruginosa UCBPP-PA14 was automatically generated using the Model SEED web service ( http://modelseed . org/genomes/ ) . Similarly , draft networks were generated for Streptococcus mitis ATCC 6249 , Streptococcus gallolyticus ICDDRB-NRC-S3 , Streptococcus oralis ATCC 49296 , Streptococcus equinus AG46 , Streptococcus pneumoniae ( PATRIC ID 1313 . 5731 ) , and Streptococcus vestibularis 22–06 S6 . Representations of media conditions ( including minimal media and cystic fibrosis sputum medium ) , and biomass representations were drawn from previous GENRE analyses of Pseudomonas aeruginosa [33 , 34] . P . aeruginosa PA14 essential genes in cystic fibrosis sputum medium were experimentally identified previously [19] . A manually curated , and thoroughly validated GENRE of P . aeruginosa UCBPP-PA14 called iPAU1129 was developed previously ( Bartell et al . In press ) , along with Biolog growth screen data for P . aeruginosa UCBPP-PA14 indicating many media conditions in which this strain will and will not grow . Growth phenotype data for six Streptococcus species was obtained from the file “Supplementary Data 1” of [20] . Small molecule amino acid binding target sequences were downloaded from the DrugBank website ( http://www . drugbank . ca/ ) [21] . After identifying homologous genes to the target sequences using BLASTP [22] , we used a custom python script to parse the results for input into Matlab ( “listPossibleTargets . py” , available in repository ) . We implemented a linear ( as opposed to binary ) gap filling algorithm in Matlab , based on the algorithms FASTGAPFILL and FastGapFilling [17 , 18] . We used the Gurobi solver version 6 . 0 . 5 for all optimization tasks ( Gurobi , Houston , TX , USA ) . To begin , we provide the algorithm with a universal database of metabolic reactions U , a universal database of exchange reactions X , a biomass reaction , and a set of growth conditions formatted as lower bounds on exchange reactions . The algorithm identifies a set of reactions from U and X that allow flux through the biomass reaction under all growth conditions . The algorithm is implemented as a linear program ( LP ) that minimizes the sum of the absolute value of all fluxes through U and X . The optimization problem takes the form: minimizesum ( ruzu ) +sum ( rxzx ) s . t . Uv+Xw=0 ( 1 ) lbu , i≤vi≤ubu , ifor alli∈[1 , Nu] ( 2 ) lbx , i≤wi≤ubx , ifor alli∈[1 , Nx] . . . zu , i≤vi≤zu , ifor alli∈[1 , Nu] ( 3 ) zx , i≤wi≤zx , ifor alli∈[1 , Nx] . . . zi≥0for alli∈[1 , Nu+Nx] ( 4 ) vbiomass , gc=j>0 . 05for allj∈[1 , Ngc] ( 5 ) zu , i≥Czu , ifor alli∈[1 , Nu] ( 6 ) zx , i≥Czx , iforalli∈[1 , Nx] . . . Where: In order to incorporate genome annotations from a specific organism , we force the inclusion of all associated reactions from those annotations using the C variables . Note that unlike a binary optimization , the LP minimizing the sum of the absolute flux values through U and X does not necessarily result in a solution with the fewest reactions , but rather the solution which requires the minimum sum of the absolute values of the fluxes through it . The LP here can be extended to utilize multiple growth conditions simultaneously ( global approach ) by duplicating the U and X matrices , once for each growth condition , but minimizing a single set of z variables across all conditions . To gap fill using multiple growth conditions sequentially , we gap fill using the first growth condition , incorporate the solution into the GENRE , then repeat the process for all growth conditions . Our Matlab function “expand ( ) ” implements this optimization problem . We implemented a binary optimization problem to trim a minimal set of reactions from a GENRE in order to prevent growth under negative growth conditions while simultaneously maintaining growth in the positive growth conditions . As input to the algorithm , we provide a GENRE , and a set of both positive and negative growth conditions . We chose to run FBA first to identify mismatches between the computational predictions and the in silico data ( negative growth conditions that erroneously predicted to support growth in silico ) . Having identified those , we then ran FBA on all the positive growth conditions to identify the top five with flux distributions most similar to the flux distribution of the negative growth condition . Flux distribution similarity was determined by calculating the Jaccard similarity between the reaction sets carrying non-zero flux values . Increasing the number of positive growth conditions increases compute time . The GENRE and the selected growth conditions are passed to the trimming problem , which takes the form: maximizesum ( ruyu ) +sum ( rxyx ) s . t . Uv+Xw=0 ( 7 ) yu , ilbu , i≤vi≤yu , iubu , iforalli∈[1 , Nu] ( 8 ) yx , ilbx , i≤wi≤yx , iubx , iforalli∈[1 , Nx] . . . vbiomass , gc=j≥0 . 05forallj∈[1 , Ngc] ( 9 ) max ( vbiomass , ngc=j ) =0forallj∈[1 , Nngc] ( 10 ) yx , i , yu , i∈{0 , 1} Where: The term max ( vbiomass , ngc = j ) = 0 requires the maximum flux through the biomass reaction for the non-growth conditions to be zero . In order to implement this constraint , we took advantage of duality theory as has been done previously [7] . Specifically , the optimal objective value of the dual of a linear program will equal the optimal value of the primal . By constraining the primal and dual objectives to equal each other , we can ensure that the flux through the biomass objective is maximized . We can replace the term max ( vbiomass , ngc = j ) = 0 ( constraint 10 ) with the following constraints: vbiomass , ngc=j=λububy−λlblby ( 11 ) UTλmets+λub−λlb=c ( 12 ) λub , λlb≥0 ( 13 ) Where λmets , λub and λlb are the dual vectors associated with the metabolites , upper and lower bounds of the primal problem . Note that the terms λubub y and λlblb y are quadratic , requiring a multiplication of the binary inclusion variable y with the dual variables . Because y is a binary variable , in this case the quadratic constraints can be converted to linear constraints through the use of additional variables: tub≤Lλyi ( 14 ) tub≥0 ( 15 ) tub≤λub , iubi ( 16 ) tub≥λub , iubi−Lλ ( 1−yi ) ( 17 ) Where tub is a stand-in for the product λubub y and L is a large number greater than or equal to the upper bound on λubub y ( e . g . 1000 ) . Similar constraints are produced for the product λlblb . The quadratic constraint above ( constraint 11 ) can then be replaced by a linear constraint: vbiomass , ngc=j=tubub−tlblb ( 18 ) Our Matlab function “trim_active ( ) ” implements this optimization problem . We implemented an iterative algorithm to integrate the LP expansion step with the binary trimming step . The algorithm first applies the expansion step to produce a GENRE that is capable of growing in all positive growth conditions . Next , the algorithm checks for negative growth conditions that allow for biomass flux and for any that do , applies the trim step as described above . The algorithm iterates between the expand and trim steps until either a completely consistent GENRE structure is identified , or it reaches a maximum attempts limit . A single attempt is completed if the GENRE structure is not yet consistent with the input growth conditions but stops making progress ( possibly stuck in a local optimum ) . In this case , a random reaction is removed from the GENRE and the search is re-initiated . If the maximum attempts limit is reached , the algorithm removes any negative growth conditions that are inconsistent with the positive growth conditions , and returns the final GENRE . This iterative algorithm is implemented in our Matlab function “build_network ( ) ” . Growth media were simulated by setting the lower bounds on exchange reactions for the appropriate nutrients to negative values . The uptake of carbon source ( s ) limited the final flux through biomass . “Growth” was determined by maximizing flux through the biomass objective . We predicted “growth” if a positive , non-zero flux could be achieved through biomass . Gene knock-outs were simulated by generating a new GENRE which was missing the reactions dependent on the knocked-out gene . The reaction-gene dependence was determined by evaluating the binary logic of the GPRs provided by Model SEED . Our custom script to evaluate GPR logic is “simulateGeneDeletion ( ) ” . We evaluated the growth predictions in terms of accuracy ( TP + TN ) / ( TP + FP + TN + FN ) , precision ( TP / TP + FP ) , and recall ( TP / TP + FN ) where TP = number of true positives , FP = the number of false positives , TN = the number of true negatives and FN = the number of false negatives . Precision indicates the fraction of positive predictions which are true positives . Recall indicates the fraction of positive events which were correctly predicted by the method . We downloaded the Drug Target Sequences for small molecules in FASTA format from DrugBank [21] . Using NCBI standalone BLASTP and an e-value cutoff of 0 . 001 , we identified homologous sequences in all six Streptococcus proteomes [22] . We downloaded KEGG subsystem annotations for the reactions in the Model SEED database ( “KEGG . pathways . tsv” ) . After predicting essential reactions for each Streptococcus species , we used the hypergeometric distribution to calculate the probability of drawing k essential reactions and finding that x or more are annotated with subsystem j , from a population of size M reactions , of which N are annotated with subsystem j . The majority of our reconstructions and simulations were performed on a 64-bit Dell Precision T3600 Desktop computer with 32 GB RAM and eight 3 . 6 GHz Intel Xeon CPUs , running Windows 7 . Incorporating negative growth information often lead to longer reconstruction times ( sometimes 2 hours per GENRE ) due to the binary optimization step . To accelerate the reconstruction time while incorporating negative growth information , we used the University of Virginia’s high performance computing cluster . Our Matlab scripts for generating an ensemble ( using the gap filling approach described in this work ) and for analyzing an ensemble are freely available in a github repository ( see Code and Data Availability ) . The Gurobi solver is required , in addition to our Matlab scripts . We have also included a tutorial script to guide the user through the necessary steps to generate and analyze an ensemble ( “test_eFBA . mat” ) .
|
Metabolism is the driving force behind all biological activity . Genome-scale metabolic network reconstructions ( GENREs ) are representations of metabolic systems that can be analyzed mathematically to make predictions about how a system will behave , as well as to design systems with new properties . GENREs have traditionally been reconstructed manually , which can require extensive time and effort . Recent software solutions automate the process ( drastically reducing the required effort ) but the resulting GENREs are of lower quality and produce less reliable predictions than the manually-curated versions . We present a novel method ( “EnsembleFBA” ) which accounts for uncertainties involved in automated reconstruction by pooling many different draft GENREs together into an ensemble . We tested EnsembleFBA by predicting the growth and essential genes of the common pathogen Pseudomonas aeruginosa . We found that when predicting growth or essential genes , ensembles of GENREs achieved much better precision or captured many more essential genes than any of the individual GENREs within the ensemble . By improving the predictions that can be made with automatically-generated GENREs , this approach enables the modeling of biochemical systems which would otherwise be infeasible .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"biotechnology",
"medicine",
"and",
"health",
"sciences",
"chemical",
"compounds",
"pneumococcus",
"pathology",
"and",
"laboratory",
"medicine",
"small",
"molecules",
"pathogens",
"metabolic",
"networks",
"microbiology",
"organic",
"compounds",
"pseudomonas",
"aeruginosa",
"network",
"analysis",
"genome",
"analysis",
"pharmacology",
"bacteria",
"drug",
"metabolism",
"bacterial",
"pathogens",
"pseudomonas",
"computer",
"and",
"information",
"sciences",
"medical",
"microbiology",
"streptococcus",
"microbial",
"pathogens",
"chemistry",
"group",
"d",
"streptococci",
"pharmacokinetics",
"organic",
"chemistry",
"genetics",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"genomics",
"gene",
"prediction",
"computational",
"biology",
"organisms"
] |
2017
|
Managing uncertainty in metabolic network structure and improving predictions using EnsembleFBA
|
Accurately measuring the neural correlates of consciousness is a grand challenge for neuroscience . Despite theoretical advances , developing reliable brain measures to track the loss of reportable consciousness during sedation is hampered by significant individual variability in susceptibility to anaesthetics . We addressed this challenge using high-density electroencephalography to characterise changes in brain networks during propofol sedation . Assessments of spectral connectivity networks before , during and after sedation were combined with measurements of behavioural responsiveness and drug concentrations in blood . Strikingly , we found that participants who had weaker alpha band networks at baseline were more likely to become unresponsive during sedation , despite registering similar levels of drug in blood . In contrast , phase-amplitude coupling between slow and alpha oscillations correlated with drug concentrations in blood . Our findings highlight novel markers that prognosticate individual differences in susceptibility to propofol and track drug exposure . These advances could inform accurate drug titration and brain state monitoring during anaesthesia .
Understanding how the human brain reversibly generates and loses consciousness , through complex interactions of neural activity at multiple spatial and temporal scales , is a grand challenge for modern neuroscience . Recent theoretical advances have argued that consciousness changes when the balance between integrated and differentiated neural activity is affected [1–4] . However , accurately tracking these changes in brain dynamics remains a key research challenge with potentially wide-ranging applications , and is complicated by the significant individual variability in the trajectory along which consciousness is lost and regained . The process of reversibly inducing unconsciousness using anaesthetic drugs like propofol is commonplace in clinical medicine [5] . However , tracking brain activity to accurately assess the depth of anaesthesia in an individual is currently not a universal component of clinical practice . Indeed , surface electroencephalography ( EEG ) is relatively easy to measure from the scalp and has long been known to index changes in brain dynamics induced by anaesthetic action [6] , but it is still not universally used in the clinical setting . This is despite the fact that intraoperative awareness during surgery continues to result in pain and distress [7] , highlighting the need for reliable depth of anaesthesia monitoring in the operating room . The absence of ubiquitous brain monitoring during general anaesthesia is , in part , due to the lack of robust EEG markers derived from current advances in neuroscience [8–12] , which can accurately track the loss and reestablishment of reportable consciousness . Monitoring of brain states is currently limited to proprietary systems with mixed results [13–15] . Crucially , one reason for this is the considerable individual variability in susceptibility to anaesthetic dosage [16] , which adversely affects the accuracy of these systems [17] . To better understand the factors underlying this variability , we combined the measurement of high-density resting state EEG from healthy volunteers sedated with propofol with measurement of drug concentrations in blood , in addition to objective assessment of behavioural responsiveness . With this aim in mind , we administered propofol at dosages expressly aimed at engendering varying degrees of mild to moderate sedation across our participant group , rather than complete unconsciousness in all of them . Employing modern functional EEG tools to assess spectral power and connectivity , we identified key changes in brain networks using graph-theoretic tools , and linked these changes to individual variability in drug concentrations and loss of behavioural acuity during sedation . Drawing upon previous research [18–21] , we hypothesised characteristic impairments in the strength and topography of EEG power and connectivity , especially manifesting in the slow and alpha frequency bands , alongside administration of propofol . In addition to confirming these hypotheses , our findings highlight valuable EEG-derived signatures that can not only track the actual amount of propofol in blood , but also predict loss of responsiveness even before any drug is administered . These findings contribute to the current interest in identifying consistent markers of the loss and recovery of consciousness during propofol sedation . In the clinical context , these findings could lead to more accurate drug titration and brain state monitoring during anaesthesia .
The behavioural changes accompanying the administration of progressively increasing amounts of propofol ( Fig 1A ) are shown in Fig 1B , which plots the hit rate of participants as a function of the level of sedation . Based on binomial modelling of their hit rates ( see Materials and Methods ) , we identified a subgroup of 7 participants who became behaviourally impaired at this simple task during moderate sedation; 13 others remained responsive throughout , though their reaction times were impaired during sedation ( Fig 1C ) . We designate these two groups as drowsy ( green triangles ) and responsive ( blue triangles ) in the following descriptions . As expected , we found a highly significant interaction between group and sedation level in hit rates ( Fig 1B; F ( 3 ) = 38 . 4 , p = 9e-09 ) . Further , in the responsive group , there was a significant effect of sedation on reaction times ( Fig 1C; F ( 2 ) = 14 . 6 , p = 0 . 0002 ) . In comparison to the relative distinction between the two groups in their hit rates , there was considerably more overlap in drug concentrations measured in blood plasma ( Fig 1D ) . We found a relatively weaker interaction between group and level of sedation in drug concentrations: F ( 2 ) = 4 . 7 , p = 0 . 0242 , and the difference between drug concentrations in the two groups reached significance only during moderate sedation ( p = 0 . 0181 ) . This finding points to the well-studied inter-individual variability in pharmacodynamic impact of propofol [16 , 17] , and motivates the development of more accurate signatures of responsiveness that can be measured passively and non-invasively during propofol sedation . Connectivity between EEG channels was assessed to directly investigate the impact of propofol on the structure of brain networks of oscillatory neural interactions , using the debiased weighted Phase Lag Index ( dwPLI , see Fig 2 and [22] ) . Here , we define brain networks as the characteristic patterns of scalp-level connectivity observable in human EEG at different frequencies , generated by underlying cortical networks [23] with firing rates oscillating at their natural frequencies [24] . We employed the dwPLI connectivity matrices in each band to construct such EEG-derived brain networks , and used graph-theoretic algorithms to quantitatively compare their topological properties . By representing the EEG channels as nodes of a network and the strength of dwPLI between them as weighted , undirected links between them , we calculated four measures that captured micro-scale ( clustering coefficient ) , meso-scale ( modularity and participation coefficient ) and macro-scale properties ( characteristic path length ) of each participant’s network at each level of sedation ( see bottom right panel of Fig 2 for a visual description of these properties ) . Importantly , these metrics were chosen a priori to summarise key network properties that we expected to be modulated during propofol sedation . In the alpha band , median dwPLI across all channel pairs was significantly more reduced in the drowsy group during mild ( p = 0 . 003 ) and moderate sedation ( p = 0 . 01 ) . Further , the clustering coefficient [25 , 26] , which measures local efficiency , was significantly lower ( Fig 3A ) in the frontal alpha networks of the drowsy group during mild ( p = 0 . 007 ) and moderate sedation ( p = 0 . 04 ) . Furthermore , within the responsive group , clustering during moderate sedation tended to decrease linearly alongside increasing reaction times ( Fig 3B ) , though this effect only approached significance . Conversely , characteristic path length ( Fig 3C ) , the inverse of global efficiency , was significantly higher during mild ( p = 0 . 0004 ) and moderate sedation ( p = 0 . 0035 ) , and tended to increase with slower reaction times among responsive participants ( Fig 3D ) . Taken together , small-worldness , a combined measure of a network’s local and global efficiency ( calculated as the ratio of clustering to path length [26 , 27] ) , was significantly reduced in the drowsy group during mild ( p = 0 . 005 ) and moderate sedation ( p = 0 . 03 ) . At the meso-scale , these drowsy alpha networks were also more modular at moderate sedation ( Fig 3E , p = 0 . 02 ) , and hence more separable into relatively disconnected topological modules [28] . Crucially , these modules lacked hub nodes that connected them into an integrated network , as evidenced by statistically lower standard deviation ( p = 0 . 002 ) of participation coefficients [29] in the drowsy group ( Fig 3F ) . Together , these network differences demonstrated that the frontal alpha connectivity in the drowsy group did not have the network capacity of the occipital alpha network commonly observed in human resting EEG during wakefulness . These changes in alpha networks can be understood more visually with Fig 4A . At baseline , both groups had prominent frontocentral and occipital modules of strong connectivity . While these modules persisted through moderate sedation in the responsive group , the structure of connectivity networks in the drowsy group shifted to qualitatively distinct state comprising of coherent , frontally centered oscillations that manifested as a frontal module ( Fig 4B ) , before reverting back to the typical pattern of baseline connectivity during recovery . On the whole , this shift in alpha connectivity mirrors the frontal shift in alpha power ( Fig 5 ) commonly observed during propofol sedation [18 , 19 , 30–32] . In contrast to these changes in alpha networks , no differences were observed between delta networks in the two groups ( see S1 Fig ) . Spectral connectivity in the alpha band identified a prospectively valuable determinant of the variability in susceptibility to propofol seen in the behavioural data . During the baseline period before sedation , though there were no differences in the topography or relative strength of alpha power between the responsive and drowsy groups ( Fig 5A and 5B ) , there were significant differences in median dwPLI ( p = 0 . 0085 ) and key network properties that captured the topological structure of connectivity in the alpha band . Specifically , alpha networks in the drowsy group were already less clustered ( Fig 3A; p = 0 . 04 ) and less small-worldy ( p = 0 . 0187 ) at baseline . They were also more modular ( Fig 3E; p = 0 . 04 ) , and had fewer hubs ( Fig 3F; p = 0 . 0018 ) . Remarkably , these baseline alpha network differences were evident when the two groups of participants were indistinguishable , both in terms of behavioural hit rates ( Fig 1B ) and occipital alpha power ( Fig 5B ) . Furthermore , this predictive value of brain connectivity was unique and specific to the alpha band , and not evident in other frequency bands ( see S1 Fig ) . In line with previous findings [31 , 33 , 34] , sedation selectively increased beta/gamma power and connectivity among responsive participants , but baseline power or connectivity in these bands was not significantly different between the two groups . To explicate this result further , Fig 6A depicts a scatter plot of alpha network small-worldness in each participant measured during pre-drug baseline , against their consequent behavioural hit rates and drug concentrations measured during moderate sedation . Though there was considerable variability in small-worldness across the responsive group at baseline , the drowsy group already had relatively lower small-worldness in comparison . To directly test whether participants who already had less robust brain networks at baseline later became drowsy or unresponsive during moderate sedation , Fig 6B plots the individual hit rate trajectories of the participants separated based on whether their baseline small-worldness was above or below the median . Those in the group with high baseline small-worldness remained responsive , and had significantly higher hit rates during moderate sedation ( Fig 6B , inset; p = 0 . 0093 ) . This predictive role of alpha brain networks in characterising individual variability in susceptibility to propofol is exemplified in Fig 6C , which depicts their evolution in two ‘drug concentration-matched’ participants . Despite registering relatively similar drug concentrations at moderate sedation , one of them remained responsive while the other became completely unresponsive . As is evident , the latter participant already had a comparatively less robust alpha network already at baseline , which then evolved into a frontally alpha module at moderate sedation . In comparison , the responsive participant had a relatively more small-worldy , less modular network at baseline , which was sustained during moderate sedation . These differences potentially explain why the drowsy group , whose alpha networks were already compromised to some degree , became behaviourally impaired while the responsive group did not , despite both groups registering overlapping levels of propofol as measured in their blood at moderate sedation . It is important to note that these differences observed in the baseline alpha networks were abolished at recovery ( see Fig 3A and 3C ) . This suggested that these differences between the two groups were essentially dependent on the latent alpha network state of the participants at the beginning of the data collection rather than any individual trait , and were ‘reset’ after the washout of the drug . We found that , at baseline , participants in both responsive and drowsy groups had similar temporal coupling between the phase of slow oscillations and alpha power , with negative values of phase-amplitude coupling ( PAC; Fig 7A ) over occipital channels ( delineated in Fig 5A , top left ) . This pattern persisted during mild sedation and only changed during moderate sedation within the drowsy group , in whom it shifted toward positive PAC values , before reverting back to negative PAC at recovery . There was a significant interaction in occipital PAC between level of sedation and group ( F ( 3 ) = 3 . 8 , p = 0 . 021 ) . Fig 7C provides more detail on this , using angular histograms of alpha power distributed over slow phase , for a pair of representative participants , one in each of the two groups , responsive and drowsy . At baseline , occipital alpha power was either evenly spread over slow phase , or was greater near the trough of the slow oscillation , resulting in a trough-max distribution and negative PAC . During moderate sedation , only the drowsy participant’s distribution shifted towards peak-max positive PAC with greater alpha power near slow oscillation peaks . At recovery , this distribution reverted back to a trough-max pattern with negative PAC . Further , we also found a highly significant positive correlation between PAC and drug concentrations in blood during moderate sedation ( Fig 7B ) . This correlation did not manifest during mild sedation or recovery , when drug concentrations were relatively low . Importantly , there was no significant correlation between PAC and reaction times . This was in contrast to the correlations between alpha power/connectivity and reaction times ( Figs 3B and 3D and 5C ) , and highlights a novel dissociation between phase-phase and phase-amplitude coupling: while the former correlated with responsiveness as measured by hit rates and reaction times , the latter correlated drug concentrations in blood . Juxtaposed with previous research , our findings are convergent with existing evidence for characteristic changes in PAC alongside propofol induction . Trough-max slow-alpha PAC has been shown to accompany transitions to unconsciousness in frontal EEG channels , which then switches to a peak-max pattern in the same channels following loss of consciousness during deep sedation [18 , 35] . While we have highlighted complementary changes in occipital channels , we also replicated these previous findings . In frontal channels , slow-alpha PAC values were close to zero at baseline , and progressed to a trough-max pattern during moderate sedation ( see S2 Fig ) . This resulted in a significant interaction between level of sedation and group in frontal PAC values ( F ( 3 ) = 4 . 1 , p = 0 . 0136 ) , with the drowsy group showing a significantly stronger trough-max pattern than the responsive group during moderate sedation ( p = 0 . 011 ) . Further , as with occipital PAC , there was a significant correlation between frontal PAC and drug concentrations in blood during moderate sedation ( S2 Fig ) .
Our experimental design used propofol sedation to engender transitional states of responsiveness that varied across participants . The levels of drug administered produced a variable pattern that spread the participant group along a spectrum of varying behavioural impairment , rather than resulting in complete unconsciousness in all of them . Using EEG to track brain activity and measuring actual levels of drug in blood alongside this spectrum of impairment has enabled us to identify neural markers that dissociate conscious report from drug exposure [2] , and makes the results presented here distinctive in their contribution to advancing understanding of the neural markers of loss of consciousness due to propofol . We have built upon previous research that has shown that while occipital alpha power progressively drops as participants become behaviourally compromised as measured by reaction times , the qualitatively dissimilar onset of frontal alpha power is a characteristic marker of the loss of consciousness [18 , 19 , 30 , 32 , 36] . Confirming our hypotheses , while this frontal alpha generates meso-synchronous modules , brain network connectivity as a whole is nevertheless impaired . Graph-theoretic measures quantify this loss of the capacity of individual brain networks in the alpha band , linking them to concomitant variability in behavioural impairment across participants . Small-worldness is commonly seen as a measure of the cost-versus-efficiency optimality of a network configuration , and our findings converge with previous evidence [37] highlighting the reduction in the efficiency of cortical networks during loss of consciousness during propofol sedation , potentially due to dysfunctional modulations in thalamocortical connectivity [8 , 38 , 39] . It is worth noting that a similar breakdown in the capacity of alpha networks has also been reported with other anaesthetic agents like sevoflurane and ketamine [40–42] . This is despite the fact that these distinct anaesthetic agents had varying effects on EEG oscillations and , unlike propofol , did not always produce increases in frontal alpha . Hence the observed changes in alpha networks due to sedation cannot be explained as a shift of alpha power and connectivity from posterior to anterior areas . Rather , our results , along with these previous findings , point toward a broader understanding of characteristic signatures of connectivity in alpha networks as potentially reliable correlates of reportable consciousness [43] . Measurement of drug concentrations at each level of sedation dissociated a principal clinical pharmacodynamics target per se ( sedation and consequent behavioural unresponsiveness ) from incidental pharmacodynamic consequences of drug exposure during propofol sedation . The considerable individual variability in the susceptibility to anaesthesia has been documented [16] , and is evident in the large overlap between blood levels of drug in our responsive and drowsy groups . While our measurement of modulations in phase-phase coupling in delta and alpha bands during sedation showed clear correlations with behavioural impairment , we have also demonstrated a latent relationship between slow-alpha phase coupling and individual variation in drug concentrations . It is important to distinguish these dynamic slow oscillations from stable slow cortical potentials observed during propofol anaesthesia [12] , and from delta oscillations during sleep [44] . This link between PAC and individual levels of drug in blood was not observed in the delta or alpha bands separately , in either power or connectivity . Analytical approaches used for estimating Bispectral Index ( BIS , see [45] that do not take phase information into account are unlikely to detect this key marker of individual drug concentration [18] . Hence our findings are relevant to the challenge of engendering an appropriate level of unconsciousness by accurately tailoring drug concentrations to individuals , a key consideration with significant implications for clinical anaesthesia . Finally , by tracking individual brain networks across levels of sedation , we have shown that the quantifiable robustness of alpha connectivity networks in the awake state before sedation predicts susceptibility to propofol . Specifically , given two behaviourally indistinguishable individuals undergoing administration of sedative , the one with the more robust , small-worldy alpha network with well-connected hubs is likely to require a greater amount of drug to render them unresponsive to the same degree . It is important to note that this latent variability in the state of alpha connectivity at baseline could be detected despite the lack of any significant differences in behavioural performance or alpha power at that time . Orthogonally , slow-alpha PAC complements this predictive capability by tracking the concentration of propofol in blood plasma . This set of results , if replicated and verified in the clinical context , could contribute to reliable applications of brain monitoring for tracking and accurately modulating consciousness with anaesthetics during routine surgery .
All healthy controls gave written informed consent . Ethical approval for testing healthy controls was provided by the Cambridgeshire 2 Regional Ethics Committee . All clinical investigations were conducted in accordance with the Declaration of Helsinki . A convenience sample of 22 neurologically healthy adults participated in the study . Data from two participants could not be used due to technical issues , leaving 20 participants ( 9 male; 11 female ) ( mean age = 30 . 85; SD = 10 . 98 ) whose data were analysed . Each experimental run began with an awake baseline period lasting 25–30 minutes ( Fig 1A ) following which a target-controlled infusion of propofol [46] was commenced via a computerized syringe driver ( Alaris Asena PK , Carefusion , Berkshire , UK ) . With such a system the anesthesiologist inputs the desired ( “target” ) plasma concentration , and the system then determines the required infusion rates to achieve and maintain the target concentration ( using the patient characteristics which are covariates of the pharmacokinetic model ) . The Marsh model is routinely used in clinical practice to control propofol infusions for general anesthesia and for sedation . Three blood plasma levels were targeted– 0 . 6μg/ml ( mild sedation ) , 1 . 2μg/ml ( moderate sedation ) , and recovery from sedation . The state of mild sedation was aimed to engender a relaxed but still responsive behavioural state . At each target level , a period of 10 minutes was allowed for equilibration of plasma propofol concentrations to attain a steady state , following which behavioural tests and EEG measurements were commenced . After cessation of infusion , plasma propofol concentration exponentially declined toward zero . Computer simulations with the TIVATrainer pharmacokinetic simulation software revealed that plasma concentration of propofol would approach zero in 15 minutes leading to behavioural recovery; hence behavioural assessment was recommenced 20 minutes after cessation of sedation . Blood samples of 1cc each were taken at the beginning and end of the mild and moderate sedation states , and once at recovery , as indicated in Fig 1A . In total , 5 blood samples were taken during the study . These samples were analysed offline for characterising the significant inter-individual variability in actual propofol levels in blood plasma . We confirmed that the samples taken at the beginning and end of mild and moderate sedation had similar values of propofol concentration . The average of the two values , along with the value at recovery , were used as distinct covariates for EEG data analysis . At each of the 4 steady-state levels above , participants were requested to perform a simple behavioural task involving a fast discrimination between two possible auditory stimuli ( Fig 1A ) . Specifically they were asked to respond with a button press to indicate whether a binaurally presented stimulus was a buzz or a noise . These stimuli constituted either broadband noise or a harmonic complex with a 150Hz fundamental frequency ( buzz ) . Forty such stimuli , twenty of each kind , were presented in random order over two blocks , with a mean inter-stimulus interval of 3 seconds . We calculated a participant’s cognitive processing of these stimuli at each sedation level based on their hit rates , i . e . , percentage of correct responses . In addition , we measured reaction times based on the delay between auditory tone onset and correct button press . We employed binomial modelling to distinguish participants who became behaviourally impaired during moderate sedation , from those who remained responsive , albeit with slower reaction times . Specifically , we fitted a binomial distribution to each participant’s hit rates at baseline and during moderate sedation . With each fitted model , the distribution parameter p , the probability of a correct response , and its 95% confidence intervals were estimated . For a given participant , if the confidence interval at moderate sedation was lower than and non-overlapping with that at baseline , they were considered to have become significantly impaired , and we designated them as drowsy . If the confidence intervals overlapped , we designated them as responsive . From each participant , approximately 7 minutes of 128-channel high-density EEG data were collected at each level of sedation . EEG was measured in microvolts ( uV ) , sampled at 250Hz and referenced to the vertex , using the Net Amps 300 amplifier ( Electrical Geodesics Inc . , Eugene , Oregon , USA ) . Participants had their eyes closed in a resting state during data collection . Data from 91 channels over the scalp surface ( Fig 2 ) were retained for further analysis . Channels on the neck , cheeks and forehead , which tended to contribute most of the movement-related noise , were excluded . Retained channels were filtered between 0 . 5–45Hz , and segmented into 10-second long epochs . Each epoch thus generated was baseline-corrected relative to the mean voltage over the entire epoch . Data containing excessive eye movement or muscular artefact were rejected by a quasi-automated procedure: abnormally noisy channels and epochs were identified by calculating their normalised variance and then manually rejected or retained by visual inspection . After pre-processing , a mean ( SD ) of 38 ( 5 ) , 39 ( 4 ) , 38 ( 4 ) and 40 ( 2 ) epochs were retained for further analysis in the baseline , mild sedation , moderate sedation and recovery conditions , respectively . An ANOVA revealed no statistically significant difference between the numbers of epochs retained . Finally , previously rejected channels were interpolated using spherical spline interpolation , and data were re-referenced to the average of all channels . These processing steps were implemented using custom MATLAB scripts based on EEGLAB [47] . Fig 2 depicts the data processing pipeline employed to calculate spectral power and connectivity measures from the clean EEG datasets . Spectral power values within bins of 0 . 25Hz were calculated using Fourier decomposition of data epochs using the pwelch method . At each channel , power values within canonical frequency bands , namely delta ( 0–4Hz ) , theta ( 4–8Hz ) , alpha ( 8–15Hz ) , beta ( 12-25Hz ) and gamma ( 25–40Hz ) , were converted to relative percentage contributions to the total power over all five bands . Alongside , cross-spectrum between the time-frequency decompositions ( at frequency bins of 0 . 49Hz and time bins of 0 . 04s ) of every pair of channels was used to calculate debiased weighted Phase Lag Index ( dwPLI , see [22] ) . For a particular channel pair and frequency band , mean dwPLI across all time at the peak frequency within each band was recorded as the ambient amount of connectivity between those channels . dwPLI is a sensitive measure of connectivity between cortical regions that has been shown to be robust against the influence of volume conduction , uncorrelated noise , and inter-subject variations in sample size [22] , and has previously be used to characterise connectivity in pathological [48] and pharmacological [49] alterations in consciousness . However , as pointed out by Vinck , Oostenveld [22] , dwPLI is relatively insensitive to true connectivity at phase differences close to 0 or 180 degrees . Further , the actual locations of brain sources producing dwPLI connectivity between a pair of sensors might not necessarily be spatially proximal to those sensors . Nevertheless , for the purposes of this study , it provides a robust measure for estimating how this indirect connectivity is affected by propofol sedation . Phase-amplitude coupling ( PAC ) , also referred to as cross-frequency coupling [50] , was used to measure the propofol-induced changes in the relationship between the phase of ongoing oscillations in the slow ( 0 . 5–1 . 5Hz ) and alpha ( 8–15Hz ) bands at each channel . Calculation of PAC was based on the Direct PAC estimator formally defined by Ozkurt and Schnitzler [51] and implemented in the Brainstorm 3 . 2 toolbox [52] . Purdon , Pierce [18] and Mukamel , Pirondini [35] previously identified changes from trough-max to peak-max PAC during propofol sedation , as determined by whether the slow oscillation is at its trough ( at a phase angle of pi ) or its peak ( phase angle of 0 ) when alpha power is maximal , respectively . Such variations were measured by assigning a negative or positive sign to the amplitude of the complex-valued Direct PAC estimator depending on whether its phase angle was closer to pi or 0 radians , to indicate trough-max and peak-max coupling respectively . The 91x91 subject-wise , band-wise dwPLI connectivity matrices were thresholded to retain between 50–10% of the largest dwPLI values . They were then represented as graphs with the channels as nodes and non-zero values as links between nodes . The lowest threshold of 10% ensured that the average degree was not smaller than 2 * log ( N ) , where N is the number of nodes in the network ( i . e . , N = 91 ) . This lower boundary guaranteed that the resulting networks could be estimated [26] . Similar ranges of graph connection densities have been shown to be the most sensitive to the estimation of ‘true’ topological structure therein [53 , 54]: higher levels of connection density result in increasingly random graphs , while lower levels result in increasingly fragmented graphs . At each step of the connection density between 50% and 10% in steps of 2 . 5% , the thresholded graphs were submitted to graph-theoretical algorithms implemented in the Brain Connectivity Toolbox [55] . These algorithms were employed to calculate metrics that captured key topological characteristics of the graphs at multiple scales , and avoided the multiple comparisons problem entailed by comparing large numbers of network connections . These included the micro-scale clustering coefficient and macro-scale characteristic path length [26] , alongside meso-scale measures like modularity and community structure [56] , and participation coefficient [29] . Here , this functional notion of modularity measures the extent to which the nodes of a graph can be parcellated into topologically distinct modules with more intra-modular links than inter-modular links [28] . Modularity as calculated by the heuristic Louvain algorithm , and all measures derived therefrom , were averaged over 50 repetitions . Next , each graph metric thus derived was normalised by the average of 50 null versions of the metric similarly derived , but after repeatedly phase-randomising the original cross-spectra and recalculating dwPLI for each channel pair . Finally , the small-worldness index of a graph was calculated as the ratio of normalised clustering coefficient to characteristic path length [57] . Metrics were compared using two-way ANOVAs with one non-repeated ( group ) measure and one repeated ( level of sedation ) measure . The obtained p-values were corrected for violations of sphericity using a Greenhouse-Geisser correction . Pairwise tests between groups were corrected for multiple comparisons using Tukey’s HSD test . The ability of graph metrics of individual participants to predict their behaviour was tested using robust linear regression [58] to calculate R2 and p-values .
|
Though scientific understanding of how brain networks generate consciousness has seen rapid advances in recent years , application of this knowledge to accurately track transitions to unconsciousness during general anaesthesia has proven difficult due to considerable variability in this gradual process across individuals . Using high-density electroencephalography , we studied changes in these networks as healthy adults were sedated using propofol . By measuring their behavioural responsiveness and amount of sedative in their blood , we found a striking pattern: the strength of their brain networks before sedation predicted why some participants lost consciousness while others did not , despite registering similar blood levels of drug . By uncovering underlying signatures of this variability , our findings could enable accurate brain monitoring during anaesthesia and minimise intra-operative awareness .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2016
|
Brain Connectivity Dissociates Responsiveness from Drug Exposure during Propofol-Induced Transitions of Consciousness
|
Age patterns in asymptomatic and symptomatic infection with Leishmania donovani , the causative agent of visceral leishmaniasis ( VL ) in the Indian subcontinent ( ISC ) , are currently poorly understood . Age-stratified serology and infection incidence have been used to assess transmission levels of other diseases , which suggests that they may also be of use for monitoring and targeting control programmes to achieve elimination of VL and should be included in VL transmission dynamic models . We therefore analysed available age-stratified data on both disease incidence and prevalence of immune markers with the aim of collating the currently available data , estimating rates of infection , and informing modelling and future data collection . A systematic literature search yielded 13 infection prevalence and 7 VL incidence studies meeting the inclusion criteria . Statistical tests were performed to identify trends by age , and according to diagnostic cut-off . Simple reversible catalytic models with age-independent and age-dependent infection rates were fitted to the prevalence data to estimate infection and reversion rates , and to test different hypotheses about the origin of variation in these rates . Most of the studies showed an increase in infection prevalence with age: from ≲10% seroprevalence ( <20% Leishmanin skin test ( LST ) positivity ) for 0-10-year-olds to >10% seroprevalence ( >20% LST-positivity ) for 30-40-year-olds , but overall prevalence varied considerably between studies . VL incidence was lower amongst 0-5-year-olds than older age groups in most studies; most showing a peak in incidence between ages 5 and 20 . The age-independent catalytic model provided the best overall fit to the infection prevalence data , but the estimated rates for the less parsimonious age-dependent model were much closer to estimates from longitudinal studies , suggesting that infection rates may increase with age . Age patterns in asymptomatic infection prevalence and VL incidence in the ISC vary considerably with geographical location and time period . The increase in infection prevalence with age and peaked age-VL-incidence distribution may be due to lower exposure to infectious sandfly bites in young children , but also suggest that acquired immunity to the parasite increases with age . However , poor standardisation of serological tests makes it difficult to compare data from different studies and draw firm conclusions about drivers of variation in observed age patterns .
The Indian subcontinent ( ISC ) appears to be on course to reach the target of elimination of visceral leishmaniasis ( VL ) as a public health problem ( less than 1 case/10 , 000 people/year at sub-district level ) in most sub-districts by 2020 [1–3] . Once this goal has been achieved , surveillance methods that require fewer resources than active case detection will be required to monitor transmission and provide early warning of possible resurgence [4] . One proposed method is monitoring age patterns in infection prevalence by serology or other diagnostic tests [4] . This approach has been used successfully for other vector-borne diseases , such as malaria [5–7] , dengue [8] and Chagas disease [9–11] , However , it is unclear whether it would be effective for VL surveillance with currently available diagnostics . Additionally , mathematical models of VL transmission dynamics are useful tools for understanding disease patterns and designing cost-effective control strategies [2 , 12–15] , and it is unknown whether age-structure is required within such models . The existence of evidence for age-related risk of infection and the viability of age-stratified serology as a post-elimination surveillance tool are the key issues we address in this review . Almost 147 million people are at risk from VL in the ISC , caused by Leishmania donovani [16] . Elimination as a public health problem is considered to be possible based on the beliefs that indoor residual spraying ( IRS ) of insecticide is effective and that transmission is exclusively anthroponotic , combined with the availability of effective diagnostics and therapeutics to curtail the infectious period [17] . As of 2016 , significant progress had been made towards reaching the elimination target , with an 82% decrease in new cases since 2011 and the target being reached in a number of previously endemic sub-districts [18] . In 2017 , sub-district level incidence in India ranged between 0 and 12 cases/10 , 000 people/year and only 72 out of 633 endemic sub-districts were above the elimination target [19]; all endemic sub-districts in Bangladesh and districts in Nepal reported incidence <1 case/10 , 000 people/year [1 , 20] . Current elimination strategies in the ISC are based on early case detection and treatment , including education of at-risk populations , and methods to reduce abundance of the Phlebotomus argentipes sandfly vector such as IRS . Research has also identified the need to strengthen existing VL epidemiological surveillance programmes in order to aid disease detection and elimination [21] . However , only three large longitudinal VL studies have been carried out in order to assess infection and disease progression . These sources of data are the KALANET bed net trial in India and Nepal ( 2006–2009 ) [22] , the Tropical Medicine Research Council ( TMRC ) study in India ( 2007- ) [23] , and the CDC- and ICDDR , B-funded study conducted by Bern and co-workers in Bangladesh ( 2002–2004 ) [24 , 25] . The scarcity of detailed , contemporary longitudinal data means that the progression , epidemiology , and transmission dynamics of the disease are still poorly understood [4] . The majority of infected individuals are asymptomatic , and never develop clinical symptoms , and estimates of the ratio of incident asymptomatic infection to incident VL vary from 4:1 [24] to 17:1 [26 , 27] . This variation may be related to differences in transmission intensity and levels of immunity in different regions and time periods ( the ratio appears to decrease as VL incidence increases [27] ) , and/or differences in the definition of asymptomatic infection between studies ( i . e . the diagnostic test ( s ) and cut-offs used ) , as there is no agreed definition or gold-standard test for asymptomatic infection . Changes in levels of immunity and spatial patterns of transmission over time will also have affected age distributions of infection and disease , so the age distributions may hold important information about transmission rates in different settings and how they have varied with time . Identification of the age groups with the highest prevalence of asymptomatic infection ( who may act as a reservoir of transmission [28–30] ) and those most at-risk of clinical VL could aid appropriate targeting of interventions to reduce transmission and disease . Positivity on serological tests is an indication of exposure rather than immunity , but may be related to protective immunity . Age patterns in infection and disease are likely related to the immunological response to L . donovani infection , many aspects of which are still unknown . In particular , it is not known how long immunity to infection/disease lasts , the extent to which this depends on whether the individual recovered from asymptomatic infection or from clinical VL following treatment , and the degree of protection afforded [31] . Immunoglobulin G ( IgG ) antibody responses to L . donovani infection are not protective against disease [32–35] , but cell-mediated immune responses are [36–40] . Thus , positivity on the leishmanin skin test ( LST ) , a delayed-type hypersensitivity test , does represent protective immunity . Hence , assessing age patterns in sero-/LST-prevalence , sero-/LST-conversion and VL incidence may yield insights into durations of immune responses , and variation in immunity with age and VL endemicity . For example , data on seroconversion incidence by age from the TMRC study [23] suggests that seroconversion rate may increase with age . If this is true , it has potentially important implications for control of VL and modelling of VL transmission , as it suggests that exposure to infected sandflies increases with age and/or that individuals living in endemic areas reconvert to seropositivity due to repeated bites from infected sandflies . Age-dependent exposure has so far only be included in one transmission model of VL [3 , 41] . Therefore , an important question for modelling is whether there is evidence of age-dependence in the infection rate in the infection age-prevalence data from other studies , and whether this effect should be included in transmission models . In this paper , we review age-stratified data on L . donovani infection prevalence and clinical VL incidence and fit simple catalytic models to the age-prevalence data with the aim of improving understanding of age trends in asymptomatic and symptomatic infection . We estimate sero- , PCR- and LST- conversion and reversion rates from the data and compare them to estimates from longitudinal studies , and assess whether the conversion rates are age-dependent . Catalytic models have been used in meta-analyses for a number of diseases , including Chagas disease [11] , malaria [6] , varicella [42] and congenital rubella syndrome [43] , to assess changes in transmission levels and identify shifts in infection prevalence towards older ages indicative of reduced transmission . Our goal is to provide insight into the epidemiology of VL and help inform improvements in interventions aimed at eliminating the disease from the ISC .
Relevant studies were identified through a systematic literature review . The search was conducted using the PubMed engine via the search terms set out in S1 Text . In addition to the PubMed search , the bibliographies of five reviews [4 , 27 , 44–46] relevant to visceral leishmaniasis incidence/infection prevalence were analysed for references eligible for this study . Only studies relevant to VL in the Indian sub-continent were included in the review . Any studies referring to cutaneous leishmaniasis , muco-cutaneous leishmaniasis or conducted in another geographical area were omitted . The identified articles were subsequently screened based on their title and abstract , with eligible articles undergoing a full-text assessment . Articles were only included if age-stratified data was available for the incidence of clinical VL and/or prevalence of seropositivity/molecular test positivity/LST positivity , or the study-population/population-at-risk and the number of sero-/molecular-test-/LST-positives/clinical VL cases . All numerical data from the identified studies were doubly entered into spreadsheets and checked . The potential risk of bias in the included studies was assessed using the Newcastle-Ottawa bias assessment scale for observational studies [47] and the Cochrane risk of bias assessment tool for intervention studies [48] . The diagnostic tests used in the identified studies were: the direct agglutination test ( DAT ) , recombinant K39 enzyme-linked immunosorbent assay ( ELISA ) , rK39 rapid diagnostic test ( RDT ) , polymerase chain reaction ( PCR ) , quantitative PCR ( qPCR ) , and the leismanin skin test ( LST ) . Brief descriptions of the typical protocols for these tests are provided in S4 Text . The tests measure different aspects of infection and/or associated immune responses , as summarised in Table 1 , so care is needed in interpreting and comparing their respective age prevalence patterns . DAT and rK39 ELISA are serological tests for antibodies ( non-protective ) against L . donovani parasites , the rK39 RDT is a rapid test form of the ELISA designed for diagnosis of clinical VL , PCR/qPCR is a molecular test for parasite DNA in the peripheral blood , and LST is a delayed-type hypersensitivity test for protective T-cell-mediated immune responses . Positivity on the different tests is believed to correlate to differing time since infection [23 , 49 , 50] . DAT and rK39 ( ELISA and RDT ) positivity represent more recent infection than LST positivity , since antibody responses are generally much shorter lived than cell-mediated immune responses [24 , 37 , 38 , 50 , 51] . The rK39 rapid test has the positive cut-off set for clinical diagnostic purposes and does not detect the low-titre antibody responses which are more common in asymptomatic infection . PCR positivity is also thought to represent recent infection [52] , due to the ability of PCR to detect low numbers of parasites in peripheral blood during active infection , although longitudinal data on persistence of PCR positivity are lacking . Although DAT and rK39 are both antibody tests , a number of important differences make it difficult to compare them , and to compare different DAT studies and different rK39 studies . These include the type of antigen used ( single recombinant for rK39 vs whole-parasite lysate for DAT ) , how the tests are standardised against known positive and negative controls ( which differs between studies ) , how the DAT is read [53] and the cut-off chosen for seropositivity ( which also varies between studies ) . The molecular tests target different genes and are performed on varying quantities of blood , affecting the diagnostic accuracy . The following data was extracted from the papers: study/institution under which the data was collected ( e . g . KALANET bed net trial , TMRC study , Indian Council of Medical Research ) ; start & end date of the study; country , state , district & ( where available ) subdistrict of the study; number of villages; total population at risk; study population; number of excluded individuals; case definitions; type ( s ) of serological/diagnostic test used; age range & population of each age-group; number of individuals who underwent diagnostic testing , number of ( sero ) positive individuals; ( sero ) positive prevalence , number of clinical VL cases; and clinical VL incidence . Freely available software for digitising data [55] was employed to obtain data values from studies that did not provide numerical values for VL incidence/positive diagnostic prevalence , but provided relevant figures for calculation . The age stratifications in the identified studies were used for statistical analyses ( see S1 Data ) . For each diagnostic test , the prevalence of infection ( the proportion of the population who were positive on the diagnostic test ) in each age group in each study was plotted with exact binomial ( Clopper-Pearson ) 95% confidence intervals ( CI ) at the mid-point of the age group , to allow visual assessment and comparison of age trends in prevalence . Similarly , VL incidence ( as number of cases/1000 study population/year ) was calculated and plotted with 95% Poisson confidence intervals for each age group for each of the incidence studies . Odds ratios ( ORs ) for the risk of being seropositive/LST-positive and risk ratios ( RRs ) for having VL in each age group compared to the youngest age group were calculated with 95% CIs and 2-tailed p-values ( using the Z-test ) to identify any statistically significant variation in sero-/LST-positivity and VL incidence with age ( at significance level 0 . 05 ) . The chi-squared test for trend was used to assess trends in VL incidence with age ( at significance level 0 . 05 ) . For studies in which multiple diagnostic tests were performed on the same individuals , the age-specific prevalence according to each test was plotted on the same graph for visual comparison and the non-parametric Friedman test was used to assess agreement in the age-specific prevalence between the different tests , using the VassarStats online computation tool [56] . Where possible , agreement between the different tests was also assessed by calculating Cohen’s kappa coefficient [57] ( a measure of agreement in classification between two tests , with 1 corresponding to perfect agreement and values ≤0 to no agreement ) , or retrieving the calculated value from the original study . The infection prevalence age distribution data was modelled using a reversible catalytic model [58] in which negative ( sero-/LST-/PCR-negative ) individuals become positive ( sero-/LST-/PCR-positive ) at a rate λ and revert back to sero-/LST-/PCR-negativity at a rate γ ( see S2 Text for full details ) . We tested different versions of the model in which λ depends on age , the study ( i . e . its location and time period ) and/or the diagnostic test used and γ depends on the study and/or test ( since the tests measure different immunological responses that may happen at different points after infection and over different timescales ) . We fitted the different versions of the model to the data to estimate λ and γ using maximum likelihood estimation and compared models using the Akaike information criterion ( AIC ) ( see S2 Text for further details ) .
A total of 19 age-stratified diagnostic and epidemiological studies that met the inclusion criteria were identified from the systematic literature review ( Fig 1 ) . Seven of these studies contained data on age-specific VL incidence ( Table 2 ) , and 13 studies contained age-stratified infection prevalence data ( Table 3 ) ( [23] contained both types of data ) . A further 8 studies containing age-stratified data are included in S1 Data ( see S2 Data for definitions of variables in S1 Data ) , but are excluded from the analysis as the data is not comparable to that in the other studies , e . g . due to differences in study design or participant inclusion criteria . The assessment of the risk of bias in the included studies is shown in Tables 1–4 in S3 Text . Sources of potential bias include the lack of sample size justification in all but one of the cross-sectional studies and their failure to demonstrate comparability of non-respondents and comparability of different outcome groups . Loss to follow up was a potential source of bias in most cohort studies . We were unable to assess the risk of publication bias due to the small number of studies that met the inclusion criteria . The age-specific VL incidence curves for the studies in Table 2 are shown in Fig 2 . The large variation in measured incidence between different studies is immediately apparent , and not unexpected given the differences in incidence between different geographical locations and different time periods . The highest average VL incidence , 13 . 2 cases/1000 people/year , was observed in the study of Bern et al [25] , in Fulbaria upazila , Mymensingh district , Bangladesh , between 1999 and 2004 . There is a general pattern of decreasing VL incidence with increasing age beyond 20 years ( Fig 2 ) , although the trend is not significant for all studies ( see p-values for chi-squared trend test in S3 Data ) . The exception ( Barnett et al [59] ) was conducted in outbreak villages in low-endemicity areas in Uttar Pradesh , India . Most of the studies for which incidence data for 0-10-year-olds is available show lower VL incidence among young children ( 0-5-year-olds ) than older children and young adults ( 5-20-year-olds ) , with a peak in incidence in the 5-20yr age group ( see RRs in S3 Data ) . Fig 3 shows age-prevalence curves for L . donovani infection from the studies in Table 3 , with results separated by test . For studies in which multiple tests were performed on the population [23 , 24 , 68 , 70] the data are plotted together in Fig 4 to allow comparison of age-specific prevalence according to the different tests .
The main conclusion that can be drawn from this review is that age patterns of L . donovani infection measured using current serological tests appear to be too variable across different settings and endemicity levels to be used to monitor levels of ongoing transmission post elimination . The extent to which this variability is due to genuine variation in the age-prevalence distribution with location and time , to properties of the tests , and/or to inconsistent test standardisation is unclear . However , the fact that significant age trends have only been observed in large studies suggests that very large sample sizes would be needed to reliably detect changes in transmission levels based on age patterns . Further longitudinal studies are required to improve understanding of the dynamics of serological responses and to determine whether serological tests can be used as a surveillance tool to monitor transmission . If such studies can demonstrate that well-standardised serological testing provides a reliable indicator of transmission , cross-sectional serological surveys may still prove to be a useful tool for achieving and sustaining elimination of VL as a public health problem in the ISC .
|
As the elimination target for visceral leishmaniasis ( VL ) in the Indian subcontinent ( <1 case/10 , 000 people/year ) is approached , there is a growing need for surveillance tools with which to monitor transmission to ensure the target is sustained , especially given the large proportion of infections which are asymptomatic ( ~75–95% ) . One potential approach to estimate underlying transmission patterns may be to track age patterns in infection or cumulative exposure using diagnostic tests . However , current understanding of age patterns in asymptomatic infection and clinical VL is poor , in particular regarding possible age-dependence of infection rates . Our systematic review and pooled-analysis of age-stratified data on infection prevalence and disease incidence suggests that available diagnostics , as currently implemented , fail to meet the requirements for a reliable tool for assessing transmission , due to inconsistent standardisation and highly variable age-prevalence patterns across different settings . It also finds weak evidence for infection rates increasing with age , though further longitudinal studies are needed to test this hypothesis and to assess whether properly standardised diagnostic tests could be used to monitor ongoing transmission .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"enzyme-linked",
"immunoassays",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"immunology",
"tropical",
"diseases",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"age",
"groups",
"protozoans",
"leishmania",
"neglected",
"tropical",
"diseases",
"antibodies",
"immunologic",
"techniques",
"research",
"and",
"analysis",
"methods",
"immune",
"system",
"proteins",
"infectious",
"diseases",
"zoonoses",
"serology",
"proteins",
"immunoassays",
"protozoan",
"infections",
"leishmania",
"donovani",
"people",
"and",
"places",
"biochemistry",
"eukaryota",
"diagnostic",
"medicine",
"physiology",
"leishmaniasis",
"population",
"groupings",
"biology",
"and",
"life",
"sciences",
"organisms"
] |
2018
|
Age trends in asymptomatic and symptomatic Leishmania donovani infection in the Indian subcontinent: A review and analysis of data from diagnostic and epidemiological studies
|
Vocal learning , the substrate of human language acquisition , has rarely been described in other mammals . Often , group-specific vocal dialects in wild populations provide the main evidence for vocal learning . While social learning is often the most plausible explanation for these intergroup differences , it is usually impossible to exclude other driving factors , such as genetic or ecological backgrounds . Here , we show the formation of dialects through social vocal learning in fruit bats under controlled conditions . We raised 3 groups of pups in conditions mimicking their natural roosts . Namely , pups could hear their mothers' vocalizations but were also exposed to a manipulation playback . The vocalizations in the 3 playbacks mainly differed in their fundamental frequency . From the age of approximately 6 months and onwards , the pups demonstrated distinct dialects , where each group was biased towards its playback . We demonstrate the emergence of dialects through social learning in a mammalian model in a tightly controlled environment . Unlike in the extensively studied case of songbirds where specific tutors are imitated , we demonstrate that bats do not only learn their vocalizations directly from their mothers , but that they are actually influenced by the sounds of the entire crowd . This process , which we term “crowd vocal learning , ” might be relevant to many other social animals such as cetaceans and pinnipeds .
Vocal learning , the ability to learn to produce vocalizations by hearing , is essential in human language acquisition , but only a few other mammals appear to possess this capability [1–8] . Some indications for the existence of vocal learning in nonhuman animals arise from the observation of group-specific vocal dialects in wild populations [9–11] . Such vocal variations can indeed stem from vocal learning of typical vocalizations by members of the group; however , it is usually impossible to completely exclude other explanations for the appearance of vocal differences between populations [12] . For instance , genetic variations may lead to unique vocal patterns , and environmental constraints may induce specific usage of vocalizations . Studies of several species of bats have indicated their vocal learning ability [4] . Early studies suggested that Phyllostomus discolor pups adapt their isolation calls to their mothers’ directive calls [13] , and P . hastatus females were shown to maintain a group-specific foraging call through vocal learning [14] . Geographic variations in vocalizations of these 2 species were also observed [15 , 16] , though genetic and environmental factors were not excluded as possible contributors to these apparent dialects . In another bat species ( Saccopteryx bilineata ) that is an important model for vocal learning , pups have been shown to learn territorial songs from adult male tutors [17] and to engage in vocal babbling behavior [18] . In a previous study [19] , we showed that depriving Egyptian fruit bat ( Rousettus aegyptiacus ) pups from hearing adults delays their vocal ontogeny . Yet we also found that these isolated pups eventually catch up with their control counterparts . Moreover , we have not shown plasticity in the vocal ontogeny of non-isolated pups . The Egyptian fruit bat is an extremely social and vocal mammal , living in colonies of dozens to thousands of individuals . In the wild , these bats are exposed to extensive vocal communication throughout their entire lives . A typical vocalization of this species is composed of a sequence of multiharmonic calls ( Fig 1; see Materials and methods for details ) . The fundamental frequency ( F0 ) in newborn pup isolation calls is high ( ca . 8–15 kHz ) ( Fig 1A ) , and it gradually decreases to ca . 0 . 2–1 . 2 kHz in adults ( Fig 1B–1D ) . We have previously shown that this process involves vocal learning [19] . A fruit bat pup is mostly exposed to adult vocalizations when in the roost . In this situation , the pup continuously hears countless vocalizations coming from the surrounding darkness and has very little , if any , interaction with most of the vocalizing individuals . It is therefore exposed to a cacophony of fruit bat vocalizations , only a slight minority of which are emitted by its mother or by nearby roostmates . In this study , we therefore set to examine whether the vocal communication of pups that grow up in such an environment is shaped by the individuals that they directly interact with or by the background vocalizations they are “passively” exposed to . We raised pups in conditions that mimic the natural acoustic conditions of a dark fruit bat cave and observed the establishment of vocal dialects through vocal learning of the entire “crowd” in the artificial cave .
We caught pregnant female Egyptian fruit bats in wild roosts in central Israel . The bats were then randomly assigned to 3 identical and acoustically isolated chambers . Each female gave birth to a single pup in these chambers ( resulting in 3 groups of 5 , 5 , and 4 pups ) . The mothers were released a few weeks after the pups were weaned ( at the age of ca . 14 weeks ) . In each of the 3 chambers , a playback of conspecific vocalizations was constantly played from day 1 and for a full year ( see Materials and methods ) . The playback intensity and frequency mimicked the vocalizations of ca . 100–200 adults . The pups were thus exposed to a situation similar to a natural roost , hearing their mothers' vocalizations embedded within the noise created by a crowd of hundreds of bats . The playbacks were sampled from a set of thousands of agonistic vocalizations previously recorded in the same setup . Agonistic calls constitute almost all of the vocalizations emitted in the roost by this species [20] . They are elicited as a response to unsolicited physical contact and are characterized by a typical range of acoustic features ( S1 Fig ) . We chose to vary the F0 of the calls after observing ( in [19] ) that this is a feature that is strongly influenced by exposure to adult vocalizations . According to the distribution of F0 across bat calls , we defined 3 groups of calls: low-fundamental calls ( Low-F0 , with F0 lower than 250 Hz ) , high-fundamental calls ( High-F0 , with F0 higher than 1 , 315 Hz ) , and intermediate-fundamental calls ( the rest and the majority of calls , see Fig 1E and Materials and methods for details ) . Playbacks were assigned to each experimental group according to their F0 content . The control group ( n = 5 ) was exposed to playbacks randomly sampled from the previously recorded repertoire with an average F0 of 564 Hz ( 1% High-F0 and 11% Low-F0 calls , black line in Fig 1F ) . The Low-F0 group ( n = 5 ) was exposed to playbacks with an average F0 of 303 Hz ( 0 . 2% High-F0 and 52% Low-F0 calls , blue line in Fig 1F ) . The High-F0 group ( n = 4 ) was exposed to playbacks with an average F0 of 1 , 871 Hz ( 26% High-F0 and 9% Low-F0 calls , red line in Fig 1F ) . Note that the High-F0 group was exposed to a playback that was highly unnatural in 2 ways: 1 ) it contained approximately 26 times more high-frequency calls than the typical adult repertoire ( the Low-F0 playback only contained approximately 4 . 5 times more Low-F0 calls ) , and 2 ) pup vocal ontogeny is typically characterized by decreasing the call fundamental , while this playback aimed to drive the pups in the opposite direction . We therefore expected that the High-F0 playback would be more difficult to mimic than the Low-F0 playback . The pups were housed in their birth chambers for the entire experiment period ( approximately 1 year ) , except for during the recording sessions , and the playbacks were constantly played in these birth chambers throughout the year of the experiment . The pups were recorded 4 times during the experiment , at the ages of 12–18 weeks , 31–35 weeks , 40–43 weeks , and 48–51 weeks . To ensure identical recording conditions , the recordings were performed in a fourth identical acoustic chamber . Each group was moved to the recording chamber for a few days in a rotating manner throughout each recording session ( which therefore lasted approximately 1 month ) . As expected , all recorded vocalizations were agonistic ( elicited as a response to unsolicited physical contact ) , and no behavioral differences were observed between the groups . The pups in the 3 treatments developed 3 distinct vocal dialects over time . In order to quantify the acoustic differences between the pups and test for any relation to the playbacks , we first calculated a set of 7 acoustic features for each recorded call and for each of the playback calls ( see Materials and methods ) . Using these features , we performed a linear discriminant analysis ( LDA ) on the calls of the 3 playbacks to obtain the 2 axes that best separated the playbacks ( S1 Table ) . We then projected the recorded pup vocalizations on these 2 axes ( Fig 2; see Materials and methods for details and S2 Table for number of analyzed calls ) . At a very young age ( after 12–18 weeks of exposure to the playback ) , a large variability was observed with no significant distinction between the groups ( Fig 2A; permutation test for linear discriminability: p = 0 . 09 ) , though some discrepancy between them may have already been present . When the pups matured , the groups became acoustically significantly separable ( Fig 2B–2D; permutation tests for linear discriminability: p = 3 . 2 × 10−5 , p = 0 . 0075 , and p = 5 . 6 × 10−5 at the ages of 31–35 , 40–43 , and 48–51 weeks , respectively ) . These findings present the formation of 3 dialects in the lab and suggest a connection between the established dialects and the auditory experience ( as explained below ) . Importantly , discriminant analysis is typically used on the “experimental data” ( in our case , the pups' vocalizations ) to examine separation; however , in the current analysis , the axes presented in Fig 2 were chosen to discriminate between the “treatments” ( i . e . , the playbacks ) and not between the pup vocalizations . This means that we did not deliberately project the data on the dimensions that separated the pups best but rather on the predefined dimensions that best separate the stimulus they were exposed to . Therefore , the axes of all of the panels in Fig 2 are the same ( as they were determined by the playbacks ) . The fact that the pup calls were clustered into 3 distinct groups along these treatment-axes strongly suggests that they were influenced by the playbacks . The acoustic features that mainly contributed to these 2 separating axes include the F0 , the energy entropy , and the spectral centroid ( S1 Table ) . It is also important to note that the pups were recorded in an environment with no playbacks ( in the recording chamber ) ; thus , they were recorded when interacting with each other after they had assimilated the conspecific vocalizations heard in their home chambers . To directly test the effects of the playbacks on the pups , we compared the acoustic parameter that we directly manipulated—i . e . , the use of different F0 in each of the groups ( Fig 3 , S2 Fig ) . The F0 distributions in the Low-F0 group and the High-F0 group were indeed biased according to their respective playbacks ( linear mixed models; A model for usage of Low-F0 calls: significant difference between the groups - p = 0 . 0004 , post-hoc test for difference between the Low-F0 and control groups - p = 0 . 0001; A model for usage of High-F0 calls: significant difference between the groups - p = 0 . 0008 , post-hoc test for difference between the High-F0 and control groups - p = 0 . 0002; see Materials and methods for details ) . All 3 groups mostly used calls with F0 around the peak of the control distribution ( approximately 600 Hz ) , suggesting an innate preference ( see Discussion ) . However , the pups in the Low-F0 group used significantly more low-fundamental ( i . e . , lower than 250 Hz ) calls than the control group from the age of ca . 31 weeks onward , in accordance with the playback they were exposed to ( blue line and blue arrow in Fig 3 , Mann–Whitney U test: p = 0 . 004 , p = 0 . 004 , and p = 0 . 016 in the second , third , and fourth recording sessions , respectively; S3J–S3L Fig ) . Similarly , the High-F0 group used significantly more high-fundamental ( i . e . , higher than 1 , 315 Hz ) calls than the control group , at least until the age of ca . 43 weeks in accordance with the playback they were exposed to ( red line and red arrow in Fig 3; Mann–Whitney U test: p = 0 . 032 , p = 0 . 032 in the second and third recording sessions , respectively; S3B and S3C Fig ) . Because of their small absolute number , the use of High-F0 calls by this group can be better seen when examining the ratio between the distribution of the High-F0 and the control groups ( Fig 3 , bottom row ) . We controlled for the possibility that the dialects we observed resulted from physiological or genetic differences . We verified that the bats within each group were not more genetically related to each other than to the bats in the other groups or to bats in the general population , i . e . , the intragroup relatedness did not significantly differ from the intergroup relatedness or the general population relatedness ( see Materials and methods ) . We also verified that there was no significant difference in F0 usage between males and females ( S3 Table , S4 Table ) and that there was no correlation between body size ( estimated by body weight ) and F0 usage in any group at any recording session ( S4 Table ) .
This study adds substantial evidence for the importance of vocal learning in the ontogeny of bat vocal communication . The highly controlled playback experiments that we performed excluded possible biasing factors such as differences in the ecological , developmental , or genetic backgrounds of the subjects or even differences in the recording conditions , all of which might lead to false reports of vocal learning . It is important to note that , in the wild , as well as in our setup , bats are exposed to an immense amount of vocalizations produced by conspecifics in the dark . Thus , young pups hear conspecifics that do not directly interact with them to an extent that quantitatively overshadows the vocalizations produced by their mothers or immediate neighbors . Accordingly , we found that our pups presented a “crowd vocal learning” phenomenon , where their vocal repertoire was shaped by the complete repertoire they heard in their colony ( mainly governed by our playbacks ) and not only by the vocalizations of a single tutor ( e . g . , their parents ) as is mostly discussed in the songbird literature [21] . Vocal learning is often assumed to include imitation [1] or at least social reinforcement of specific vocalizations [8] . The bats in our study did not interact with their models and hence were not subject to reinforcement , and we cannot assert that they imitated specific calls . It may be in line with recent views , which dispute the dichotomous definition of ( presence or absence of ) vocal learning abilities and rather find varying levels of this skill among different species [22] . Furthermore , when syllables are not readily categorized into specific types , as in the case of fruit bat vocalizations [20] , it might be more difficult to identify imitation than when clear syllable types are recognized ( as in the case of many birdsongs ) . Yet the bat crowd vocal learning demonstrates some degree of imitation , with an apparent tendency to social conformity . We hypothesize that such crowd vocal learning may be employed by other species that are exposed to many vocalizations of conspecifics without directly interacting with them . Such auditory exposure occurs , for instance , in many cetaceans , whose calls travel very long distances , or in congregating species such as pinnipeds and some sea birds ( in which vocal learning has so far not been described ) . Several aspects of the behavior of the High-F0 group suggest that innate preferences also play a role in vocal ontogeny: 1 ) The bats have not adopted calls with F0 above 2 kHz , although these were abundant in the playback . Such high F0 calls characterize subadults and are very rarely emitted by adults , and 2 ) They reduced the use of high F0 calls when reaching sexual adulthood . At the age of 43 weeks ( approximately 300 days ) , the bats are already mature , and the use of high-fundamental calls at this age is extremely rare in fruit bats ( possibly due to physical constraints ) . Hence , it seems that a bias that is related to the animal’s physiology overrides learning of too-high-fundamental calls after a certain age ( High-F0 group , Mann–Whitney U test: p = 0 . 14 in the fourth recording session; S3D Fig ) . Note also that the High-F0 bats also included more low F0 calls in their repertoire relative to the controls ( red outlined arrow , Fig 3 ) . We can only hypothesize that this was due to their lesser exposure to calls around the control peak ( approximately 600 Hz ) . Importantly , even if the High-F0 bats reduced the excess of high-frequency calls in their repertoire towards the end of the year , they still exhibited their unique vocal dialect that was also driven by additional acoustic properties . This can be learnt from the forming of separable groups in the time period of the last recordings ( Fig 2B–2D , note that the probability of getting a separable group by chance is extremely low; see for example 4 random permutations in S4 Fig and exact p-values above ) . One acoustic feature that contributed to the unique dialect of the High-F0 group was the energy entropy ( S5 Fig; also conforming to the LDA analysis in S1 Table ) . To conclude , in a tightly controlled acoustic environment , we observed the formation of vocal dialects as a result of crowd vocal learning . When such dialects are found in the wild , it is often difficult to exclude nonsocial factors , but in this study , the pups were raised and recorded in identical settings except for the playback they heard . Notably , shared intragroup behaviors acquired and transmitted through social learning are generally referred to as culture [12 , 23] . Furthermore , evidence for nonhuman culture is occasionally based on learned vocal behaviors of birds [24–26] and mammals [27 , 28] , with specific emphasis on vocal dialect variations between wild populations [29–31] . In our study , though pups did not directly learn from conspecifics , they were actually exposed to a conspecific stimulus that is very similar to that available to them in the wild ( i . e . , a stimulus that includes sound without vision or touch ) . Hence , our results demonstrate the assimilation of shared behavioral phenotypes , which were acquired by social vocal learning from a conspecific stimulus and thus might be considered as in-lab establishment of ( vocal ) culture in a mammalian model .
Adult , heavily pregnant female bats ( R . aegyptiacus ) were captured in 2 wild roosts in central Israel and were randomly mixed . The bats were kept in 3 identical acoustic chambers ( length: 190 cm; width: 90 cm; height: 82 cm ) large enough to allow flight and fed with a variety of fruit ad lib . The light/dark regime was 12 h/12 h . The bats were randomly assigned to 3 groups , each housed in 1 chamber: 5 bats in the High-F0 group , 5 bats in the Low-F0 group , and 5 bats in the control group . All bats gave birth inside the chambers . One pup of the High-F0 group and 1 pup of the control group died few days after birth . Subsequently , 1 mother with a pup approximately 1 . 5 months old ( caught in the wild roost ) was added to the control group when the pups were ca . 1 . 5 months old . All experiments were reviewed and approved by the Animal Care Committee of Tel Aviv University ( Number L-13-016 ) and were performed in accordance with its regulations and guidelines regarding the care and use of animals for experimental procedures . The use of bats was approved by the Israeli National Park Authority . In previous studies in this exact setup , we have recorded hundreds of thousands of bat vocalizations . Examining the distribution of the F0 among the recorded adult and subadult vocalizations ( Fig 1E ) , we defined 2 extreme groups of calls—High-F0 ( above 1 , 315 Hz , 2 SD above the mean ) and Low-F0 ( below 250 Hz , which is the minimum between the 2 modes in the bimodal distribution , 1 . 1 SD below mean ) . For the playbacks ( Fig 1F ) , we sampled the original dataset with 2 biased samples: one containing a high proportion of Low-F0 calls , which was played to the Low-F0 group , and one containing a high proportion of High-F0 calls ( including subadult vocalizations ) , which was played to the High-F0 group . For the control group , we used a random sample ( see diamond shapes in Fig 2; see also S3 Fig and lines in the middle row of Fig 3 for the F0 content of the playbacks ) . We used raw recordings ( audio files ) without any editing to keep the stimulus as natural as possible . All in all , 105 , 227 , and 191 different recordings were included in the High-F0 , Low-F0 , and control playbacks , respectively ( each group was exposed to the same number of played recordings during the entire experiment period , where each recording included a sequence of calls and represented a full vocal interaction that was recorded between adult bats; see below ) . The playback vocalizations were played around the clock with a timing distribution mimicking the natural vocal behavior of this species , where many of the vocalizations are emitted at dawn and dusk and more vocalizations are emitted during the night than during the day [20] . In each playback event , 1 vocalization ( a raw recording of a sequence of calls ) was selected randomly for each group , and these vocalizations were played concurrently in their corresponding chambers , i . e . , the playbacks were played in a random , nonrepeating order . The rate of the playbacks was 14 , 057 call-sequences ( i . e . , recordings ) per day and was the same in all 3 groups . Because not all sequences had the same number of calls , the groups heard 69 , 931 , 48 , 651 , and 129 , 715 calls per day on average for the Low-F0 , High-F0 , and control groups , respectively ( to clarify the difference between a recording and a call , see Fig 1C , where a recording with 4 calls is shown , and Fig 1D , depicting a recording with 3 calls ) . These might seem like large differences , but even in the treatment with the fewest calls ( i . e . , 48 , 651 calls per day ) , the pups were exposed to a playback rate that was approximately 16 times higher than the calling rate of 5 adult bats [20] . Thus , pups heard ( at least ) 16–30 times more playback vocalizations per day than the vocalizations produced by their mothers during the first 14 weeks of the experiment ( when the mothers were still present ) . We recorded the pups’ vocalizations in 4 recording sessions , when the pups were at the ages of 12–18 weeks , 31–35 weeks , 40–43 weeks , and 48–51 weeks . All ages are reported with an accuracy of ±15 days . During a recording session , each group of pups was transferred into a recording chamber , which was similar to the housing chambers . All pups in a group were transferred together ( except for part of the first recording session in which the pups were recorded in triplets; see S5 Table ) , recorded for 1–5 days , and returned to their home chamber . This transfer was repeated for each group in rotation until the end of the recording session , which lasted for 21–45 days , resulting in all groups being recorded for approximately the same time and no more than a few days apart ( see S5 Table for the detailed schedule ) . The recording chamber was continuously monitored with IR-sensitive cameras and omnidirectional electret ultrasound microphones ( Avisoft-Bioacoustics Knowles FG-O; 2 microphones in a cage , 1 in each side of the cage ) . Audio was sampled using Avisoft-Bioacoustics UltraSoundGate 1216H A/D converter with a sampling rate of 250 kHz . Raw audio recordings were automatically segmented and filtered for noises and echolocation clicks , leaving only bat social communication calls ( see [19] for details of this process ) . The video was synchronized to the audio , resulting in a short movie accompanying each audio recording . Videos were then analyzed by L . A . , who identified the emitter of each call . The bats were individually marked using fur bleaching . An emitter bat was recognized by its mouth movements , and 2–3 cameras could be used to verify a distinct assignment . If there was any doubt regarding the emitter's identity , we excluded the vocalization from the analysis . Social vocalizations of R . aegyptiacus are composed of sequences of separated calls ( in our analysis , we regarded a call as a vocalized segment of a duration of at least 20 ms that is separated by at least 4 ms of silence from other vocalized segments ) . The vocal sequences commonly contain between 1 to 20 calls , with an average length of 2 . 7 calls ( ±2 . 6 , SD ) per sequence ( see examples in Fig 1C and 1D ) and an average duration of 119 . 1 ms ( ±69 . 3 ms , SD ) per call . These calls are typically broadband ( with 90% of the energy spread between approximately 3–45 kHz ) , generally harmonic squawks , with an average F0 of 544 Hz for an adult bat ( F0 for a single call was defined as the geometric mean of the F0 content in that call ) . The calls are not readily clustered into different acoustic syllables ( in the past , we have tested many more features than were used in this paper ) . They rather appear to rest on an acoustic continuum ( see S1 Fig for a description of different acoustic features across the repertoire ) . They can thus all be considered as variations of one large “acoustic cloud” of agonistic calls . For each call , 7 acoustic features were extracted: log F0 , Shannon entropy of the power spectrum , Wiener entropy , spectral centroid , frequency with peak energy , amplitude entropy , and duration . The features were measured with a sliding window of 20 ms ( 19 ms overlap ) and were averaged for each call ( except for the duration , which was measured for the entire call ) . The F0 was calculated using the YIN algorithm [32] . This processing was computed over all recorded calls as well as all playback calls . We first examined the differences between the groups and their relation to the playbacks using LDA ( Fig 2 ) . To this end , we performed an LDA on the features extracted from the 3 playbacks , obtaining the 2 discriminant functions ( a projection of the 7 acoustic features onto a new 2-dimensional space , S1 Table ) that best discriminate between the playbacks . We then plotted the average of the calls of each pup in each recording session in these new 2 dimensions . The features were scaled prior to the application of the LDA by subtracting the mean and dividing by the SD , for both the playbacks and the pup vocalizations . The separation between the groups , which is clearly visible from the second recording session onwards , was evaluated for statistical significance ( using permutations ) as follows: For each recording session ( each panel in Fig 2 ) , we tested the linear separation between the group , i . e . , how many pups are correctly assigned to their group if straight lines are drawn to best separate the groups ( this was done using a second LDA applied to obtain the separation significance ) . We then tested all possible permutations of group assignments for the pups , keeping the number of pups in each group constant , and computed an exact p-value ( correct assignments in best separation: 10/14 , 14/14 , 12/14 , and 14/14 , with p-values: 0 . 09 , 3 . 2 × 10−5 , 0 . 0075 , and 5 . 6 × 10−5 , for recording sessions 1–4 , respectively ) . To control for possible sex biases ( i . e . , differences between males and females ) , we repeated these permutations while also keeping the male/female compositions of the groups , obtaining similar results ( p = 0 . 1 , p = 6 . 8 × 10−5 , p = 0 . 0076 , and p = 2×10−4 , for recording sessions 1–4 , respectively ) . In order to assess the statistical significance of the use of different F0 ( S3 Fig ) , we performed a mixed linear model analysis , testing the effect of the group on the development of Low-F0 usage or High-F0 usage . We also tested for a possible effect of the sex of the pups ( including it in the models ) and found no such significant effect ( see S3 Table ) . After finding an overall group effect , we used 1-tailed Mann–Whitney U tests to demonstrate the differences between the manipulation groups and the control group at each recording session ( S2 Fig ) . The mixed model analysis was performed in SPSS . All other processing and the analysis of the data were performed using Matlab 8 .
|
The spontaneous acquisition of speech by human infants is considered a keystone of human language , but the ability to reproduce vocalizations acquired by hearing is not commonly described in other mammals . This skill , termed vocal learning , is challenging to study in nonhuman animals since such investigation requires the detection and exclusion of innate developmental effects . The recognition of vocal dialects among different populations can open a window on the vocal learning abilities of animals , but such findings in the wild may reflect genetic or ecological differences between groups rather than the learning of group-specific vocal behavior . In this study , we used a playback-based lab experiment to induce vocal dialects in fruit bat pups . By exposing groups of pups to different playbacks of conspecific calls , we could establish separate dialects , demonstrating the vocal learning skill of these bats . Furthermore , while songbirds , for instance , learn their songs directly from a specific tutor , our bats showed the ability to pick up vocal variations from the surrounding crowd , without direct interaction with any given tutor .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"learning",
"acoustics",
"vocalization",
"linguistics",
"linear",
"discriminant",
"analysis",
"social",
"sciences",
"vertebrates",
"neuroscience",
"animals",
"mammals",
"learning",
"and",
"memory",
"animal",
"signaling",
"and",
"communication",
"cognitive",
"psychology",
"animal",
"behavior",
"mathematics",
"statistics",
"(mathematics)",
"zoology",
"discrete",
"mathematics",
"fruit",
"bats",
"combinatorics",
"research",
"and",
"analysis",
"methods",
"behavior",
"mathematical",
"and",
"statistical",
"techniques",
"short",
"reports",
"physics",
"psychology",
"eukaryota",
"permutation",
"bats",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"language",
"acquisition",
"cognitive",
"science",
"amniotes",
"statistical",
"methods",
"organisms"
] |
2017
|
Crowd vocal learning induces vocal dialects in bats: Playback of conspecifics shapes fundamental frequency usage by pups
|
Fragile X-associated tremor/ataxia syndrome ( FXTAS ) , a late-onset neurodegenerative disorder , has been recognized in older male fragile X premutation carriers and is uncoupled from fragile X syndrome . Using a Drosophila model of FXTAS , we previously showed that transcribed premutation repeats alone are sufficient to cause neurodegeneration . MiRNAs are sequence-specific regulators of post-transcriptional gene expression . To determine the role of miRNAs in rCGG repeat-mediated neurodegeneration , we profiled miRNA expression and identified selective miRNAs , including miR-277 , that are altered specifically in Drosophila brains expressing rCGG repeats . We tested their genetic interactions with rCGG repeats and found that miR-277 can modulate rCGG repeat-mediated neurodegeneration . Furthermore , we identified Drep-2 and Vimar as functional targets of miR-277 that could modulate rCGG repeat-mediated neurodegeneration . Finally , we found that hnRNP A2/B1 , an rCGG repeat-binding protein , can directly regulate the expression of miR-277 . These results suggest that sequestration of specific rCGG repeat-binding proteins could lead to aberrant expression of selective miRNAs , which may modulate the pathogenesis of FXTAS by post-transcriptionally regulating the expression of specific mRNAs involved in FXTAS .
Fragile X syndrome ( FXS ) , the most common form of inherited mental retardation , is caused by expansion of the rCGG trinucleotide repeat in the 5′ untranslated region ( 5′ UTR ) of the fragile X mental retardation 1 ( FMR1 ) gene , which leads to silencing of its transcript and the loss of the encoded fragile X mental retardation protein ( FMRP ) [1]–[6] . Most affected individuals have more than 200 rCGG repeats , referred to as full mutation alleles [7] . Fragile X syndrome carriers have FMR1 alleles , called premutations , with an intermediate number of rCGG repeats between patients ( >200 repeats ) and normal individuals ( <60 repeats ) [8] . Recently , the discovery was made that male and , to a lesser degree , female premutation carriers are at greater risk of developing an age-dependent progressive intention tremor and ataxia syndrome , which is uncoupled from fragile X syndrome and known as fragile X-associated tremor/ataxia syndrome ( FXTAS ) [9] , [10] . This is combined with cognitive decline associated with the accumulation of ubiquitin-positive intranuclear inclusions broadly distributed throughout the brain in neurons , astrocytes , and in the spinal column [11] , [12] . At the molecular level , the premutation is different from either the normal or full mutation alleles . Based on the observation of significantly elevated levels of rCGG-containing FMR1 mRNA , along with either no detectable change in FMRP or slightly reduced FMRP levels in premutation carriers , an RNA-mediated gain-of-function toxicity model has been proposed for FXTAS [13]–[17] . Several lines of evidence in mouse and Drosophila models further support the notion that transcription of the CGG repeats leads to this RNA-mediated neurodegenerative disease [11] , [15] , [17]–[19] . The hypothesis is that specific RNA-binding proteins may be sequestered by overproduced rCGG repeats in FXTAS and become functionally limited , thereby contributing to the pathogenesis of this disorder [15] , [17] , [19] , [20] . There are three RNA-binding proteins found to modulate rCGG-mediated neuronal toxicity: Pur α , hnRNP A2/B1 , and CUGBP1 , which bind rCGG repeats either directly ( Pur α and hnRNP A2/B1 ) or indirectly ( CUGBP1 , through the interaction with hnRNP A2/B1 ) [21] , [22] . MicroRNAs ( miRNAs ) are small , noncoding RNAs that regulate gene expression at the post-transcriptional level by targeting mRNAs , leading to translational inhibition , cleavage of the target mRNAs or mRNA decapping/deadenylation [23] , [24] . Mounting evidence suggests that miRNAs play essential functions in multiple biological pathways and diseases , from developmental timing , fate determination , apoptosis , and metabolism to immune response and tumorigenesis [25]–[31] . Recent studies have shown that miRNAs are highly expressed in the central nervous system ( CNS ) , and some miRNAs have been implicated in neurogenesis and brain development [32]–[34] . Interest in the functions of miRNAs in the CNS has recently expanded to encompass their roles in neurodegeneration . Investigators have begun to reveal the influence of miRNAs on both neuronal survival and the accumulation of toxic proteins that are associated with neurodegeneration , and are uncovering clues as to how these toxic proteins can influence miRNA expression [35] . For example , miR-133b is found to regulate the maturation and function of midbrain dopaminergic neurons ( DNs ) within a negative feedback circuit that includes the homeodomain transcription factor Pitx3 in Parkinson's disease [36] . In addition , reduced miR-29a/b-1-mediated suppression of BACE1 protein expression contributes to Aβ accumulation and Alzheimer's disease pathology [37] . Moreover , the miRNA bantam is found to be a potent modulator of poly-Q- and tau-associated degeneration in Drosophila [38] . Other specific miRNAs have also been linked to other neurodegenerative disorders , such as spinocerebellar ataxia type 1 ( SCA1 ) and Huntington's disease ( HD ) [39] , [40] . Therefore , miRNA-mediated gene regulation could be a novel mechanism , adding a new dimension to the pathogenesis of neurodegenerative disorders . Here we show that fragile X premutation rCGG repeats can alter the expression of specific miRNAs , including miR-277 , in a FXTAS Drosophila model . We demonstrate that miR-277 modulates rCGG-mediated neurodegeneration . Furthermore , we identified Drep-2 , which is associated with the chromatin condensation and DNA fragmentation events of apoptosis , and Vimar , a modulator of mitochondrial function , as two of the mRNA targets regulated by miR-277 . Functionally , Drep-2 and Vimar could modulate the rCGG-mediated neurodegeneration , as well . Finally , we show that hnRNP A2/B1 , an rCGG repeat-binding protein , can directly regulate the expression of miR-277 . These data suggest that hnRNP A2/B1 could be involved in the transcriptional regulation of selective miRNAs , and fragile X premutation rCGG repeats could alter the expression of specific miRNAs , potentially contributing to the molecular pathogenesis of FXTAS .
Given the important roles of miRNAs in neural development and human neurological disorders , we investigated the role of miRNAs in rCGG-mediated neurodegeneration . To determine whether fragile X premutation rCGG repeats could influence the expression of miRNAs , we profiled the expression of 72 known miRNAs using rCGG repeat transgenic flies that we generated previously [15] . In rCGG repeat transgenic flies , the severity of their phenotype depends on both dosage and length of the rCGG repeat . Moderate expression of ( CGG ) 90 repeats exclusively in the eyes have an effect on morphology and histology; however , expression of ( CGG ) 90 repeats in the neurons leads to lethality at the embryonic stage , preventing analysis at the adult stage [15] . Therefore , we used a shorter repeat length , r ( CGG ) 60 , which allowed us to examine the gene expression in adults . To analyze the effect of rCGG repeats in adult brains , we used RNAs isolated from the age- and sex-matched brains of control flies ( elav-GAL4 ) and flies expressing rCGG60 repeats in neurons ( elav-GAL4;UAS-CGG60-EGFP ) for miRNA profiling experiments ( Figure 1 ) . We identified a subset of miRNAs that consistently displayed altered expression in rCGG repeat flies versus the control group . Seven miRNAs with a ≥two-fold increase and two miRNAs with expression decreased by ≥1 . 5-fold have been found in rCGG repeat flies . These results suggest that fragile X premutation rCGG repeats could lead to the dysregulation of a subset of specific miRNAs . To assess the potential involvement of the miRNAs that showed altered expression in FXTAS fly brain , we examined the genetic interaction between specific miRNAs and rCGG-mediated neuronal toxicity based on the fragile X premutation rCGG repeat-mediated neurodegenerative eye phenotype we observed previously [15] . We generated UAS fly lines that could overexpress Drosophila miR-277 , bantam , let-7 , or miR-1 , as well as bantam mutant lines ( ban12 and ban20 ) that we generated previously [41] . We then crossed these transgenic lines with gmr-GAL4 , UAS- ( CGG ) 90-EGFP transgenic flies that exhibit photoreceptor neurodegeneration to determine the role of specific miRNAs in rCGG-mediated neurodegeneration . As shown in Figure 2 , flies co-expressing miR-277 and rCGG90 consistently showed an aggravated eye phenotype , with enhanced disorganized , fused ommatidia compared with flies expressing rCGG90 alone ( Figure 2B and 2D ) . Flies overexpressing miR-277 alone displayed a very mild rough eye phenotype ( Figure 2C ) . Alterations of the levels of bantam , let-7 , or miR-1 by either a gain of function or loss of function had no effect on rCGG-mediated neurodegeneration ( Figure 2E–2N ) . These data together suggest that miR-277 could be involved in rCGG-mediated neurodegeneration . The role of miR-277 in rCGG-mediated neurodegeneration seems specific , since the other miRNAs we found with altered expression in the presence of fragile X premutation rCGG repeats , including bantam , let-7 , and miR-1 , had no effect on the rCGG90 eye phenotype . The rest of our work focused on the role of miR-277 and its potential mechanisms in modulating rCGG-mediated neurodegeneration . Our miRNA profiling and genetic interaction studies indicated that an increase in miR-277 expression in rCGG repeat flies could alter the expression of specific cellular mRNAs by miR-277 , resulting in the enhanced rCGG-induced eye phenotype . To further explore the potential regulatory effect of miR-277 on rCGG-mediated neurodegeneration , we generated a transgenic miR-277 sponge ( miR-277SP ) line , which could block the activity of miR-277 , to test for any blocking effect on the rCGG-induced neurodegenerative eye phenotype . We generated the miRNA sponge transgenic construct as described previously [42] , [43] . In brief , we placed 10 repetitive sequences complementary to miR-277 with mismatches at positions 9–12 into the 3′ UTR of EGFP in a pUASP expression vector ( Figure 3A ) . We crossed miR-277SP transgenic flies with the flies expressing 90 CGG repeats and found that the expression of miR-277 sponge could consistently suppress rCGG-mediated neuronal toxicity ( Figure 3B ) . MiR-277 sponge alone or scramble control sponge had no effect on eye morphology ( Figure 3B and Data not shown ) . This result suggests that blocking the activity of miR-277 could mitigate the neurodegeneration caused by fragile X premutation rCGG repeats . To seek the mechanisms by which miR-277 modulates rCGG-mediated neurodegeneration , we searched the RNA network and referenced TargetScanFly 5 . 1 to identify potential miR-277 targets [44] , [45] . We selected the top candidates for miR-277 target mRNAs with the mutant alleles available for further analyses ( Table 1 ) . We then carried out a genetic screen on the rCGG90 neurodegenerative eye phenotype to identify potential miR-277 targets that could modulate rCGG-mediated neuronal toxicity . We crossed gmr-GAL4 , UAS- ( CGG ) 90-EGFP transgenic flies with fly mutants in genes coding for the top candidates for miR-277 target genes ( Table 1 ) . The progenies were then tested for potential suppression or enhancement of the disorganized eye phenotype versus flies expressing rCGG90 alone . Through this screen , we identified two modifiers of rCGG-mediated neurodegeneration , Drep-2 and Vimar . As shown in Figure 4 , partial loss of Drep-2 could enhance the rCGG90 eye phenotype by increasing ommatidial disorganization ( Figure 4B ) . Overexpression of Drep-2 could suppress the rCGG-induced eye phenotype ( Figure 4C ) . Flies carrying the Drep-2 mutation alone or Drep-2 overexpression alone displayed normal eyes ( Figure 4D and 4E ) . We also found a heterozygous loss-of-function mutant of Vimar that aggravated the rCGG90 eye phenotype ( Figure 4F ) . Eyes of control flies carrying the same Vimar mutation but no rCGG90 repeats are normal ( Figure 4G ) . These data together indicate that Drep-2 and Vimar could modulate rCGG-mediated neurodegeneration . Since both Drep-2 and Vimar were predicted to be regulated by miR-277 , by introducing 3′-UTR dual luciferase assays , we tested whether miR-277 could indeed target to Drep-2 or Vimar . We cloned the 3′-UTR of Drep-2 or Vimar containing the predicted miR-137 target sites from fly cDNA into a dual luciferase ( R-Luc and F-Luc ) reporter construct , allowing for the assessment of protein translation of these targets regulated via their 3′-UTR ( Figure 5A ) . These 3′-UTR dual constructs were transfected into HEK293 cells . We found that overexpression of miR-277 could suppress the R-Luc activity at 48 h post-transfection ( Figure 5B and 5C ) . Furthermore , when we mutated the seed regions of miR-277 located within the Drep2-3′-UTR reporter and Vimar-3′-UTR reporter , we saw that the mutation alleviated the miR-277-mediated suppression of luciferase activity ( Figure 5A–5C ) , suggesting the action of miR-277 is specific to the miR-277 seed region within the Drep-2 3′-UTR and Vimar 3′-UTR . Together , these data demonstrate that the effect of miR-277 on Drep-2 and Vimar expression is repressive and specific . Importantly , they also suggest that Drep-2 and Vimar are the functional targets of miR-277 . Next we went on to examine the steady-state levels of Drep-2 and Vimar mRNA in rCGG repeat flies ( elav-GAL4;UAS-CGG60-EGFP ) , in which the expression of miR-277 is increased . We saw a significant reduction of endogenous Drep-2 mRNA in rCGG repeat flies relative to control flies ( elav-GAL4 ) , whereas the Vimar mRNA expression in rCGG repeat flies remained similar to control flies ( elav-GAL4 ) ( Figure 5D ) . Furthermore , ectopic expression of miR-277 or miR-277 sponge could alter the endogenous mRNA level of Drep2 while have no apparent effect on Vimar mRNA level ( Figure 5E ) . These observations suggest that miR-277 could regulate Drep-2 and Vimar mRNAs differentially , with miR-277 regulating the expression of Drep-2 mainly at the mRNA level , and Vimar via translational suppression instead . Two rCGG repeat-binding proteins , Pur α and hnRNP A2/B1 , were previously found to bind rCGG repeats directly and modulate rCGG-mediated neuronal toxicity [21] , [22] . Intriguingly , recent studies have shown that multiple heterogeneous nuclear ribonucleoproteins ( hnRNPs ) could interact with heterochromatin protein 1 ( HP1 ) to bind to genomic DNA and modulate heterochromatin formation [46] . Thus we tested whether hnRNP A2/B1 could interact directly with genomic regions proximal to miR-277 . We performed hnRNP A2/B1-specific chromatin immunoprecipitation ( ChIP ) followed by real-time quantitative PCR across a six-kb region surrounding miR-277 . Immunoprecipitation of chromatin chemically cross-linked to DNA with an hnRNP A2/B1-specific antibody demonstrated that a region 1 . 5 kb upstream of miR-277 was enriched ∼seven-fold relative to IgG control and adjacent regions ( Figure 6A and 6B ) . Furthermore , the ectopic expression of hnRNP A2/B1 in fly brain could reduce the expression of miR-277 ( Figure 6C ) . These results together suggest that hnRNP A2/B1 could directly bind to the upstream region of miR-277 and regulate its expression . In the presence of fragile X premutation rCGG repeats , hnRNP A2/B1 will be sequestered , leading to the de-repression of the miR-277 locus .
Fragile X-associated tremor/ataxia syndrome ( FXTAS ) is a neurodegenerative disorder that afflicts fragile X syndrome premutation carriers , with earlier studies pointing to FXTAS as an RNA-mediated neurodegenerative disease . Several lines of evidence suggest that rCGG premutation repeats may sequester specific RNA-binding proteins , namely Pur α , hnRNP A2/B1 , and CUGBP1 , and reduce their ability to perform their normal cellular functions , thereby contributing significantly to the pathology of this disorder [15] , [21] , [22] . The miRNA pathway has been implicated in the regulation of neuronal development and neurogenesis [32] , [47]–[49] . A growing body of evidence has now revealed the role of the miRNA pathway in the molecular pathogenesis of neurodegenerative disorders [35] . Here we demonstrate that specific miRNAs can contribute to fragile X rCGG repeat-mediated neurodegeneration by post-transcriptionally regulating target mRNAs that are involved in FXTAS . We show that miR-277 plays a significant role in modulating rCGG repeat-mediated neurodegeneration . Overexpression of miR-277 enhances rCGG repeat-induced neuronal toxicity , whereas blocking miR-277 activity could suppress rCGG repeat-mediated neurodegeneration . Furthermore , we identified Drep-2 and Vimar as the functional miR-277 targets that could modulate rCGG repeat-induced neurodegeneration . Finally , we show that hnRNP A2/B1 , an rCGG repeat-binding protein , can directly regulate the expression of miR-277 . Our biochemical and genetic studies demonstrate a novel miRNA-mediated mechanism involving miR-277 , Drep-2 , and Vimar in the regulation of neuronal survival in FXTAS ( Figure 7 ) . Several lines of evidence from studies in mouse and Drosophila models strongly support FXTAS as an RNA-mediated neurodegenerative disorder caused by excessive rCGG repeats [11] , [15] , [17]–[19] . The current working model is that specific RNA-binding proteins could be sequestered by overproduced rCGG repeats in FXTAS and become functionally limited , thereby contributing to the pathogenesis of this disorder [15] , [17] , [19] , [20] . Three RNA-binding proteins are known to modulate rCGG-mediated neuronal toxicity: Pur α , hnRNP A2/B1 , and CUGBP1 , which bind rCGG repeats either directly ( Pur α and hnRNP A2/B1 ) or indirectly ( CUGBP1 , through the interaction with hnRNP A2/B1 ) [21] , [22]; how the depletion of these RNA-binding proteins could alter RNA metabolism and contribute to FXTAS pathogenesis has thus become the focus in the quest to understand the molecular pathogenesis of this disorder . Nevertheless , the data we present here suggest that the depletion of hnRNP A2/B1 could also directly impact the transcriptional regulation of specific loci , such as miR-277 . We know that hnRNPs can interact with HP1 to bind to genomic DNA and modulate heterochromatin formation [46] . Our results indicate that hnRNP A2/B1 could participate in the transcriptional regulation of miR-277; however , it remains to be determined whether other loci could be directly regulated by hnRNP A2/B1 , as well . Identifying those loci will be important to better understand how the depletion of rCGG repeat-binding proteins could lead to neuronal apoptosis . In recent years , several classes of small regulatory RNAs have been identified in a range of tissues and in many species . In particular , miRNAs have been linked to a host of human diseases . Some evidence suggests the involvement of miRNAs in the emergence or progression of neurodegenerative diseases . For example , accumulation of nuclear aggregates that are toxic to neurons have been linked to many neurodegenerative diseases , and miRNAs are known to modulate the accumulation of the toxic proteins by regulating either their mRNAs or the mRNAs of proteins that affect their expression . Moreover , miRNAs might contribute to the pathogenesis of neurodegenerative disease downstream of the accumulation of toxic proteins by altering the expression of other proteins that promote or inhibit cell survival [35] . Our genetic modifier screen revealed that miR-277 could modulate rCGG repeat-mediated neurodegeneration . By combining our genetic screen and reporter assays , we identified Drep-2 and Vimar as the functional targets of miR-277 that could modulate rCGG-mediated neurodegeneration . The closest ortholog of miR-277 in human is miR-597 based on the seed sequence . It would be interesting to further examine the role of miR-597 in FXTAS using mammalian model systems . Drep-2 is associated with the chromatin condensation and DNA fragmentation events of apoptosis [50] , [51] . Drep-2 is one of four Drosophila DFF ( DNA fragmentation factor ) -related proteins . While Drep-1 is a Drosophila homolog of DFF45 that can inhibit CIDE-A mediated apoptosis [50] . Drep-2 has been shown to interact with Drep-1 and to regulate its anti-apoptotic activity [50] . Vimar is a Ral GTPase-binding protein that has been shown to regulate mitochondrial function via an increase in citrate synthase activity [52] . In the presence of fragile X premutation rCGG repeats , overexpression of miR-277 will suppress the expression of both Drep-2 and Vimar , thereby altering anti-apoptotic activity as well as mitochondrial functions , which have been linked to neuronal cell death associated with neurodegenerative disorders in general ( Figure 7 ) . Interestingly , we saw a significant reduction of Drep-2 mRNA in the flies expressing rCGG repeats , while Vimar mRNA levels remained similar to control flies . This observed difference may be due to the fact that miRNA could be involved in different modes of action , including mRNA cleavage , translational inhibition and mRNA decapping/deadenylation its target mRNAs [23] , [24] . In summary , here we provide both biochemical and genetic evidence to support a role for miRNA and its selective mRNA targets in rCGG-mediated neurodegeneration . Our results suggest that sequestration of specific rCGG repeat-binding proteins can lead to aberrant expression of selective miRNAs that could modulate the pathogenesis of FXTAS by post-transcriptionally regulating the expression of specific mRNAs involved in this disorder . Identification of these miRNAs and their targets could reveal potential new targets for therapeutic interventions to treat FXTAS , as well as other neurodegenerative disorders .
All flies were maintained under standard culture conditions . The rCGG repeat transgenic flies ( UAS-CGG60-EGFP and UAS-CGG90-EGFP ) were generated in our lab as described previously [15] . Flies mutant in genes coding for different candidate miR-277 targets were obtained from Bloomington Stock Center . The UAS-miR-277-Sponge transgenic flies were generated as described previously [43] . We introduced 10 repetitive miR-277 sponge sequences ( TGTCGTACCAGGCGTGCATTTA ) with a 4-nt linker between each repeat downstream of EGFP in a pUASP expression vector . Similarly a scramble control construct ( GTTCACGGATAGTGCCTGTACT ) was generated as well . Both constructs were confirmed by DNA sequencing and then injected in the w1118 strain by standard methods . Total RNAs were isolated from the control ( elav-GAL4 ) and rCGG60 fly heads using Trizol . TaqMan MicroRNA Assays detecting 72 known individual Drosophila miRNAs were obtained from ABI ( ABI ) . cDNA was prepared with High-Capacity cDNA Reverse Transcription Kits ( ABI; Cat#437496 ) . The 15-µl reverse transcription reactions consisted of 10 ng of total RNA , 5 U MultiScribe Reverse Transcriptase , 0 . 5 mM of each dNTP , 1× reverse transcription buffer , 4 U RNase inhibitor , and nuclease-free water . This was performed at 16°C for 30 min and at 42°C for 30 min , terminated at 85°C for 5 min and 4°C until use in TaqMan assays . For real-time PCR of TaqMan MicroRNA Assays , we used 0 . 5 ul 20×TaqMan MicroRNA Assay Primer , 1 . 33 ul undiluted cDNA , 5 ul 2×TaqMan Universal PCR Master Mix , 3 . 17 ul nuclease-free water . Each PCR reaction was performed in triplicate with MicroAmp optical 96-well plates using a 7500 Fast Real-Time PCR System ( ABI ) , with reactions incubated at 95°C for 10 min , followed by 40 cycles of 95°C for 15 s , and 60°C for 1 min . Fluorescence readings were taken during the 60°C step . RQs were calculated using the ΔΔCt method , with 2S RNA TaqMan miRNA control assay as the endogenous control , and calibrated to the control samples . The fly heads from control ( elav-Gal4 ) and rCGG60 flies were collected . Trizol ( Invitrogen; Cat# 15596-026 ) was used to isolate total RNA from each genotype . RNA samples were reverse-transcribed into cDNA with oligo ( dT ) 20 and SuperScript III ( Invitrogen; Cat#18080051 ) . Real-time PCR was performed with gene-specific primers and Power SYBR Green PCR Master Mix ( Applied Biosystems; Cat# 4367660 ) using the 7500 Standard Real-Time PCR System ( Applied Biosystems ) . RpL32 ( Qiagen; Cat# QT00985677 ) was used as an endogenous control for all samples . Primers for Drep2 and Vimar transcripts were designed using Primer Express 3 . 0 software ( Applied Biosystems ) and were as follows . Drep2: forward , 5′-TGGAACGCCTCAACTCCAA-3′; and reverse , 5′-TCGGACTCGCGATCCAA-3′ . Vimar: forward , 5′-GCACCCGCCGAACAGA-3′; and reverse , 5′-TGCGATCGTAGTCTTGCGTTA -3′ . All real-time PCR reactions were performed in triplicate , and RQs were calculated using the ΔΔCt method , with calibration to control samples . Drep-2 3′-UTR and Vimar 3′-UTR sequences were PCR-amplified directly from w1118 fly brain first-strand cDNA generated from 5 ug TRIZOL-isolated total RNA using oligo-dT SuperScript III reverse transcription according to the manufacture's protocol ( Invitrogen; Cat . #1808-093 ) . The PCR products were then cloned into psiCHECK-2 dual luciferase vector ( Promega; Cat# C8021 ) . The miR-277 target sites in the Drep-2 3′-UTR and Vimar 3′-UTR were deleted using QuikChange Site-Directed Mutagenesis Kits ( Stratagene; Cat . #2000518 ) . Target sites deletions were verified by Genewiz sequencing . Briefly , 293FT cells were co-transfected by Attractene transfection reagent ( Qiagen; Cat . #301005 ) with psiCHECK-2-3′UTR or psiCHECK-2-3′ UTRΔmiR-277 and miR-277 duplex RNA ( Qiagen; Cat# MSY0000338 ) or control miRNA duplex ( Qiagen; Cat# 1027280 ) . All co-transfections used a total of 600 ng of plasmid DNA and 120 nmol of duplex RNA . Luciferase expression was detected using the Dual-Luciferase Reporter 1000 System ( Promega; Cat# E1980 ) according to the manufacturer's instructions . At 48 h after transfection , R-Luc activity was normalized to F-Luc activity to account for variation in transfection efficiencies , and miR-277-mediated knockdown of R-Luc activity was calculated as the ratio of normalized R-Luc activity in the miR-277 duplex treatments to normalized R-Luc activity in the negative control duplex treatments . Luciferase experiments were repeated three times . ChIP was performed using a ChIP Assay Kit ( Millipore ) . S2 cells were cross-linked with 1% formaldehyde ( Sigma-Aldrich ) for 10 min at room temperature . Chromatin was fragmented to an average size of 500 bp by sonication ( Sonicator 3000; Misonix ) and immunoprecipitated with anti-Flag M2 antibody ( sigma ) . Immunoprecipitated and purified DNA fragments were diluted to 1 ng/µl in nuclease-free water . We used 8 ng of DNA in 20-µl SYBR Green real-time PCR reactions consisting of 1× Power SYBR Green Master Mix and 0 . 5 µM forward and reverse primers . Reactions were run on an SDS 7500 Fast Instrument ( Applied Biosystems ) . Primers were designed using Primer Express 3 . 0 software ( Applied Biosystems ) and were as follows . 4 . 5 kb upstream: forward , 5′-CAGAAAACAGGCGTGCAAAC; and reverse , 5′-GAATTTGCATTGGCTTTGGAA . 3 . 5 kb upstream: forward , 5′- TTACAATTGGATGGGCTTCGT; and reverse , 5′-AAGCTGACGGCCTGACTAAAAA . 2 . 5 kb upstream: forward , 5′-GTTGGCTGCTGCGTCAATT; and reverse , 5′- GCCCCAGCGGCATTTATA . 1 . 5 kb upstream: forward , 5′- TTCTGGCACTGGCAGCTTT; and reverse , 5′- CATCGTGCTGGCCAACAC . 1 . 0 kb upstream: forward , 5′-TGTACGGGCATGTGTATGCA; and reverse , 5′- TCAACGAACACGCTGCGTAT . 0 . 5 kb upstream: forward , 5′- GGGCATTTTCATTTCATTCCA; and reverse , 5′- CGGGCAGCGTAATTTAAGCT . 0 . 5 kb downstream: forward , 5′- CGCCCACAAGAGCTTTTGA; and reverse , 5′- TTTCCACGGTATGCTGCTTTT . 1 . 5 kb downstream: forward , 5′- CGTTTCCATTTAGTTGGATTTTTGT; and reverse , 5′- GGCAAACCACACATTTTAACATACA . DNA relative enrichment was determined by taking the absolute quantity ratios of specific IPs to nonspecific IPs ( normal mouse IgG only ) , IP/IgG , and normalizing to control ( pUAST only ) . Independent chromatins were prepared for all ChIP experiments , and real-time PCR reactions were performed in triplicate for each sample on each amplicon . For scanning electron microscopy ( SEM ) images , whole flies were dehydrated in gradient concentration ethanol ( 25% , 50% , 75% , 100% ) , dried with hexamethyldisilazane ( Sigma; Cat# 16700 ) , and analyzed with an ISI DS-130 LaB6 SEM/STEM microscope .
|
Fragile X–associated tremor/ataxia syndrome ( FXTAS ) is an adult-onset neurodegenerative disorder , usually affecting males over 50 years of age . FXTAS patients are the carriers of fragile X premutation alleles . Using a FXTAS Drosophila model , we previously demonstrated that fragile X premutation rCGG repeats alone could cause neurodegeneration . Pur α and hnRNP A2/B1 were identified as specific premutation rCGG repeat-binding proteins ( RBPs ) that could bind and modulate fragile X permutation rCGG-mediated neuronal degeneration . MiRNAs are sequence-specific regulators of post-transcriptional gene expression . Here we show that fragile X premutation rCGG repeats could lead to aberrant expression of selective miRNAs , which may modulate the pathogenesis of FXTAS by post-transcriptionally regulating the expression of specific mRNAs involved in FXTAS .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"neurological",
"disorders",
"neurology",
"cerebellar",
"disorders"
] |
2012
|
MicroRNA-277 Modulates the Neurodegeneration Caused by Fragile X Premutation rCGG Repeats
|
The lipid kinase phosphatidylinositol 4-kinase III alpha ( PI4KIIIα ) is an essential host factor of hepatitis C virus ( HCV ) replication . PI4KIIIα catalyzes the synthesis of phosphatidylinositol 4-phosphate ( PI4P ) accumulating in HCV replicating cells due to enzyme activation resulting from its interaction with nonstructural protein 5A ( NS5A ) . This study describes the interaction between PI4KIIIα and NS5A and its mechanistic role in viral RNA replication . We mapped the NS5A sequence involved in PI4KIIIα interaction to the carboxyterminal end of domain 1 and identified a highly conserved PI4KIIIα functional interaction site ( PFIS ) encompassing seven amino acids , which are essential for viral RNA replication . Mutations within this region were also impaired in NS5A-PI4KIIIα binding , reduced PI4P levels and altered the morphology of viral replication sites , reminiscent to the phenotype observed by silencing of PI4KIIIα . Interestingly , abrogation of RNA replication caused by mutations in the PFIS correlated with increased levels of hyperphosphorylated NS5A ( p58 ) , indicating that PI4KIIIα affects the phosphorylation status of NS5A . RNAi-mediated knockdown of PI4KIIIα or pharmacological ablation of kinase activity led to a relative increase of p58 . In contrast , overexpression of enzymatically active PI4KIIIα increased relative abundance of basally phosphorylated NS5A ( p56 ) . PI4KIIIα therefore regulates the phosphorylation status of NS5A and viral RNA replication by favoring p56 or repressing p58 synthesis . Replication deficiencies of PFIS mutants in NS5A could not be rescued by increasing PI4P levels , but by supplying functional NS5A , supporting an essential role of PI4KIIIα in HCV replication regulating NS5A phosphorylation , thereby modulating the morphology of viral replication sites . In conclusion , we demonstrate that PI4KIIIα activity affects the NS5A phosphorylation status . Our results highlight the importance of PI4KIIIα in the morphogenesis of viral replication sites and its regulation by facilitating p56 synthesis .
Worldwide about 170 million people are chronically infected with hepatitis C virus ( HCV ) , a positive-strand RNA virus belonging to the Flaviviridae family , frequently leading to severe liver disease . The viral genome encompasses 9 . 6 kb and encodes mainly for a polyprotein of about 3 , 000 amino acids in length , flanked by nontranslated regions , which is cleaved into ten mature proteins by cellular and viral proteases ( reviewed in [1] , [2] ) : core , envelope glycoprotein 1 ( E1 ) and E2 , p7 and the six nonstructural ( NS ) proteins NS2 , NS3 , NS4A , NS4B , NS5A and NS5B . The structural proteins core , E1 and E2 , which are the major constituents of the viral particle , p7 , a presumed viroporin , and NS2 , which is part of the protease mediating NS2/NS3 cleavage , are mainly involved in the generation of infectious virions , whereas NS3 to NS5B are required for RNA replication . NS3 comprises helicase and NTPase activities in the C-terminal part and an N-terminal protease , which is constitutively bound to its cofactor NS4A . NS4B plays a major role in inducing membrane alterations that are required for viral replication ( reviewed in [3] ) . NS5A is a phosphoprotein consisting of three subdomains with functions in viral RNA replication and virus assembly ( reviewed in [4] ) and NS5B is the viral RNA-dependent RNA-polymerase ( RdRP ) . Viral RNA replication takes place in vesicular membrane alterations designated the membranous web ( MW ) [5] , [6] . The morphology and biogenesis of the MW are still poorly understood , but it is believed that NS4B is the most important determinant , since sole expression of NS4B induces vesicular structures [6] . Models based on biochemical evidence and related viruses furthermore suggested that RNA synthesis takes place in membrane invaginations connected to the cytoplasm [7]–[9] . However , more recent results point to a far more complex morphology by showing that the MW mainly consists of double membrane vesicles ( DMVs ) and multimembrane vesicles ( MMVs ) [10] , [11] , probably including autophagosomes [10] , [12] , but it is currently not clear how these structures are topologically linked to RNA synthesis . The complexity of these membrane alterations , which are distinct from the vesicles induced by NS4B , suggests that the MW is generated by a concerted action of the nonstructural proteins and host factors . Indeed , NS5A and the cellular lipid kinase phosphatidylinositol 4-kinase type III alpha ( PI4KIIIα ) have recently been identified to contribute to the morphology and complexity of the MW [11] . PI4KIIIα ( PIK4CA , PI4KA ) is an ER-resident enzyme of 230–240 kD in size converting phosphatidylinositol to phosphatidylinositol 4-phosphate ( PI4P ) . In mammalian cells , the family of PI4-kinases comprises two types with two isoforms each ( PI4KIIα , PI4KIIβ , PI4KIIIα and PI4KIIIβ ) differing in subcellular localization and being responsible for the synthesis of distinct PI4P pools ( reviewed in [13] ) , in case of PI4KIIIα those in the ER , the plasma membrane [14] and parts of Golgi PI4P [15] . PI4P , besides being the precursor of other important phosphatidylinositides , is believed to play a role in trafficking by serving as a membrane code for vesicles and organelles and several proteins have been identified specifically binding to this lipid ( reviewed in [16] ) . PI4KIIIα has been identified as an essential host factor of HCV RNA replication by a number of studies [17]–[23] arguing for a pivotal role in viral RNA synthesis . An involvement of PI4KIIIβ has been discussed as well , but might be restricted to genotype 1 and more pronounced for other steps of the viral replication cycle [20] , [24]–[26] . Importantly , PI4KIIIβ and PI4P are also intimately linked to replication of enteroviruses [22] , suggesting that dependence on PI metabolism and particularly PI4P is a common theme for many virus groups and opening the possibility to create broadly acting antivirals ( reviewed in [27] ) . Some inhibitors with specificity towards PI4KIIIα [15] , [28] and PI4KIIIβ [20] are available , however , since mice with conditional knockout of the PI4KIIIα gene developed lethal gastrointestinal disorders , severe side effects might be expected also for chemical inhibition of these enzymes in vivo , thereby restricting their use in humans [28] . The mechanistic role of PI4KIIIα in the HCV replication cycle has been addressed by several studies [11] , [24] , [29] . Silencing of PI4KIIIα results in a “clustered” distribution of nonstructural proteins in IF , accompanied by an altered ultrastructure of the MW that contains smaller DMVs and lacks MMVs [11] . These changes in viral replication sites have been attributed to elevated PI4P levels correlating with an altered localization of PI4P observed in the presence of HCV nonstructural proteins in cell culture and in vivo . PI4KIIIα directly interacts with NS5A and NS5B and in vitro assays suggested that lipid kinase activity is stimulated by NS5A [11] , [29] . Recently , it was furthermore shown that HCV not only activates the kinase in cell culture , but also causes a depletion of PI4P in the plasma membrane , arguing for a general reorganization of cellular PI4P metabolism induced by the virus [15] . However , the molecular mechanism of PI4KIIIα activation and PI4P redistribution have not been clarified yet and it is unclear how these changes impact on MW morphology and what the precise role of NS5A is in this process . NS5A consists of three domains ( D1 , 2 and 3 ) separated by two low complexity sequences ( LCS I and LCS II ) [30] and an N-terminal amphipathic helix essential for its membrane association [31] . D1 is capable of binding RNA and involved in replication [32] , [33] , whereas D2 is to the most part dispensable for replication [34] . D3 is important for the generation of infectious virus , probably due to an interaction with core that is regulated by NS5A phosphorylation [34]–[36] . NS5A can be found in two distinct phosphorylated forms: a basal ( hypo- ) and hyperphosphorylated state , designated according to their apparent molecular weight p56 and p58 , respectively [37] . The phosphorylation status is directly or indirectly modulated by NS3 , NS4A , NS4B and NS5B [38]–[41] . However , cellular kinases and mechanistic details regulating NS5A phosphorylation are still poorly defined . Basal phosphorylation seems to depend on kinases of the CMGC family ( e . g . casein kinase ( CK ) II [35] , [42] ) and involves mainly sequences in NS5A D2 and D3 ( reviewed in [4] ) . In contrast , p58 synthesis is mediated by the CKI protein kinase family , particularly CKIα [43] , [44] and recently an additional role of Polo-like kinase 1 has been identified [45] . A cluster of serine residues encompassing amino acids ( aa ) 222–235 at the C-terminus of D1 and in LCS I has been shown to be involved in hyperphosphorylation of NS5A and in the regulation of RNA replication by adaptive mutations [40] , [46]–[48] . Mutations in this region typically decrease p58/p56 ratio and increase RNA replication of HCV genotype 1 isolates , probably by modulating the interaction with the host factor hVAP-A [46] , [48] , [49] . However , a complete loss of hyperphosphorylation typically abrogates RNA replication indicating that low amount of p58 is essential for RNA replication [50] , [51] . The fact that mutations causing reduced p58/p56 levels enhance replication of genotype 1 isolates [46] , [48] , [52] , but in turn repress particle morphogenesis , raised the concept that p56 is mainly involved in RNA replication . This is corroborated by the finding that kinase inhibitors reducing NS5A hyperphosphorylation stimulate HCV replication [53] . In contrast , p58 might be a negative regulator of RNA replication and/or is required for assembly [35] , [36] , [54] , [55] . In this study , we provide a detailed characterization of the mechanism of action of PI4KIIIα in HCV RNA replication . We identified a site in NS5A D1 that is involved in functional PI4KIIIα interaction and RNA replication . Mutations in this site phenocopied PI4KIIIα silencing , including abrogation of PI4P induction and alterations in MW morphology . Importantly the same mutations resulted in increased NS5A p58 levels and we show that enzymatic activity of PI4KIIIα plays a vital role in the modulation of NS5A phosphorylation . Transcomplementation studies suggest that alterations of NS5A phosphorylation status rather than elevation of PI4P abundance are the main determinant for efficient HCV RNA replication . These results reveal a complex mechanism of action of PI4KIIIα in HCV replication and provide novel insights into the biogenesis of the viral replication sites .
Recently , we found that PI4KIIIα directly interacts with HCV NS5A D1 but not with D2 or D3 [11] . To narrow down the NS5A binding region we designed several deletions within NS5A D1 in the context of the polyprotein NS3 to NS5B ( NS3-5B ) of genotype 2a ( isolate JFH-1 [56] ) , all of them retaining the N-terminal amphipathic alpha-helix ( Fig . 1A ) . We used a transient expression model based on plasmids encoding the HCV NS3 to NS5B polyprotein under transcriptional control of the T7-promoter , which were transfected into Huh7-Lunet cells constitutively expressing T7-RNA polymerase ( Huh7-Lunet T7 ) . This approach enabled us to examine effects of mutations on PI4KIIIα interaction independent of the requirement for functional NS5A for HCV RNA replication . Polyproteins bearing deletions within NS5A D1 were coexpressed with HA-tagged PI4KIIIα and subjected to immunoprecipitation using HA- and NS5A-specific antibodies ( Fig . 1B ) . The amount of PI4KIIIα coprecipitating with NS5A ( Fig . 1B lower panel , HA-PI4K ) was quantified and normalized to PI4KIIIα input levels ( Fig . 1B upper and lower panel , respectively ) to determine the impact of deletions within NS5A D1 on the interaction of both proteins . None of the deletions completely abrogated NS5A- PI4KIIIα binding , indicating that determinants outside D1 might contribute to PI4KIIIα interaction . However , deletion of the entire D1 severely impaired PI4KIIIα binding , confirming our previous findings ( Fig . 1B , [11] ) . A similar phenotype was observed upon deletion of the C-terminal part of NS5A D1 ( ΔS3 , Δ151–214aa ) , in contrast to deletions encompassing the N-terminus ( ΔS1 ) or the center ( ΔS2 ) of NS5A D1 , which did not impair , but apparently increased PI4KIIIα binding , pointing to complex determinants contributing to the interaction of both proteins . To further narrow down the PI4KIIIα interaction site within NS5A we generated smaller deletions within aa151–214 of NS5A ( ΔS3A , B and C , respectively ) and found that removal of aa187–199 as well as 200–213 both reduced PI4KIIIα binding to about 10% of wt NS5A ( wt ) . These results suggested that the very C-terminal 28aa of NS5A D1 contained sequence elements involved in the binding of PI4KIIIα , which we designated PI4KIIIα binding region ( PBR , Fig . 2A ) . Our next aim was to identify more subtle mutations within this region , allowing a detailed analysis of the functional role of the NS5A-PI4KIIIα interaction in HCV replication . Therefore , we generated overlapping triple alanine substitutions within the NS3-5B polyprotein ( Fig . 2A ) and first analyzed their impact on NS5A-PI4KIIIα binding ( Fig . 2B , C , D ) . Interestingly , all mutations except mutSQL , LPC , VLR and ETA significantly reduced PI4KIIIα coprecipitation efficiency , supporting our assumption that this whole region was involved in PI4KIIIα binding . However , the effect of most NS5A triple alanine mutations was not very strong and reduced the efficiency of PI4KIIIα binding by about 50% . Only a set of mutations spanning a region of nine aa reduced PI4KIIIα coprecipitation to about 25% of wt NS5A ( Fig . 2C , Table 1 ) . We furthermore assessed the impact of the triple alanine mutations on NS5A coprecipitating with HA-PI4KIIIα ( Fig . 2D and not shown ) . However , the efficiency of coprecipitation was very low and was only detectable after long overexposures of the gels ( Fig . 2D and not shown ) , most likely due to a consistently high molar excess of NS5A to HA-PI4KIIIα in all experiments ( 20–100fold , data not shown ) . Interestingly , both phosphoisoforms were precipitated with similar efficiency in case of NS5A wt . However , the signal intensities were not high enough to allow a thorough quantitation of NS5A binding and p58/p56 ratios of the NS5A mutants to draw firm conclusions ( data not shown ) . In essence , none of the triple alanine mutations within the PBR of NS5A abrogated PI4KIIIα binding entirely , suggesting that the interaction between both proteins relies on multiple determinants . Next , we quantified the impact of the triple alanine mutations on HCV RNA replication using monocistronic reporter replicons of the JFH-1 isolate ( Fig . 3A ) using luciferase reporter assays ( Fig . 3B ) and direct quantitation of viral genomes after transfection of replicon RNA ( Fig . 3C , Fig . S1A ) . Both assays revealed very similar results , except that the dynamic range of the RNA detection was lower due to large amounts of input RNA still present at 72 hours after transfection ( Fig . S1A ) . Surprisingly , only mutations showing a strong reduction in PI4KIIIα binding were entirely incapable of replication ( repMLT , TDP , PPH and HIT , Fig . 3B , 3C ) . Based on their replication phenotype we assumed that the region covered by mutants mutMLT , TDP , PPH and HIT was most critical for functional PI4KIIIα interaction and therefore termed this sequence motif PI4KIIIα functional interaction site ( PFIS , indicated in Fig . 2A ) . The fact that several mutant replicons were not impaired in RNA replication at all , despite significant reductions in PI4KIIIα binding efficiency ( e . g . repCEP , PEP and PDA ) suggested that a minimum threshold level of PI4KIIIα interaction of more than 25% of NS5A wt was necessary and sufficient for viral RNA replication ( Table 1 ) . However , the moderately delayed replication kinetics of the two mutants flanking the PFIS ( RSM , TAE; Fig . 3B ) compared to their strongly reduced interaction with PI4KIIIα binding ( Fig . 2C , Table 1 ) , suggested that the determinants of functional PI4KIIIα interaction were not entirely reflected by the binding efficiency measured by co-immunoprecipitation . In addition to the PFIS mutants , repLPC also was severely impaired in replication , but still replicated at a very low level , indicated by the increase in luciferase activity at 48 h and 72 h after transfection . However , the neighbouring mutants repSQL and repCEP were not impaired in replication at all . In addition , the PI4KIIIα binding efficiencies of these three mutants with overlapping triple alanine mutations were variable with mutLPC showing no significant reduction in PI4KIIIα binding ( Fig . 2C ) , arguing for a role of proline188 in the function of NS5A independent from PI4KIIIα interaction . To support our assumption that replication defects in the PFIS region indeed were functionally linked to PI4KIIIα , we next analyzed whether overexpression of PI4KIIIα would rescue replication of the PFIS mutants . Therefore , we used Huh7-Lunet cells constitutively overexpressing HA-PI4KIIIα upon lentiviral transduction ( Fig . 3E ) and compared replication efficiency of the different mutants to naïve Huh7-Lunet cells ( Fig . 3D , Fig . S1B , C ) . Interestingly , repPPH , carrying mutations in the center of the PFIS region and being replication deficient in naïve Huh7-Lunet cells , was strongly stimulated by overexpression of PI4KIIIα , suggesting that the phenotypes observed for this mutant were indeed functionally linked to PI4KIIIα ( Fig . 3D , Fig . S1B ) . In addition , the delayed replication kinetics of mutants RSM and TAE , flanking the PFIS , were fully rescued by overexpression of PI4KIIIα and both mutants replicated to wildtype levels already at 24 hours after electroporation ( Fig . S1C ) . In contrast , replication efficiency of mutant LPC was not significantly enhanced by PI4KIIIα overexpression , again suggesting that the replication defect of this mutant was independent of PI4KIIIα . PFIS mutants MLT , TDP and HIT were not rescued by PI4KIIIα . This might either be due to an insufficient PI4KIIIα overexpression level , a stronger impairment of functional PI4KIIIα interaction compared to mutant PPH and/or due to defects unrelated to NS5A- PI4KIIIα interactions , like NS5A RNA binding , dimerization , replicase interactions etc . However , mutants MLT , TDP and HIT exhibited almost identical phenotypes in all kinds of analyses throughout our study compared to mutant PPH ( Table 1 ) , which was partially rescued by overexpression of PI4KIIIα . Therefore , it seemed likely that all phenotypes of the PFIS mutants besides RNA replication were indeed primarily caused by defects in functional PI4KIIIα interaction . Triple alanine mutations in a sequence motif designated PFIS , reduced PI4KIIIα binding to about 25% of wt NS5A and completely abolished RNA replication , arguing for a mechanistic role of NS5A-PI4KIIIα interaction in viral replication . A mechanistic link between replication defects caused by mutations in the PFIS and PI4KIIIα was furthermore supported by a strong enhancement of replication efficiency of one of the PFIS mutants upon PI4KIIIα overexpression . Previous studies identified strongly enhanced PI4P levels in HCV-positive cells due to an activation of PI4KIIIα by NS5A [11] , [29] . Silencing of PI4KIIIα abolished induction of PI4P synthesis by HCV and abrogated RNA replication . These phenotypes coincided with a so called “clustered” intracellular accumulation of HCV nonstructural proteins in immunofluorescence and an altered architecture of HCV-induced membrane alterations , suggesting a fundamental role of PI4P in the formation of the HCV replication sites [11] . To test whether our triple alanine mutations would mimic some of these PI4KIIIα knockdown phenotypes , we expressed mutant NS3-5B polyproteins in Huh7-Lunet T7 cells and costained for NS5A and PI4P to judge the localization of NS5A and to quantify the levels of PI4P . We first quantified intracellular PI4P levels in cells expressing NS3-5B compared to mock transfected cells ( Fig . 4A , middle panels , Fig . 4C , Table 1 ) . PI4P primarily localized to the Golgi-apparatus in mock-transfected cells . Expression of NS3-5B wt resulted in strongly enhanced levels of PI4P in most cells , was widely spread in the cytoplasm and partially colocalized with NS5A , as expected ( Fig . 4A , C; [11] ) . The degree of PI4P enhancement was 4 . 4-fold on average , but varied widely amongst individual cells in agreement with our previous analysis [11] . Most triple alanine mutants still gave rise to significantly increased PI4P levels compared to mock-transfected cells , but to lower levels than detected in wt polyprotein-expressing cells ( Fig . 4A , C , Table 1 ) . Importantly , PI4P levels in cells transfected with mutants in the PFIS ( mutMLT , TDP , PPH , HIT ) were not increased compared to mock-transfected cells . The same mutations completely abrogated RNA replication ( Fig . 3B , C; Table 1 ) , suggesting that increased intracellular PI4P levels might be critical for viral RNA replication , as indicated by previous studies [11] . However , the determinants for PI4KIIIα activation could not be directly correlated to the efficiency of NS5A binding , since on one hand , mutants like LPC , which were not impaired in PI4KIIIα binding but replication deficient did not induce PI4P synthesis , whereas on the other hand mutants with strongly reduced PI4KIIIα binding capability ( e . g . RSM , Fig . 2C ) activated PI4KIIIα . Therefore , the determinants of PI4KIIIα activation by NS5A appeared complex and required an interaction which was not directly reflected by the binding efficiency . In addition , viral RNA replication seemed to be compatible to a wide range of intracellular PI4P levels , indicated by the huge variations already observed upon expression of NS3-5B wt and by several mutants , which were not impaired in RNA replication despite a significantly reduced ability to induce PI4P ( mutSQL , CEP , PEP , PDA , ADV ) . To confirm this result in a replication competent model , we analyzed PI4P induction by wt and two representative mutant replicons ( repCEP and PDA ) , replicating to the same levels , despite significant differences in mean PI4P induction levels in NS3-5B expressing cells ( 4 . 4-fold vs . 2 . 3- and 2 . 7-fold , respectively ) . Interestingly , results obtained with the replicons were very similar to the expression of NS3-5B . Again , the level of induction was highly variable in individual cells , but the mean value remained significantly higher for wt compared to repCEP and repPDA ( 3 . 5-fold vs 2 . 4- and 2 . 8-fold , respectively , Fig . S3B , C ) , arguing for a wide range of intracellular PI4P levels being compatible with efficient HCV replication . We next analyzed the impact of the triple alanine mutations on the intracellular distribution of NS5A . In cells expressing NS3-5B wt , NS5A appeared mainly in a dot-like , dispersed pattern ( Fig . 4A , upper panel ) , reminiscent of the distribution in replicon cells ( Fig . S3A , [5] ) . However , ca . 3% of cells overexpressing NS3-5B wt but none of the cells harboring the wt replicon gave rise to the clustered NS5A distribution . Expression of almost all mutant polyproteins , except mutETA , resulted in a higher number of cells exhibiting the clustered phenotype ( Fig . 4A , upper panel , mutCEP , PDA and HIT , Fig . 4B ) , coinciding with reduced levels of PI4P induction ( Fig . 4C , Table 1 ) . Since the clustered phenotype was also observed upon silencing of PI4KIIIα [11] , these results suggested that a clustered distribution of NS5A was indeed linked to reduced PI4P levels . The same clustered phenotype was found using antisera directed against NS3 , NS4B and NS5B , and all these nonstructural proteins co-localized to NS5A , irrespective of a dot-like or clustered distribution ( Fig . S2 ) . Therefore , cluster formation seemed to rely mainly on a reorganization of membrane structures rather than on a general disturbance of nonstructural protein interactions , which might have resulted in a loss of co-localization . The proportion of cells with NS5A clusters varied from 7–57% for the mutants compared to 3% for the wt and several mutants showing predominantly the clustered phenotype replicated like wt ( e . g . mutPDA , Table 1 ) , suggesting that this phenotype was not necessarily associated with defects in RNA replication . Therefore , we analyzed the NS5A staining pattern of two of the replication competent triple alanine mutations ( mutCEP , mutPDA ) in replicon-containing cells and surprisingly found no NS5A clusters at all , as in case of the wt replicon cells ( Fig . S3 ) . Probably , protein expression levels in replicon cells were not high enough to induce clusters and/or clusters were not compatible with RNA replication . Hence , these data suggest that the formation of NS5A clusters is favored by high ectopic NS3-5B polyprotein expression combined with reduced levels of PI4P , thereby pointing to mutants with reduced capability to activate PI4KIIIα . Collectively , our data demonstrate that mutations in the PBR region of NS5A resulted in a higher abundance of a clustered NS5A distribution , as observed for PI4KIIIα knockdown , but this phenotype was not correlated with RNA replication competence and was not observed in replicon cells , suggesting that protein overexpression and reduced PI4P induction both contributed to this phenotype . We furthermore found that most mutations in the PBR affected activation of PI4KIIIα , resulting in a less pronounced enhancement of intracellular PI4P pools . RNA replication was not affected over a wide range of PI4P induction levels . However , mutations of the PFIS strongly reducing PI4KIIIα binding and not giving rise to any induction of PI4P synthesis abolished RNA replication . Our previous study identified an altered morphology of MW structures upon PI4KIIIα knockdown , suggesting that PI4P was critically involved in web integrity [11] . Therefore , we next aimed to analyze MW morphology induced by triple alanine mutants of the PBR , giving rise to different levels of PI4P induction . Membrane alterations induced by expression of NS3-5B wt were heterogeneous , consisting of double membrane vesicles ( DMVs ) with an average diameter of 200 nm interspersed by multi-membrane vesicles ( MMVs ) ( Fig . 5A , C [11] ) , Both vesicle types did not accumulate in distinct areas , but rather were dispersed throughout the cytoplasm . Very similar web structures were found in wt replicon cells , but with lower abundance , probably due to lower protein expression levels ( Fig . S3D , upper panel , [11] ) . Silencing of PI4KIIIα resulted in more homogenous web structures , lacking MMVs , with DMVs of an average diameter of 133 nm ( Fig . 5B , C ) , confirming our previous results [11] . For further ultrastructural analysis we chose one representative mutant in the PFIS ( mutHIT ) devoid of RNA replication and PI4P induction and two fully replication competent mutants with intermediate levels of PI4P induction ( mutCEP and PDA ) . MW morphology induced by expression of NS3-5B mutHIT resembled the morphology obtained after expression of NS3-5B wt in PI4KIIIα-silenced cells: membrane alterations only contained DMVs , which were very homogeneous , but even smaller in size ( 91+/−20 nm , Fig . 5A , C ) , thus further confirming that the major phenotypes associated with this mutant indeed relied on its defects in PI4KIIIα interaction . MutCEP and mutPDA both induced web structures with an intermediate phenotype containing DMVs smaller than wt , but larger than mutHIT ( 136+/−21 and 142+/−40 nm , respectively ) and MMVs . Expression of both mutants often resulted in vesicles accumulating in larger clusters like in shPI4KIIIα knockdown cells , in line with the high abundance of NS5A clusters detected in immunofluorescence ( IF ) ( Fig . 4B ) . Membrane alterations induced by replicons harboring the CEP or PDA mutations were less abundant and more dispersed , again concordant with the IF phenotype , showing no NS5A clusters ( Fig . S3A ) . In addition , the average diameter of the DMVs was similar to the one observed with wt replicon cells ( Fig . S3D , E ) . This result suggests that DMVs have to adopt a minimal size to allow active RNA replication , since in case of the wt , average DMV diameter was identical for the replicon compared to the expression model . In summary , mutations in the PFIS ( e . g . mutHIT ) resulted in phenotypes largely resembling and even more pronounced than knockdown of PI4KIIIα , with MW structures solely consisting of DMVs of reduced average diameter . Expression of replication-competent mutants causing a moderate reduction in PI4KIIIα activation ( mutCEP and PDA ) resulted in intermediate phenotypes . These results provided evidence for a critical involvement of functional NS5A-PI4KIIIα interaction in MW morphology . NS5A exists in two phospho-isoforms , p56 ( hypophosphorylated ) and p58 ( hyperphosphorylated ) , which can be distinguished by their electrophoretic mobility . The balance between the two different NS5A variants has been implicated in the regulation of the viral replication cycle , with p56 thought to favor RNA replication and p58 involved in assembly [46] , [48] , [55] . Since NS5A was interacting with PI4KIIIα and is critically involved in kinase activation , we wondered whether the phosphorylation state of NS5A might be involved in this process . Therefore , we analyzed NS5A containing triple alanine mutations in the PBR expressed in the context of the NS3-5B polyprotein for the ratio of p58/p56 ( Fig . 6 ) . Surprisingly , we found a huge variability of p58/p56 ratios among the NS5A variants , ranging from 0 . 2 to 1 . 5 ( Fig . 6A , B and Table 1 ) . The differences were indeed due to variable phosphorylation , because phosphatase treatment resulted in a reduced apparent molecular weight of all bands ( Fig . S4 ) , in contrast to a generally higher apparent molecular weight of mutants VLR , RSM and MLT , which was likely due to a change in protein conformation . NS5A wt typically gave rise to less p58 compared to p56 ( roughly 40% to 60% ) and had a p58/p56 ratio of 0 . 6 , whereas most mutants with no ( mutSQL , ETA , LPC ) or a moderate impairment of PI4KIIIα binding ( mutLPC , CEP , PEP , PDA , ADV ) exhibited slightly reduced p58 levels resulting in a lower p58/p56 ratio ( Fig . 6B , Fig . S3F ) . Importantly , all mutations in the PFIS resulting in an abrogation of RNA replication caused a strong enhancement of p58/p56 levels ( Fig . 6A , B ) , arguing for an impact of PI4KIIIα on the phosphorylation state of NS5A . To our knowledge , these are the first reported mutations in NS5A causing an increase rather than a decrease in NS5A hyperphosphorylation . The strongly enhanced p58/p56 levels observed for the PFIS mutants furthermore suggested that impaired functional interaction of NS5A resulted in higher p58 levels , whereas a regular interaction of NS5A with PI4KIIIα might favor p56 synthesis . To analyze whether the changes in NS5A phosphorylation observed for the PFIS mutants were directly associated with PI4KIIIα , we addressed the impact of PI4KIIIα knockdown and overexpression on NS5A phosphorylation ( Fig . 7 ) . We used cells with stable knockdown of PI4KIIIα by constitutive expression of sh-RNA ( shPI4K ) ; cells expressing a non-targeting control ( shNT ) served as negative control . For further control we restored PI4KIIIα expression in shPI4K cells by transduction with a variant of the PI4KIIIα gene resistant to knockdown due to silent mutations in the shPI4K binding site ( sh+Esc ) . Efficient knockdown and reconstitution was confirmed by western blot for PI4KIIIα and by functional replication assays using subgenomic reporter replicons ( data not shown ) . Indeed , silencing of PI4KIIIα expression resulted in a significant increase in the p58/p56 ratio compared to shNT cells ( Fig . 7A , B , JFH-1 wt ) . Importantly , the wt phenotype was restored by expression of the knockdown-resistant PI4KIIIα gene , strongly arguing for a specific impact of PI4KIIIα on the regulation of NS5A phosphorylation . The same result was obtained for two variants of the genotype 1b isolate Con1: Con1 wt , barely replicating in cell culture and Con1 ET , harboring cell culture adaptive mutations in NS3 and NS4B [52] , demonstrating that modulation of the NS5A phosphorylation status by PI4KIIIα is a general feature of different HCV genotypes . We furthermore generated and tested the HIT triple alanine mutant in the PFIS of the Con1 isolate to confirm the general importance of this motif ( Fig . 7A , B , Con1 mutHIT ) . The HIT mutation in Con1 resulted in a significantly increased p58/p56 ratio and was associated with a strongly decreased binding to PI4KIIIα . This phenotype was similar to the one found for JFH-1 ( Fig . 7E , and data not shown; note the low amount of HA-PI4K coprecipitating with Con1 mutHIT ) , supporting the contribution of the PFIS to NS5A-PI4KIIIα binding . In contrast to JFH-1 , silencing of PI4KIIIα had no impact on the p58/p56 ratio of Con1 mutHIT , which was in line with our conclusion that the increase in p58 levels of this mutant NS5A was caused by its reduced binding to PI4KIIIα . To further confirm the role of PI4KIIIα on the regulation of NS5A phosphorylation , we tested the impact of a recently identified specific inhibitor of PI4KIIIα , AL-9 [15] , on the NS5A phosphorylation status . We found very similar inhibitory concentrations of AL-9 on replication of Con1 and JFH-1 based replicons as reported [15] ( IC50 ca . 0 . 1 and 0 . 5 µM , respectively ) . AL-9 treatment generally enhanced the abundance of NS5A , which was probably due to subtle effects of PI4KIIIα on NS5A stability , which we also observed in other experiments ( e . g . Fig . 7E , compare the impact of PI4KIIIα wt overexpression on NS5A abundance to mock and inactive mutant D1957A for Con1 wt and ET ) . Importantly , we also found a dose-dependent increase in the p58/p56 ratios of NS5A from JFH-1 , Con1 wt and Con1 ET upon AL-9 treatment ( Fig . 7C , D ) , as expected from the knockdown experiments , thereby providing more evidence for the involvement of PI4KIIIα in regulating the phosphorylation state of NS5A . We could furthermore rule out that AL-9 had an impact on binding of NS5A to PI4KIIIα ( Fig . S5 ) , thereby supporting our assumption that indeed inhibition of the enzymatic activity of PI4KIIIα by AL-9 resulted in an increased p58/p56 ratio . The silencing experiments and the phenotype of the PFIS mutants suggested that PI4KIIIα somehow facilitated the synthesis of p56 or suppressed the synthesis of p58 . Overexpression of PI4KIIIα should therefore result in a lowering of p58/p56 ratios and we tested this hypothesis by ectopic expression of PI4KIIIα ( Fig . 7E , F ) . In parallel to the wt gene we expressed an inactive mutant of PI4KIIIα ( D1957A , [11] ) to address whether kinase activity was involved in the regulation of NS5A phosphorylation . Ectopic expression of PI4KIIIα wt , but not of the D1957A mutant , indeed resulted in reduced p58/p56 ratios for JFH-1 wt and Con1 wt ( Fig . 7E , F ) , arguing for a role of enzymatically active PI4KIIIα in NS5A phosphorylation . P58/p56 ratios were also reduced in case of Con1 mutHIT coexpressed with PI4KIIIα wt and , surprisingly , also with mutant D1957A ( Fig . 7E , F ) . Overexpression of PI4KIIIα might therefore compensate the binding defects of the PFIS mutants to some extent , thereby reducing p58/p56 ratios , but probably also triggering changes in NS5A phosphorylation by different mechanisms as compared to NS5A wt , e . g . by preventing the access of cellular kinases involved in p58 synthesis . In case of Con1 ET , ectopic expression of PI4KIIIα wt had no impact on p58/p56 ratios , which might be due to the generally low p58/p56 ratio of NS5A Con1 ET in this experiment , consistent with a recent report [48] . Similar effects were found for some JFH-1 PBR mutants with low p58/p56 ratios ( Fig . S6 ) . In this case , p58/p56 ratios were slightly increased upon ectopic expression of PI4KIIIα for some mutants ( SQL , LPC , PEP ) , whereas the high p58/p56 ratios of mutants in the PFIS ( MLT , TDP , PPH and HIT ) were reduced , as for NS5A wt ( Fig . S6 compared to Fig . 6 ) . The reduction of p58/p56 ratios of PFIS mutants by PI4KIIIα overexpression might also explain the partial rescue of RNA replication upon PI4KIIIα overexpression in case of mutant PPH ( Fig . 3D ) . In summary , mutants of the PFIS impaired in PI4KIIIα binding exhibited increased ratios of p58/56 . The same phenotype was observed upon knockdown or pharmacological inhibition of PI4KIIIα , whereas overexpression of enzymatically active PI4KIIIα resulted in decreased p58/56 ratios . Thus , PI4KIIIα appears to affect NS5A phosphorylation either by facilitating p56 synthesis or by blocking NS5A hyperphosphorylation . Triple alanine mutations in the PFIS impaired NS5A-PI4KIIIα binding and completely abolished HCV RNA replication . Abrogation of RNA replication was associated with a loss of PI4P induction and increased p58/p56 ratios . The same phenotypes were observed upon PI4KIIIα knockdown or inhibition of PI4KIIIα activity by AL-9 , suggesting that replication deficiency of PFIS mutants was indeed mechanistically linked to PI4KIIIα activity ( Table 1 ) . This assumption was furthermore supported by the partial rescue of RNA replication of mutant PPH upon PI4KIIIα overexpression ( Fig . 3D ) . However , it was not clear whether induction of PI4P synthesis or regulation of NS5A phosphorylation were the key function of PI4KIIIα required for HCV replication . To get a hand on mutants with probably more distinct phenotypes we generated point mutations affecting serine and threonine residues within or surrounding the PFIS , which could be potential phosphorylation sites involved in the altered p58/p56 ratios ( Fig . S7 A–D ) . Among those only T2185A ( T210A in NS5A , affecting the threonine in mutHIT ) was impaired in RNA replication , correlating with a slight reduction of PI4P induction and an increase in p58/p56 ratio . We furthermore assessed whether a phosphomimetic mutation at this site would rescue PI4KIIIα interaction and RNA replication of mutHIT , in case this threonine was phosphorylated in vivo . Therefore , we replaced threonine or alanine at position 210 by glutamic acid in the wildtype sequence and in mutHIT , respectively ( Fig . S7E–G , mutHIE and mutAAE , respectively ) . However , both variants strongly interfered with PI4KIIIα binding and completely abrogated RNA replication , arguing against a phosphorylation event at this site . Since alteration of NS5A phosphorylation and PI4P induction seemed to be intimately linked in case of the PFIS mutants , we next tried to dissect the requirements for both parameters by using transcomplementation assays . Previous studies have shown that replication-deficient mutants of NS5A , in particular those with defects in phosphorylation , can be rescued by expression of a wt protein in trans [50] , [51] , [57] . Minimal requirement for transcomplementation was the expression of NS5A in the context of a NS3-5A polyprotein , which was necessary and sufficient for NS5A hyperphosphorylation . In contrast , the sole expression of NS5A was not capable of rescuing deficient mutants [50] , most likely due to aberrant phosphorylation , indicated by the lack of p58 [38] . Therefore , we generated six Huh7-Lunet derivatives , either containing a subgenomic replicon or constitutively expressing NS3-5A or NS5A , each based on JFH-1 or Con1 ET , respectively , to analyze which of these settings was capable of rescuing replication of a representative PFIS mutant ( HIT ) in the context of JFH-1 and Con1 ET reporter replicons ( Fig . 8A ) . Wt replicons of both genotypes as well as a NS5B mutant ( ΔGDD ) , which cannot be complemented in trans [50] , were included as positive and negative controls , respectively . We first analyzed each rescue setting for induction of PI4P synthesis and NS5A phosphorylation in the absence of the transfected mutant replicons ( Fig . 8B–D ) . Both replicon cell lines contained functional NS5A by definition , although in case of Con1 ET only p56 was clearly detectable ( Fig . 8D ) . Both cell lines expressing NS3-5A contained two NS5A species of the expected sizes , in contrast to the cells expressing NS5A only ( Fig . 8D ) . The replicon cell lines also contained elevated levels of PI4P ( Fig . 8B , C ) . Surprisingly , neither expression of NS3-5A nor of NS5A increased intracellular PI4P abundance ( Fig . 8B , C ) , despite similar NS5A levels compared to replicon cells ( Fig . 8D ) . The lack of PI4P induction in cells expressing NS3-5A suggested that activation of PI4KIIIα requires not only NS5A , but also NS5B , which has been shown to interact with this lipid kinase as well [11] . Next , we addressed the rescue profile of the mutant replicons in the various transcomplementation settings ( Fig . 8E , F ) . Successful rescue was judged by comparison with the ΔGDD mutant , which does not replicate and cannot be complemented in trans [50] . RepHIT JFH-1 ( Fig . 8E ) and repHIT Con1 ET ( Fig . 8F ) did not replicate in naïve Huh7-Lunet cells nor in cells expressing only NS5A , providing neither properly phosphorylated NS5A nor PI4P . Replication of repHIT JFH-1 was rescued by expression of NS3-5A and replicons of both genotypes ( Fig . 8E ) , suggesting that the defect caused by the mutant was complemented by providing NS5A rather than by high PI4P levels , which were not provided by expression of NS3-5A ( Fig . 8C ) . In contrast , replication of mutant repHIT Con1 ET was only rescued in cells harboring a replicon or NS3-5A of the same genotype ( Fig . 8F ) . Expression of JFH-1 NS3-5A was even inhibitory for replication of Con1 ET wt for unknown reasons , whereas JFH-1 replicon cells supported Con1 ET wt replication , but did not rescue Con1 ET repHIT ( Fig . 8F ) . The latter result clearly demonstrated that increased PI4P levels provided by the JFH-1 replicon cells were not sufficient to compensate for the defects caused by mutations in the PFIS and strongly argued for the necessity to provide wt NS5A , probably due to the aberrant phosphorylation induced by PFIS mutations . However , we cannot rule out the possibility that mutHIT causes additional defects in NS5A , which might require complementation by functional NS5A of the same genotype . We next analyzed the impact of restored replication upon transcomplementation on intracellular PI4P levels . We used wt and HIT-mutant replicons with an eGFP inserted in NS5A [58] to unequivocally detect cells with active replication of the mutant replicon . This experiment was focused on JFH-1 replicons due to the limited replication and transcomplementation efficiency of Con1 ET and only included rescue conditions not inducing PI4P synthesis ( NS5A and NS3-5A of Con1 and JFH-1 , Fig . 9A ) . The transcomplementation pattern of JFH-1 repHIT-eGFP was identical to the non-GFP tagged variant ( Fig . 9B ) : Replication of repHIT was restored by expression of NS3-5A , but not NS5A , of Con1 and JFH-1 . Quantification of PI4P in NS5A-positive cells revealed a strong , but variable , induction of PI4P for the JFH-1 wt-eGFP replicon ( Fig . 9C , D ) , which was about 6-fold on average ( Fig . S8A ) and very similar to the JFH-1 wt replicon ( Fig . S3B , C ) and to expression of NS3-5B wt ( Fig . 4C , Table 1 ) . Overall , the quantity of wt NS5A-eGFP correlated significantly with the amount of PI4P in individual cells ( Fig . S8B , blue dots ) , in line with the assumption that NS5A and NS5B activate PI4KIIIα , thus giving rise to elevated levels of PI4P . In contrast , PI4P induction was much weaker upon transcomplementation of JFH-1 repHIT-eGFP ( Fig . 9C , D ) , only 2–3 fold on average ( Fig . S8A ) , although replication levels of the wt-eGFP replicon were identical to repHIT-eGFP in cells expressing NS3-5A JFH , as judged by luciferase counts ( Fig . 9B ) . Many cells with bona fide HCV replication even had no detectable induction of PI4P synthesis ( Fig . 9C , D , repHIT-eGFP ) , suggesting that strongly elevated levels of PI4P are no prerequisite of HCV replication . However , local changes in PI4P levels at the replication sites would clearly be below the detection limit of the IF based quantitation . Therefore , we cannot exclude and it even seems likely that a local elevation of PI4P levels at the replication sites is critical for viral replication , since active replication always generated conditions capable of activating PI4KIIIα . This was indicated by a clear induction of PI4P in some of the cells containing a repHIT-eGFP replicon ( Fig . 9D ) . The amount of NS5A-eGFP and PI4P did not correlate significantly with PI4P in this case ( Fig . S8B ) , most likely due to the fact that PI4KIIIα was activated only by wt NS5A , provided by expression of NS3-5A , but not by the mutant NS5AGFP-mutHIT , which we quantified in our analysis . In summary , our transcomplementation analysis revealed that activation of PI4KIIIα required both NS5A and NS5B . Conditions capable of providing properly phosphorylated NS5A rescued replication of a PFIS mutant . However , elevated PI4P levels were not sufficient for successful transcomplementation and no prerequisite for rescue of a replicon with mutations in the PFIS . Still , our results indicated that active HCV replication generated conditions capable of activating PI4KIIIα . Therefore , modulation of NS5A phosphorylation as well as of PI4P metabolism by PI4KIIIα seem to be intimately linked processes and might both be required for HCV replication .
Our initial deletion analysis identified aa 187–213 of NS5A to be involved in PI4KIIIα binding ( PBR ) , which could be narrowed down to a sequence of 7–9 aa ( PFIS ) crucial for PI4KIIIα interaction ( Fig . 10A ) . The PFIS encompasses the very C-terminus of NS5A D1 ( aa 202–210 ) and is highly conserved among all HCV genotypes , in line with the essential function of PI4KIIIα in HCV replication and with previous mapping studies [11] , [61] . No function has been assigned yet to this region [4] and , unfortunately , this motif is not included in published crystal structures of NS5A D1 [33] , [62] . However , a very recent study suggests that the PFIS partially overlaps with a region adopting an α-helical structure , which might be induced upon interaction with other proteins [63] . It also seems likely that the interaction of NS5A with PI4KIIIα is not restricted to the PFIS since almost all mutations in the PBR affected binding to PI4KIIIα to various extents and NS5A lacking the entire PBR or even D1 retained some PI4KIIIα binding ( Fig . 1 , ΔS3 and ΔD1 ) . Our results furthermore strongly indicate that the interaction with NS5B is also essential for functional PI4KIIIα interaction , since expression of an NS3 to NS5A polyprotein was not capable of activating PI4KIIIα . Still , according to our data , the PFIS is important for PI4KIIIα binding and indispensable for the activation of the lipid kinase activity . All triple alanine mutations in this region had almost identical phenotypes , very similar to PI4KIIIα knockdown ( Table 1 ) , correlated with a strong impairment in PI4KIIIα binding and blocked HCV replication . This correlation argued for a replication defect mediated by interference with functional NS5A-PI4KIIIα interaction , which was furthermore supported by the partial rescue of RNA replication of PFIS mutant PPH upon PI4KIIIα overexpression , probably compensating the reduced binding efficiency . Replication defects of other PFIS mutants ( MLT , TDP and HIT ) were not compensated by PI4KIIIα overexpression , suggesting that the functional interaction of these NS5A mutants with PI4KIIIα was more severely impaired as for mutant PPH . Indeed , the PPH mutant was slightly less impaired in PI4P induction ( Fig . 4C ) , induced MW clusters in a lower number of cells ( Fig . 4B ) and had a slightly reduced p58/p56 ratio ( Fig . 6 ) compared to the other replication dead PFIS mutants . Alternatively , mutants MLT , TDP and HIT could require higher PI4KIIIα expression levels to compensate the functional interaction defects , which cannot be achieved by our lentiviral transduction system . We furthermore cannot rule out that mutations in the highly conserved PFIS motif might affect other important functions of NS5A independent from PI4KIIIα , like RNA binding , dimerization etc . and therefore cannot be rescued solely by PI4KIIIα overexpression . M202A and T210A conferred intermediate phenotypes to the adjacent triple alanine mutants ( RSM and TAE , respectively , Table 1 ) and therefore seem important , however not essential for the NS5A-PI4KIIIα interaction . Other mutations outside the PFIS were less consistent in their phenotypes by interfering with PI4KIIIα binding and PI4P synthesis only slightly , but not affecting RNA replication or p58/p56 ratios ( e . g . mutants CEP , PEP and PDA; Table 1 ) . This limited correlation argues for a complex interaction and suggests some functional flexibility until the defects fall below a threshold , as in case of the PFIS mutants . This flexibility was also indicated by the lack of strong phenotypes in case of the single mutants within the PFIS , despite the high degree of conservation at these positions ( T204A and T210A , Fig . S7 , Fig . 10A ) . Interestingly , alterations of D205 have even been identified as replication enhancing adaptive mutations in Con1 [46] , [52] , although this residue is invariant in natural isolates . The most striking and surprising phenotype observed for the PFIS mutants was the increase in p58/p56 ratios . Impairment of hyperphosphorylation is a much more common phenotype , which has been described upon sole expression of NS5A [38] , [39] , [50] , drug treatment [43] , [53] , [64] and for many mutants [46] , [48] , [50] . Loss of hyperphosphorylation is often associated with abrogation of replication by general disturbance of the NS5A structure , as observed for mutations in the N-terminal amphipathic helix [31] , [65] or in NS3 , 4A and 4B [37] , [38] , [66] . To our knowledge , triple alanine mutants of the PFIS are the first NS5A variants reported with increased p58 levels as compared to wt NS5A . This phenotype strongly argues against a general impairment of NS5A structure of the PFIS mutants and for a specific effect mediated by PI4KIIIα . Importantly , the same phenotype was observed upon knockdown of PI4KIIIα , whereas overexpression of an active PI4KIIIα resulted in a reduced p58/p56 ratio . Collectively , these data clearly point to a specific regulation of NS5A phosphorylation by PI4KIIIα in a way favoring p56 synthesis or suppressing hyperphosphorylation , congruent with the essential role of PI4KIIIα in HCV RNA replication . Several mechanisms could be envisaged mediating this phenotype . First , NS5A phosphorylation might be altered upon changes in the lipid environment of the viral replication sites induced by the activation of PI4KIIIα , thereby sequestering p56 and preventing p58 synthesis . This scenario seems rather unlikely , since the majority of viral nonstructural proteins ( >95% ) is not shielded by the MW , as judged by the accessibility to proteases [7] , [8] . In addition , p58 synthesis starts immediately after polyprotein processing and is completed within 20–60 minutes [40] , [67] , which might not be compatible with the time lines required for sequestration in membrane rearrangements . However , a concise analysis of the kinetics of p56 and p58 synthesis of different virus isolates in presence and absence of PI4KIIIα will be required to shed light on this important issue . It will also be very interesting to analyze whether p58/p56 ratios differ in total cellular lysates versus protease resistant fractions for wildtype and PFIS mutants , although these experiments are technically challenging due to the low amounts of protease resistant nonstructural proteins . Second , enhanced PI4P levels might recruit kinases or phosphatases involved in the regulation of p58/p56 ratios . Up to now only a few proteins have been found to specifically bind PI4P ( reviewed in [16] ) , which might be involved in the recruitment of enzymes modulating NS5A phosphorylation . Among those , oxysterol-binding protein 1 ( OSBP ) and ceramide transfer protein ( CERT ) have already been shown to be involved in the secretion of virions [68] , [69] . However , recent proteomic data on the composition of the HCV replication sites neither identified significant accumulations of any of these proteins nor kinases or phosphatases which might be involved in regulation of NS5A phosphorylation [70] . Still , a more comprehensive proteomic study involving conditions of PI4KIIIα knockdown might help to identify host factors involved in NS5A phosphorylation recruited by PI4P or by PI4KIIIα . Third , PI4KIIIα might shield NS5A from interaction with kinases such as CKIα or Plk1 [44] , [45] , thereby preventing hyperphosphorylation . Interestingly , the PFIS motif ( aa 202–210 ) identified in our study to be essential for PI4KIIIα interaction is adjacent to the serine cluster critical for p58 synthesis ( aa 222–235 ) . Therefore , binding of PI4KIIIα to the PFIS might passively prevent the access of enzymes promoting hyperphosphorylation and/or block the phosphorylation sites , thereby favoring p56 synthesis . However , this hypothesis is not in line with the requirement for PI4KIIIα enzymatic activity to reduce p58/p56 ratios in case of JFH-1 and Con1 wt . Fourth , PI4KIIIα might directly phosphorylate NS5A , thereby blocking the interaction with CKIα/Plk1 and preventing hyperphosphorylation . Although no protein kinase activity has been demonstrated for PI4KIIIα yet , the closely related isoform PI4KIIIβ has been shown to autophosphorylate in vitro , thereby inhibiting lipid kinase activity [71] . A more distinctly related class of PI kinases , phosphoinositide 3-kinases ( PI3K ) , also contain lipid kinase as well as protein kinase activities , which are both essential for their function [72] , [73] , and share structural similarities to protein kinases [74] . Interestingly , phosphatidylinositol 3-kinase-related kinases ( PIKKs ) like mTOR comprise a related family of Ser/Thr kinases without lipid kinase activity [75] . All PI3Ks , PIKKs , PI4KIIIα and PI4KIIIβ are sensitive to wortmannin [13] , [76] , arguing for some structural relatedness in the active center and leaving the possibility that PI4KIIIα might give rise to protein phosphorylation as well . Preliminary data using PI4KIIIα from commercial sources ( Invitrogen , Millipore ) and proteins purified by ourselves indeed revealed some protein kinase activity in vitro ( D . Radujkovic , V . Lohmann , unpublished data ) , but these preparations were not of sufficient purity to draw firm conclusions , as judged by the presence of contaminating protein kinase activity also in preparations of inactive kinase mutants . Therefore , a comprehensive analysis of PI4KIIIα protein kinase activity in vitro including inactive mutants and a thorough analysis of substrate requirements are needed to verify a direct role of PI4KIIIα in NS5A phosphorylation . Several experimental results pointed to a role of PI4KIIIα in regulating the abundance of NS5A , irrespective of the phosphorylation state ( Fig . 7 ) . First , AL-9 treatment generally and dose dependently enhanced the abundance of NS5A ( Con1 and JFH-1 ) . Second , overexpression of PI4KIIIα wt reproducibly reduced the abundance of NS5A ( Con1 ) , whereas an inactive PI4KIIIα mutant rather seemed to increase the apparent expression levels of NS5A . Third , NS5A PFIS mutants , devoid of functional PI4KIIIα interaction , seemed to be expressed to higher levels than the wildtype counterparts ( preferentially Con1 ) . These results indicate that active PI4KIIIα probably reduces the stability of NS5A , which might be important for the regulation of RNA replication . We have not yet analyzed , whether PI4KIIIα also affects the abundance of other nonstructural proteins , but this important question will be addressed in future experiments . However , it is currently unlikely that the impact of PI4KIIIα on NS5A stability/abundance is the main role of this host factor in regulating HCV RNA synthesis , since this phenotype was less pronounced for JFH-1 NS5A ( apart from AL-9 treatment ) and not found at all upon silencing of PI4KIIIα . Still , a regulatory role of PI4KIIIα in the half-life and turnover of HCV nonstructural proteins and/or replication complexes is an interesting hypothesis , which needs to addressed in more detail in subsequent studies . Previous studies have shown that NS5A was critically involved in the activation of PI4KIIIα [11] , [29] , which was confirmed by the failure of the PFIS mutants to give rise to PI4P induction . Interestingly , an additional mutant ( LPC ) , which was severely impaired in RNA replication , most likely by a mechanism independent of PI4KIIIα , failed to activate PI4KIIIα , suggesting that activation of PI4KIIIα might be disturbed by a number of pleiotropic mutations interfering with the overall integrity of NS5A . Our new data furthermore indicate that expression of NS5A or NS3-5A is not sufficient for PI4KIIIα activation ( Fig . 8C ) , suggesting that a complex of NS5A and NS5B is involved in the regulation of PI4KIIIα . This is in line with previous data identifying an interaction of PI4KIIIα with both viral proteins [11] , but in striking contrast to results showing PI4P induction and redistribution by the sole expression of NS5A [11] , [29] . Here , we used lentiviral transduction to reach physiological expression levels comparable to replicon cells , whereas former studies relied on T7-based expression [11] or tet-inducible expression in osteosarcoma cells [29] , giving rise to much higher protein abundance , which might explain this discrepancy . Although NS5A was shown to be sufficient for activation of PI4KIIIα in vitro [11] , [29] , we found no evidence for a stronger activation of PI4KIIIα in vitro by simultaneous addition of NS5A and NS5B purified from E . coli ( D . Radujkovic , V . Lohmann , unpublished data ) . This observation points to a complex regulation , probably depending on the phosphorylation state of NS5A or on the presence of additional host factors like protein kinase D , which activates the related PI4KIIIβ by phosphorylation [77] . Indeed , protein kinases seem to be physically associated even with purified NS5A [47] , [78] , which could be involved in the activation of PI4KIIIα . Anyhow , a thorough analysis in vitro and in cell culture will be required to unravel the intricate determinants of PI4KIIIα activation by HCV and the role of NS5A and NS5B in this process . Generally , little is known about the regulation of PI4KIIIα [13] , but virus-induced increases in PI4P levels are not unique to HCV [27] . Enteroviruses activate PI4KIIIβ by a yet to be defined mechanism , generating a binding platform for the viral polymerase important for RNA replication [22] . In case of HCV , a more recent study indicates that HCV not only activates PI4KIIIα but also prevents transport of PI4P to the plasma membrane by an unknown mechanism and both mechanisms seem to contribute to the intracellular accumulation of PI4P [15] . However , our novel data suggest that the gross changes in PI4P levels are not essential for HCV RNA replication but might rather be a consequence of viral replication/protein expression . This assumption is based on the huge variations in PI4P levels observed in HCV positive cells , particularly since some of the mutants gave rise to significantly reduced PI4P levels but were not impaired in RNA replication at all ( e . g . mutCEP and PDA , Table 1 ) . Furthermore , the transcomplementation analysis revealed a number of cells with detectable rescue of HCV replication but lacking detectable alterations in intracellular PI4P levels and substantial colocalization of PI4P and NS5A ( Fig . 9C , D ) . Finally , elevated levels of PI4P alone were not capable of rescuing the defects of the PFIS mutants , as demonstrated by the lack of replication of Con1 repHIT in JFH-1 wt replicon cells , although this phenotype might be due to additional defects in NS5A caused by the HIT mutations . However , previous data demonstrated that only 20% of intracellular PI4P colocalized with NS5A [11] and that expression of the PI4P phosphatase Sac1 had only a limited impact on HCV replication [22] , [24] . Collectively , these results argue against a distinct role of globally enhanced intracellular PI4P concentrations in viral RNA replication . We therefore speculate that activation of PI4KIIIα by HCV might be required to generate locally enhanced PI4P amounts at the site of replication , resulting in global disturbance of PI4P metabolism as a collateral effect . Alternatively , the new PI4P pools might be required at a later step of the viral replication cycle like assembly or release of virions , as suggested by involvement of PI4P binding proteins CERT and OSBP in HCV secretion [68] , [69] . Enhancement of intracellular and depletion of plasma membrane PI4P pools might furthermore impact on signaling events and/or contribute to viral pathogenesis in vivo . However , these important questions can only be addressed in adequate animal models , which will hopefully be developed in the future . Accumulating evidence suggests that PI4KIIIα is important for the generation of the MW . Silencing and inhibition of PI4KIIIα was associated with clustered web structures in IF [11] , [15] , [18] , which contained smaller DMVs and lacked MMVs [11] . These phenotypes have been attributed to reduced PI4P levels in absence of PI4KIIIα activity . Clustered distribution of NS5A in IF as well as more distinct accumulation of vesicles with smaller DMVs was also observed upon expression of the mutant NS3-5B proteins , confirming that these phenotypes are linked to PI4KIIIα . However , mutHIT , being devoid of PI4KIIIα activation , induced even smaller DMVs than silencing of PI4KIIIα , although PI4P levels were comparable and not significantly different for both conditions ( Table 1 ) . These data argue for PI4KIIIα-mediated mechanisms modulating the morphology of the HCV replication sites independent from PI4P and probably mediated by NS5A phosphorylation . Interestingly , a very recent study demonstrated that ectopic expression of NS5A induced the formation of vesicles containing several lipid bilayers and occasionally vesicles containing a pair of membranes morphologically identical to DMVs , whereas sole expression of NS3/4A , NS4B and NS5B generated only single membrane vesicles , emphasizing the role of NS5A as a key regulator of MW morphogenesis [79] . Importantly , expression of NS3-5A did not significantly induce PI4P synthesis in our study , but generated DMVs and some more tubular double membrane structures with an average diameter similar to NS3-5B wt [79] and larger than mutHIT or silencing of PI4KIIIα . It is therefore tempting to speculate that changes of NS5A phosphorylation rather than PI4P induced by PI4KIIIα are involved in modulation of the MW , since mutHIT also generated a stronger change in p58/p56 ratio than PI4KIIIα silencing ( 1 . 5 vs . 1 . 0 respectively , compared to 0 . 6 for wt ) . The small DMV size and the lack of MMVs observed upon silencing of PI4KIIIα and expression of mutHIT might therefore at least in part be due to higher p58/p56 ratios associated with these conditions ( Table 1 ) , probably mediated by a host factor . Interestingly , hVAP-A , a cellular protein involved in vesicle trafficking and essential for HCV replication [80] , [81] , has been shown to bind p56 and not p58 [49] and thereby represents a promising candidate factor for phosphorylation dependent regulation of MW morphology , particularly since hVAP-A is enriched in viral replication sites [81] . However , a more detailed ultrastructural and biochemical analysis of PI4P induction and NS5A phosphorylation will be required to further dissect the contribution of both factors to the composition of HCV-induced membrane alterations and to clarify the role of other cellular proteins . Our novel finding of PI4KIIIα modulating NS5A phosphorylation adds an additional layer of complexity to the mechanism of action of this essential host factor of HCV RNA replication . However , we favor a model in which interaction of PI4KIIIα with NS5A p56 triggers a phosphorylation event preventing further hyperphosphorylation ( Fig . 10B ) . P56 or a distinct ratio of p58/p56 is then involved in the morphogenesis of the MW , probably by recruiting a host factor like hVAP-A . In addition , p58/p56 ratios might also have a more direct impact on viral replication , e . g . by regulating RNA synthesis . Interaction of PI4KIIIα with NS5A in concerted action with NS5B furthermore activates the lipid kinase , resulting in increased intracellular PI4P levels . Activation of PI4KIIIα might again be supported by a host factor , e . g . a protein kinase , recruited by NS5A or NS5B . Locally increased PI4P levels might be involved in MW morphology as well , but our data argue against a vital role of globally increased PI4P concentrations . In contrast , we speculate that the activation of the lipid kinase is rather a consequence of the interaction with the viral proteins and might impact on viral pathogenesis . Conclusively , our study provides important novel insights into the role of PI4KIIIα in HCV RNA replication . Still , we are far from a comprehensive view on the entire mechanism of action , in particular regarding the role of NS5B in its activation , the determinants of NS5A phosphorylation and further host factors involved . However , unraveling the complexities of HCV-PI4KIIIα interactions will be of central importance for our understanding of the viral replication machinery and might also help to shed light on the distinct cellular functions of PI4KIIIα , which are barely understood .
The Huh-7 cell clone Huh7-Lunet , highly permissive for HCV RNA replication [82] was used for electroporation assays . Huh7 cell lines bearing subgenomic replicons of either JFH-1 isolate [56] or adapted Con1 ET [7] have been described recently . Huh7-Lunet T7 cells [83] were used for transient expression of plasmids coding for HCV proteins analyzed in immunofluorescence and immunoprecipitation assays . Huh7-Lunet cells served for the generation of cell lines stably overexpressing HCV proteins . Following cell lines have been created by lentiviral transduction as described elsewhere [83] using lentiviral pWPI plasmid vectors under selection of blasticidin or zeocin: Huh7-Lunet NS3-5A-JFH1 , Huh7-Lunet NS3-5A-Con1ET , Huh7-Lunet NS5A-JFH1 and Huh7-Lunet NS5A-Con1ET were used in transcomplementation assays . Huh7-Lunet T7 cells stably overexpressing either wildtype PI4KIIIα or inactive D1957A mutant were established by lentiviral transduction using pWPI-HA-PI4KIIIα or pWPI-HA-PI4KIIIαD1957A , respectively [11] . Cells with stable knockdown of PI4KIIIα ( shPI4KIIIα ) or non-targeting control cells ( shNT ) were prepared according to a recently published protocol [84] . The shRNA targeting sequences were 5′-CAG TGG AAG GAC AAC GTG-3′ ( PI4KIIIα ) and 5′-TCT CGC TTG GGC GAG AGT AAG-3′ ( NT ) . Huh7-Lunet T7 cells with stable PI4KIIIα knockdown expressing an shRNA escape variant of PI4KIIIα ( si+esc ) have been generated by lentiviral transduction using pWPI HA-PI4KIIIα-shEsc . The lentiviral vector pWPI-BLR [83] has been used for cloning of following plasmids for the generation of stable cell lines under blasticidin selection: pWPI-NS3-5AJFH1 , pWPI-NS5AJFH1 , pWPI-NS3-5ACon1ET and pWPI-NS5ACon1ET . The gene encoding the >230 kDa full-length PI4KIIIα isoform 2 was originally obtained from Kazusa DNA Research Institute , Chiba , Japan ( product ID FXC00322 , corresponding to GenBank accession number AB384703 , numberings refer to this GeneBank ID ) . For construction of pWPI-PI4KIIIα-shEsc , a synthetic gene fragment was obtained from GeneArt ( Regensburg , Germany ) , which comprised the region targeted by the different shRNAs and exhibited silent mutations in shRNA-targeting sequences . The silent escape mutations were based on the following nucleotide exchanges: g4644a , g4650a , c4653t , c4656t , g4659c , g5085a , c5088t , c5091a , c5097t , c4437g , t4441a , c4442g , t4443c , g4446a , a4449g , c4452g . Detailed cloning protocols for lentiviral constructs can be obtained upon request . All amino acid and nucleotide numbers refer to the position of the corresponding amino acid in the complete HCV genomes or in the NS5A protein of JFH1 and Con1 ( GenBank accession no . AB047639 and AJ238799 , respectively ) . PTM vectors allowing expression of the HCV nonstructural proteins NS3 to 5B or individual HCV NS proteins as well as the HA-tagged PI4KIIIα have been described recently [11] , [83] . PTMNS3-5B served as vector plasmid for internal NS5A deletions as well as for triple alanine NS5A mutations . PTMNS3-5B/NS5AΔD1 was described recently [83] . Following subdeletions of NS5A within the context of the polyprotein were generated by three-fragment ligation by using the BstXI/SpeI digested PCR fragments and vector fragments obtained by restriction with BstXI and SpeI or SpeI and RsrII . Overlap-PCR were performed using primers F_MfeI and R_RsrII_JFH as well as primers indicated in table S1 spanning the crossover sites . PTM NS3-5B/5AΔS1 contains an in frame deletion of aa2002–2068 established by overlap-PCR using primers F_del5AD1a and R_del5AD1a . PTM NS3-5B/5AΔS2 contains an in frame deletion of aa2069–2126 established by overlap-PCR primers F_del5AD1b and R_del5AD1b . PTM NS3-5B/5AΔS3 contains an in frame deletion of aa2127–2190 within NS5A domain I established by overlap-PCR using primers F_BssHII and R_SacI as well as the primers F2_Del_alles_small and R124_Del_alles spanning the crossover site . Plasmid was generated by three-fragment ligation by using the BssHII/SacI digested PCR fragment and vector fragments obtained by restriction with BssHII and SpeI or SpeI and SacI . Further subdeletions ΔS3A ( Δaa2127–2162 ) , ΔS3B ( Δaa2163–2175 ) and ΔS3C ( Δaa2176–2189 ) were established by overlap-PCR using the primers F_BssHII and R_SacI as well as the primers for the according single crossover sites which can be found in table S1 . Plasmids were generated by three-fragment ligation by using the BstXI/RsrII digested PCR fragment and vector fragments obtained by restriction with BstXI and SpeI or SpeI and RsrII . Triple alanine mutants pTMNS3-5B/NS5AmutXXX ( XXX refers to SQL , LPC , CEP , PEP , PDA , ADV , VLR , RSM , MLT , TDP , PPH , HIT , TAE , ETA ) as well as pTMNS3-5B containing single amino acids substitutions in NS5A ( S201A , T204A , T210A and T213A ) were created by overlap-PCR using the primers F_BssHII and R_SacI as well as the primers for the according single crossover sites which can be found in table S1 . Plasmids were generated by three-fragment ligation by using the BssHII/SacI digested PCR fragment and vector fragments obtained by restriction with BssHII and SpeI or SpeI and SacI . According plasmid constructs of genotype 1b ( Con1 or Con1ET ) containing the triple alanine mutation mutHIT have been created as well by overlap-PCR and fragment substitutions: pTMNS3-5B/NS5AmutHIT Con1/Con1ET were created by overlap-PCR using the primers S/1B/6803 and A/1B/7263 as well as the primers for the according single crossover sites which can be found in table S1 . Plasmids were generated by fragment ligation by using the EcoRI/XhoI digested PCR fragment and vector fragment ( either from pTMNS3-5B Con1 or Con1ET , respectively ) obtained by restriction with EcoRI and XhoI . Subgenomic reporter replicons have been have been described elsewhere: the monocistronic genotype 2a ( JFH-1 ) reporter pFKI389-Lucubi-NS3-3′/JFH1wt_δg [85]; the bicistronic adapted genotype 1b ( Con1ET ) reporter pFKI341-Luc-EI/NS3-3′/Con1ET_δg [52]; the luciferase reporter replicon of genotype 2a with an additional eGFP insertion in NS5A domain III pFKI389-Luc-NS3-3′/δg/JFH1-5A-eGFP [58] . Triple alanine mutations within pFKI389-Lucubi-NS3-3′/JFH1mutXXX_δg monocistronic replicons were created by replacing the NsiI/HindIII fragment of pFKI389-Lucubi-NS3-3′/JFH1wt_δg with the NsiI/HindIII fragment of the according pTMNS3-5B/NS5AmutXXX plasmid . PFKI341-PILuc-NS3-3′ET/NS5AmutHIT_δg bicistronic replicon was created by replacing the MluI/XhoI fragment of PFKI341-PILuc-NS3-3′ET_δg with the MluI/XhoI fragment of the pTMNS3-5B/NS5AmutHIT Con1ET plasmid . The triple alanine mutation mutHIT within the reporter replicon pFKI389-Luc-NS3-3′/δg/JFH1-NS5AmutHIT-eGFP containing an eGFP insertion in NS5A domain III was cloned by three fragment ligation by using the AgeI/SacI vector fragment of pFKI389-Luc-NS3-3′/δg/JFH1-5A-eGFP and fragments obtained from AgeI/NsiI cleavage of pFKI389-Luc-NS3-3′/δg/JFH1-5A-eGFP or NsiI/SacI cleavage of pTMNS3-5B . All PCR-derived sequences were confirmed by sequencing ( GATC , Konstanz , Germany ) . Transient HCV RNA replication assays were performed as described previously [86] . In brief , replicon encoding plasmid DNA harboring hepatitis delta virus ribozymes were restricted with MluI ( JFH1 ) or Spe I ( Con1 ) prior to in vitro transcription . Ten µg of run-off transcripts were used for electroporation of 4×106 Huh7-Lunet cells or Huh7-Lunet cells overexpressing PI4KIIIα , that were resuspended in 12 ml culture . Two ml aliquots were seeded per well of a 6-well plate and replication was determined by measuring luciferase activity in case of genomes containing the luciferase reporter gene at 4 h , 24 h , 48 h and 72 h post electroporation . Since luciferase activity measurable 4 h post transfection is derived from transfected input RNA , these values were used to normalize for transfection efficiency . For transcomplementation assays , Huh7-Lunet cells containing a stably selected subgenomic replicon or stably overexpressing NS3 to 5B or NS5A were transfected with 10 µg of replicon RNA that contain a luciferase reporter gene and are derived from pFKI plasmid DNA . Electroporated cells were seeded as described above and values obtained 4 h post electroporation were used to determine the transfection efficiency . To assess replication efficiency of different replicon constructs by direct measurement of viral RNA , 0 . 25 µg of replicon RNA was electroporated into Huh7-Lunet cells and HCV RNA was quantified by RT-PCR as described recently [87] . In brief , viral RNA was isolated from transfected cells using the Nucleo Spin RNAII kit ( Macherey-Nagel , Düren , Germany ) as recommended by the manufacturer . 200 ng of RNA sample was used for quantitative RT-PCR analysis using an ABI PRISM 7000 sequence detector system ( Applied Biosystems , Foster City , CA ) . HCV-specific RT-PCRs were conducted in triplicates with the One Step RT-PCR kit ( QIAGEN , Hilden , Germany ) using the following JFH1-specific probe ( TIB Molbiol , Berlin , Germany ) and primers ( MWG-Biotech , Martinsried , Germany ) : A-195 , 5′-6-carboxyfluorescein-AAA GGA CCC AGT CTT CCC GGC AAT T-tetrachloro-6-carboxyfluorescein-3′; S-146 , 5′-TCT GCG GAA CCG GTG AGT A-3′; and A-219 , 5′-GGG CAT AGA GTG GGT TTA TCC A-3′ . The amount of HCV RNA was calculated by comparison to serially diluted in vitro transcripts . For overexpression of HCV proteins , Huh7-Lunet T7 cells were transfected with Effectene ( Qiagen , Hilden Germany ) transfection reagent according to manufacturer's instructions and fixed 24 h post-transfection . HCV replicons were transfected by electroporation into Huh7-Lunet cells and fixed after 48 hours . Immunofluorescence protocol was performed as described elsewhere [11] . In brief: Cells were fixed in 4% PFA for 20 min and permeabilized with 50 µg/ml Digitonin for another 15 min . Primary antibodies were incubated in 3% BSA for 1 h at RT . NS5A was detected by using either NS5A-specific monoclonal mouse antibody ( 9E10 , generous gift from Charles M Rice ) with a final concentration of 3 µg/ml or using a polyclonal NS5A rabbit antiserum ( #4952 , [83] ) at a dilution of 1∶300 . NS3 , NS4B and NS5B were detected by polyclonal antisera previously described [83] . PI4P was stained using monoclonal mouse IgM anti-P4P antibody ( Echelon , Z-P004 ) with a final concentration of 5 µg/ml . eGFP signals were enhanced by a polyclonal rabbit anti-GFP antibody ( Abcam , ab290 ) with a dilution of 1∶200 . Alexa 488 or 546 conjugated secondary antibodies ( Invitrogen , Molecular Probes ) were incubated in 3% BSA for 45 min at RT with a dilution of 1∶1000 . Nuclei were stained using DAPI for 1 min . at a dilution of 1∶4000 . Cells were mounted with Fluoromount G ( Southern Biotechnology Associates , Birmingham , USA ) and pictures were acquired with a Nikon C1Si spectral imaging confocal laser scanning system on a Nikon Ti fully automated inverted microscope equipped with 60× objective . For PI4P quantitation , a 40× objective was used and z projections of confocal z stacks were generated with the “sum slices” option of ImageJ . After thresholding of the signal intensity of PI4P staining , PI4P amount of 35 different cells was determined by defining cell areas and taking the IntDen-value obtained with the “analyze particles” functions of Image J . Quantitation of membranous web phenotypes was based on a qualitative visual judgment of 350 randomly chosen NS5A positive cells , which were either assigned to the wt or cluster phenotype , based on the representative images shown in Fig . 4A . The percentage relies on the relative number of cells grouped into the wt or cluster phenotype for each construct . For overexpression of HCV proteins , Huh7-Lunet T7 cells were transfected with TransIT-LT1 ( Mirus , Madison USA ) transfection reagent according to manufacturer's instructions and fixed 24 hours post transfection . HCV replicons were transfected by electroporation into Huh7-Lunet cells and fixed after 48 hours . For fixing , cells were washed 3 times with 1× PBS and fixed for 30 min with pre-warmed 2 . 5% glutaraldehyde in 50 mM sodium cacodylate buffer ( pH 7 . 2 ) containing 1 M KCl , 0 . 1 M MgCl2 , 0 . 1 M CaCl2 and 2% sucrose . Cells were washed thoroughly 5 times with 50 mM cacodylate buffer and post-fixed on ice in the dark with 2% OsO4 in 50 mM cacodylate buffer for 40 min . Cells were washed with H2O overnight , treated with 0 . 5% uranyl acetate in H2O for 30 min , rinsed thoroughly with H2O and dehydrated in a graded ethanol series at RT ( 40% , 50% , 60% , 70% and 80% ) for 5 min each and 95% and 100% for 20 min each . Cells were immersed in 100% propilene oxid and immediately embedded in an Araldite-Epon mixture ( Araldite 502/Embed 812 Kit; Electron Microscopy Sciences ) . After polymerization at 60°C for 2 days coverslips were removed and the embedded cell monolayers were sectioned using a Leica Ultracut UCT microtome and a diamond knife . Sections with a thickness of 65 nm were counter-stained with 3% uranyl acetate in 70% methanol for 5 min and 2% lead citrate in H2O for 2 min , and examined with the transmission electron microscope Philips CM120 TEM ( Biotwin , 120 kV ) . For metabolic labeling a total of 4×105 Huh7-Lunet cells constitutively expressing the T7 RNA-polymerase ( Huh7-Lunet T7 ) were seeded in each well of a 6-well cell culture plate in complete DMEM . One day later , cells were transfected with pTM vectors [88] allowing protein expression under transcriptional control of the T7 promoter . 2 µg of pTM vectors supporting expression of the HCV nonstructural proteins NS3 to NS5B with deletions of NS5A subdomains or mutations within NS5A were cotransfected with 2 µg of either pTM HA-PI4KIIIα or an empty pTM vector ( mock ) by using Lipofectamine2000 ( Invitrogen ) according to the instructions of the manufacturer . After 7 h , cells were washed with methionine/cysteine-free medium and incubated in this medium for 1 h . For radiolabeling cells were incubated for 16 h in 1 ml methionine/cysteine-free medium , supplemented with 10 mM glutamine , 10 mM Hepes , and 100 µCi/ml of Express Protein labeling mix ( Perkin Elmer , Boston ) . In assays using AL-9 ( generous gift from R . De Francesco , P . Neddermann and F . Peri , Milan , Italy ) , the drug was diluted in DMSO to indicated concentrations and supplemented to the methionine/cysteine-free medium and incubated overnight . Cells were lysed by incubation of the cell pellets in NPB ( 50 mM Tris-Cl [pH 7 . 5] , 150 mM NaCl , 1% Nonidet P-40 , 1% sodium deoxycholate , 0 . 1% SDS , protease inhibitors ) for 1 h on ice . Lysates were cleared by centrifugation at 14 , 000 g for 10 min at 4°C and used for immunoprecipitation with antibodies of the following specificities: NS5A of genotype 1a ( H77 ) cross-reacting with Con1 and JFH-1 ( sheep polyclonal , a generous gift of M . Harris , Leeds university , U . K . ) ( Macdonald et al . , 2003 ) ; 7 µg of anti-HA tag ( mouse H3663 , Sigma ) . After 3 h incubation at 4°C immunocomplexes were captured by using protein-G-sepharose beads ( Sigma ) for an additional 3 h incubation at 4°C . Where indicated , complexes were treated after washing in buffer 3 ( NEB ) with 1 U CIP ( NEB , #M0290S ) for 1 h at 30°C . Immunocomplexes were dissolved in protein sample buffer , separated by 10% polyacrylamide-SDS gel electrophoresis and detected by autoradiography . Proteins were quantified by phosphoimaging using the Quantity One software ( Bio-Rad , Munich ) . HA-PI4KIIIα binding to NS5A was determined by normalizing the amount of HA-PI4KIIIα co-precipitating with NS5A to the total amount of HA-PI4KIIIα determined by direct IP using anti-HA antibodies . This option was chosen since quantitation of total PI4KIIIα input levels by western blot was not reproducible due to variable and inconsistent transfer rates of the 240 kDa PI4KIIIα band . Data were furthermore not normalized to input NS5A levels due to a consistent 20–100fold molar excess of NS5A compared to HA-PI4KIIIα ( data not shown ) . For western blotting 1/10 cells of a T25 cell culture flask were denatured and heated in 2× Laemmli-buffer and loaded onto an 8% polyacrylamide-SDS gel . After separation and transfer to a PVDF membrane , immunoblotting was performed detecting NS5A using a monoclonal mouse antibody ( 9E10 ) at a concentration of 0 . 1 µg/ml and β-actin with a monoclonal mouse antibody ( Sigma , A5441 ) and a dilution of 0 . 5 µg/ml . LI-COR secondary antibodies conjugated to IRDye were used 1∶10000 . Bound antibodies were detected with the LI-COR Infrared Imaging System . General statistical analyses as indicated in the corresponding figures were performed using Microsoft Excel software . Correlations of experimental data were obtained by linear regression analysis using GraphPad Prism software .
|
Hepatitis C virus ( HCV ) infections affect about 170 million people worldwide and often result in severe chronic liver disease . HCV is a positive-strand RNA virus inducing massive rearrangements of intracellular membranes to generate the sites of genome replication , designated the membranous web . The complex biogenesis of the membranous web is still poorly understood , but requires the concerted action of several viral nonstructural proteins and cellular factors . Recently , we and others identified the lipid kinase phosphatidylinositol-4 kinase III alpha ( PI4KIIIα ) , catalyzing the synthesis of phosphatidylinositol 4-phosphate ( PI4P ) , as an essential host factor involved in the formation of the membranous web . In this study , we characterized the virus-host interaction in greater detail using a genetic approach . We identified a highly conserved region in the viral phosphoprotein NS5A crucial for the interaction with PI4KIIIα . Surprisingly , we found that PI4KIIIα , despite being a lipid kinase , appeared to regulate the phosphorylation status of NS5A , thus contributing to viral replication . Our results furthermore suggest that the morphology of the membranous web is regulated by NS5A phosphorylation , providing novel insights into the complex regulation of viral RNA replication .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"medicine",
"enzymes",
"microbiology",
"hepatitis",
"lipid",
"metabolism",
"hepatitis",
"c",
"infectious",
"diseases",
"lipids",
"virulence",
"factors",
"and",
"mechanisms",
"proteins",
"enzyme",
"regulation",
"membranes",
"and",
"sorting",
"biology",
"viral",
"replication",
"complex",
"viral",
"replication",
"biochemistry",
"virology",
"viral",
"diseases",
"molecular",
"cell",
"biology",
"metabolism"
] |
2013
|
The Lipid Kinase Phosphatidylinositol-4 Kinase III Alpha Regulates the Phosphorylation Status of Hepatitis C Virus NS5A
|
The loss of HIV-specific CD8+ T cell cytolytic function is a primary factor underlying progressive HIV infection , but whether HIV-specific CD8+ T cells initially possess cytolytic effector capacity , and when and why this may be lost during infection , is unclear . Here , we assessed CD8+ T cell functional evolution from primary to chronic HIV infection . We observed a profound expansion of perforin+ CD8+ T cells immediately following HIV infection that quickly waned after acute viremia resolution . Selective expression of the effector-associated transcription factors T-bet and eomesodermin in cytokine-producing HIV-specific CD8+ T cells differentiated HIV-specific from bulk memory CD8+ T cell effector expansion . As infection progressed expression of perforin was maintained in HIV-specific CD8+ T cells with high levels of T-bet , but not necessarily in the population of T-betLo HIV-specific CD8+ T cells that expand as infection progresses . Together , these data demonstrate that while HIV-specific CD8+ T cells in acute HIV infection initially possess cytolytic potential , progressive transcriptional dysregulation leads to the reduced CD8+ T cell perforin expression characteristic of chronic HIV infection .
CD8+ T cells play a central role in the control of HIV replication . During acute infection the emergence of HIV-specific CD8+ T cells correlates with resolution of peak viremia [1 , 2] , and in the nonhuman primate model experimental depletion of CD8+ T cells prior to infection with simian immunodeficiency virus delays resolution of acute viremia until the CD8+ T cell pool is reconstituted [3] . Further evidence of the immunologic pressure exerted by CD8+ T cells is manifest by CTL escape mutations throughout all phases of HIV infection and the association of certain MHC class I alleles with superior control of viral replication [4–9] . However , for the vast majority of infected individuals control is incomplete and ultimately fails in the absence of therapy . A better understanding of the CD8+ T cell response to HIV may inform the design of vaccines , therapeutics , or eradication strategies designed to stimulate or potentiate the natural response to infection resulting in better , if not complete , control . The CD8+ T cell response to viral infection is multifaceted , including the ability to proliferate , produce multiple cytokines and chemokines , degranulate , and induce cytolysis upon contact with infected targets [10] . During chronic progressive infection , HIV-specific CD8+ T cells have impaired proliferative potential [11–13] , are less capable of multifunctional responses [14 , 15] , and have reduced cytotoxic capacity [16–20] . The primary mechanism by which CD8+ T cells kill virally infected cells is via exocytosis of granules containing the cytolytic proteins perforin and granzyme B [21 , 22] . Control of HIV viremia has been associated with the ability of CD8+ T cells from chronically HIV-infected donors to upregulate these cytotoxic effector molecules following in vitro culture [18] , and we have shown that CD8+ T cell cytotoxic potential , defined by the ability to rapidly upregulate perforin following brief stimulation ex vivo , correlates inversely with viral load [16] . Effector CD8+ T cell development is coordinated by an array of transcription factors [23] . Murine studies have identified the T-box transcription family members T-bet and eomesodermin ( Eomes ) as important regulators of the differentiation and function of cytotoxic effector T cells [24–26] . T-bet positively regulates genes associated with effector functions including perforin , granzyme B , and IFN-γ [27 , 28] , whereas Eomes is associated with the expression of perforin as well as proteins involved in maintenance of memory CD8+ T cells [24 , 26 , 29 , 30] . While previous studies suggested a level of redundancy in the gene targets of these transcription factors , recent data show that the balance of T-bet and Eomes expression within a cell is a determinant of the differentiation pathway and functionality of the cell [30–34] . In the context of chronic HIV infection , HIV-specific CD8+ T cells with high levels of T-bet demonstrate greater overall functionality and maintain the ability to express perforin whereas cells with a T-betLoEomesHi phenotype are less differentiated , less functional , exhausted , and express little to no perforin [28 , 32] . Notably , during chronic progressive infection the T-betLoEomesHi phenotype dominates the HIV-specific CD8+ T cell pool [32] . It remains unclear if low T-bet levels and the associated deficiency in perforin expression results from progressive loss on the part of responding HIV-specific CD8+ T cells or if responding cells are inherently dysfunctional throughout the infection period . Much of our current knowledge regarding the dynamics of CD8+ T cell responses during acute infection is derived from murine models , particularly following infection with lymphocytic choriomeningitis virus , gammaherpesvirus , or influenza [35–37] . Infection by these viruses induces rapid and substantial activation and expansion of antigen-specific CD8+ T cells . Following resolution of acute viremia , the virus-specific population contracts , giving rise to memory cells that provide long-term protection . Human antiviral CD8+ T cell responses have primarily been assessed in the context of chronic infection , after the memory pool has been established [10 , 38–41] . Recent studies have examined development of human CD8+ T cell responses to a range of primary infections , including attenuated yellow fever virus , attenuated vaccinia virus , influenza , tick-borne encephalitis virus ( TBEV ) , hantavirus , and Epstein-Barr virus [42–47] , demonstrating that antigen-specific cells have immediate cytotoxic capacity directly ex vivo during the acute phase of these infections . The few studies to examine the earliest responses to HIV showed that HIV-specific CD8+ T cells have limited functionality during the acute phase of infection but did not assess cytotoxic potential or regulation by T-bet or Eomes [48 , 49] , leaving the question unresolved as to whether these effector molecules are induced during acute infection . Here , we examined the temporal dynamics of the CD8+ T cell effector response in peripheral blood of subjects experiencing acute primary HIV infection . We found that infection elicited a robust and highly activated response with immediate cytotoxic potential within the peripheral CD8+ T cell pool and that cells responding to short in vitro stimulation with HIV peptides were able to degranulate and rapidly upregulate perforin de novo . However , HIV-specific CD8+ T cells rapidly lost the ability to upregulate perforin following resolution of peak viremia . Loss of perforin expression coincided with a concurrent reduction in the expression of T-bet , but not Eomes , on a per-cell basis . Our data provide evidence of a robust and physiologically appropriate response during the earliest phase of acute HIV infection that is rapidly lost during progressive chronic infection , due in part to an inability to express sufficient levels of T-bet to properly drive effector differentiation .
Longitudinal samples were obtained from 32 subjects experiencing primary HIV infection ( Fig 1A ) , 28 of whom had at least one acute time point ( 36 time points total; median 54 d from infection , range 23–100 d ) and 23 with at least one chronic time point ( 40 time points total; median 551 d , range 367–880 d ) . Samples were drawn from three separate cohorts of acutely infected individuals: the CHAVI001 acute-infection cohort , the Montreal Primary Infection cohort , and the RV217/ECHO cohort . These cohorts provided broad geographical representation including North America , East Africa , Malawi , and Thailand ( S1 Table ) . Subjects were antiretroviral therapy naïve at all time points , consistent with the standard of care at the time of study , and none controlled viral load to undetectable levels ( Fig 1B ) . The mean peak viral load was 5 . 2 log10 RNA copies/ml for the entire study population ( 7 . 0 log10 RNA copies/ml for the better-characterized RV217 donors ) and 4 . 42 log10 RNA copies/ml at set point . Peripheral blood CD4+ T cell counts and CD8+ T cell counts both declined over the study period ( average rates of 80 cells/mm3 per year and 75 cells/mm3 , respectively; Fig 1C and 1D ) . Samples from 41 seronegative healthy donors , including pre-infection time points for the 11 RV217 acute subjects ( median -210 d from infection , range -41 to -478 d; Fig 1A and S1 Table ) , were analyzed for comparison . To determine if different phases of infection were associated with changes in circulating CD8+ T cell differentiation and activation , we assessed the size and composition of the memory CD8+ T cell pool ( S1 Fig ) . Relative to HIV-negative donors , HIV-infected subjects had a significantly larger memory ( non-CCR7+CD45RO- ) CD8+ T cell pool in both the acute and chronic phases of infection ( Fig 2A and 2B ) . Of note , the frequency of total memory CD8+ T cells at the earliest post-infection time points inversely correlated with peak viral load , but not with set point viral load ( Figs 2C and S2A ) . In addition to the larger memory pool we also observed a shift in the distribution of memory subsets in infected subjects , with significantly higher proportions of central memory ( CCR7+CD45RO+ ) and , predominately , effector memory ( CCR7-CD45RO- ) subsets during acute infection ( Fig 2D ) . Only the effector memory pool remained significantly elevated into the chronic phase . There was no difference in the proportion of the effector cell pool ( CCR7-CD45RO- ) during either phase of infection , although the relative frequency of these cells did appear to be larger as infection progressed ( Fig 2D ) . When we examined the activation state of the memory pool for four RV217 subjects by measuring surface expression of HLA-DR , we found massive levels of activation within the memory CD8+ T cell compartment following HIV infection ( Fig 2E ) , in agreement with recent data from Ndhlovu et al . [48] . To determine if this population of highly activated cells expressed cytolytic molecules directly ex vivo we measured perforin content . We found that almost all HLA-DR+ cells expressed perforin during the acute phase ( Fig 2F ) . In addition , we observed a significantly greater proportion of perforin+ cells in both acute and chronic phases of infection compared to healthy donors ( Fig 2G ) . There was , however , no significant association between the frequency of perforin+ CD8+ T cells and viral load at any time point ( S2B Fig ) . Together , these data show that during acute HIV infection a large proportion of the peripheral CD8+ T cell pool is highly activated and primed to exert cytotoxic effector activity but the absolute magnitude of total cytotoxic CD8+ T cells does not predict set point viral load . We next examined if the large frequency of cytotoxic CD8+ T cells observed during acute HIV infection was consistent across other acute viral infections . We compared the total CD8+ T cell responses of subjects from the RV217 cohort with those of HIV-negative individuals who were vaccinated with attenuated vaccinia virus ( VV ) or attenuated yellow fever virus ( YFV ) -17D , or experimentally infected with a H1N1 strain of influenza virus ( S3A–S3C Fig ) . Vaccination with VV or YFV elicits a robust and highly specific CD8+ T cell response that peaks approximately two weeks after inoculation and is largely resolved by four weeks [46] . The peripheral CD8+ T cell response to influenza is less robust , peaks at 1–2 weeks , and resolves by four weeks post-infection [43] . Consistent with the comparison between healthy donors and acute phase HIV infection ( Fig 2B ) , both the total memory CD8+ T cell pool and the effector memory subset increased significantly from pre- to acute HIV infection ( Figs 3A and S4E ) . There was also a significant increase in the proportion of perforin+ cells over the first thirty days of infection , with almost all ( >90% ) circulating memory CD8+ T cells expressing perforin in some donors ( Fig 3E ) . When we examined the CD8+ T cell responses to in vivo stimulation following vaccination with VV or YFV , or infection with influenza , we did not observe significant changes in the size or distribution of the peripheral memory pool ( Figs 3B–3D and S4 ) . We did find increased levels of activated HLA-DR+ cells in some donors after vaccination with VV and YFV , but frequencies of perforin+ cells remained relatively stable throughout the entire vaccine course ( Figs 3F , 3G and S5A–S5D ) . Only infection with influenza resulted in a slight but significant increase in perforin+ cells at d28 post-infection ( Fig 3H ) . While these models of acute viral infections do have limitations in their use as comparators for our HIV-infected donors ( e . g . different antigen loads , different localizations , and more precise timing of infection ) , overall these data show the dramatic increase in cytotoxic cells that takes place in the peripheral blood of HIV acutely infected subjects is significantly more pronounced compared to live-attenuated vaccination or influenza infection . We next sought to determine if the cytotoxic potential of HIV-specific cells demonstrated similar dynamics to the total memory CD8+ T cell pool during acute to chronic HIV infection . To identify HIV-specific cells we focused on the detection of IFN-γ production and CD107a-marked degranulation following a short-term in vitro stimulation with peptides derived from the HIV-1 Gag and Nef proteins [15 , 48–50] . In agreement with previous studies that evaluated HIV-specific cells longitudinally by functional responses or tetramer staining [49 , 51] , we found no difference in the absolute magnitude of responding cells for either protein over time ( Figs 4A and S6A ) . Consistent with the memory distribution of the total CD8+ T cell pool , Gag-specific cells largely had an effector memory phenotype in the acute phase of infection but became more equally distributed between effector and effector memory subsets for early chronic time points ( Fig 4B and 4C ) . Also in agreement with previous data , cells tended to degranulate more readily than upregulate IFN-γ in the acute phase of infection ( Figs 4D and S6B ) [48 , 49] . The high proportion of degranulating cells suggested that the HIV-specific response might be cytotoxic over the course of infection , as analysis of the total CD8+ T cell pool had indicated . However , degranulation is not an absolute surrogate of cytolytic potential [16 , 52] , nor does it indicate whether the cells will continue to be cytotoxic following the initial granule release [53] . To assess cytotoxic potential more directly , we measured perforin expression levels within the Gag- and Nef- specific cells ( Figs 4E and S6C ) . The majority of cells that responded to direct ex vivo stimulation rapidly upregulated perforin during the earliest time points following infection , suggesting that the early HIV-specific response was likely highly cytotoxic . In contrast to the bulk memory CD8+ T cell pool , however , as acute viremia was resolved there was a rapid loss of perforin expression by both HIV-1 Gag- and Nef-specific CD8+ T cells ( Figs 4F and S6D ) . A large proportion of HIV-specific CD8+ T cells have previously been shown to upregulate β-chemokines independently of degranulation during acute HIV infection [49] . To determine if β-chemokine-producing cells similarly expressed perforin , we assessed expression of MIP-1α by responding cells in a subset of subjects . Inclusion of MIP-1α did not significantly change the overall magnitude of Gag-specific cells detected over time , though it did identify a subset of cells not captured by IFN-γ or CD107a ( S7A–S7C Fig ) . Importantly , the dynamics with which expression of perforin by Gag-specific cells was lost was the same with or without MIP-1α ( S7D and S7E Fig ) . Combined , these data show similarities in the total and Gag-specific CD8+ T cell responses in both differentiation state and cytotoxic potential , suggesting the bulk of activated cells during acute HIV infection could be comprised of HIV-specific CD8+ T cells . Studies in both murine models and humans have strongly linked the transcription factors T-bet and Eomes to the regulation of effector CD8+ T cell differentiation and function , including the expression of perforin [24–26 , 28 , 30 , 42 , 54] . To gain further insight into the evolution of the cytotoxic CD8+ T cell response to HIV we assessed the expression of T-bet and Eomes over the course of infection . For healthy donors , including HIV pre-infection time points , perforin expression was directly associated with T-bet and/or Eomes expression such that the majority of perforin+ cells were either T-bet+Eomes+ or T-bet+Eomes- ( Fig 5A and 5B ) . In contrast , acutely HIV-infected individuals showed marked dissociation between perforin and both T-bet and Eomes resulting in significantly lower proportions of T-bet+Eomes+ and T-bet+Eomes- perforin+ cells ( Fig 5A and 5B ) , and an expansion of perforin+ cells expressing neither T-bet nor Eomes . By the chronic stage these subsets had largely , though incompletely , returned to their normal distributions . When we analyzed T-bet and Eomes expression longitudinally for perforin+ CD8+ T cells within the HIV-infected cohort we found the proportion of T-bet+Eomes+ cells decreased over the first 30 days of infection and T-bet-Eomes- cells increased over the first 60 days before gradually returning to pre-infection levels ( Fig 5C and 5D ) . We have previously shown that the level of T-bet expression within peripheral CD8+ T cells is directly associated with perforin expression , where perforin was found predominantly within T-betHi cells [28] . Consistent with those findings , perforin was most highly associated with a T-betHiEomes+ expression pattern in HIV negative donors and this subset experienced the largest drop during acute HIV ( S8 Fig ) . Despite these shifts in expression patterns that appeared to coincide with the rise and fall plasma viremia , there was no association between the acute frequencies of T-bet or Eomes subsets and acute or set point viral loads ( S9A–S9D and S10A–S10D Figs ) . However , frequencies of T-bet+ and T-bet-Eomes- CD8+ T cells at set point time points were inversely or directly associated with set point viral load , respectively ( S9A and S9D Fig ) . To determine if the dissociation between perforin , T-bet , and Eomes was unique to HIV , we examined T-bet and Eomes expression within total perforin+ cells following YFV and VV vaccination . While we found almost no dissociation for YFV , there was a transient dissociation following vaccination with vaccinia , although not to the same extent as observed during acute HIV ( S11A and S11B Fig ) . We next examined expression of T-bet and Eomes within HLA-DR+ cells throughout the different vaccine courses . As noted above , during acute HIV infection the vast majority of HLA-DR+ cells are also perforin+ ( Fig 2E ) ; thus , it was unsurprising to find that perforin+ and HLA-DR+ cells showed almost identical dynamics in the loss of T-bet and Eomes expression for HIV ( S11C Fig ) . Similarly , for both YFV and VV , activated cells showed a transient increase in the frequency of T-bet-Eomes- cells at day 14 post-vaccination . Together these data suggest that the transient expansion of highly activated bulk effector CD8+ T cells during acute viral infection in humans may not require expression and/or maintenance of T-bet and Eomes . To determine if the transient loss of T-bet and Eomes within the bulk activated CD8+ T cell memory pool during acute HIV infection extended to HIV-specific CD8+ T cells , we assessed expression of these transcription factors in Gag-specific CD8+ T cells . In marked contrast to the highly activated bulk CD8+ T cell effector population during acute HIV infection , HIV-specific CD8+ T cells expressed T-bet and/or Eomes at the earliest detectable time point and throughout the course of infection ( Fig 6A–6C ) . This indicates that despite their phenotypic similarities total and HIV-specific CD8+ T cells may be primed quite differently during acute infection and raises the possibility that the majority of expanded effector CD8+ T cells in early HIV infection may not be specific for HIV . We next examined whether loss of perforin expression was related to changes in the level of T-bet expression during early HIV infection . Interestingly , the distribution of T-bet within Gag-specific CD8+ T cells changed over time from acute to chronic infection ( Fig 6D ) . In the acute phase , responding cells were equally distributed between T-betHiEomes+ and T-betLoEomes+ expression patterns , which during the chronic phase began to be dominated by T-betLoEomes+ cells ( Fig 6D ) . Furthermore , T-betHiEomes+ HIV-specific CD8+ T cells continued to express perforin as infection progressed , whereas T-betLoEomes+ cells gradually lost perforin expression over time ( Fig 6E and 6F ) . Finally , in contrast to the recent findings by Ndholuvu , et al . [48] , we did not find the magnitude , proportion perforin+ , or any T-bet- or Eomes-expressing subset of responding HIV-1 Gag-specific CD8+ T cells to be predictive of peak or set point viral load ( S12 and S13 Figs ) . Despite this , our data suggest that in the earliest phase of infection , HIV-specific CD8+ T cells have both the transcriptional and functional properties associated with long-term control of HIV replication [16 , 28] , and that the inability to durably maintain high-level T-bet expression contributes to a qualitatively inferior response as infection progresses .
Mechanisms underlying the inability of CD8+ T cells to fully control HIV replication have remained unclear . Failure of antiviral immunity has been attributed in part to qualitative defects in total and HIV-specific CD8+ T cells [15 , 16 , 20 , 55 , 56] . However , the dysfunction observed within the CD8+ T cell pool has largely been defined in the context of chronic infection when the success or failure of the presumed response has already been determined . The question of whether CD8+ T cells in progressive infection were intrinsically less functional from the outset or if dysfunction arose over time has remained unanswered . To address this issue , we assessed the longitudinal CD8+ T cell responses of a diverse cohort of individuals experiencing acute/early HIV infection . We show that acute HIV infection elicits a robust cytotoxic CD8+ T cell response characterized by cells that express the cytolytic effector molecule perforin and the effector-associated transcription factors T-bet and Eomes . Importantly , the quality of the response quickly waned following the resolution of acute viremia , with a significant decrease in perforin expression by HIV-specific CD8+ T cells that was at least partially accounted for by a shift from T-betHiEomes+ cells to T-betLoEomes+ cells . The attenuation of the cytolytic response may help explain the failure of CD8+ T cells to control HIV replication in the long-term . It is well documented that CD8+ T cell responses are elicited early in HIV infection and are associated with control of viral replication [1 , 2 , 48 , 57] . Some of the strongest evidence of the CD8+ T cell-mediated immunologic pressure exerted during this period is the rapid emergence of viral escape mutations within known CD8+ T cell epitopes [4 , 6 , 9] . We found that HIV-specific cells had high cytotoxic potential at the earliest time points following HIV infection , but rapidly lost this function as disease progressed . This suggests a mechanism through which CD8+ T cells may exert a strong direct selective pressure on the virus resulting in the rapid selection of escape variants early in infection that ultimately have a reduced capacity to stimulate cytolytic CD8+ T cell responses [6 , 9 , 58 , 59] . It should be noted that whereas perforin expression was lost over time almost all HIV-specific responding cells continued to produce MIP-1α . Thus , while cytotoxic CD8+ T cells play an important role in the resolution of acute viremia , as they lose their ability to express perforin they may be able to keep the virus partially in check through a combination of the remaining cytotoxic response and non-cytotoxic inhibitory effects exerted via the continued expression of β-chemokines or other non-cytolytic mechanisms [60] . This would be consistent with models suggesting CD8+ T cell cytotoxic mechanisms do not account for the entirety of CD8+ T cell-mediated viral suppression during chronic progressive SIV infection [61 , 62] . It remains unclear if maintenance of perforin expression following acute infection would further enhance the level of control over viral replication CD8+ T cells provide as we would predict it should based on studies of CD8+ T cell responses in the chronic phase of infection [16 , 18 , 19] . Unfortunately , we were unable to find any direct associations between HIV-1 Gag-specific perforin , T-bet , or Eomes expression and the level of plasma viremia or CD4+ T cell numbers . T-bet and Eomes are important regulators of effector CD8+ T cell differentiation and function for both mice and humans [24–26 , 28 , 30 , 31 , 33 , 34 , 42 , 54] . Expression patterns of these transcription factors have been described for CD8+ T cells in the context of various human viral infections , including CMV , EBV , HBV , HCV , HIV , and TBEV [28 , 32 , 34 , 42 , 54 , 63–66] . These studies demonstrated a high degree of variability in the relative levels of T-bet and Eomes expressed by virus-specific CD8+ T cells depending on time from infection , whether the infection was controlled , and tissue localization . CMV-specific cells express T-bet and Eomes during both acute and chronic phases of infection , but control of viral replication in the acute phase is associated with a higher ratio of T-bet+ versus Eomes+ cells [64 , 66] . EBV- and TBEV-specific cells also express T-bet and Eomes during the earliest phase of their respective infections , but EBV-specific cells lose expression of both during convalescence whereas TBEV-specific cells retain T-bet expression and show a gradual reduction in Eomes [42 , 63] . HCV-specific cells are T-bet+ in acute/resolving HCV infection and T-bet-Eomes- during acute/non-resolving infection . Post-acute phase , HCV-specific cells in the peripheral blood are T-bet-Eomes- for both resolved and non-resolved HCV infection , but T-bet+ within the livers of subjects with resolved infection and Eomes+ in livers of chronically infected subjects [34 , 65] . Together , these results suggest expression of T-bet during the acute phase is a critical determinant of viral infection outcome . The differential outcomes associated with Eomes were also reflective of the relative expression level of T-bet , suggesting Eomes may not be as important for the resolution of acute viremia . Rather , Eomes expression may determine whether antigen-specific cells are fated to form a stable memory pool or become exhausted subsequent to the acute phase , dependent on whether or not the infection is ultimately cleared [34 , 67] . Similar associations between T-bet , Eomes , and outcome have been demonstrated in chronic HIV infection . In this context , a high level of T-bet expression was associated with greater overall functionality of HIV-specific CD8+ T cells , including cytotoxic potential , and relative control of viral replication , whereas low T-bet levels and continued Eomes expression has been associated with lower overall functionality and persistent viremia [28 , 32] . Our data show that HIV-specific cells have high cytotoxic potential during acute infection , but lose the ability to express or rapidly upregulate perforin in chronic infection . This loss of cytotoxic potential over time can at least partially be explained by a change in the relative expression levels of T-bet and Eomes: HIV-specific cells were equally T-betHiEomes+ and T-betLoEomes+ during acute infection and both subsets efficiently upregulated perforin initially but the proportion of T-betLoEomes+ cells increased significantly as infection progressed and cells with this phenotype had an inferior capacity to express perforin compared to T-betHiEomes+ cells . The expression of perforin by either phenotype during acute infection may be reflective of the high degree of inflammation and activation during this phase , a differential role for Eomes at different stages of infection , and/or the result of additional transcription factors not assessed here . Whatever the case may be , T-betHiEomes+ HIV-specific CD8+ T cells retain the ability to upregulate perforin following resolution of acute viremia and this subset declines during chronic progressive infection . Recent data from Ndhlovu et al . suggests HIV infection elicits a massive antigen-specific CD8+ T cell response with limited bystander activation [48] . Similar observations have been reported after vaccination with vaccinia and yellow fever virus [46] . The similarities in differentiation state , activation , and immediate cytotoxic potential between total peripheral memory and Gag-specific cells reported here support the idea of a robust and specific response to HIV infection . However , we found a significant discrepancy between transcriptional control of HIV-specific CD8+ T cells versus the bulk activated perforin+ memory CD8+ T cell population . The degree to which these differences reflect a true lack of specificity , dysfunction on the part of the bulk activated cells , an inability to identify an appropriate functional marker , or an attempt by the host to mitigate immune-mediated pathology remains unclear . It is likely there area many more circulating HIV-specific CD8+ T cells than indicated by our findings using in vitro stimulation with only two HIV-1 proteins and a limited number of functional parameters to identify responding cells . However , it should be noted that CD8+ T cell bystander activation has been reported during acute HIV and EBV infection in humans and it is possible at least a subset of CD8+ T cells are activated non-specifically in our cohort [44 , 68] . T cell receptor stimulation is required for upregulation of T-bet [69] , but a large proportion of bulk activated perforin+ cells during acute HIV infection appear to express neither T-bet nor Eomes whereas all Gag-specific cells expressed one or the other . In addition , perforin can be upregulated in the absence of direct antigenic stimulation via exposure to IFN-α [70] , levels of which are highly elevated during acute HIV infection [71] . Thus , the difference in T-bet and Eomes expression we observed between bulk perforin+ and responding HIV-specific CD8+ T cells raises the possibility that a significant number of bystander-activated cells are being induced in response to HIV infection . Alternatively , given the association between activation and the size of the T-bet-Eomes- pool across infections with vaccinia , yellow fever , and HIV , the absence of T-bet and Eomes expression in the bulk perforin+ CD8+ T cell pool may be a characteristic of the contraction phase that typically follows the initial CD8+ T cell response . This would be consistent with the pro-apoptotic phenotype of the majority of cells following peak HIV viremia and the timing of our samples [48] . Whether HIV-specific or bystander , the lack of T-bet and Eomes expression by these cells suggests they would be unable to sustain perforin expression upon encountering infected target cells . This may in part explain the inability of bulk peripheral CD8+ T cells from acutely HIV infected individuals to efficiently inhibit viral replication in vitro and further suggests they would not make a meaningful contribution to long-term control of viral replication in vivo [72 , 73] . These data show how the peripheral CD8+ T cell response to HIV evolves over the course of progressive infection . HIV-specific CD8+ T cells are able to upregulate perforin and T-bet initially but begin to lose this capacity soon after peak viremia , demonstrating for the first time that there is not an initial intrinsic inability of HIV-specific CD8+ T cells to upregulate these molecules . It remains unclear how or if these responses differ from those of CD8+ T cells from subjects who go on to spontaneously control viral replication to very low levels in the chronic phase . While we did find frequencies of T-bet+ and T-bet-Eomes- total memory CD8+ T cells at set point time points were inversely or directly associated with set point viral load , respectively , we did not find any associations between viral load and the size of the total peripheral perforin+ pool or the magnitude or cytotoxic potential of HIV-1 Gag-specific cells at any time point . Nor did we find any subset of total memory or Gag-specific cells to be predictive of set point viral load for this group of subjects , possibly due to the limited number of very early time points and relatively narrow range of viral loads at set point . However , the fact that the initial phenotype of HIV-specific cells is similar to that associated with control during the chronic phase of infection suggests induction and maintenance of cells capable of upregulating high levels of T-bet and perforin could lead to subsequent control . Eliciting HIV-specific cells with these characteristics might serve as an important target for vaccination or therapeutic modalities seeking to fully control early viral replication or eradicate the chronic viral reservoir .
Blood specimens were acquired with the written informed consent of all study participants and with the approval of the institutional review board at each respective institution where patient materials were collected: University of Pennsylvania ( IRB# 809316 ) , McGill University Health Centre ( REB# GEN-10-084 ) , Human Subjects Protection Branch ( RV217/WRAIR#1373 ) , Kenya Medical Research Council ( KEMRI/RES/7/3/1 ) , The United Republic of Tanzania Ministry of Health and Social Welfare ( MRH/R . 10/18/VOLL . VI/85 ) , Tanzanian National Institute for Medical Research ( NIMR/HQ/R . 8aVol . 1/2013 ) , Royal Thai Army Medical Department ( IRBRTA 1810/2558 ) , Uganda National Council for Science and Technology–National HIV/AIDS Research Committee ( ARC 084 ) , Uganda National Council of Science and Technology ( HS 688 ) , East London and City and the Southwest and Southwest Hampshire Ethics Review Committees , Duke University ( IRB# Pro00006579 and IRB# Pro00007558 ) , Emory University ( IRB# 00009560 ) , Oregon Health and Science University ( IRB# 2470 and IRB# 2832 ) . The study was conducted in accordance with the principles expressed in the Declaration of Helsinki . Eleven HIV-1 acutely infected participants were enrolled as part of the RV217 Early Capture HIV cohort , nine were enrolled in the CHAVI 001 acute infection cohort , and twelve were enrolled in the Montreal Primary Infection cohort . Participant demographics are summarized in S1 Table . Acute HIV-1 infection was determined by measuring plasma HIV RNA content and HIV-specific antibodies using ELISA and Western blot . Fiebig staging [74] immediately following the first positive visit or at the screening visit was used to characterize the timing of infection for RV217 and CHAVI participants , respectively . The only exception was RV217 donor 40067 for which the estimated date of infection was taken as the midpoint between the last negative and first positive visit . For the Montreal Primary Infection cohort the following guidelines proposed by the Acute HIV Infection Early Disease Research Program sponsored by the National Institutes of Health were used to estimate the date of infection: the date of a positive HIV RNA test or p24 antigen assay available on the same day as a negative HIV enzyme immunoassay ( EIA ) test minus 14 days; or the date of the first intermediate Western blot minus 35 days . In addition , information obtained from questionnaires addressing the timing of high-risk behavior for HIV transmission was taken into account in assigning a date of infection when consistent with biological tests . The timing of visits relative to estimated date of infection for all acutely HIV infected donors used in this study is provided in Fig 1A . Study participants were antiretroviral therapy naïve at all time points analyzed , consistent with the standard of care at the time of study . HIV-1 viral loads were measured using the Abbot Real-Time HIV-1 assay ( RV217; Abbot Laboratories , Abbott Park , IL ) , COBAS AMPLICOR HIV-1 monitor test , version 1 . 5 ( CHAVI; Roche Diagnostics , Branchburg , NJ ) , or the UltraDirect Monitor assay ( Montreal; Roche Diagnostics , Branchburg , NJ ) . HIV set point viral loads were defined as the average of all viral load measurements between 90 and 365 days post-infection in the absence of therapy with the requirement for at least two viral load measurements during this period . For HIV-negative cohorts , volunteers were administered the live-attenuated YFV-17D vaccine ( YF-Vax , Sanofi Pasteur ) , the live vaccinia smallpox vaccine ( Dryvax , Wyeth Laboratories ) , or challenged with influenza A/Brisbane/59/07 . YF-Vax was administered subcutaneously in the arm , Dryvax was administered by scarification of the upper arm with three pricks of a bifurcated needle , and influenza A virus was administered intra-nasally . Peripheral blood mononuclear cells ( PBMCs ) from pre-vaccination or pre-infection time points were available for most donors along with several time points post-vaccination or infection ( S3A–S3C Fig ) . Pre-infection time points from all cohorts , including RV217 participants , along with PBMCs obtained from fifteen healthy human subjects through the University of Pennsylvania’s Human Immunology Core were combined for a total of 41 healthy donor data points . Potential T cell epitope ( PTE ) peptides corresponding to the HIV-1 Gag and Nef proteins were obtained from the NIH AIDS Reagent Program ( NIH , Bethesda , Maryland , USA ) . PTE peptides are 15 amino acids in length and contain naturally occurring 9 amino acid sequences that are potential T cell determinants embedded in the sequences of circulating HIV-1 strains worldwide , including subtypes A , B , C , D and circulating recombinant forms ( CRF ) . As such , these peptide pools provided the coverage necessary for the T cell stimulation assays performed in this study given the broad geographical distribution of our study participants and diversity of infecting viruses ( S1 Table ) . Lyophilized peptides were dissolved in dimethyl sulfoxide ( DMSO , Sigma-Aldrich , St Louis/Missouri , USA ) , combined into two pools at 400 μg/ml , and stored at -20°C . Cryopreserved PBMCs were thawed and rested overnight at 2x106 cells/ml in RPMI medium supplemented with 10% fetal bovine serum , 2 mM L-glutamine , 100 U/ml penicillin , and 100 mg/ml streptomycin . Cell viability was checked both immediately after thawing and after overnight rest by trypan blue exclusion . Costimulatory antibodies ( anti-CD28 and anti-CD49d , 1 μg/mL each; BD Biosciences ) and pre-titrated fluorophore conjugated anti-CD107a was included at the start of all stimulations . PBMCs were incubated for 1 hour at 37°C and 5% CO2 prior to the addition of monensin ( 1 μg/mL; BD Biosciences ) and brefeldin A ( 10 μg/mL; Sigma-Aldrich ) followed by an additional 5 hour incubation at 37°C and 5% CO2 . For peptide stimulations , peptides from the two Gag PTE pools were added to a single tube of cells such that each individual peptide was at a final concentration of 1 μg/ml . As a negative control , DMSO was added to the cells at an equivalent concentration to the one used for peptide stimulation . Antibodies for surface staining included CCR7 APC-Cy7 ( clone G043H7; Biolegend ) , CCR7 APC-eFluor780 ( clone 3D12; eBioscience ) , CD4 PE-Cy5 . 5 ( clone S3 . 5; Invitrogen ) , CD8 BV711 ( clone RPA-T8; Biolegend ) , CD8 Qdot 605 ( clone 3B5; Invitrogen ) , CD14 BV510 ( clone M5E2; Biolegend ) , CD14 Pacific Blue ( clone M5E2; custom ) , CD14 PE-Cy5 ( clone 61D3; Abcam ) , CD14 PE-Cy7 ( clone HCD14; Biolegend ) , CD16 Pacific Blue ( clone 3G8; custom ) , CD16 PE-Cy5 ( clone 3G8; Biolegend ) , CD16 PE-Cy7 ( clone 3G8; Biolegend ) , CD19 BV510 ( clone HIB19; Biolegend ) , CD19 Pacific Blue ( clone HIB19; custom ) , CD19 PE-Cy5 ( clone HIB19; Biolegend ) , CD19 PE-Cy7 ( clone HIB19; Invitrogen ) , CD45RO ECD ( clone UCHL1; Beckman Coulter ) , CD45RO PE-CF594 ( clone UCHL1; BD Biosciences ) , CD107a PE-Cy5 ( clone eBioH4A3; eBioscience ) , CD107a PE-Cy7 ( clone H4A3; Biolegend ) , and HLA-DR Pacific Blue ( clone LN3; Invitrogen ) . Antibodies for intracellular staining included CD3 BV570 ( clone UCHT1; Biolegend ) , CD3 BV650 ( clone OKT3; Biolegend ) , CD3 Qdot 585 ( clone OKT3; custom ) , CD3 Qdot 650 ( clone S4 . 1; Invitrogen ) , Eomes Alexa 647 ( WD1928; eBioscience ) , Eomes eFluor 660 ( WD1928; eBioscience ) , IFN-γ Alexa 700 ( clone B27; Invitrogen ) , Perforin BV421 ( clone B-D48 , Biolegend ) , Perforin Pacific Blue ( clone B-D48; custom ) , Perforin PE ( clone B-D48 , Cell Sciences ) , T-bet FITC ( clone 4B10; Biolegend ) , and T-bet PE ( clone 4B10; eBioscience ) . At the end of the stimulations , cells were washed once with PBS prior to be being stained for CCR7 expression for 15 min at 37°C in the dark . Cells were then stained for viability with aqua amine-reactive viability dye ( Invitrogen ) for 10 min at room temperature in the dark followed by addition of a cocktail of antibodies to stain for surface markers for an additional 20 min . The cells were washed with PBS containing 0 . 1% sodium azide and 1% BSA , fixed and permeabilized using a Cytofix/Cytoperm kit ( BD Biosciences ) , and stained with a cocktail of antibodies against intracellular markers for 1 h at room temperature in the dark . The cells were washed once with Perm Wash buffer ( BD Biosciences ) and fixed with PBS containing 1% paraformaldehyde . Fixed cells were stored at 4°C in the dark until acquisition . Antibody capture beads ( BD Biosciences ) were used to prepare individual compensation controls for each antibody used in the experiment . ArC Amine Reactive beads ( ThermoFisher Scientific ) were used to generate a singly stained compensation control for the aqua amine-reactive viability dye . For each stimulation condition , a minimum of 250 , 000 total events were acquired using a modified LSRII ( BD Immunocytometry Systems ) . Data analysis was performed using FlowJo ( TreeStar ) software . Gating strategy is provided in the supplementary materials ( S1 Fig ) . Reported antigen-specific data have been corrected for background based on the negative ( no peptide ) control , and only responses with a total frequency twice the negative control and above 0 . 01% of total memory CD8+ T cells ( after background subtraction ) were considered to be positive responses . By analyzing the data in this way , we examined cytolytic protein production resulting from antigen-specific stimulation and ensured that its expression was considered only within responding CD8+ T cells expressing at least one other functional parameter . Whereas IFN-γ , CD107a , and MIP-1α were used to identify antigen-specific CD8+ T cells for some donors , only IFN-γ and CD107a were used consistently for all donors and figures depicting antigen-specific data were derived from analysis of cells expressing these two markers unless otherwise noted . All statistical analysis was performed using Stata ( version 14 . 0 ) . Graphs were generated using Stata or GraphPad Prism ( version 5 . 0a ) . Generalized estimating equations ( GEEs ) with robust variances were used to test for changes while adjusting for repeated measurements on the same individuals [75] . In instances where many values were at 100% a random-effects tobit regression model was used to do a combined analysis of the percent of data points at 100% versus differences in values for data points below 100% . P values were Holm-adjusted for multiple comparisons . Bars represent approximations of the means generated by the models . Lowess smoothers were used to represent the mean over time for longitudinal data . Correlations were determined using Spearman’s rank correlation test ( non-parametric; two-tailed ) .
|
Previous studies have demonstrated that HIV-specific CD8+ T cells are critical for the initial control of HIV infection . However , this control is typically incomplete , being able to neither clear infection nor maintain plasma viremia below undetectable levels . Mounting evidence has implicated CD8+ T cell cytotoxic capacity as a critical component of the HIV-specific response associated with spontaneous long-term control of HIV replication . CD8+ T cell cytotoxic responses are largely absent in the vast majority of HIV chronically infected individuals and it is unclear when or why this functionality is lost . In this study we show that HIV-specific CD8+ T cells readily express the cytolytic protein perforin during the acute phase of chronic progressive HIV infection but rapidly lose the ability to upregulate this molecule following resolution of peak viremia . Maintenance of perforin expression by HIV-specific CD8+ T cells appears to be associated with the expression level of the transcription factor T-bet , but not with the T-bet paralogue , Eomes . These findings further delineate qualitative attributes of CD8+ T cell-mediated immunity that may serve as targets for future HIV vaccine and therapeutic research .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"blood",
"cells",
"hiv",
"infections",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"viral",
"transmission",
"and",
"infection",
"pathogens",
"immunology",
"microbiology",
"cloning",
"neuroscience",
"learning",
"and",
"memory",
"retroviruses",
"viruses",
"immunodeficiency",
"viruses",
"rna",
"viruses",
"cognition",
"memory",
"cytotoxic",
"t",
"cells",
"molecular",
"biology",
"techniques",
"viral",
"load",
"research",
"and",
"analysis",
"methods",
"infectious",
"diseases",
"white",
"blood",
"cells",
"memory",
"t",
"cells",
"animal",
"cells",
"medical",
"microbiology",
"hiv",
"t",
"cells",
"microbial",
"pathogens",
"molecular",
"biology",
"cell",
"biology",
"virology",
"viral",
"pathogens",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"viral",
"diseases",
"lentivirus",
"cognitive",
"science",
"organisms"
] |
2016
|
Temporal Dynamics of CD8+ T Cell Effector Responses during Primary HIV Infection
|
Recent studies suggest that some monoclonal antibodies ( mAbs ) specific for ebolavirus glycoprotein ( GP ) can protect experimental animals against infections . Most mAbs isolated from ebolavirus survivors appeared to target the glycan cap or the stalk region of the viral GP , which is the envelope protein and the only antigen inducing virus-neutralizing antibody response . Some of the mAbs were demonstrated to be protective in vivo . Here , a panel of mAbs from four individual survivors of ebolavirus infection that target the glycan cap or stem region were selected for investigation of the mechanisms of their antiviral effect . Comparative characterization of the inhibiting effects on multiple steps of viral replication was performed , including attachment , post-attachment , entry , binding at low pH , post-cleavage neutralization of virions , viral trafficking to endosomes , cell-to-cell transmission , viral egress , and inhibition when added early at various time points post-infection . In addition , Fc-domain related properties were characterized , including activation and degranulation of NK cells , antibody-dependent cellular phagocytosis and glycan content . The two groups of mAbs ( glycan cap versus stem ) demonstrated very different profiles of activities suggesting usage of mAbs with different epitope specificity could coordinate inhibition of multiple steps of filovirus infection through Fab- and Fc-mediated mechanisms , and provide a reliable therapeutic approach .
Filoviruses are enveloped , filamentous-like viruses with non-segmented RNA genome of negative polarity . The Ebolavirus genus of the Filoviridae family includes five species: Ebola ( EBOV ) , Sudan ( SUDV ) , Bundibugyo ( BDBV ) , Taï Forest ( TAFV ) and Reston ( RESTV ) viruses . Most of these viruses are responsible for highly lethal disease outbreaks , for example the occurrence of 11 , 323 human fatalities during the 2013–2016 EBOV epidemic in West Africa [1 , 2] . Despite intense international collaborative efforts , there is still no licensed therapeutic available against filovirus disease . GP is the sole ebolavirus envelope protein responsible for cell entry and , hence , serves as the primary target for antibody-based therapies and as antigen for vaccine development [3] . The primary nucleotide sequence of the GP gene encodes soluble glycoprotein ( sGP ) , which shares its 295 N-terminal amino acid residues with GP , whereas GP mRNA synthesis requires the insertion of an extra adenosine into the nascent mRNA via stuttering of the EBOV RNA-dependent RNA polymerase over the transcriptional editing site [4] . The mature GP at the surface of nascent virions represents a 450 kDa trimer assembled from GP1/GP2 heterodimers [3] . The GP1 subunit mediates cellular attachment of viral particles and includes base domain interacting with the GP2 subunit , and a chalice-like structure formed by the receptor-binding domain ( RBD ) , glycan cap and heavily N- and O-glycosylated mucin-like domain ( MLD ) . The RBD is sequestered in the chalice bowl , whereas the glycan cap and MLD are exposed and covered by a thick glycan layer that likely shields much of GP from effective humoral immune recognition [5 , 6] . The GP2 subunit forms a GP stalk containing the hydrophobic internal fusion loop ( IFL ) , two heptad repeats ( HR1 and HR2 ) , the membrane-proximal external region ( MPER ) , the transmembrane anchor and the short cytoplasmic domain . This subunit is responsible for fusion of the viral and host cell membranes during the entry . EBOV attachment to the cell surface occurs via two types of low affinity interactions . First , using a set of N- and O-linked glycans on the MLD and the glycan cap of GP1 , virus can bind to multiple C-type lectins . Second , EBOV uses phosphatidylserine molecules incorporated into viral envelope to bind TIM/TAM receptors ( reviewed in reference [7] ) . After adherence , virions internalize to the cell by macropinocytosis , and subsequently traffic through the labyrinth of endosomal compartments , where critical pH-dependent GP priming by cathepsin proteases takes place . The consecutive processing of GP by cathepsins L and B results in the excision of most of the GP1 subunit , which includes the glycan cap and MLD , and exposure of the RBD for interaction with the intracellular filovirus receptor , the cholesterol transporter protein NPC1 [8–11] . The first human EBOV neutralizing mAb , KZ52 , was generated from RNA isolated from the bone marrow of a survivor of natural infection [12] . This mAb protected guinea pigs from lethal EBOV challenge [13] , but failed to protect non-human primates ( NHPs ) [14] . The feasibility of post-exposure prophylaxis with antibodies in monkeys was demonstrated six years ago with total IgG purified from convalescent serum of macaques [15] . Several mAb cocktails that protect NHPs from EBOV infection have been developed subsequently: MB-003 ( human or human/mouse chimeric mAbs c13C6 , h13F6 and c6D8 ) , ZMAb ( murine mAbs m1H3 , m2G4 and m4G7 ) and ZMapp ( human/mouse chimeric mAbs c13C6 , c2G4 and c4G7 ) [16 , 17] . The latter cocktail , which showed a beneficial effect , however , failed to demonstrate the pre-specified statistical threshold for efficacy in a clinical trial performed during the West Africa epidemic [18] . Human mAbs from survivors of natural ebolavirus infection , rather than antibodies raised in experimentally vaccinated or infected animals , are preferable for the development of therapeutics against filovirus infections . Such antibodies have a full compatibility of Fc fragments with the receptors on human immune cells , which is expected to make them more effective due to Fc-mediated protective mechanisms . While several published studies demonstrate binding of filovirus mAbs from human survivors to GP at the atomic level [5 , 19–24] , none of them are characterized for the ability to affect multiple steps of viral replication . Here , we present a comprehensive comparative study of Fab- and Fc-mediated biological functions of a panel of ebolavirus mAbs from human survivors [20] targeting epitopes in the GP glycan cap and stalk region . The results indicate that both types of mAbs interfere with and target different steps of viral replication , including virus entry , egress , cell-to-cell transmission , secondary infection and facilitate destruction of infected cells through antibody-dependent cellular cytotoxicity ( ADCC ) and antibody-dependent cellular phagocytosis ( ADCP ) mechanisms . However , important differences between the two groups also were observed , suggesting complementary effects of various antibodies generated during natural filovirus infections .
In previous work , we isolated and characterized multiple mAbs from the blood of human survivors of natural BDBV infection [20] . To study mechanisms of inhibition of filovirus replication by antibodies , we selected a panel of mAbs from four donors , with differing virus neutralization properties and affinity to GPs of EBOV , BDBV and SUDV: BDBV52 , BDBV270 , BDBV41 , BDBV289 , BDBV259 , BDBV317 and BDBV223 ( Fig 1 ) . Identification of epitopes demonstrated that most of the mAbs can be grouped into those recognizing two major antigenic sites: those specific for glycan cap and those specific for stem [20 , 25 , 26] . To test whether any of the mAbs inhibited attachment , we incubated BDBV virus-like particles ( VLPs ) and mAbs at 37°C for 1 hour , added the mixtures to Vero-E6 cell culture monolayers in chambered slides , and incubated on ice for 1 hour . Then , cells were fixed , and cell-bound VLPs were immunostained . Confocal microscopic analysis of cell monolayers demonstrated a strong binding inhibition only by mAb BDBV289 ( Figs 2A and S1 ) . Agreeing with the confocal microscopy results , flow cytometric analysis demonstrated 2-fold inhibition of viral binding by BDBV289 , some enhancement of binding by BDBV52 , BDBV270 , BDBV259 , and strong enhancement for BDBV223 ( Figs 2B and S2 ) . We did not observe any enhancement of viral binding in our confocal microscopic assay for the BDBV223-treated samples . The reasons for this are unclear and may be related to the use of attached cells for the confocal microscopy and suspension cells for flow cytometry . An irrelevant human mAb 2D22 of the IgG1 isotype , specific to dengue virus envelope protein in the dimeric structure [27] , was used as a control . A significant difference was not observed between the 2D22 and the no-mAb groups ( S3 Fig ) . For the post-attachment inhibition assay , BDBV was adsorbed first on Vero-E6 cell culture monolayers for 20 min at 4ºC . Then , mAbs were added , incubated for 20 min at 4ºC , and viral plaques were developed at 37ºC . The BDBV223 mAb strongly reduced plaque numbers , suggesting that MPER-targeting mAbs can effectively block post-attachment steps of virus replication ( Fig 2C ) . The inhibiting effect of the other MPER-specific mAb , BDBV317 , was comparable to that of BDBV289 and BDBV259 in this assay . Only marginal post-attachment inhibition was demonstrated for BDBV41 and BDBV270 mAbs from the glycan cap-targeting group . To assess the total impact of mAbs on inhibition of virus entry ( binding and post-attachment steps ) , we used a chimeric replication-competent EBOV in which GP was replaced with its counterpart from BDBV and that expresses eGFP from an added transcriptional cassette to visualize infected cells ( EBOV/BDBV-GP ) [28] . EBOV/BDBV-GP was incubated with mAbs for 1 hour and adsorbed on Vero-E6 cell culture monolayers for 40 min at 4ºC . Then , cells were incubated for 24 hours at 37ºC , and the percentages of infected eGFP+ cells were determined by flow cytometry ( Figs 2D and S4A ) . As expected , MPER-specific mAbs completely abolished virus entry in cells . High levels of inhibition also were demonstrated by BDBV270 , BDBV41 , BDBV289 and BDBV259 mAbs . Unexpectedly , the non-neutralizing BDBV52 mAb slightly increased virus entry into the cells . To investigate the effect of mAbs on intracellular steps of virus life cycle , trafficking of mAb-treated VLPs through the cell organelle network was analyzed by confocal microscopy . BDBV or EBOV VLPs were mixed with mAbs , placed on Vero-E6 cell culture monolayers , incubated for 30 or 60 min and immunostained for EBOV VLPs and for endosomal markers . Unexpectedly , BDBV259 and BDBV317 , but not the other mAbs , caused accumulation of VLPs in late endosomes , as evidenced by GP/Rab7 co-localization ( Figs 2E , S5 , S6 and S7 ) . However , the effect of BDBV317 was relatively short , as the co-localization disappeared after one hour of incubation ( S5 and S6 Figs ) , probably suggesting instability of BDBV317/GP complexes in the acidic pH of endosomes . When treated with BDBV223 , but not the other mAbs , VLPs were found to be co-localized with the lysosomal-associated membrane protein 1 ( LAMP-1 ) marker of lysosomes as early as 30 min after infection ( Figs 2E and S6 ) , which was still observed at 60 min ( S5 and S6 Figs ) . We next tested binding of mAbs to BDBV GP at low or neutral pH by ELISA ( Fig 2F ) . Binding of BDBV317 mAb was impaired at low pH compared to neutral pH , consistent with the short duration of BDBV317/Rab7 co-localization . In contrast , binding of BDBV259 was 1 . 5 times higher at low pH compared to neutral pH , whereas the difference for BDBV223 mAb was as much as 3 . 6 times higher . Hence , we propose that an acidic pH environment stabilizes BDBV223/GP complexes , allowing this antibody to retain viral particles inside lysosomal compartments and prevent nucleocapsid entry into the cytoplasm . We hypothesized that the accumulation in acidic compartments observed for BDBV223 , BDBV317 and BDBV259 mAbs was caused by inhibition of binding of GP to NPC1 in the late endosomes . To test this hypothesis , we developed Förster resonance energy transfer ( FRET ) analysis using NPC1 fused to red fluorescent protein ( NPC1-RFP ) and GP immunostained with AlexaFluor 647 . Vero-E6 cells were transfected with NPC1-RFP-expressing plasmid and incubated overnight . A modified EBOV/BDBV-GP that does not express eGFP ( EBOV/BDBV-GP_no eGFP ) was pre-incubated with selected mAbs for 60 min at 37ºC . NPC1-RFP-transfected Vero-E6 cell culture monolayers then were inoculated with virus-mAb complexes at an MOI of 10 PFU/cell for 2 hours , fixed , and GP was immunostained . FRET analysis was performed by scanning confocal microscopy; the NPC1-GP interaction was quantified by changes in FRET efficiency when compared with virus in the absence of mAbs ( S8 Fig ) . We analyzed BDBV223 , BDBV259 and BDBV317 in comparison with the glycan cap-specific mAb BDBV289 as a negative control , since BDBV289 inhibits attachment and entry ( Fig 2A and 2B ) and is expected not to reach endosomes . The FRET efficiencies for virus samples treated with mAbs were equivalent to those without mAb , suggesting that the tested antibodies did not affect binding of the virus to NPC1 . We also compared the numbers of FRET-positive events , and observed a dramatic increase with BDBV223 and a more modest increase with BDBV259 . The increase in the number of virus-associated events was consistent with the increased trapping of the VLPs treated with these two mAbs in endosomes ( Fig 2E ) . Notably , BDBV223 and BDBV259 , but not the other mAbs tested , were found to bind to GP at low pH ( Fig 2F ) . As described above , GP is processed by cysteine proteases , cathepsins B and L , resulting in the removal of glycan cap and MLD from GP1 subunit followed by interaction of the exposed RBD with the C-loop of NPC1 . Treatment of EBOV GP with the bacterial metalloproteinase thermolysin also results in deletion of the glycan cap and MLD , thus mimicking endosomal proteolysis of GP mediated by cathepsins [29 , 30] . To test if selected mAbs interfere with late stages of virus cell entry by interacting with GP after its cleavage , we treated sucrose gradient-purified replication-competent vesicular stomatitis virus enveloped with BDBV GP ( VSV/BDBV-GP ) [31] with thermolysin and compared its neutralization with non-treated virus . As shown in Fig 3A , thermolysin treatment of virions resulted in complete proteolysis of GP1 subunit , and slight reduction of the virus titer ( 3 . 8-fold ) . Incubation of intact VSV/BDBV-GP with neutralizing glycan cap-specific antibodies led to a dramatic reduction of virus titers , whereas thermolysin-processed virus was resistant to BDBV270 , BDBV41 and BDBV289 , with no effect shown for the non-neutralizing BDBV52 mAb against either virus preparation . In contrast , GP cleavage with thermolysin did not reduce virus sensitivity to GP2-specific BDBV259 , BDBV317 or BDBV223 mAbs , suggesting that these mAbs interact with the full-sized and processed fusion-active form of GP equally well , and , thus can inhibit multiple steps of virus entry . However , no mAb prevented the proteolysis of GP after treatment with cathepsin B and cathepsin L , as the 20 kDa GP1 fragment band resulting from the digestion of GP was present in all samples treated with cathepsin regardless of the mAb used ( Fig 3B ) . The differences in the band intensity observed with different mAbs are probably caused by binding of mAbs to additional cathepsin cleavage sites , which are not involved in generation of the 20 kD fragment . Secondary infection of cells by transfer of virions and intermediate products of viral replication ( genome copies , viral proteins or the whole vRNP complexes ) across the cytoplasmic bridges between infected and uninfected cells was shown to play an important role in the pathogenesis of HIV [32] , influenza virus [33 , 34] and EBOV [35] . Such cell-cell contacts can increase the effective viral MOI at the sites of transmission , making this route of infection spread 70-fold [35] to 2–3 orders of magnitude [32] more efficient compared to cell-free dissemination . Moreover , use of the alternative intercellular gateway for direct access to the cytoplasm of a new host cell allows virus to escape from antibodies targeting initial steps of cell entry and/or virus egress . To analyze the effects of mAbs on cell-to-cell transmission , we used a flow cytometry-based approach previously described for HIV studies [36] . THP-1 monocytes ( donor cells ) infected with EBOV/BDBV-GP ( which expresses eGFP ) were incubated with mAbs for 1 hour , placed at the top of Vero-E6 cell culture monolayers ( acceptor cells ) pre-stained with CellTrace Far Red , and incubated for 72 hours . Since all mAbs but BDBV52 are strong suppressors of viral entry ( Fig 2D ) , their constant presence in the cell medium was expected to prevent spread of infection through the medium . Indeed , titration of supernatant aliquots harvested from co-cultures of THP-1 and Vero-E6 cells on days 3–5 after the inoculation of monocytes showed an absence of detectable live viral particles in samples containing BDBV270 , BDBV289 , BDBV223 or BDBV317 mAbs , but not in those with 2D22 or no mAb ( S9 Fig ) . To measure cell-to-cell virus transmission , CellTrace FarRed+/eGFP+ cells were quantified by flow cytometry and expressed as percentages of the total CellTrace FarRed+ population ( Figs 3C and S4B ) . Consistent with the previous experiments ( Figs 2C–2F and 3A ) , the MPER-specific BDBV223 appeared to be the most potent mAb , as it completely suppressed the infection of acceptor cells at all concentrations tested . The other MPER-specific BDBV317 , along with one glycan cap-specific BDBV270 , demonstrated a clear dose-dependent inhibition of viral transmission , and the glycan cap-specific BDBV289 showed a somewhat lesser inhibition . The non-neutralizing glycan cap-specific BDBV52 mAb did not cause any detectable inhibitory effect . The overall ability of mAbs to inhibit virus transmission better corresponded to their ability to inhibit viral replication at higher ( 0 . 1 PFU/cell ) than the lower ( 0 . 01 PFU/cell ) MOI ( S10 and S11 Figs ) , as was previously demonstrated for HIV [32] . The overall antibody potency in the cell-to-cell transmission assay was much lower compared to the mAb effects on primary virus entry ( Fig 2D ) suggesting that higher mAb doses may be required to overcome secondary virus infection through the intercellular connections . Upon completion of replication cycle inside the cell , progeny virions release into the extracellular matrix and spread the infection to bystander cells . The budding of MARV can be prevented by antibodies even in the absence of virus neutralization detectable in plaque reduction assays , possibly by bivalent cross-linking of newly formed virions to each other and to the viral proteins exposed on the cell membrane [37] . We therefore analyzed the effect of mAbs on virus release from infected Vero-E6 cells . Since the presence of the neutralizing mAbs in the medium would interfere with analysis of released viral particles by plaque assay , we quantitated both live and neutralized virus released from infected cells and present in the medium by quantifying viral genomic RNA by droplet digital RT-PCR . The egress of virus was strongly and dose-dependently inhibited by all glycan cap-specific mAbs ( Fig 3D ) . In contrast , stalk-specific mAbs BDBV259 and BDBV317 increased the viral load in the supernatants when provided in low doses . Only high doses of these mAbs ( 100 μg/ml ) strongly reduced release of viral particles , which did not reach , however , the level of inhibition seen for the glycan cap-specific mAbs . Taken together , these data suggest that retention of produced virions on the cells is a common mechanism of antibodies targeting external , well-exposed domains of the viral envelope proteins involved in an interaction with a cell surface . Strikingly , the single tested non-neutralizing BDBV52 mAb abolished release of virus , even at concentrations as low as 1 μg/ml . The same inhibition level was demonstrated by the highly neutralizing BDBV223 mAb , highlighting the lack of correlation between suppression of virus egress and in vitro neutralization . To ensure that viral RNA detected in the cell supernatants resulted from a bona fide viral egress process , but not from the exit of RNA from cells via the exosome pathway , we conducted an additional experiment with depletion of exosomes in supernatant samples ( S12 Fig ) . Indeed , for each of the tested mAb , regardless of its concentration , incubation of supernatants with exosome-binding beads did not result in a significant change of the viral RNA level ( p > 0 . 05 , paired Student’s t-test ) . Next , we tested if mAbs can suppress virus replication when added at different time points after infection ( Figs 3E , S4A and S13 ) . Administration of BDBV270 , BDBV289 , BDBV223 or BDBV317 , which belong to different epitope recognition groups , 3 hours prior to or during cell inoculation completely blocked replication of virus . Infection inhibition by BDBV41 and BDBV259 was less prominent . In contrast , the two MPER-specific mAbs BDBV317 and BDBV223 caused the strongest reduction of infected cell numbers when added up to 3 hours after inoculation . When added at 24 hours post-inoculation , none of the mAbs tested could prevent virus replication by more than 20% . Addition of BDBV52 at any time point did not change percentages of eGFP+ cells ( Fig 3E ) , despite the fact that this mAb efficiently inhibited virus egress from infected cells ( Fig 3D ) . This finding can be explained by direct virus dissemination between cells skipping the step of virion release into the extracellular space , since BDBV52 did not block cell-to-cell transmission of the virus ( Fig 3C ) . The results indicate that MPER-specific antibodies are important for control viral replication , as they effectively prevent virus replication when added late . Glycosylation is one of the most common posttranslational modifications of viral surface proteins . Glycosylation of viral proteins may target key epitopes at the surface of virions , masking them from antibody recognition . A decade ago , an unusual post-translational modification , C-mannosylation , was found at the residue Trp288 of EBOV sGP , which was the first demonstration of this type of glycosylation in a viral protein [38] . Since then , no evidence of any biological significance of sGP C-mannosylation has been reported . BDBV GP possesses the same W288AFW291 motif in the glycan cap , which is considered to be the most immunogenic region of filoviral GP [3 , 17] . We therefore hypothesized that C-mannosylation of BDBV GP can impact antibody virus neutralization ( Fig 4 ) . To test this hypothesis , we disabled the mannosylation site in EBOV/BDBV-GP by introduction of the mutation W291A ( Fig 1 ) . To avoid interference of sGP with neutralization of viruses and to assess the pure effect of spike GP C-mannosylation on their resistance to mAbs , we also disabled the expression of sGP by stabilization of GP gene editing site [28] . BDBV270 mAb demonstrated a striking difference between neutralization of W291- and A291-bearing ( ΔC-mann ) mutants . At a concentration 0 . 8 μg/ml , the levels of neutralization of ΔsGP and ΔsGP/W288 ΔC-mann viruses by this mAb were 7 . 2% and 72 . 5% , respectively . Neutralization of the two viruses by MPER-specific BDBV223 ( Fig 4 ) or by other mAbs included in this study performed for comparison did not show any difference . These results demonstrated for the first time shielding of a viral epitope by C-mannosylation . Next , we sought to compare protection by glycan cap- and MPER-specific antibodies . Since BDBV52 , BDBV41 and BDBV259 mAbs are BDBV species-specific and do not bind EBOV GP [20] , and no mouse-adapted BDBV exists so far , they were not included in animal studies . We have shown previously that BDBV289 , BDBV223 and BDBV317 mAbs protect mice when given as a single 100 μg dose the day after challenge with 1 , 000 PFU of mouse-adapted EBOV delivered by the intraperitoneal route [20 , 25] . Here , we extended the study by testing BDBV270 ( S14 Fig ) . The in vivo activity of the antibody was similar to that of another glycan cap-specific mAb BDBV289 with 80% protection ( 4 out of 5 mice ) , with similar dynamics of changes in weight and disease score , although the difference in animal survival between BDBV270 and 2D22 mAb groups did not reach statistical significance ( p = 0 . 0644 , Mantel-Cox test ) . Thus , the two MPER-specific mAbs demonstrated complete protection and two glycan cap-specific mAbs demonstrated a high but not absolute protection in our present and previously reported studies [20 , 25] with the selected dose and regiment of treatment . Besides direct blocking of viral entry and/or exit through interaction with virions via Fab domains , mAbs also provide a second level of defense by cross-linking the viral proteins exposed on the surface of infected cells and Fc receptors on multiple immune cells to activate ADCC or ADCP mechanisms . Natural killer ( NK ) cells play a pivotal role in elimination of infected cells by ADCC . Activation of FcγRIIIa on NK cells causes the release of cytotoxic granules , which causes apoptotic death of target cells , as well as secretion of cytokines ( IFNγ and TNFα ) and chemokines ( MIP-1α and MIP-1β ) , which correlate with their activation . Phagocytosis through engulfment of infected cells represents another important mechanism of rapid clearance of infection , which is mediated by FcγR-bearing immune cells including monocytes , macrophages , dendritic cells , neutrophils , and mast cells known as professional phagocytes [39] . Since the expression of CD107a correlates with cytokine production and cytotoxicity and it is used as a marker of NK cell degranulation [40] , we used it as a marker of NK cell activation . In our experiments , the only two mAbs of IgG3 subclass , BDBV259 and BDBV223 , both are stalk-specific , induced a high level of surface expression of CD107a and intracellular production of IFNγ and MIP-1β in NK cells directed against BDBV GP ( Figs 5A–5C , S15A and S16A ) that is consistent with the higher affinity of IgG3 , compared to IgG1 , for binding to FcγRs [41] . BDBV259 and BDBV223 also induced ADCP of GP-covered beads by THP-1 monocytes and neutrophils ( Figs 5D , 5E , S15B , S16B and S16C ) . Interestingly , however , another stalk-specific mAb , BDBV317 , belonging to the IgG1 subclass , showed only a slight increase in NK cell activation compared to glycan cap-specific mAbs , yet induced neutrophil phagocytosis similarly to BDBV259 and BDBV223 . As interaction with FcRs can be modulated by both IgG subclass and Fc glycans structures , analysis of the glycans on the Fc domain was performed for each mAb ( Fig 5F–5I ) . Interestingly , the stalk-specific IgG3 mAbs , BDBV259 and BDBV223 , and the stalk-specific IgG1 , BDBV317 , were all characterized by higher sialylation of the Fc domain . As increased sialyation has been typically associated with anti-inflammatory activity [42 , 43] , the IgG3 subclass of BDBV259 and BDBV223 may underlie the enhanced functional activity associated with these mAbs . However , the IgG1 BDBV317 mAb was characterized by increased levels of galactose and bisecting N-acetylglucosamine ( GlcNAc ) glycan structures , and elevated levels of bisecting GlcNAc has been previously associated with greater phagocytic activity [44] and enhanced interaction with FcγRIIIa and ADCC activity [41] . The level of fucosylation , which negatively impacts binding of all IgG subclasses to FcγRIIIa and induction of ADCC [42] , was equally high for all tested mAbs . Altogether , these data suggest that while IgG3 induced the highest level of Fc-mediated effects , the epitope location also contributed to some of the Fc-mediated effects , consistent with previously published studies with influenza virus [45–50] . Finally , we selected BDBV223 mAb , which has the broadest spectrum of inhibitory activities against different steps of viral infection in vitro , to address the physiological relevance of the observed Fc-mediated effects for MPER mAbs ( Fig 5 ) . We introduced the L234A/L235A ( LALA ) mutation , which impairs binding of antibodies to FcγRs [51–54] , into the Fc region of the antibody , and compared the efficacy of mutated and non-mutated recombinant mAbs in a mouse model of EBOV infection ( Fig 6 ) . Human IgG1 and IgG3 have been shown previously to interact with mouse FcγRs [55] . Human IgG1 induces mouse innate immune effector functions at the levels equivalent to that induced by the most functional mouse subclass , IgG2a , while human IgG3 shows reduced activity with murine cells compared to human IgG1 [55] . Thus , it is possible that the human IgG3 mAbs cannot fully leverage the mouse innate immune system to maximize in vivo protective efficacy . We therefore generated the recombinant BDBV223 mAbs of IgG1 subclass , although the original BDBV223 subclass is IgG3 ( Fig 1 ) . Groups of BALB/c mice ( 5 animals per group ) were inoculated with 1 , 000 PFU of mouse-adapted EBOV , strain Mayinga , and 24 hours later treated by the intraperitoneal route with 40 or 100 μg of wild-type rBDBV223-IgG1 or rBDBV223-IgG1-LALA . At both doses tested , wild-type antibody , but not the LALA mutant , provided complete protection of mice from the lethal EBOV infection . The differences between survival of animals in rBDBV223-IgG1 and rBDBV223-IgG1-LALA groups were statistically significant: 40 μg , p = 0 . 0158 , 100 μg , p = 0 . 0494 ( Mantel-Cox test ) . These data suggest that Fc-FcγR interactions can play a critical role in protection against EBOV infection mediated by MPER mAbs in vivo .
The unprecedented epidemic of EBOV in West Africa in 2013–2016 demonstrated the urgent need for treatments against this and related highly pathogenic filoviruses . Antibody-based therapy remains the only available effective strategy against the infection . Further progress in development of more broad and effective filovirus mAbs requires identification of the mechanism of the protective effect of these mAbs . The glycan cap and MLD are excised by cathepsins during endosomal GP processing and , therefore they are dispensable for virus entry into the cytoplasm . It has been proposed that antibodies targeting these domains of GP are generally non-neutralizing , with some of them being able to confer protection likely through Fc-mediated mechanisms , such as ADCC or ADCP of infected cells [39] . In contrast , antibodies targeting the GP base could prevent membrane fusion [56] by blocking GP cleavage [57] or fusion-triggering conformational changes in proteolytic primed GP bound to NPC1 [58] , and therefore are mostly neutralizing . However , we isolated glycan cap-specific mAbs from the blood of survivors of natural ebolavirus infection that protect mice and guinea pigs from lethal EBOV challenge [20] . Murine m8C4 mAb targeting the glycan cap was reported to neutralize EBOV and SUDV and confer partial protection of mice against these viruses; induction of ADCP by neutrophils , monocytes and dendritic cells was proposed as one of the mechanisms of protection [44] . The discovery of novel antibody epitopes in RBD [44 , 59] , glycan cap/RBD interface [60] , IFL [19 , 59 , 61] , and epitopes proximal to the viral membrane [19 , 20] have substantially extended the concept of vulnerability sites on EBOV GP . Murine 6D6 cross-neutralizing mAb targeting the tip of the IFL prevented GP-mediated membrane fusion and protected mice against EBOV and SUDV [61] . Inhibition of cathepsin-cleaved EBOV GP binding to its endosomal receptor NPC1 was demonstrated to be the major mechanism of protection by human antibody mAb114 [57] and macaque-derived FVM04 mAb [21] . MAb114 recognizes an epitope spanning both the glycan cap and RBD , while FVM04 binds to the tip of the RBD crest . Interestingly , although antibody access to RBD is considered to be largely restricted by the surrounding glycan cap and MLD domains [5] , the epitope of FVM04 is exposed in the full-sized GP . Thus , prevention of endosomal membrane fusion remains the only demonstrated mechanism of EBOV neutralization by RBD- , IFL- and GP base-specific antibodies , whereas antiviral mechanisms employed by antibodies targeting the glycan cap and novel epitopes proximal to the viral membrane are not clear . Here , we investigated antiviral mechanisms for a diverse panel of human antibodies isolated from several human survivors of natural ebolavirus infections . Generation of escape mutant viruses resulted in mutations in the glycan cap of GP1 or in the IFL/stalk region of the GP2 subunit [26] . Glycan cap represents a well-exposed portion of the GP trimer in its native conformation , and therefore is a common target of the antibody response [5 , 22] , while the GP areas proximal to the viral membrane are less accessible , and have been only recently identified as a novel group of mAb epitopes [20 , 62] . Since filoviruses attach to the cell surface through low-affinity interactions with multiple types of molecules , none of the filovirus-specific mAbs , including those described in the present study , were shown to completely inhibit cell attachment and infection . Moreover , all of the neutralizing mAbs studied here showed dose-dependent inhibition of viral replication when added after virus attachment to cells , suggesting they inhibit intracellular steps of entry . Interestingly , the non-neutralizing BDBV52 mAb caused an enhanced viral attachment and entry into cells , which perhaps can be mediated by the re-uptake of de novo synthesized viral particles retained at the cell surface by BDBV52 at the budding step . We next analyzed mAb effects on VLP trafficking through the endosomal network . The tested mAbs did not prevent cathepsin cleavage of GP ( Fig 3B ) and had no effect on GP/NPC1 interaction ( S8 Fig ) . Therefore , the co-localization of viral particles with LAMP-1 and Rab7 endosomal markers observed in the presence of stalk-binding mAbs is likely a consequence of events accompanying the merge of viral and endosomal membranes , such as conformational rearrangements of GP2 subunit after interaction of cleaved GP with NPC1 . Other than blocking of virus entry , mechanisms of restriction of infection can include inhibition of cell-to-cell transmission or budding of nascent virions from infected cells . Both steps of virus infection were found to be inhibited by glycan cap and MPER mAbs in our study . However , these mechanisms are not mutually exclusive , and , moreover , could be mediated at least in part by direct virus neutralization . The latter mechanism seems to pertain for the most potent neutralizer , BDBV223 , which completely blocked virus transmission and egress , presumably by trapping it inside LAMP-1+ vesicles during cell entry . Unexpectedly , a comparable effect on egress inhibition was demonstrated for the non-neutralizing mAb BDBV52 , with no impact on virus transmission to neighboring cells observed . Overall , mAbs demonstrated differing patterns of cell-to-cell transmission and virus egress inhibition , which could not be explained by simple differences in their neutralization activity , and is probably determined by a combination of factors , such as the location of epitope and affinity to GP at differing pH conditions . The addition of N-linked glycans to envelope proteins is a commonly used strategy of immune evasion employed by HIV , influenza , Nipah and other viruses , which , at the same time , does not interfere with their attachment to the cell surface [63] . We found that C-mannosylation can also make virus less sensitive to a glycan cap-specific antibody . The C-mannosylation motif is located in the region shared by GP and sGP and is conserved in all known ebolavirus species: EBOV , SUDV , TAFV , BDBV and RESTV . Despite the fact that this modification was found in EBOV sGP protein [38] , the results of comparative neutralization of viruses with intact or disrupted C-mannosylation site and the lack of sGP produced by the viruses used in the assay suggests that envelope GPs of BDBV , and likely of all other ebolaviruses , are also subjected to C-mannosylation . The neutralization kinetics showed that the mannose residue on W288 is likely to restrict epitope access for at least some of the glycan cap-specific mAbs . Antibodies mediate antiviral effects both by binding epitopes on targeted pathogens by Fv region interactions and by activating Fc receptor-bearing effector cells , such as NK cells , neutrophils , macrophages and dendritic cells by Fc domain interactions . The spectrum of Fc-mediated effects induced by an antibody depends on its affinity for binding to particular FcγRs , which , in turn , depends on the IgG subclass and Fc region glycosylation . The conformational nature of the epitope recognized also impacts the efficiency of immune cell engagement . The disruption of Fc-FcγR linkage through either introduction of a D265A mutation in the Fc region or using knockout mice with disabled FcγRs leads to a complete loss of in vivo protection from influenza virus by broadly neutralizing HA stalk-targeting mAbs , but not by strain-specific mAbs binding to HA head domain [45 , 64] . From this insight , it was interesting to observe a substantial activation of NK cells and induction of monocyte- and neutrophil-mediated phagocytosis by stalk-specific mAbs BDBV259 and BDBV223 in our study compared to the glycan cap-specific mAbs . While the observed increased activation may be due to their IgG3 subclass , which have higher affinity for FcγRIIIa and FcγRIIa compared to IgG1 antibodies [65] , the stalk-specific BDBV317 IgG1 mAb also induced greater ADCP activity by neutrophils and stimulation of NK cells compared to the glycan cap-specific mAbs tested here , which also belong to the IgG1 isotype . Therefore , it is of interest to test if the direct contact between antibody-bound filovirus GP and the effector cell is required for optimal triggering of Fc mechanisms . The biological effects of mAbs demonstrated in this study are summarized in Fig 7 . In general , stalk-specific mAbs have greater Fab- and Fc-mediated effects , with the noticeable exception of the inhibition of viral egress , which was highly pronounced for all glycan cap-specific mAbs tested , and the greater level of protection in vivo . The current approach for treatment of filovirus infections with antibody cocktails demonstrated in animal models uses the principle of targeting of non-overlapping epitopes [20 , 44 , 59 , 60 , 66–68]; for example , our recent study demonstrated synergistic effects of the MPER-specific mAb BDBV223 and the glycan cap-specific mAb BDBV289 [20] . The data presented here suggest that there may be cooperative or synergistic effects of antibodies that block varying steps of viral replication , and cocktails based on combining such effects also should be tested . As the two contrast groups of mAbs tested in this study have different biological effects ( Fig 7 ) , the beneficial effects of cocktails of non-overlapping epitopes may be related not only to targeting different epitopes , but also to the ability of these antibodies to inhibit different steps of viral replication .
Wild-type BDBV , strain 200706291 Uganda , which was originally isolated from the serum of a patient during the first known outbreak [69] was passaged three times in Vero-E6 cells . The EBOV/BDBV-GP virus enveloped with glycoprotein of Bundibugyo strain , and EBOV/BDBV-GPΔsGP virus lacking sGP production were generated as described earlier [28] . To generate an EBOV/BDBV-GP derivative not expressing eGFP , the full-length clone was digested with BsiWI restriction endonuclease to remove eGFP gene , and then re-ligated . The resulting plasmid was transfected into 293T cell monolayers to rescue EBOV/BDBV-GP_no eGFP virus . To obtain EBOV/BDBV-GP ΔsGP derivative with disabled C-mannosylation site , we subjected pEBOwtΔBamHI-SbfI , AscI-PspOMI subclone with the ORF for the GP of BDBV with stabilized RNA editing site to PCR mutagenesis using the QuikChange site-directed mutagenesis kit ( Stratagene ) . Amino acid substitution W291A in BDBV GP was introduced into the construct to disrupt C-mannosylation of W288 residue in W288AFW291 motif . For generation of full-length construct , ApaI-SacI restriction endonuclease fragment from the resulting subclone was used to replace those in pEBO-eGFP plasmid . The obtained construct was transfected into 293T cell monolayers to rescue chimeric virus with disrupted C-mannosylation site - EBOV/BDBV-GP ΔsGP/W288 ΔC-mann . Neutralization of viruses by mAbs was tested in high-throughput screening assay based on the detection of residual eGFP fluorescence [28] . To generate VLPs enveloped with BDBV GP , glycoprotein ORF in pWRG7077:64755-2010-233-1_GP_optGP was substituted with that of BDBV . First , BamHI restriction endonuclease sites were disabled in pEBOwtΔBamHI-SbfI , AscI-PspOMI subclone with the ORF for the GP of BDBV with stabilized RNA editing site ( ΔsGP ) by introduction of silent mutations using the QuikChange site-directed mutagenesis kit ( Stratagene , La Jolla , CA ) . Then , BDBV GP ORF was amplified from the resulting construct with following primers: direct , AGTCACGTGCGGCCGCCACCATGGTTACATCAGGAATTCT; and reverse , AGTCACGTGGATCCTTATCATCAGAGTAGAAATTTGCAAA ( the NotI or BamHI restriction endonuclease sites are underlined , and the start of the BDBV GP ORF direct sequence and the end of the BDBV GP ORF complementary sequence are italicized ) . The obtained PCR product was used to replace EBOV GP ORF in EBOV VLP GP plasmid by NotI and BamHI sites to get the final GP-bearing plasmid for BDBV VLP production . EBOV NP and codon optimized VP40 were cloned into the pCEZ vector [70] . pCEZ-NP was a kind gift from Drs . Kawaoka and Feldmann . The plasmids were transfected to 293T cells using TransIT-LT1 transfection reagent ( Mirus ) . VLPs were harvested after 72 hours of the transfection , purified by sucrose gradient and quantified using ViroCyt Virus Counter ( VC ) 2100 ( ViroCyt ) . For confocal microscopy , Vero-E6 cell cultures ( American Type Culture Collection ) were grown in monolayers in chambered slides . BDBV VLPs were incubated in the presence of mAbs ( 200 μg/ml ) for 1 hour at room temperature , added to Vero-E6 cell culture monolayers at a ratio of 500 VLP/cell , and cells were placed on ice for 1 hour . Then , cells were fixed in formalin ( ThermoFisher Scientific ) for 15 min , permeabilized with 0 . 5% Triton X-100 in phosphate buffered saline ( PBS ) for 15 min to increase sensitivity of the subsequent immunostaining of viral proteins . Cells then were blocked with 5% donkey serum diluted in PBS with 1% BSA and 0 . 1% Triton X-100 ( PBS-T-BSA ) for 1 hour . Next , VLPs were stained using rabbit immune serum raised against EBOV VLPs ( IBT Bioservices ) supplemented with rabbit polyclonal antibodies specific for BDBV GP ( IBT Bioservices; all antibodies for virus staining were diluted at 1:100 in PBS-T-BSA ) . The slides then were incubated with donkey anti-rabbit antibodies conjugated with AlexaFluor 647 ( ThermoFisher Scientific ) for 1 hour at room temperature . Next , the slides were washed 3 times in PBS with 0 . 1% Triton X-100 ( PBS-T ) , fixed in 10% formalin and incubated with 4' , 6-diamidino-2-phenylindole dihydrochloride ( DAPI ) ( Invitrogen ) at 1 μg/ml for 2 min . Then , slides were washed 5 times in PBS and mounted onto coverslips using PermaFluor mounting medium ( ThermoFisher Scientific ) . The slides were analyzed by laser scanning confocal microscopy using an Olympus FV1000 confocal microscope housed in the Galveston National Laboratory . Lasers with 405 nm wavelength were used for DAPI excitation , and 635 nm for Alexa Fluor 647 . All images were acquired using a 60x oil objective . For quantification , five representative randomly selected images were acquired and the AlexaFluor 647 fluorescence was analyzed using the FV1000 software image measurement tool . Statistical analysis was performed using ANOVA with Tukey post hoc test . For flow cytometric analysis of virus binding , Vero-E6 cells were plated in U-bottom 96-well plates ( ThermoFisher Scientific ) at 106 cells per well and placed on ice . EBOV/BDBV-GP_no eGFP was incubated with mAbs ( 200 μg/ml ) at 37ºC for 1 hour followed by 15 min on ice and used to inoculate cells at an MOI of 5 PFU/cell . Cells were incubated for 2 hours on ice and washed with 2% fetal bovine serum ( FBS ) in PBS . Thereafter , cells were immunostained with rabbit immune serum against EBOV VLPs ( IBT Bioservices ) supplemented with anti-BDBV GP rabbit polyclonal antibody ( IBT Bioservices ) ; both the immune sera and antibody were added at 1:100 dilution in PBS with 2% FBS and incubated for 30 min at room temperature . After staining , cells were washed three times with 2% FBS in PBS , fixed in 10% formalin for 15 min , stained with donkey anti-rabbit antibodies labeled with Alexa Fluor 647 ( ThermoFisher Scientific ) and washed again 3 times with 2% FBS in PBS . Flow cytometry was performed using an LSRII Fortessa cytometer ( BD Biosciences ) . For each sample , 10 , 000 events were acquired . BDBV was adsorbed on Vero-E6 monolayer cell cultures in 24-well plates at an MOI of 0 . 1 PFU/cell for 20 min at 4ºC . Cells were washed 3 times with cold PBS , incubated with four-fold serial dilutions of mAbs for 20 min at 4ºC , washed again and covered with a 0 . 45% methylcellulose overlay in minimal essential medium ( MEM ) with 2% fetal bovine serum . Cells were incubated for 6 days at 37ºC , and plaques were visualized by immunostaining with BDBV52 mAb [20] followed by secondary goat anti-human IgG conjugated with horseradish peroxidase and 4CN two-component peroxidase substrate system ( KPL ) . Post-attachment inhibition was calculated as a percent reduction of numbers of viral plaques developed after incubation with antibody compared to no mAb control , as previously described [71 , 72] . For the no-mAb control samples , the average number of plaques per well was 263 . Three million PFU of eGFP-expressing EBOV/BDBV-GP were incubated with various mAbs at the final concentration 100 μg/ml for 1 hour at 37ºC and then adsorbed on Vero-E6 cell culture monolayers for 40 min at 4ºC . Cells were washed 3 times with MEM containing 10% FBS and incubated in fresh medium for 24 hours . Then , cells were treated with trypsin , harvested , washed twice with PBS and fixed with 4% paraformaldehyde for 24 hours for virus inactivation . Cells were analyzed by flow cytometry using an Accuri C6 cytometer ( BD Biosciences ) to determine the percentages of infected eGFP+ cells and their mean fluorescence intensity ( MFI ) . On average , 7 , 728 events were acquired per sample . BDBV VLPs were generated as described above . EBOV VLPs were purchased from IBT Bioservices . BDBV or EBOV VLPs were incubated with 200 μg/ml of mAbs for 60 min at 37ºC . Monolayers of Vero-E6 cells were inoculated with VLP/mAb complexes , incubated for 30 or 60 min and fixed with 4% paraformaldehyde for 15 min . Monolayers were washed and permeabilized with 0 . 5% Triton-X100 solution in PBS for 15 min . Monolayers were blocked with 5% donkey serum diluted in PBS-T-BSA for 30 min . Cell monolayers were stained with mouse mAb specific for lysosomal marker LAMP-1 ( Santa Cruz ) at a 1:50 dilution and goat polyclonal antibodies specific for late endosome marker Rab7 ( Santa Cruz ) at a 1:50 dilution . VLPs were stained with rabbit immune serum against EBOV VLPs or the same rabbit immune serum supplemented with rabbit anti-BDBV GP polyclonal antibody ( IBT Bioservices ) at a 1:100 dilution for each antibody . Slides were incubated for 1 hour at 37ºC , washed 3 times as above , and incubated with a mixture of three secondary antibodies , each at 1:200 dilution in PBS-T-BSA: donkey anti-mouse conjugated with Alexa Fluor 488 , donkey anti-goat conjugated with Alexa Fluor 594 and donkey anti-rabbit conjugated with AlexaFluor 647 ( ThermoFisher Scientific ) . Next , cells were washed 3 times in PBS-T , and nuclei were stained with DAPI , as described above . Slides were analyzed by laser scanning confocal microscopy using an Olympus FV1000 confocal microscope with 405 nm wavelength laser for DAPI excitation , 488 nm for Alexa Fluor 488 , 543 nm for Alexa Fluor 594 , and 635 nm for Alexa Fluor 647 . VSV/BDBV-GP was propagated in Vero-E6 cells; at 48 hours after inoculation , the virus suspension was harvested and clarified from cell debris by low-speed centrifugation . To purify the virus , supernatants were placed atop a 25% sucrose cushion and pelleted in an ultracentrifuge for 2 hours at 175 , 000 x g , 4ºC . Pellets were resuspended in 1x STE buffer ( 10 mM Tris , 1 mM EDTA , 0 . 1 M NaCl ) and further purified by ultracentrifugation in 20–60% sucrose gradient ( 1 . 5 hours at 288 , 000 x g , 4ºC ) . The virus-containing band was harvested , and VSV/BDBV-GP virions were washed from sucrose by final ultracentrifugation in 1x STE buffer ( 1 hour , 4ºC , 175 , 000 x g ) . The obtained viral particles were resuspended in 1x STE buffer . Flat-bottom high-binding 96-well microplates ( Greiner Bio-One ) were coated overnight with purified VSV/BDBV-GP particles diluted in PBS . Bound antigen was blocked with 1% bovine serum albumin ( Sigma-Aldrich ) in PBST buffer ( 0 . 1% Tween-20 in PBS ) , and treated for 20 min with 20 mM sodium citrate , pH 5 . 0 ( Sigma-Aldrich ) , or PBS for 20 min . MAbs were added at 1 μg/ml in 0 . 1% Tween-20 containing 20 mM sodium citrate , pH 5 . 0 , or PBST buffer , respectively , and incubated for 1 hour at 37ºC . Plates were washed three times in PBST buffer , secondary goat anti-human IgG conjugated with horseradish peroxidase ( KPL ) were added at a 1:2 , 000 dilution in PBST buffer , and plates were incubated for 1 hour at 37ºC . Next , plates were washed three times in PBST buffer , 1-component SureBlue Reserve TMB Microwell Peroxidase Substrate ( KPL ) was added , and plates were incubated for 20 min at room temperature and scanned in a Synergy microplate reader ( BioTek ) at the emission wavelength 630 nm . VSV/BDBV-GP purified as described above was resuspended in thermolysin digestion buffer ( 50 mM Tris , pH 8 . 0 , 0 . 5 mM CaCl2 ) and divided into two aliquots; one aliquot was treated with 0 . 5 mg/ml of thermolysin ( Promega ) and another one with an equal volume of thermolysin digestion buffer ( mock-treated virus ) for 40 min at 37ºC . The reactions were stopped by addition of EDTA up to the final concentration 10 mM . Virus samples were re-pelleted through a 25% sucrose cushion as described above , and washed by ultracentrifugation in 10 mM Tris , 0 . 1 M NaCl for 1 hour at 175 , 000 x g , 4ºC . The resulting preparations were resuspended in 10 mM Tris , 0 . 1 M NaCl , incubated with 100 μg/ml mAbs for 1 hour at 37ºC , or mock-incubated , and titrated on triplicate Vero-E6 cell culture monolayers using plaque reduction assay . Aliquots of thermolysin-treated or mock-treated purified virions were heated for 10 min at 95ºC and separated in Nu-PAGE 4 to 12% Bis-Tris gel with Novex Sharp Pre-Stained Protein Standard used as a molecular weight marker . Proteins were transferred to a nitrocellulose membrane using the iBlot Gel transfer system ( Life Technologies ) . The membrane was incubated with primary rabbit polyclonal antibodies against BDBV GP ( 1:500; IBT Bioservices ) and secondary goat anti-rabbit IgG antibodies conjugated with horseradish peroxidase ( 1:500; KPL ) . Protein bands were visualized using the chromogenic 4CN two-component peroxidase substrate system ( KPL ) . EBOV VLPs alone or in the presence of 200 μg/ml of mAbs were incubated in sodium acetate buffer , pH 5 . 0 , with 0 . 1 μg/μl of cathepsin B and cathepsin L at 37ºC overnight . Thereafter , VLPs were denatured in Laemmli buffer ( Novex ) in reducing conditions , and GP cleavage was confirmed by immunoblotting with a pan-filovirus GP-specific monoclonal antibody ( IBT Bioservices ) . Densitometry was performed using ImageJ gel analyzer plug-in . For normalization , we used VP40 as a housekeeping protein and the 20 kDa band of GP as the target protein . THP-1 monocytic cells ( American Type Culture Collection ) were inoculated with EBOV/BDBV-GP virus expressing eGFP at MOI of 2 PFU/cell , incubated for 48 hours , washed two times to remove unbound virus , and incubated with 100 μg/ml of mAbs or no mAb . Following a one hour-long incubation , cells were placed atop of monolayers of Vero-E6 cells pre-stained with CellTrace Far Red ( ThermoFisher Scientific ) according to the manufacturer’s recommendations , incubated for 72 hours and fixed with 4% paraformaldehyde . Cells were analyzed by flow cytometry to determine the percentages of cells double-positive for CellTrace Far Red and eGFP of total cells positive for CellTrace Far Red . The percentage of double-positive cells indicated the percentage of cells that became infected due to cell-to-cell transmission of virus . For each sample , 30 , 000 events were counted . In a separate experiment , supernatant aliquots were harvested from co-cultures of THP-1 and Vero-E6 cells on days 3–5 after the inoculation of monocytes and then titrated on Vero-E6 cell monolayers . Vero-E6 cell culture monolayers were inoculated with EBOV/BDBV-GP expressing eGFP at an MOI of 0 . 1 PFU/cell , incubated for 1 hr , washed 3 times to remove non-attached viral particles , and covered with medium containing 1 , 10 or 100 μg/ml of mAbs or no mAb . Cells were incubated for 48 hours , supernatants were collected , and RNA was isolated . Viral genomes were quantitated by one-step reverse transcription droplet digital RT-PCR ( Bio-Rad ) according the manufacturer’s instructions . Sequences of primers are available upon request . In a separate experiment , cell supernatants were incubated with exosome removal beads ( Exosome-Human CD63 Isolation/Detection Reagent , ThermoFisher Scientific ) for 30 min at ambient temperature , or mock-incubated , centrifuged for 5 min at low speed for sedimentation of beads , transferred to the clean tubes and subjected to RNA isolation and droplet digital RT-PCR analysis . Vero-E6 cell culture monolayers in 24-well plates were inoculated with EBOV/BDBV-GP expressing eGFP at an MOI of 0 . 1 PFU/cell , with mAbs added at final concentration 100 μg/ml 3 hours prior to , at the moment of infection , or 3 or 24 hours after virus inoculation . Forty-eight hours after inoculation , cells were washed twice with PBS , treated with trypsin , harvested , fixed with 4% paraformaldehyde , and infected ( eGFP+ ) cells were quantified by flow cytometry . For each sample , 10 , 000 events were counted . Seven-week-old BALB/c mice ( Charles River Laboratories ) were placed in the ABSL-4 facility of the Galveston National Laboratory . Groups of mice at five animals per group were injected intraperitoneally with 1 , 000 PFU of the mouse-adapted EBOV . Twenty-four hours later , animals were injected with mAbs at indicated amounts by the intraperitoneal route . Animals treated with the 2D22 mAb specific for dengue virus served as controls . The recombinant versions of BDBV223 mAb with or without LALA mutation in the Fc fragment ( rBDBV223-IgG1-LALA and rBDBV223-IgG1 , respectively ) were generated as described elsewhere [25 , 26 , 73] . The animal observation procedure was performed as previously described [20] . The extent of illness was scored using the following parameters: dyspnea ( possible scores 0–5 ) , recumbence ( 0–5 ) , unresponsiveness ( 0–5 ) , and bleeding/hemorrhage ( 0–5 ) . Moribund mice were euthanized as per the protocol approved by the UTMB Institutional Animal Care and Use Committee . The humane endpoint for weight loss was 20% . The overall observation period lasted for 28 days . The NPC1-encoding plasmid was purchased from OriGene . NPC1 was amplified by PCR and cloned into the p3xFLAG-CMV9 plasmid ( Sigma-Aldrich ) . To add the red fluorescent protein ( RFP ) at the N-terminus , RFP gene cDNA was PCR-amplified from pcDNA3-mRFP ( Addgene ) and added upstream of the NPC1 coding sequence using NotI and BamHI restriction endonuclease sites . Vero-E6 cell culture monolayers were electroporated with a P3X_NPC1-RFP using Neon transfection system ( ThermoFisher Scientific ) with 2 pulses of 20 msec at 1 , 150 V , placed in chambered slides ( Nalge Nunc International ) and incubated overnight at 37ºC . EBOV/BDBV-GP_no eGFP was incubated with 200 μg/ml of mAbs for 1 hour at 37ºC and used for inoculation of transfected cells at an MOI of 10 PFU/cell for 30 min . Thereafter , cells were fixed with 4% paraformaldehyde for 15 min . Cell monolayers were washed 3 times in PBS-T , and viruses were incubated with rabbit immune serum against EBOV VLPs ( IBT Bioservices ) supplemented with rabbit anti-BDBV GP polyclonal antibody ( IBT Bioservices ) at a 1:100 dilution for both antibodies for 1 hour . Next , cells were washed 3 times with PBS-T and incubated with donkey anti-rabbit antibody conjugated with Alexa Fluor 647 ( ThermoFisher Scientific ) diluted 1:200 in PBS-T-BSA for 30 min . Next , the slides were washed 3 times in PBS-T , fixed in 10% formalin for 72 hours and removed from the BSL-4 . The slides were washed 3 times in PBS and mounted onto coverslips using PermaFluor mounting medium ( ThermoFisher Scientific ) . FRET analysis was performed by scanning confocal microscopy using an Olympus FV1000 confocal microscope with the 543 nm laser for excitation and a far-red emission filter for detection . FRET efficiency ( E ) was calculated using Olympus FV1000 software . The effect of mAbs on NPC1-GP interaction was measured by changes of FRET efficiency when compared with the effect of virus inoculated in the absence of mAbs . Human NK cells were enriched from peripheral blood by negative selection using RosetteSep negative selection kit ( Stem Cell Technologies ) followed by Ficoll separation . NK cells were rested overnight in the presence of 1 ng/ml recombinant IL-15 ( PeproTech ) . 3 μg/ml of BDBV GP ( IBT Bioservices ) was coated on a Maxisorp ELISA plate ( Nunc ) at 4°C overnight , and plates were blocked with 5% BSA prior to addition of antibodies ( 5 μg/ml ) in PBS for 2 hours at 37°C . The control EBOV-specific mAb c13C6 was purchased from IBT Bioservices . Unbound antibodies were removed by washing wells 3X with PBS prior to addition of NK cells . The NK cells were added at 5 x 104 cells/well in the presence of brefeldin A ( Sigma Aldrich ) , GolgiStop ( BD Biosciences ) , and anti-CD107a PE-Cy5 antibody ( BD Biosciences clone H4A3 ) and incubated for 5 hours at 37°C . NK cells were stained with flow cytometry antibodies for the following surface markers: CD3 AlexaFluor700 ( BD Biosciences clone UCHT1 ) , CD56 Pe-Cy7 ( BD Biosciences clone B159 ) , and CD16 APC-Cy7 ( BD Biosciences clone 3G8 ) , followed by intracellular staining for IFNγ ( FITC , BD Biosciences clone B27 ) and MIP-1β ( PE , BD Biosciences clone D21-1351 ) to detect the production of cytokines and chemokines . Cells were analyzed by flow cytometry on a BD LSRII flow cytometer and data was analyzed using FlowJo software . Recombinant BDBV GP was biotinylated and conjugated to streptavidin-coated Alexa488 beads ( Life Technologies ) . BDBV-coated beads were incubated with antibodies at 5 μg/ml in culture medium for 2 hours at 37°C . Human THP-1 cells ( ATCC ) were added at a concentration of 2 . 5 x 104 cells/well and incubated for 18 hours at 37°C in 96-well plates . Cells were fixed with 4% paraformaldehyde and analyzed by flow cytometry on a BD LSRII using Diva software and FlowJo analysis software . The phagocytic score was determined using the following calculation: ( % of AlexaFluor488+ cells ) * ( AlexaFluor488 geometric MFI of AlexaFluor488+ cells ) /10 , 000 . Recombinant BDBV GP was biotinylated and conjugated to streptavidin-coated Alexa488 beads ( Life Technologies ) . BDBV-coated beads were incubated with antibodies at 5 μg/ml in culture medium for 2 hours at 37°C . Human white blood cells were isolated from peripheral blood by lysis of red blood cells using ammonium chloride potassium lysis buffer . Cells were washed with PBS , and 5 . 0 x 104 cells/well were added to bead-antibody immune complexes , and then incubated for 1 hour at 37°C . Cells were stained with the following antibodies to identify neutrophils: CD66b Pacific Blue ( BioLegend clone G10F5 ) , CD14 APC-Cy7 ( BD Biosciences clone MφP9 ) and CD3 AlexaFluor700 ( BD Biosciences clone UCHT1 ) . Cells were fixed with 4% paraformaldehyde and were analyzed on a BD LSRII flow cytometer . A phagocytic score was determined as described above . 20 μg of antibodies were digested with 120 U of IDEZ ( NEB ) for 1 hour at 37°C to separate the F ( ab′ ) 2 and Fc regions . The Fc region was purified by incubating digested antibodies with magnetic protein G beads ( NEB ) for an additional hour at room temperature . Beads were washed with 2X with distilled water . Beads were then incubated with PNGaseF ( ThermoFisher Scientific ) to remove the N-linked glycan at 50°C for 1 hour . Released glycans were purified and labeled using the GlycanAssure APTS labeling kit ( ThermoFisher Scientific ) according to manufacturer’s instructions . Labeled glycans were analyzed on a 3500xL Genetic Analyzer ( Applied Biosystems ) using a POP7 polymer . Glycan peaks and relative abundance of glycan content was analyzed using the GlycanAssure Data Analysis Software v1 . 0 ( Applied Biosystems ) . Factorial ANOVA and two-sided t-test were used for statistical analysis of in vitro data . Animal survival data were analyzed by log-rank ( Mantel-Cox ) test . The animal protocol for testing of mAbs in mice was approved by the UTMB Institutional Animal Care and Use Committee ( protocol №1307033 ) in compliance with the Animal Welfare Act and other applicable federal statutes and regulations relating to animals and experiments involving animals . Challenge studies were conducted under maximum containment in an animal biosafety level 4 ( ABSL-4 ) facility of the Galveston National Laboratory .
|
Recent progress in isolation of mAbs from survivors of filovirus infections suggests that the human adaptive immune system is capable of producing strong antibody responses . However , the effects of mAbs with different epitope specificity on individual steps of filovirus infection are still unclear . We evaluated a panel of mAbs obtained from survivors of natural filovirus infections , specific for the glycan cap or stem region of GP , for their effects on the attachment of viral particles to the cell surface , intracellular traffic of viral particles , proteolytic processing of GP , its interaction with the NPC1 receptor , cell-to-cell virus transmission , virus egress from infected cells , activation of natural killer cells and antibody-dependent cellular phagocytosis through Fc-mediated mechanisms . We found that antiviral activity of glycan cap-specific antibodies results from inhibition of attachment , cell-to-cell transmission and inhibition of virion budding . In contrast , the antiviral mechanisms of stem-specific antibodies were found to be inhibition of virus release from endosomal network to the cytoplasm , and also activation of natural killer cells and phagocytosis mediated by monocytes and neutrophils . The data provide new insight into the development of immune protective mechanisms during natural human infection , and have important implications for the treatment of filovirus infections by passively-transferred antibodies and vaccine design .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"blood",
"cells",
"flow",
"cytometry",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"immune",
"cells",
"viral",
"transmission",
"and",
"infection",
"immunology",
"cell",
"processes",
"microbiology",
"viral",
"structure",
"fluorophotometry",
"antibodies",
"research",
"and",
"analysis",
"methods",
"immune",
"system",
"proteins",
"fluorescence",
"resonance",
"energy",
"transfer",
"white",
"blood",
"cells",
"viral",
"entry",
"animal",
"cells",
"proteins",
"viral",
"replication",
"phagocytosis",
"virions",
"spectrophotometry",
"biochemistry",
"cytophotometry",
"cell",
"biology",
"nk",
"cells",
"virology",
"physiology",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"spectrum",
"analysis",
"techniques"
] |
2018
|
Asymmetric antiviral effects of ebolavirus antibodies targeting glycoprotein stem and glycan cap
|
Entry into mitosis is triggered by cyclinB/Cdk1 , whose activity is abruptly raised by a positive feedback loop . The Greatwall kinase phosphorylates proteins of the endosulfine family and allows them to bind and inhibit the main Cdk1-counteracting PP2A-B55 phosphatase , thereby promoting mitotic entry . In contrast to most eukaryotic systems , Cdc14 is the main Cdk1-antagonizing phosphatase in budding yeast , while the PP2ACdc55 phosphatase promotes , instead of preventing , mitotic entry by participating to the positive feedback loop of Cdk1 activation . Here we show that budding yeast endosulfines ( Igo1 and Igo2 ) bind to PP2ACdc55 in a cell cycle-regulated manner upon Greatwall ( Rim15 ) -dependent phosphorylation . Phosphorylated Igo1 inhibits PP2ACdc55 activity in vitro and induces mitotic entry in Xenopus egg extracts , indicating that it bears a conserved PP2A-binding and -inhibitory activity . Surprisingly , deletion of IGO1 and IGO2 in yeast cells leads to a decrease in PP2A phosphatase activity , suggesting that endosulfines act also as positive regulators of PP2A in yeast . Consistently , RIM15 and IGO1/2 promote , like PP2ACdc55 , timely entry into mitosis under temperature-stress , owing to the accumulation of Tyr-phosphorylated Cdk1 . In addition , they contribute to the nuclear export of PP2ACdc55 , which has recently been proposed to promote mitotic entry . Altogether , our data indicate that Igo proteins participate in the positive feedback loop for Cdk1 activation . We conclude that Greatwall , endosulfines , and PP2A are part of a regulatory module that has been conserved during evolution irrespective of PP2A function in the control of mitosis . However , this conserved module is adapted to account for differences in the regulation of mitotic entry in different organisms .
Entry into mitosis in eukaryotic cells is driven by cyclin-dependent kinases ( CDKs ) bound to B-type cyclins . Many targets of cyclinB/CDKs have been identified in different organisms and include proteins involved in mitotic spindle formation/elongation , nuclear envelope breakdown , chromosome condensation and segregation . Several mechanisms contribute to the rapid raise in cyclinB/CDKs activity at the onset of mitosis: i ) cyclin accumulation , through transcriptional activation and inhibition of their proteolysis; ii ) phosphorylation of their catalytic subunit Cdk1 by Cdk1-activating kinases ( CAKs ) ; iii ) removal of Cdk1 inhibitory phosphorylations on Thr14 and Tyr15 ( reviewed in [1] . Phosphorylation on Thr14 and Tyr15 of Cdk1 is carried out by Wee1 ( Swe1 in budding yeast ) and Myt1 protein kinases and is reversed by Cdc25-like phosphatases ( Mih1 in budding yeast ) that promote entry into mitosis . Polo kinase , which is activated in mitosis by cyclinB/Cdk1 [2] , [3] , activates in turn Cdc25 [4] , [5] . In addition , the Wee1 kinase is phosphorylated and downregulated by cyclinB/Cdk1 [6] , [7] , [8] , [9] . Together , these data have led to the idea of a positive autoregulatory loop for cyclinB/Cdk1 activation at the onset of M phase [10] , [11] . In budding yeast a positive feedback loop for mitotic CDKs also exists , but differs from that of other organisms in many respects . The mitotic CDKs Clb1-4/Cdk1 drive a mitotic transcriptional program that leads to accumulation of several mitotic proteins , including the Clb1-4 cyclins themselves and the polo kinase Cdc5 ( reviewed in [12] ) . Clb2-Cdk1 initially activates Swe1 through direct phosphorylation , thus leading to its own inhibition [6] . However , this phosphorylation primes further phosphorylations on Swe1 by multiple kinases , including Cdc5 , that eventually target it to degradation [13] . Another key aspect of mitotic entry concerns the downregulation of phosphatases counteracting cyclinB/Cdk1 activity . In budding yeast , the Cdc14 phosphatase promotes inactivation of cyclinB/Cdk1 at the end of mitosis and reverses their phosphorylations [14] , [15] . In interphase Cdc14 is kept inactive in the nucleolus [16] , [17] . In all other eukaryotic cells analysed to date , phosphorylation events by mitotic CDKs are reversed mainly by PP2A and , to a lesser extent , by PP1 phosphatases ( reviewed in [18] ) . Phosphatases are often protein complexes containing catalytic and regulatory subunits that confer substrate specificity . Core PP2A complexes are made of one catalytic ( C subunit ) , one scaffold ( A subunit ) and one of many regulatory subunits ( B subunit ) [19] . Additional proteins , like the conserved Tap42 , can also bind to PP2A and modulate its activity ( reviewed in [20] ) . The B55 regulatory subunit , which exists in several isoforms , confers specificity of PP2A complexes towards Cdk1-dependent phosphorylation sites [21] , suggesting that PP2A-B55 is the most relevant phosphatase in many eukaryotes for reversing Cdk1-dependent phosphorylations . Consistently , PP2A-B55δ prevents mitotic entry in Xenopus egg extracts [22] , [23] and PP2A-B55α promotes mitotic exit in human cells [24] . In the past few years , the Greatwall protein kinase has emerged as a key factor for restraining PP2A-B55 activity in mitosis and allowing mitotic entry . In Xenopus egg extracts Greatwall is required for mitotic entry and to maintain the mitotic state [23] , [25] , [26] , [27] . Depletion of Greatwall in Drosophila neuroblasts leads to mitotic defects that are compatible with a role of Greatwall in promoting entry into mitosis [28] , [29] . Two closely related regulatory proteins , α-endosulfine ( ENSA ) and Arpp19 , bind and inhibit PP2A-B55δ upon phosphorylation by Greatwall on a specific serine residue [30] , [31] . Inactivation of Drosophila endosulfine leads to mitotic defects similar to those caused by Greatwall inactivation [32] , while depletion of Arpp19 arrests Xenopus egg extracts in G2 [30] , [31] . Indeed , Greatwall-dependent phosphorylation of α-endosulfine and Arpp19 promotes mitotic entry both by maintaining high levels of phosphorylated Cdk1 substrates and by feeding the Cdk1 autoregulatory loop ( reviewed in [33] , [34] ) . In stark contrast to other organisms , budding yeast PP2ACdc55 promotes , rather than prevents , timely entry into mitosis by participating in the positive feedback loop for Cdk1 activation ( reviewed in [20] ) . Indeed , PP2ACdc55 opposes the initial Swe1 phosphorylation by Cdk1 , which stimulates Cdk1-Tyr19 inhibitory phosphorylation [35] . In addition , PP2ACdc55 dephosphorylates and activates Mih1 [36] . Consistently , mutants defective in PP2ACdc55 activity accumulate high levels of Tyr19-phosphorylated Cdk1 [35] , [37] , [38] , [39] . Budding yeast possesses two redundant endosulfines , called Igo1 and Igo2 ( Initiation of G zero ) , that are phosphorylated by the Greatwall-related kinase Rim15 on the same serine ( Ser64 of Igo1 and Ser63 of Igo2 ) that in Xenopus Arpp19 and ENSA is phosphorylated by Greatwall [40] . Deletion of IGO1 and IGO2 , as well as of RIM15 , does not affect cell viability but compromises the establishment of the quiescence ( G0 ) state . Indeed , Igo proteins phosphorylated by Rim15 bind to an mRNA decapping factor to shelter degradation of specific mRNAs during initiation of the quiescence program [40] . Furthermore , recent data indicate that Igo proteins help establishing the quiescence-specific transcriptional program through binding and inhibition of PP2A bound to its B subunit Cdc55 [41] . Similar to yeast , depletion of α-endosulfine by RNAi or deletion of its aminoacid sequence targeted by Greatwall has no obvious consequence on cell division and fertility in the worm C . elegans [42] . In this manuscript we report the roles of Rim15 and Igo proteins in the control of mitosis . In agreement with recently published data [41] , we find that upon phosphorylation by Rim15 Igo1 binds to yeast PP2ACdc55 and can inhibit its activity in vitro . Consistently , phosphorylated Igo1 promotes mitotic entry in Xenopus egg extracts similarly to ENSA and Arpp19 , suggesting that endosulfines from different species are interchangeable . However , phenotypic analyses of yeast mutants lacking Rim15 or Igo proteins , together with biochemical data , indicate that yeast endosulfines behave in vivo as activators , rather than inhibitors , of PP2ACdc55 and contribute to timely activation of mitotic CDKs . Our data indicate that Igo proteins help establishing the positive feedback loop for Cdk1 activation and retaining PP2ACdc55 in the cytoplasm [43] , thereby promoting mitotic entry . We propose that endosulfines are accessory subunits of PP2A-B55 complexes that can contribute to their activation while inhibiting their phosphatase activity , to meet specific mitotic features in different organisms .
The budding yeast paralogous proteins Igo1 and Igo2 share significant homology with vertebrate endosulfines [44] , especially around the serine that in endosulfines is phosphorylated by Greatwall ( Gwl ) ( Fig . 1A ) . Indeed , Ser64 of Igo1 was shown to be phosphorylated by the yeast Greatwall-like kinase Rim15 [40] . To investigate if yeast endosulfines interact physically with PP2A in yeast cells , endogenous Igo1 was tagged at the C-terminus with 3 Pk epitopes and immunoprecipitated from yeast cells extracts . We then assessed the presence of various PP2A subunits in the immunoprecipitates . As shown in Fig . 1B , we found the catalytic subunit Pph21 and the Cdc55 regulatory subunit of PP2A to co-immunoprecipitate with Igo1-Pk3 . In addition , Rts3 and Tap42 , which associate with different PP2A phosphatases [45] , [46] , were also found in the immunoprecipitates . In contrast , we could not detect Rts1 , the other major PP2A regulatory subunit , bound to Igo1-Pk3 ( Fig . S1A ) . Based on several independent experiments , we estimated that about 1% of total Cdc55 is bound to Igo1 in asynchronous cycling cells . Interestingly , Igo1 association with PP2A subunits was dramatically impaired upon deletion of RIM15 , suggesting that it is enhanced by Igo1 phosphorylation ( Fig . 1B ) . Consistently , the interaction between HA-tagged Cdc55 and the mutant protein Igo1-S64A tagged with myc epitopes was markedly reduced compared to its wild type counterpart ( Fig . 1C ) . Thus , efficient interaction between Igo1 and PP2A requires Rim15-dependent phosphorylation of Igo1 on Ser64 . The residual binding between Igo1-S64A and Cdc55 appeared to be further impaired by RIM15 deletion ( Fig . 1C ) , suggesting that Rim15 might phosphorylate additional Igo1 residues besides S64 to stabilize the Igo1-PP2ACdc55 complex . Indeed , recombinant Igo1-S64A could be still phosphorylated in vitro by human Greatwall , albeit inefficiently ( Fig . S1B ) . We next asked if Igo1 interaction with PP2ACdc55 is regulated during the cell cycle . To this end , wild type cells co-expressing Pk-tagged Igo1 and HA-tagged Cdc55 were arrested in G1 by alpha-factor and released into fresh medium at 25°C . At different time points after release we analysed Igo1-Cdc55 interaction by co-immunoprecipitation , as well as other cell cycle parameters like budding , formation and elongation of bipolar spindles . As shown in Fig . 1D , a basal level of Cdc55 binding to Igo1 was detectable during most cell cycle stages but sharply peaked at 60 minutes from the G1 arrest . Based on the kinetics of bipolar spindle formation ( starting at 60 minutes after release ) and the appearance of the mitotic cyclin Clb2 and the polo kinase Cdc5 ( peaking at 75 minutes after release ) , we conclude that Igo1-Cdc55 interaction is maximal in late S/G2 phase . The rise and fall in Igo1-Cdc55 interaction during the cell cycle was very sharp , as shown by a tighter time course , and depended on Rim15 ( Fig . 1E ) . In addition , it was present at relatively high levels in cells treated with the replication inhibitor hydroxyurea , but not in cells arrested in G1 by alpha factor or arrested in mitosis upon spindle depolymerization by nocodazole ( Fig . 1F ) . In spite of its cell cycle-regulated interaction with Cdc55 , phosphorylation of Igo1 on Ser64 appeared to be constant throughout the cell cycle , as shown by Phos-tag phosphate affinity gel electrophoresis ( Fig . 1G ) . Thus , Rim15-dependent phosphorylation of Ser64 is necessary for Igo1 binding to Cdc55 , but other factors must be involved for this interaction to be maximal in late S or G2 phase . To test if yeast endosulfines inhibit PP2A catalytic activity like in other organisms , we affinity-purified PP2ACdc55 complexes from yeast cells expressing HA-tagged Cdc55 and measured its associated phosphatase activity on the P-Ser/P-Thr substrate phosphorylase a [47] . Recombinant GST-tagged Igo1 , but not its S64A mutant variant , could be readily phosphorylated by a hyperactive version of the Gwl kinase ( Gwl-K72M ) purified from baculovirus-infected insect cells ( Fig . 2A ) , suggesting that Gwl and Rim15 are interchangeable for Igo1 phosphorylation in vitro . Consistent with previous results [41] , addition of phosphorylated Igo1 to PP2ACdc55 complexes inhibited their activity in vitro in a dose-dependent manner ( Fig . 2B ) . The S64A mutation reduced , but did not abolish , the ability of Igo1 to inhibit PP2A activity ( Fig . 2B ) . To establish if Igo1 and Igo2 are functional orthologs of endosulfines , we asked if they could promote mitotic entry in Xenopus egg extracts . Addition of recombinant Igo1 phosphorylated in vitro by Gwl to Xenopus interphase extracts induced mitotic entry , as assessed by the appearance of phosphorylated forms of Gwl and Cdc25 , the disappearance of Cdk1 inhibitory phosphorylation and raise in histone H1 kinase activity . In stark contrast , addition of Igo1-S64A had no effect ( Fig . 2C ) . Since endosulfine-mediated mitotic entry under these conditions depends on inhibition of PP2A/B55 , we asked if Igo1 could bind to PP2A in Xenopus egg extracts . Strikingly , wild type Igo1 , but not Igo1-S64A , pulled down the catalytic ( C ) , structural ( A ) and regulatory ( B55 ) subunit of PP2A from CSF-induced mitotic extracts ( Fig . 2D ) . Thus , yeast endosulfines can inhibit the phosphatase activity of PP2A complexes and fulfill the role of their vertebrate counterparts in promoting mitotic entry . To further investigate the possible role of Rim15 and Igo1 , 2 in cell cycle progression , we analysed the latter upon deletion of RIM15 or IGO1 and IGO2 . Deletion of RIM15 or IGO1 and IGO2 had no significant effect on the kinetics of cell division at physiological temperatures ( 25°C and 30°C , data not shown ) . However , rim15Δ and igo1Δ igo2Δ mutant cells where sensitive to thermal stress ( 14°C , 16°C and 38°C ) and were hypersensitive to cell wall stressors , such as caffeine , calcofluor white and SDS ( Fig . S2A ) . We compared the cell cycle progression of rim15Δ and igo1Δ igo2Δ mutants to that of wild type cells at high and low temperatures . Cells were grown at 25°C , arrested in G1 by alpha-factor and then released in the cell cycle at 38°C . Although the kinetics of budding were similar in the three strains , mitotic events , such as spindle formation , spindle elongation and nuclear division , were delayed by 10–30 minutes in rim15Δ and igo1Δ igo2Δ cells relative to wild type . In addition , accumulation of the polo kinase Cdc5 , as well as of the mitotic cyclin Clb2 and its associated kinase , were affected by RIM15 deletion and , even more pronouncedly , by IGO1 and IGO2 deletion ( Fig . 3A ) . A similar experiment showed that deletion of RIM15 and IGO1 and IGO2 delayed mitosis also at 16°C , as shown by analysis of the same cell cycle markers above ( Fig . 3B ) . Also at low temperature the mitotic defects of igo1Δ igo2Δ cells were more pronounced than those of rim15Δ cells . Taken together , these data indicate that Rim15 and its targets Igo1 and Igo2 are required for timely mitotic entry and mitotic progression under temperature stress , with Igo proteins having a more prominent role than Rim15 in this process . Consistent with a conserved function for Igo proteins and endosulfines in promoting mitotic entry , at least under certain conditions , expression of human Arpp19 or human ENSA from the IGO1 promoter partially rescued the temperature-sensitivity of igo1Δ igo2Δ cells at 37°C ( Fig . S2B ) . Since yeast PP2ACdc55 promotes timely entry into mitosis , the genetic data above raised the possibility that in vivo Igo proteins contribute to PP2ACdc55 activation . We therefore set out to measure PP2ACdc55 catalytic activity in wild type and igo1Δ igo2Δ cells . To this end , we immunoprecipitated HA-Cdc55 and tested the activity of the associated phosphatase using phosphorylase a and histone H1 as substrates . As shown in Fig . 4A , anti-HA immunoprecipitates contained the Pph21 catalytic and the Tpd3 scaffold subunits of PP2A . Remarkably , lack of Igo1 and Igo2 caused a small ( 15–20% ) but significant decrease on PP2ACdc55 activity on both phosphorylase a and histone H1 ( Fig . 4B ) . No further decrease in PP2ACdc55 activity was observed by incubating igo1Δ igo2Δ cells at the restrictive temperature ( 38°C , data not shown ) . Thus , although Igo proteins are PP2ACdc55 inhibitors , in vivo they sustain full PP2ACdc55 activity . However , they did not seem to affect the basic composition of the hetero-trimeric PP2ACdc55 complex , as shown by the similar levels of interaction between endogenous HA-tagged Cdc55 and the PPh21 and Tpd3 subunits in wild type and igo1Δ igo2Δ cells ( Fig . 4C ) . Deletion of CDC55 causes an accumulation of Tyr19-phoshorylated Cdk1 and delays mitotic entry [37] , [39] . Our finding that Igo1 and Igo2 are required for full PP2ACdc55 activity and for timely mitotic entry under temperature stress conditions prompted us to test if Rim15 and Igo proteins regulate Cdk1 phosphorylation on Tyr19 . To test if the levels of Tyr19 phosphorylation on Cdk1 were misregulated in the absence of Rim15 or Igo proteins , wild type , igo1Δ igo2Δ and rim15Δ mutant cells , as well as cdc55Δ cells used as control , were arrested in G1 and released in the presence of the microtubule-depolymerizer nocodazole at 25°C , i . e . at a temperature where RIM15 and IGO1/2 are not required for mitotic entry ( data not shown ) . The phosphorylation status of Cdk1 was monitored by western analysis using a phospho-specific antibody that recognizes phosphorylated Cdk1-Y19 . In agreement with previous reports [36] , [39] , [43] , phosphorylation of Cdk1-Y19 was barely detectable in wild-type cells and high in cdc55Δ mutant cells under these conditions ( Fig . 4D ) . Similarly , Cdk1-Y19 phosphorylation was increased in igo1Δ igo2Δor rim15Δ mutant cells relative to wild type cells and was fully abolished by SWE1 deletion ( Fig . 4D ) . Similar results were obtained under conditions of thermal stress ( data not shown ) . Thus , Rim15 and Igo proteins are required for timely Cdk1 dephosphorylation , even in conditions where they do not appear to regulate mitosis . We therefore analysed the levels and phosphorylation state of Swe1 during the cell cycle . Swe1 is phosphorylated in early mitosis by Clb2-Cdk1 , which stimulates Swe1's ability to bind , phosphorylate and inhibit mitotic CDKs [6] . This initial Swe1 phosphorylation is opposed by PP2ACdc55 [35] . Later on during mitosis , Swe1 gets hyperphosphorylated and eventually degraded [13] , thus feeding the positive feedback loop . We assayed Swe1 levels and phosphorylation in synchronized rim15Δ and igo1Δ igo2Δ cells released from G1 in the presence of nocodazole at 25°C . As shown in Fig . S3A , HA-tagged Swe1 ( Swe1-HA3 ) accumulated at intermediate phosphorylation levels and got more slowly hyperphosphorylated in the absence than in the presence of Rim15 ( Fig . S3A ) . Similar data were obtained in cells lacking Igo1 and Igo2 ( data not shown ) . Delayed appearance of Swe1 hyperphosphorylated forms did not seem to be further affected by deletion of RIM15 or IGO1 and IGO2 upon release of G1 cells at the restrictive temperature of 38°C ( Fig . S3B ) . Thus , Rim15 and Igo proteins contribute , like PP2ACdc55 , to timely Swe1 phosphorylation even in conditions where these proteins are apparently not required for timely mitotic entry . Cdk1-Y19 phosphorylation is reversed by the Mih1 phosphatase , which in turn undergoes a PP2ACdc55-dependent dephosphorylation that can be visualized by an increase in its electrophoretic mobility and correlates with mitotic entry [36] . We therefore asked if the levels of Mih1 phosphorylation are affected by deletion of RIM15 or IGO1 and IGO2 . To this purpose , we expressed HA-tagged Mih1 ( Mih1-HA3 ) in wild type , rim15Δ and igo1Δ igo2Δ mutant cells , as well as cdc55Δ cells used as control , and analysed Mih1 phosphorylation by western blot . Consistent with a previous report [36] , most Mih1 was present in the cells in phosphorylated forms ( Fig . 4E–F ) . Deletion of RIM15 or IGO1 and IGO2 led to accumulation of hyperphosphorylated forms already at 25°C , albeit not to the same levels as deletion of CDC55 ( Fig . 4E ) . To analyse the transient appearance of the dephosphorylated , and presumably active , form of Mih1 during the cell cycle , wild type , rim15Δ and igo1Δ igo2Δ cells were arrested in G1 by alpha factor and released in the cell cycle at 38°C . Whereas dephosphorylated Mih1-HA3 started accumulating in wild type cells at 70–80 minutes after the release , coincident with the time of mitotic entry , its appearance was delayed in the absence of Rim15 or Igo proteins ( Fig . 4F ) , in agreement with reduced PP2ACdc55 activity . Thus , the phosphorylation state of both Swe1 and Mih1 is affected by deletion of RIM15 or IGO1/2 . Consistent with a role of Rim15 and Igo proteins in the regulation of Cdk1 phosphorylation , genetic analyses showed that lack of Swe1 fully rescued the temperature-sensitive growth defect of igo1Δ igo2Δ and rim15Δ mutant cells at 38°C ( Fig . 5A ) . Expression of non-phosphorylatable Cdc28-Y19F also rescued the cold- and temperature-sensitivity of igo1Δ igo2Δ and rim15Δ mutants , although somewhat less efficiently than SWE1 deletion ( Fig . S4 ) . Finally , MIH1 deletion enhanced the temperature-sensitivity of igo1Δ igo2Δ cells at 37°C and caused a cold-sensitive growth phenotype to rim15Δ cells at 16°C ( it should be noted that igo1Δ igo2Δ cells are sensitive to this temperature ontheirown , Fig . 5B ) . We then asked if SWE1 deletion could rescue the mitotic entry delay of igo1Δ igo2Δ under temperature stress . To address this question , we arrested wild-type , igo1Δ igo2Δ and igo1Δ igo2Δswe1Δ cells in G1 and released them into fresh medium at 38°C . Under these conditions , igo1Δ igo2Δ cells showed a marked delay in the accumulation of Clb2 and Cdc5 and in spindle elongation relative to wild type cells ( Fig . 3A and 5C ) . Strikingly , SWE1 deletion suppressed these mitotic defects and restored the normal kinetics of Clb2 and Cdc5 accumulation during the cell cycle ( Fig . 5C ) . Altogether , these data suggest that Rim15 and Igo proteins regulate the phosphorylation state of Cdk1 for mitotic entry , presumably by affecting the PP2ACdc55–dependent regulation of Swe1 and Mih1 , and indicate that excessive Cdk1 inhibitory phosphorylation is responsible for the mitotic delay of igo1Δ igo2Δ cells . If budding yeast endosulfines were required for full PP2ACdc55 phosphatase activity , deletion of RIM15 or IGO1 and IGO2 should rescue the toxic effects caused by CDC55 overexpression , which prevents mitotic progression [48] . Consistent with our previous findings , cells carrying four integrated copies of a galactose-inducible GAL1-CDC55 construct were mostly unable to divide on galactose-containing plates ( YPGal , Fig . 5D ) . Loss of Rim15 or Igo1/2 efficiently suppressed this lethality ( Fig . 5D ) without affecting the levels of CDC55 overexpression ( data not shown ) . Altogether , these data indicate that Rim15 and Igo proteins contribute to activation of PP2ACdc55 , at least for what concerns some of its mitotic targets , thereby tuning the levels of Cdk1 phosphorylation to promote mitotic entry . The apparently opposite impact of Igo proteins on PP2A regulation in vitro and in vivo ( i . e . inhibition versus activation , respectively ) , raised the possibility that in yeast Igo1/2 modulate PP2A activity at a different/additional level . For instance , the Zds1 and Zds2 proteins , which bind PP2ACdc55 in a stoichiometric complex and inhibit its activity in vitro [49] , [50] , were recently shown to regulate the subcellular localization of Cdc55 . More specifically , Zds1/2 induce the nuclear export of Cdc55 into the cytoplasm , which in turn promotes mitotic entry [43] . We therefore asked if Rim15 and Igo proteins might play a similar role in controlling Cdc55 localization . HA-tagged Cdc55 was detected by indirect immunofluorescence in wild type and igo1Δ igo2Δ cells at various cell cycle stages , namely in G1 ( unbudded cells ) , in S , G2 and early M phases ( budded mononucleated cells ) and in middle/late M phase ( budded binucleated cells ) . As previously shown [43] , Cdc55 was localized in both the nucleus and the cytoplasm . However , it was significantly more concentrated in the nucleus of rim15Δand igo1Δ igo2Δ cells than in the wild type in all cell cycle stages ( Fig . 6A ) . Strikingly , deletion of SWE1 restored the normal nuclear/cytoplasmic ratio of Cdc55 ( Fig . 6A ) , strongly indicating that the altered subcellular distribution of Cdc55 in mutants lacking Rim15 or Igo proteins is a consequence , rather than a cause , of misregulated positive feedback loop for Cdk1 activation . The similar phenotype of igo1Δ igo2Δ and zds1Δ zds2Δ cells with respect to Cdc55 localization raised the possibility that Igo and Zds proteins work in concert and/or are part of the same complex . We therefore decided to analyse the possible interdependence between Igo and Zds proteins for their interaction with Cdc55 . As previously shown , Pk-tagged Zds1 efficiently co-immunoprecipitated HA-tagged Cdc55 from extracts obtained from cycling cells . Interaction between Zds1 and Cdc55 was not affected by deletion of IGO1 and IGO2 ( Fig . 6B ) . Similarly , co-immunoprecipitation of HA-tagged Cdc55 with Pk-tagged Igo1 was unaffected by deletion of ZDS1 and ZDS2 . It is worth noting that in zds1Δ zds2Δ cells the mobility shift of Cdc55 , which was previously shown to be due to phosphorylation [47] , disappears , indicating another possible way for Zds proteins to regulate PP2ACdc55 activity besides its subcellular localization . Altogether , these data suggest that Igo and Zds proteins bind independently to PP2ACdc55 and might regulate its activity in an independent manner . Consistent with this conclusion , deletion of IGO1 and IGO2 caused synthetic sickness at high temperatures when combined with deletion of ZDS1 and ZDS2 ( Fig . S5 ) .
In several organisms endosulfine-like proteins bind to PP2A-B55 upon Greatwall-dependent phosphorylation of a conserved serine and inhibit it [33] , [34] . Recent data showed that in quiescent yeast cells Rim15 phosphorylates the paralogous endosulfines Igo1 and Igo2 , which in turn promote the transcription of specific nutrient-regulated genes by direct inhibition of the phosphatase PP2ACdc55 [41] . We show here that Igo1 phosphorylated on Ser64 by Rim15 interacts also during the unperturbed cell cycle with the catalytic subunit of PP2A ( Pph21 ) , the B regulatory subunit Cdc55 , Tap42 and Rts3 and that phosphorylated Igo1 can bind PP2A complexes in yeast and Xenopus egg extracts . PP2A has multiple functions during the cell cycle ( reviewed in [20] ) . In yeast it associates with two major and alternative B subunits , Cdc55 and Rts1 . Rts1 does not appear to interact with Igo1 in our co-immunoprecipitations . Cdc55 has been involved in many cellular processes , such as mitotic entry and exit , morphogenesis and cytokinesis , mitotic checkpoints and stress response [35] , [36] , [38] , [39] , [48] , [51] , [52] , [53] , [54] , [55] , [56] . Of particular interest is the presence in our Igo1 immunoprecipitates of the essential Tap42 subunit , the orthologue of human PP2A-associated α4/IgBP1 . Tap42 associates to the catalytic subunit of PP2A and PP2A-like phosphatases independently of the A and B subunit during logarithmic growth as opposed to stationary phase , suggesting that its interaction with PP2A phosphatases is regulated by nutrients [57] . Consistently , Tap42 phosphorylation , which mediates its interaction with PP2As , depends on the phosphatidylinositol-related kinases Tor1 and Tor2 ( Target of rapamycin ) , which regulate cell growth in response to nutrient availability and cell stress [58] . The last PP2A subunit that we found in Igo1 immunoprecipitates is Rts3 , a poorly characterized protein that was found to interact with different PP2A and PP2A-like complexes [45] , [59] , [60] and whose deletion causes sensitivity to caffeine [61] , which in turn inhibits the Tor complex TORC1 [62] . The peak of interaction between Igo1 and Cdc55 is cell cycle-regulated and peaks in late S or G2 phase ( note that it is not possible to discriminate between late S and G2 phase in budding yeast due to the lack of specific markers ) , i . e . when PP2ACdc55 participates to the positive feedback loop for Cdk1 activation ( see below ) . In agreement with recently published data [41] , we show that mutation of Igo1 Ser64 , which is targeted by Rim15 [40] , and deletion of RIM15 reduce significantly Igo1 interaction with PP2ACdc55 , indicating that , like in higher organisms , endosulfine phosphorylation is required for its interaction with PP2A . In spite of its cell cycle-regulated interaction with PP2ACdc55 , Igo1 phosphorylation on Ser64 appears to be constitutive during the cell cycle , raising the interesting possibility that cell cycle-controlled factors are involved to make the Igo1-Cdc55 binding periodic . Along the same line , it is also interesting to notice that phosphorylation of Ser64 of Igo1 and Ser63 of Igo2 is stimulated after inhibition of Cdk1 [63] , suggesting that Cdk1 activation in early mitosis may cause the dissociation of Igo-PP2ACdc55 complexes . The evolutionary conservation of the Greatwall-Endosulfine pathway is further strengthened by the finding that yeast Rim15 can phosphorylate in vitro human ENSA and Arpp19 [40] , whereas human and Xenopus Greatwall can phosphorylate yeast Igo1 ( this manuscript ) , indicating that Rim15 and Igo proteins are the true counterparts of vertebrate Greatwall and endosulfines , respectively . Thus , the Greatwall-Endosulfine-PP2A regulatory module appears to be conserved in all organisms analysed to date , with the notable exception of the nematode C . elegans where an obvious Greatwall-like kinase seems to be missing [42] . ENSA and Arpp19 inhibit PP2A activity in Xenopus egg extracts [30] , [31] . Inhibition of PP2A by endosulfines is in turn essential for mitotic entry in Xenopus and human cells and for mitotic progression in Drosophila [23] , [27] , [29] , [30] , [31] , [64] . Similarly , we and others [41] find that phosphorylated Igo1 inhibit PP2A activity in vitro and induces interphase Xenopus egg extracts to enter mitosis , indicating that budding yeast endosulfines can inhibit PP2A , like their vertebrate counterparts . PP2ACdc55 inhibition in rapamycin-treated yeast cells is important to establish a quiescence-specific transcriptional program [41] . We provide several lines of evidence indicating that Rim15 and Igo proteins also contribute to activate PP2ACdc55 for mitotic entry in vivo . First , lack of Igo1 and Igo2 causes a slight reduction , rather than an increase , in PP2ACdc55 phosphatase activity . Taking into account that only about 1% of Cdc55 is bound to Igo1 ( and presumably a similar fraction of Cdc55 is bound to Igo2 ) , the 15–20% reduction in PP2ACdc55 activity in cells lacking Igo proteins would imply a prominent role for this proteins in the full activation of PP2ACdc55 complexes . Second , deletion of RIM15 or IGO1 and IGO2 delays accumulation of active cyclinB/Cdk1 and dephosphorylation of Cdk1 Tyr19 , similarly to inactivation of PP2ACdc55 [35] , [39] . This is accompanied by a misregulation in the phosphorylation of the Swe1 kinase and the dephosphorylation of the Mih1 phosphatase during the cell cycle , which are both controlled by PP2ACdc55 [35] , [36] . Third , the toxic effects caused by CDC55 overexpression [48] are rescued by RIM15 or IGO1 and IGO2 deletion . Yet , not only phosphorylated Igo1 can inhibit PP2A activity in vitro , but expression of human ENSA or Arpp19 rescues the temperature-sensitivity of igo1Δ igo2Δ cells , suggesting that vertebrate and yeast endosulfines are interchangeable for their mitotic function ( s ) . Therefore , the apparently opposite modes of PP2A regulation by endosulfines in yeast versus vertebrates are unlikely linked to intrinsic differences in the structure of endosulfines and/or in their ability to bind and inhibit PP2A . Several other examples of proteins that behave as activators or inhibitors of their binding partner ( s ) depending on the organism and/or the conditions have been reported . Similarly to Igo1 and Igo2 , budding yeast Zds proteins were shown to inhibit PP2ACdc55 phosphatase activity [49] and to promote efficient mitotic entry through nuclear export of Cdc55 and Cdc55-dependent regulation of the Cdk1 positive feedback loop [43] , [50] , [65] . Tap42 has been proposed to work as activator and inhibitor of PP2A [57] , [66] , [67] , [68] . Finally , securin is both an inhibitor and a chaperone of separase for the regulation of sister chromatid splitting in anaphase . Depending on the organism , securin depletion/inactivation prevents sister chromatid separation or causes premature anaphase onset ( reviewed in [69] ) . We speculate that Igo proteins might be chaperones for PP2ACdc55 and assist its proper folding while keeping the complex inhibited . The presence of phosphorylated substrates could competitively bind to PP2ACdc55 and overcome its inhibition . The exact biochemical mechanism of PP2A-B55 inhibition by endosulfines is currently unknown and will be likely clarified by structural and biochemical data . Consistent with a positive role of Igo proteins in PP2ACdc55 regulation for mitosis , we find that the phosphatase activity of PP2ACdc55 complexes in igo1Δ igo2Δ cells is reduced compared to wild type cells . This results in delayed phosphorylation of Swe1 and Mih1 , which are known targets of PP2ACdc55 [35] , [36] , thus preventing the sharp rise in CDK activity during mitotic entry . Indeed , rim15Δ and igo1Δ igo2Δ cells are unable to timely dephosphorylate Tyr19 of Cdk1 and to efficiently raise cyclinB/Cdk1 activity in mitosis , resulting in a delayed spindle elongation , which requires high cyclinB/Cdk1 levels [70] . The observation that rim15Δ and igo1Δ igo2Δ mutants display lower activity of PP2ACdc55 complexes , nuclear retention of Cdc55 and accumulation of Tyr19-phosphorylated Cdk1 at physiological temperatures ( e . g . 25°C ) , while progressing through mitosis with normal timing , is somewhat puzzling . Since mitotic defects of rim15Δ and igo1Δ igo2Δ mutants become apparent only under stress conditions , it is possible that the physiological state of cells can make them differentially responsive to increased levels of Tyr19-phosphorylated Cdk1 . Deletion of SWE1 or expression of the non-phosphorylatable Cdc28-Y19F mutant protein rescues the temperature-sensitive growth and mitotic defects of rim15Δ and igo1Δ igo2Δ cells , whereas deletion of MIH1 aggravates them . The role of PP2ACdc55 in the nucleolar retention and inhibition of Cdc14 [52] , which is the major CDK-counteracting phosphatase in yeast [15] , might also be affected by lack of endosulfines and contribute to timely mitotic entry . The activity of PP2ACdc55 has been recently proposed to be spatially regulated by nuclear export of Cdc55 , which is regulated by Zds1 and Zds2 and promotes mitotic entry [43] . In this respect it is important to notice that a key determinant in Swe1 downregulation is its proteolysis , which requires Swe1 recruitment to the bud neck ( i . e . out of the nucleus ) [71] , [72] . Thus , the cytoplasmic pool of PP2ACdc55 might be instrumental to trigger efficient Swe1 degradation [39] . We show that the nuclear/cytoplasmic ratio of Cdc55 is increased throughout the cell cycle upon deletion of RIM15 or IGO1 and IGO2 . The increased concentration of Cdc55 in the nuclei of rim15Δ and igo1Δ igo2Δ cells might be a consequence of decreased PP2ACdc55 activity or be caused by a more direct function of yeast endosulfines in the control of Cdc55 nuclear-cytoplasmic shuttling . Our finding that SWE1 deletion restores proper subcellular localization of Cdc55 in the absence of Rim15 and Igo proteins argues against the latter possibility . We thus favor the idea that yeast endosulfines only impact on PP2ACdc55 activity , participating in the positive feedback loop for Cdk1 activation like their Xenopus counterparts [27] . CDK activity could in turn drive , directly or indirectly , Cdc55 export from the nucleus to the cytoplasm ( Fig . 7 ) . It is nevertheless possible that nuclear retention of Cdc55 in cells lacking Rim15 or endosulfines has a greater impact than the slight decrease in PP2A activity by itself on dephosphorylation of PP2ACdc55 substrates , including Swe1 and Mih1 . In fact , Cdc55 mislocalization could keep PP2ACdc55 spatially segregated from its substrates . Further experiments will be required to understand the links between PP2A activity and its subcellular localization . Similar to endosulfines , the Zds1 and Zds2 proteins bind tightly to PP2ACdc55 [49] , [50] , affect its subcellular localization [43] and target PP2ACdc55 activity to Mih1 [65] , thus promoting timely mitotic entry [43] , [50] , [65] . In spite of the similarity , Igo and Zds proteins seem to work independently . Indeed , binding of Igo1 to Cdc55 does not require Zds1/2 , whereas Zds1 interaction with Cdc55 does not require Igo proteins . In addition , deletion of IGO1 and IGO2 causes synthetic growth defects when combined to deletion of ZDS1 and ZDS2 . In conclusion , our data emphasize the plasticity of the Greatwall-Endosulfine-PP2A module . This module is conserved , but it has adapted to account for the differences in mitotic regulation depending on the cellular context . In cells where PP2A antagonizes Cdk1 activity the Greatwall-endosulfine module only inhibits PP2A , whereas in budding yeast where PP2A promotes timely mitotic entry the same module also stimulates PP2A activity , resulting in both cases in a sharp mitotic switch [73] .
All yeast strains ( Table S1 ) were derivatives of W303 ( ade2-1 , trp1-1 , leu2-3 , 112 , his3-11 , 15 , ura3 , ssd1 ) , except for strains used for phosphatase assays that were derivatives of S288C . Cells were grown in either synthetic minimal medium ( SD ) supplemented with the appropriate nutrients or YEP ( 1% yeast extract , 2% bactopeptone , 50 mg/l adenine ) medium supplemented with 2% glucose ( YEPD ) or 2% galactose ( YEPG ) . Unless differently stated , alpha factor was used at 4 µg/ml , hydroxyurea ( HU ) at 200 mM and nocodazole at 15 µg/ml . Standard techniques were used for genetic manipulations [74] , [75] . Gene deletions were generated by one-step gene replacement [76] . One-step tagging techniques [77] , [78] were used to tag at their C-terminus Igo1 ( Igo1-Pk3 ) , Swe1 ( Swe1-HA3 ) and Mih1 ( Mih1-HA3 ) . In situ immunofluorescence was performed on formaldehyde-fixed cells expressing HA-tagged Cdc55 using anti-HA monoclonal antibody ( 16B12 , Covance Research Products ) , followed by indirect immunofluorescence using Cy3-conjugated goat anti-mouse antibody ( GE Healthcare ) . To detect spindle formation and elongation , anti-tubulin immunostaining was performed with the YOL34 monoclonal antibody ( Serotec ) followed by indirect immunofluorescence using rhodamine-conjugated anti-rat antibody ( Pierce Chemical Co ) . Digital images were taken with an oil 63X 1 , 4-0 , 6 HCX Plan-Apochromat objective ( Zeiss ) with a Coolsnap HQ2-1 charge-coupled device camera ( Photometrics ) mounted on a Zeiss AxioimagerZ1/Apotome fluorescence microscope controlled by the MetaMorph imaging system software . Fluorescence intensity of HA3-Cdc55 in the nucleus and the cytoplasm was quantified with ImageJ on a single focal plane . Significance of the differences between fluorescence intensities was statistically tested by means of a two-tailed t-test , assuming unequal variances . TCA protein extracts were prepared as previously described [79] for Phos-tag phosphate affinity gel electrophoresis ( Wako , Fig . 1G ) and to analyse the electrophoretic mobility of HA-tagged Swe1 and Mih1 . For immunoprecipitations of Igo1-Pk3 or Zds1-Pk6 , pellets from 50 ml yeast cultures ( 107 cells/ml ) were lysed at 4°C with acid-washed glass beads in lysis buffer ( 50 mM Tris-Cl pH 7 . 5 , NaCl 150 mM , 10% glycerol , 1 mM EDTA , 1% NP40 , supplemented with protein inhibitors ( Complete , Roche ) , 1 mM Na-orthovanadate and 60 mM ß-glycero-phosphate ) . Total extracts were cleared by spinning at 12000 rpm for 10 minutes and quantified by NanoDrop . Same amounts of protein extracts were subjected to immunoprecipitation with an anti-Pk antibody ( from AbD serotec ) pre-adsorbed to protein A-sepharose . Inputs represent 1/50th of the IPs final volumes . For Clb2/Cdk1 kinase assays protein extracts were prepared as previously described [80] . 50 µg of extract were used for measuring kinase activity on histone H1 [81] and 30 µg for Western blot analysis . Hyperactive human Gwl ( Gwl-K72M ) purified from baculovirus-infected cells or endogenous Gwl immunoprecipitated from CSF-treated Xenopus egg extracts [30] were used to phosphorylate in vitro GST-Igo1 purified from E . coli for the in vitro phosphorylation assay ( Fig . 2A and Fig . S1B ) and the mitotic entry experiment ( Fig . 2C ) , respectively . Yeast protein extracts were prepared according to [82] for western blot analysis . Proteins transferred to Protran membranes ( Schleicher and Schuell ) were probed with monoclonal anti-HA 12CA5 , anti-Pgk1 ( Molecular Probes ) , anti-Pk ( AbD serotec ) , anti-Myc 9E10 or polyclonal Cdc5 antibodies ( sc-6733 Santa Cruz ) , anti-Clb2 ( sc-9071 Santa Cruz ) or anti-phospho-cdc2 ( pTyr15 Cell Signalling ) . Anti-Tap42 , anti-Pph21 and anti-Rts3 antibodies were described [47] . Affinity-purified antibodies against Gwl and Cdc25 were also previously described [30] . Monoclonal anti-PP2A/C subunit ( 1D6 ) and anti-PP2A/A ( 6G3 ) antibodies were obtained from Upstate/Millipore and Cell Signalling , respectively . Polyclonal antibodies against PP2A/B55 were raised in rabbits and affinity-purified . Secondary antibodies were purchased from Amersham and proteins were detected by an enhanced chemiluminescence system according to the manufacturer . For immunoprecipitation assays , yeast whole-cell extracts were prepared as described previously [83] except that lysis was performed using a Fastprep ( MP Biomedicals , 1×40 s , 6 m/s ) . HA-tagged Cdc55 containing a flexible glycine linker ( GL ) after the HA epitope ( GGGSGGGGS ) was expressed from the strong constitutive TPI1 promoter on an episomal centromeric plasmid [47] . HA-GL-Cdc55 was immunoprecipitated with anti-HA ( clone 12CA5 ) antibodies cross-linked to BSA-coated protein A–Sepharose beads ( GE Healthcare ) . Phosphatase activity of PP2A immunoprecipitates was assayed toward 32P-labeled phosphorylase a ( Figs . 2B , 4A ) or towards 32P-labeled histone H1 ( Fig . 4B ) , as previously described [47] , [83] . Immunoprecipitates were analyzed by 10% SDS-PAGE , immunoblotted and incubated with specific antibodies against Pph21 ( rabbit pAB ) , Cdc55 ( clone 9D3-H6 ) or Tpd3 ( clone 5G2 ) . The assay values ( average of at least five independent experiments ) are presented as a percentage of the wild-type strain activity , which was set at 100% . Values were normalized to the amount of Pph21 co-immunoprecipitated with HA-Cdc55 as determined by immunoblot and densitometer analysis using an Odyssey Infrared Imaging System ( LI-COR , http://www . licor . com ) . For Figure 2B immunoprecipitates were split into 7 aliquots and in-vitro Greatwall-phosphorylated recombinant GST-Igo1 or GST-Igo1S64A ( where kinase assays were carried out in 50 mM Tris pH7 . 2 , 10 mM MgCl2 , 1 mM ATP , 30°C for 30′ ) was added as indicated and phosphatase activity assays toward 32P-labeled phosphorylase a were conducted . The activities of the different samples are presented as percent of activity with respect to the activity of the untreated control , which was set to 100% . Data are presented as mean values ± standard deviation ( SD ) , and were analyzed using two-tailed Student's t-test , assuming unequal variances . Differences with p-values lower than 0 . 05 were considered statistically significant ( * p<0 . 05 ; ** p<0 . 01 ; *** p<0 . 001 ) . Nuclear division was scored with a fluorescence microscope on cells stained with propidium iodide ( Sigma Aldrich ) . Flow cytometric DNA quantification was performed according to [84] on a Becton-Dickinson FACSCalibur .
|
In all eukaryotic cells chromosome partition during mitosis requires a number of processes , including the formation of the mitotic spindle , i . e . the machinery that drives chromosome segregation to the daughter cells . Mitotic entry requires a delicate balance between protein phosphorylation , driven by cyclin-dependent kinases ( CDKs ) , and protein dephosphorylation , carried out by specific phosphatases that counteract CDK activity . A critical threshold in CDK activity is indeed required for mitotic entry . In the past few years the Greatwall kinase has also been implicated in mitotic entry through phosphorylation of proteins of the endosulfine family , which in turn inhibit the activity of the PP2A phosphatase that would otherwise dephosphorylate CDK targets . Whether Greatwall and endosulfines have a mitotic function in budding yeast , where PP2A promotes , rather than inhibits , mitotic entry has not been established . Here we show that the Greatwall-endosulfine-PP2A regulatory module is conserved also in budding yeast and that endosulfines from different species are interchangeable for their mitotic function . However , in budding yeast cells endosulfines contribute to full activation and proper localization of PP2A , suggesting that they act as both inhibitors and activators of PP2A . Our data emphasize how the same regulatory module is adapted to meet specific mitotic features in different organisms .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"molecular",
"cell",
"biology",
"genetics",
"biology",
"model",
"organisms"
] |
2013
|
Budding Yeast Greatwall and Endosulfines Control Activity and Spatial Regulation of PP2ACdc55 for Timely Mitotic Progression
|
Vascular endothelial cells act as gatekeepers that protect underlying tissue from blood-borne toxins and pathogens . Nevertheless , endothelial cells are able to internalize large fibrin clots and apoptotic debris from the bloodstream , although the precise mechanism of such phagocytosis-like uptake is unknown . We show that cultured primary human endothelial cells ( HUVEC ) internalize both pathogenic and non-pathogenic Listeria bacteria comparably , in a phagocytosis-like process . In contrast with previously studied host cell types , including intestinal epithelial cells and hepatocytes , we find that endothelial internalization of Listeria is independent of all known pathogenic bacterial surface proteins . Consequently , we exploited the internalization and intracellular replication of L . monocytogenes to identify distinct host cell factors that regulate phagocytosis-like uptake in HUVEC . Using siRNA screening and subsequent genetic and pharmacologic perturbations , we determined that endothelial infectivity was modulated by cytoskeletal proteins that normally modulate global architectural changes , including phosphoinositide-3-kinase , focal adhesions , and the small GTPase Rho . We found that Rho kinase ( ROCK ) is acutely necessary for adhesion of Listeria to endothelial cells , whereas the actin-nucleating formins FHOD1 and FMNL3 specifically regulate internalization of bacteria as well as inert beads , demonstrating that formins regulate endothelial phagocytosis-like uptake independent of the specific cargo . Finally , we found that neither ROCK nor formins were required for macrophage phagocytosis of L . monocytogenes , suggesting that endothelial cells have distinct requirements for bacterial internalization from those of classical professional phagocytes . Our results identify a novel pathway for L . monocytogenes uptake by human host cells , indicating that this wily pathogen can invade a variety of tissues by using a surprisingly diverse suite of distinct uptake mechanisms that operate differentially in different host cell types .
Vascular endothelial cells prevent free movement of material from the bloodstream into underlying tissues by tight regulation of cellular internalization pathways and robust cell-cell junctions . Nonetheless , in a process termed “angiophagy” , endothelial cells lining small-diameter capillaries in the brain , heart , lung , and kidney have been shown to internalize large fibrin or cholesterol clots that are subsequently released into the underlying parenchyma[1 , 2] . Furthermore , light and electron microscopy studies have established that liver endothelial cells can internalize apoptotic particles and latex beads in situ[3 , 4] . While this phenomenon is well documented , its molecular regulators have not been elucidated , making it difficult to establish a bona fide role for angiophagy in vivo . Additionally , it is unclear whether an endothelial phagocytosis-like process could be exploited by pathogens to access underlying tissue . The food-borne bacterium Listeria monocytogenes can disseminate from the initial site of infection at the intestinal epithelium to cause meningitis , encephalitis , sepsis , and spontaneous abortion by crossing different types of vascular endothelia[5] . In fact , L . monocytogenes infects human endothelial cells themselves in vivo[6] , but the mechanism of such infection is unknown . L . monocytogenes can directly invade intestinal epithelial cells and hepatocytes , using the bacterial surface proteins internalin[7 , 8] ( InlA ) and InlB [9 , 10] , respectively , which interact with host cell proteins . Once internalized into a membrane-bound compartment , L . monocytogenes expresses the pore-forming toxin listeriolysin O ( LLO ) , which promotes release of the bacterium into the cytosol , where it replicates[11 , 12] . Previous studies have conflictingly suggested that invasion of endothelial cells in culture requires InlA[13] , InlB[14 , 15] or neither[16 , 17] . We therefore sought to clarify whether L . monocytogenes uses internalins to invade endothelial cells or , alternatively , might use a distinct pathway , perhaps an angiophagy- or phagocytosis-like process , i . e . a process by which the bacterium does not trigger its own uptake through specific molecular recognition between its own surface proteins and those of the host cell . If L . monocytogenes exploits phagocytosis-like uptake in endothelial cells , then identifying regulators of L . monocytogenes entry may not only elucidate the myriad strategies of this model bacterial pathogen but may also provide mechanistic insight into how other large objects , such as stroke-causing clots in small-diameter blood vessels , are internalized by endothelial cells . We examined L . monocytogenes infection in human umbilical vein endothelial cells ( HUVEC ) , as these are human primary cells that are amenable to physical and genetic perturbation . We found that invasion was independent of pathogenic bacterial factors , suggesting that L . monocytogenes does indeed exploit a phagocytosis-like process for entry . We perturbed host cell signaling to identify specific regulators of such entry and determined that adhesion of L . monocytogenes to HUVEC requires the activity of the Rho GTPase effector kinase ROCK , and that efficiency of subsequent internalization was modulated by signaling from cell-substrate adhesions and by the formin family of actin nucleators . Furthermore , we found that these same regulators modulate phagocytosis-like uptake of non-pathogenic bacteria by HUVEC , but do not dramatically affect macrophage phagocytosis of L . monocytogenes . Our results demonstrate that endothelial cells internalize L . monocytogenes using a mechanism that is distinct from that employed by epithelial cells , hepatocytes , or professional phagocytes . Furthermore , endothelial phagocytosis-like uptake may be a previously unappreciated mechanism for systemic spread of pathogenic bacteria and viruses and for modulation of traffic from the bloodstream to the underlying parenchyma .
A number of cell types , including PtK2[18] , MDCK[19 , 20] , mouse embryonic fibroblasts[21] and L2 cells[22] , can tolerate exposure to high titers of L . monocytogenes ( >100 bacteria per host cell ) in culture; however , we found that exposing HUVEC to L . monocytogenes under such conditions resulted in dramatic and highly variable HUVEC death ( S1A–S1D Fig ) . Neither the closely related bacterium Listeria innocua , which lacks the pathogenic apparatus of L . monocytogenes[23] , nor an L . monocytogenes strain lacking the pore-forming toxin LLO ( hly mutant , JAT314 ) caused HUVEC death ( S1D Fig ) . Indeed , purified 6-His-LLO[24 , 25] induced early HUVEC death at low concentrations ( S1F Fig ) . Notably , monocyte-like U937 cells did not display increased death in response to either L . monocytogenes or to purified 6-His-LLO ( S1E and S1G Fig ) . These data collectively suggest that HUVEC are particularly sensitive to LLO and that extracellular LLO causes HUVEC death during initial exposure to high bacterial titers in culture . An LLO point mutant , LLOG486D ( JAT745 ) has previously been reported to exhibit decreased hemolysis relative to the wild-type protein , while still supporting bacterial escape from the phagocytic vacuole[26 , 27]; LLOG486D does not cause early cell death in HUVEC ( S1H Fig ) . To determine whether LLOG486D supported invasion and vacuolar escape in HUVEC , we constructed an LLOG486D strain ( LLOG486D actAp::mTagRFP , JAT983 ) that expressed RFP only when in the host cell cytoplasm[28]; we found that LLOG486D mutants could invade HUVEC and escape the vacuole ( Fig 1A ) . In most cell types , L . monocytogenes replicates in the cytoplasm and expresses the protein ActA , which activates the Arp2/3 complex to promote actin polymerization at the surface of the bacterium[29 , 30]; addition of new actin subunits at the bacterial surface pushes the bacterium forward[31] . When a moving bacterium reaches the cell membrane , it can spread from cell to cell by extending a long membrane-bound protrusion that can be taken up by an adjacent cell into a double-membraned vacuole , from which the bacterium can again escape[12 , 19] . We found that LLOG486D supported the ability of L . monocytogenes to move freely within cells ( compare S1 and S2 Movies with L . monocytogenes expressing wild-type LLO in S3 Movie ) and form bacterial protrusions that could extend from one endothelial cell and be internalized by an adjacent endothelial cell ( S4 and S5 Movies ) . To confirm that LLOG486D could propagate infection within an endothelial sheet , we employed the gentamicin protection assay , in which HUVEC were exposed to L . monocytogenes ( JAT983 ) and , after 1 hour , the antibiotic gentamicin was added to specifically kill extracellular bacteria[21 , 32]; subsequent infection could occur only by cell-to-cell spread . We found an exponential increase in frequency of infected cells as a function of time ( S2A , S2B and S2C Fig ) , indicating that L . monocytogenes expressing only LLOG486D could indeed spread from an infected cell to an uninfected cell , escape from the secondary vacuole , and replicate in the newly infected cell . To quantify the extent of cell-to-cell spread , we evaluated the size of clusters of adjacent infected cells , termed foci . These foci represent an initial uptake event in a single cell , followed by subsequent cell-to-cell spread to neighboring uninfected cells ( Fig 2A ) . The median focus remained stable for the first 6 hours of infection , then grew between 6 and 8 hours after infection , most likely representing the first successful cycle of cell-to cell-spread ( S2D Fig ) . The significant motility of HUVEC in culture ( S5 Movie ) tended to fragment foci after 8 hours , so continuous spread was most evident by tracking the size of the largest decile of foci ( S2D Fig ) . To quantify the contribution of cell-to-cell spread to overall infection of an endothelial sheet , we compared infection of the LLOG486D mutant ( JAT983 ) to an LLOG486D ΔactA mutant ( JAT985 ) , which cannot polymerize actin and , therefore , cannot move within or between cells ( S2E and S2F Fig ) . The number of foci , representing the number of distinct invasion events , was indistinguishable between JAT983 and JAT985 ( S2H Fig ) , as expected , given that ActA is primarily expressed by intracellular bacteria and is not involved in invasion[29 , 33 , 34] . Compared to the ActA-deficient strain , JAT983 exhibited lower bacterial density in infected cells ( S2I and S2J Fig ) and larger focus size ( S2K and S2L Fig ) , strongly suggesting that LLOG486D supports cell-to-cell spread . Notably , these larger foci likely contribute to the higher percentage of cells infected with JAT983 versus JAT985 ( S2G Fig ) . Collectively , these data demonstrate conclusively that LLOG486D supported invasion , vacuolar escape , actin-based motility , and cell-to-cell spread in HUVEC without causing early cell death . We therefore used this mutant for all subsequent experiments in HUVEC . Using the LLOG486D mutant ( JAT983 ) , we determined that HUVEC are highly susceptible to L . monocytogenes invasion; at a high multiplicity of infection ( MOI ) , more than 90% of HUVEC in a confluent monolayer harbored bacteria 8 hours after infection ( Fig 1B ) . The amount of HUVEC infection is strongly dependent on MOI; thus , minor variations in MOI may result in substantially different frequencies of infected cells . Surprisingly , an LLOG486D inlAB ( JAT1119 ) mutant exhibited comparable invasiveness to JAT983 across the entire range of MOI examined; thus , neither of the canonical bacterial invasion proteins that promote uptake by intestinal epithelial cells and hepatocytes is required for internalization of L . monocytogenes into HUVEC ( Fig 1B ) . Therefore , we suspected that either L . monocytogenes uses a different internalin-like protein to invade HUVEC or L . monocytogenes capitalizes on an intrinsic uptake mechanism in endothelial cells . To distinguish between these possibilities , we exposed HUVEC to L . innocua , which lacks most putative internalin family members and lacks all members with a known pathogenic role[23] , or to polystyrene beads , which lack all bacterial factors . HUVEC were comparably susceptible to L . monocytogenes and L . innocua ( Fig 1C and 1D ) . Surprisingly , HUVEC internalized polystyrene beads comparably to bacteria ( Fig 1E and 1F ) . Concurrent exposure to L . innocua did not alter the frequency of HUVEC that internalized beads , suggesting that bacterial factors neither are required for nor enhance phagocytosis-like uptake by HUVEC ( Fig 1F ) . Thus , L . monocytogenes likely exploits a generic constitutive uptake process in HUVEC without bacterial- or pathogen-specific requirements; such uptake may exhibit more similarity to a process like angiophagy or macrophage phagocytosis than to internalin-mediated invasion of epithelial cells[7 , 8] . To identify molecular regulators of endothelial phagocytosis-like uptake and L . monocytogenes infection , we performed a targeted siRNA screen , for which in vitro diced pools of siRNAs were generated , each targeting a distinct gene of interest ( S1 Table ) [35–37] . This method of generating complex siRNA pools , containing hundreds of different individual siRNAs , has been shown to reduce off-target effects often seen with single synthetic siRNAs by diluting the off-target effects of individual siRNAs in the pool[38] . We included genes that had previously been shown to modulate L . monocytogenes phagocytosis by macrophage-like Drosophila S2 cells[39 , 40] to compare that process to endothelial uptake . We also included components of cell-substrate adhesions and cell-cell contacts , as well as genes known to modulate collective motility , endocytic processes , intracellular trafficking , or membrane fusion . Endothelial monolayers were infected with JAT983 in a gentamicin protection assay[21 , 32] . In normal infection , images of infected monolayers reveal multiple infection foci ( Fig 2A and 2G ) . siRNA pools that specifically decrease uptake of bacteria should decrease the number of foci and the fraction of cells infected , but not focus size or the density of bacteria per infected cell ( Fig 2A and 2H ) . Pools that specifically decrease cell-to-cell spread should decrease focus size while increasing the density of bacteria per infected cell , without changing the number of foci ( Fig 2A and 2J ) . By combining multiple image-based metrics , distinct infection phenotypes may be extracted ( Fig 2B ) . Notably , siRNAs could affect endothelial cell density , for instance by decreasing endothelial cell viability; if endothelial density affects L . monocytogenes internalization or spread , these siRNAs would have indirect effects on infection , but would be classified as significant in the screen . To correct our morphological metrics of infection for effects from changes in host cell density , we infected HUVEC that had been plated at varying densities and found that the frequency of infected cells ( S3A Fig ) , bacterial density per infected cell ( S3C Fig ) , and the size of the largest quartile of foci ( S3D Fig ) were uncorrelated with endothelial cell density . In contrast , the number of foci was linearly correlated with endothelial cell density ( S3B Fig ) ; therefore , we used the density of foci ( number of foci divided by number of HUVEC ) to quantify invasion independent of host cell density . A number of siRNA pools caused phenotypes consistent with increased or decreased invasion ( Fig 2C and 2D ) , while far fewer altered cell-to-cell spread ( Fig 2E and 2F ) . To confirm that some siRNA pools specifically affected cell-to-cell spread , we examined the effect of 85 siRNAs from the original screen on infection of endothelial cells with an ActA-deficient strain ( JAT1045 ) that is incapable of cell-to-cell spread; we included siRNAs that exhibited increased bacterial density per infected cell ( siCAPZA2 , siACTR2 ) or decreased focus size ( siRACGAP1 , siSTX16 , siMAP1LC3A ) in the original screen , expecting that these siRNA pools should not have a significant phenotype in this assay . We analyzed infection by flow cytometry ( S4 Fig ) , which provided an orthogonal confirmation of the morphological metrics used in the initial screen . The candidates identified as likely to affect cell-to-cell spread in the original screen were not significantly different from controls in the ΔactA screen ( S2 Table ) , confirming this interpretation . Notably , siITGB1 significantly increased infection with the ActA-deficient strain ( S2 Table ) , consistent with its phenotype of increased bacterial invasion in the original screen ( Fig 2C , 2D and 2I ) . Given that most of the candidate factors that appear to be involved only in cell-to-cell spread did not exhibit invasion phenotypes in this follow-up screen , we suspect that bacterial uptake and cell-to-cell spread are likely differentially regulated processes in endothelial cells . We were surprised to find that depletion of Arp2 yielded a phenotype consistent with a defect exclusively in cell-to-cell spread ( Fig 2C , 2D , 2E and 2J and S2 Table ) , because previous studies have indicated that the Arp2/3 complex is the primary actin nucleator when L . monocytogenes invades epithelial cells and macrophages [40 , 41] . We confirmed this cell-to-cell spread-specific phenotype in HUVEC using synthetic siRNA pools that targeted distinct Arp2/3 subunits and successfully depleted the Arp complex ( S5 Fig ) . Our phenotype was consistent with the known role of Arp2/3 in promoting L . monocytogenes actin-based motility and cell-to-cell spread [30] but demonstrated that bacterial uptake in HUVEC likely requires less Arp2/3 activity . Local actin polymerization is required by many cell types to internalize micron-sized objects , such as bacteria [42–44] , and a subset of hits from the screen ( underlined in Fig 2C and 2D ) , including DLG1[45] , NCK1[46 , 47] , PFN1[48] , RAC2 , and MYO9A [49 , 50] , encode proteins that modulate actin assembly; depleting these proteins might alter the availability of cortical actin for local actin polymerization during bacterial internalization . Therefore , we examined whether actin polymerization during L . monocytogenes internalization by endothelial cells might be primarily controlled by the formin family of actin nucleators . Formin proteins contain multiple domains , including the formin homology-2 ( FH2 ) domain , which binds to actin filaments and promotes elongation , and the FH1 domain , which modulates the activity of the FH2 domain by interacting with the actin monomer-binding protein profilin[48] . In our siRNA screen , depletion of profilin ( PFN1 ) decreased the frequency of infected cells , consistent with an invasion defect ( Fig 2C and S6A Fig ) . A cell-permeable small molecule inhibitor of the FH2 domain ( SMIFH2 ) broadly inhibits formin- but not Arp2/3-mediated actin polymerization[51] . We exposed HUVEC to the drug either during or after uptake of L . monocytogenes and assayed infection by flow cytometry ( S4D Fig ) . HUVEC infection decreased when SMIFH2 was present during uptake ( Fig 3A ) but was comparable to the control when SMIFH2 was added after uptake ( Fig 3B ) . These results suggest that formins normally promote uptake of L . monocytogenes by HUVEC but that their activity is not essential for cell-to-cell spread . The human genome encodes 15 formins with distinct expression patterns , localizations , and functions[48 , 52] . Two of these formins , diaphanous-related formins 1 and 2 ( DIAPH1 and DIAPH2 ) , were examined in our original screen , but did not exhibit a significant phenotype ( S1 Table ) . To determine which formins were involved in uptake of L . monocytogenes by endothelial cells , we screened a targeted siRNA library that included each mammalian formin , and assayed infection by flow cytometry . We found that siRNAs targeting FHOD1 , FMNL3 , GRID2IP ( Delphilin ) , or INF2 exhibited significantly lower levels of L . monocytogenes infection than the control distribution ( Fig 3C ) . To confirm our results , we also examined infection by microscopy after depletion of FHOD1 , FMNL3 , GRID2IP or INF2; all four decreased bacterial uptake , though depletion of FMNL3 had the weakest effect ( Fig 3D ) . By quantitative reverse transcriptase PCR , we reliably amplified FHOD1 , FMNL3 , and INF2 in HUVEC , but did not detect expression of GRID2IP , and identical expression results have previously been reported for HUVEC and other endothelial primary cells [53] and in an endothelial-derived cell line ( The Human Protein Atlas [54 , 55] ) . We suspect that this protein is not expressed in HUVEC and may not play a significant role in infection . We confirmed that the siRNAs targeting FHOD1 and FMNL3 reliably depleted their target mRNAs , however the siRNA targeting INF2 minimally depleted INF2 mRNA ( S6B Fig ) . siRNAs targeting INF2 and GRID2IP did not decrease levels of FHOD1 or FMNL3 mRNAs ( S6C and S6D Fig ) , so their phenotype is most likely caused by other off-target effects . We therefore conclude that FHOD1 and FMNL3 , and not Arp2/3 , are the primary actin nucleators involved in internalization of L . monocytogenes by HUVEC . FMNL3 and FHOD1 modulate actin dynamics in a number of critical cellular processes; in particular , both have been shown to interact with or modulate focal adhesions[56–58] , large protein complexes that transduce mechanical and chemical signals between the cytoplasm and the extracellular matrix . Notably , the most robust invasion-specific hit in our screen came from siRNA pools targeting the focal adhesion protein integrin beta-1 ( ITGB1 ) , which increased the fraction of HUVEC infected and the density of foci in our original screen ( Fig 2C , 2D and 2I ) and also significantly increased infection of HUVEC with an ActA-deficient strain ( S2 Table ) . Focal adhesions have not previously been implicated in L . monocytogenes invasion in non-phagocytic cell types , and depletion of focal adhesion proteins did not alter phagocytosis of L . monocytogenes by macrophage-like S2 cells[39 , 40] . To complement siRNA experiments , which cause long-term depletion , we used small molecules to acutely perturb focal adhesions during bacterial uptake ( S4D Fig ) . Furthermore , such pharmacological perturbations do not share the same off-target effects as siRNAs and , in particular , are independent from the entire process of RNA interference . Therefore , as with formins , use of both pharmacological and siRNA perturbations could provide independent confirmation of the role of focal adhesions in L . monocytogenes internalization by HUVEC . MnCl2 , which promotes the formation of focal adhesions by activating integrins[59] , decreased uptake of L . monocytogenes ( Fig 4A ) . Treating HUVEC with the focal adhesion kinase ( FAK ) inhibitors FAK-14 or PF573228 increased the frequency of abnormally large adhesions ( S7 Fig ) , and therefore likely inhibited adhesion turnover . Both FAK inhibitors also inhibited uptake of L . monocytogenes in a dose-dependent manner ( Fig 4B ) . The siRNA pools targeting FAK in our screen failed to deplete FAK mRNA ( S6A Fig ) ; thus it is not surprising that they did not exhibit a significant phenotype in the screen ( S1 Table ) . Collectively , these data confirm that modulation of focal adhesions can inhibit uptake of L . monocytogenes by endothelial cells , as suggested by our siRNA screen ( Fig 2C and 2D ) . When endothelial cells are exposed to laminar shear ( as might result from fluid in the bloodstream ) , a signaling pathway initiated at the apical surface promotes phosphoinositide 3-kinase ( PI3K ) -dependent reinforcement of focal adhesions , which signal through the small GTPase RhoA to increase cellular stiffness and cell-substrate adhesion[60–62] . PI3K has been shown to regulate L . monocytogenes invasion in other cell types[63] , and our data demonstrate a clear role for FAK in promoting L . monocytogenes uptake in HUVEC . Therefore , we wondered if other elements of the shear-stress responsive pathway might be involved as well . Indeed , the PI3K inhibitors LY294002 and wortmannin both decreased uptake of L . monocytogenes in a dose-dependent manner ( Fig 4C and 4D ) . RhoA activity is decreased by GTPase activating proteins ( GAPs ) , which promote GTP hydrolysis , and is increased by guanine nucleotide exchange factors ( GEFs ) , which promote the exchange of GDP for GTP[64] . In our screen , the siRNA pool targeting the RhoGAP MyosinIX ( MYO9A ) increased the fraction of cells infected and the density of foci ( Fig 2C and 2D ) . While RhoA has been implicated in L . monocytogenes invasion of epithelial cells[65] , depletion of RhoA increased uptake of L . monocytogenes by S2 cells[40] , exactly the opposite of the result suggested by siMyo9A in our screen in HUVEC . Nascent focal adhesions inhibit Rho activity via p190RhoGAP ( ARHGAP5 ) [66]; however , FAK can also promote Rho activity via the RhoGEF GEF-H1[61 , 62] . To distinguish between these pathways , we exposed cells to siRNAs targeting p190RhoGAP or GEF-H1 . When p190RhoGAP was depleted , L . monocytogenes internalization by HUVEC was still decreased by FAK inhibition ( Fig 5A ) . In contrast , depleting GEF-H1 ( S6B Fig ) reduced the frequency of infected HUVEC , and FAK inhibition did not affect uptake of L . monocytogenes when GEF-H1 was depleted ( Fig 5A ) , indicating that GEF-H1 acts downstream of FAK in this pathway . Thus , we concluded that FAK signaling normally increases Rho activity via GEF-H1 to promote uptake of L . monocytogenes . When we acutely treated cells with Y27632[67] , which inhibits the major Rho effector , Rho kinase ( ROCK ) , the frequency of infected cells decreased in a dose-dependent fashion ( Fig 5B ) , indicating that ROCK activation is necessary during uptake of bacteria . As with FAK , the siRNA pools in the original screen did not exhibit a significant phenotype ( S1 Table ) , but also only moderately depleted ROCK mRNA ( S6A Fig ) . Furthermore , an acute perturbation in ROCK activity may be more indicative of a specific role in bacterial uptake than long-term depletion , which may be accompanied by other cytoskeletal remodeling . The siRNA pool targeting RhoA effectively depleted its target mRNA ( S4A Fig ) but did not exhibit a significant phenotype in the screen ( S1 Table ) . Redundant function of RhoA , B , and C may contribute to the lack of phenotype when only RhoA is depleted [68] . It is also probable that RhoA has multifaceted effects on L . monocytogenes infection; it could influence both global and local actin dynamics , which might have opposing effects on L . monocytogenes internalization . Our data indicate that FAK- and GEF-H1-dependent ROCK activity and also formin-mediated actin polymerization promote uptake of L . monocytogenes by HUVEC; however these data were all obtained using strains with the LLOG486D point mutation , given the substantial susceptibility of HUVEC to LLO . To verify that the presence of wild-type LLO would not significantly change the process of internalization , we examined the effects of pharmacological FAK and ROCK inhibition and FHOD1 depletion during very low dose infection with wild-type L . monocytogenes ( JAT607 ) and found that internalization of wild-type L . monocytogenes by HUVEC is strongly FAK- , ROCK- , and FHOD1-dependent ( Fig 6A , 6B and 6C ) . If L . monocytogenes capitalizes on an intrinsic , constitutive , phagocytosis-like process , then this same signaling pathway should be necessary for uptake of other large objects by HUVEC . First , we examined whether inhibition of ROCK or FAK could substantially reduce the ability of HUVEC to internalize L . innocua as well as L . monocytogenes and found this to be true ( Fig 6D ) . Similarly , formin inhibition reliably reduced internalization of L . innocua ( Fig 6D ) . Phagocytosis-like uptake could be modulated by changes in bacterial adhesion to cells , in the internalization process itself , or in changes in global cellular state , such as cell stiffness , that might indirectly affect adhesion or internalization . To differentiate between these possibilities , we quantified adhesion of L . innocua in the presence of ROCK , FAK , or formin inhibitors and found , surprisingly , that ROCK inhibition , but not inhibition of formins or FAK , dramatically reduced L . innocua adhesion to HUVEC ( Fig 6E ) . Furthermore , inhibition of ROCK did not affect the ability of HUVEC to internalize beads , which adhere non-specifically ( Fig 6F ) , consistent with a role for ROCK specifically in bacterial adhesion , rather than internalization . Formin inhibition did decrease internalization of beads by about 50% ( Fig 6F ) , comparable to its effect on L . innocua internalization ( Fig 6D ) and to the effect of siFHOD1 and siFMNL3 on L . monocytogenes internalization ( Fig 3D ) ; however , formin inhibition did not inhibit bacterial adhesion ( Fig 6E ) . Therefore , we conclude that formins are involved in actin remodeling specifically during phagocytosis-like uptake . Surprisingly , FAK inhibition decreased L . innocua and L . monocytogenes internalization without affecting L . innocua adhesion , but had no effect on internalization of beads ( Fig 6F ) . Finally , we examined whether these regulators of bacterial uptake by HUVEC affected macrophage phagocytosis of L . monocytogenes . Only inhibition of FAK disrupted phagocytosis of L . monocytogenes by activated U937 cells , a macrophage-like cell line , whereas ROCK and formin inhibition had no significant effect ( Fig 6G ) . Furthermore , it has previously been shown that ROCK inhibition does not affect uptake of L . monocytogenes by a variety of macrophage-like cell lines[69] . Therefore , we conclude that endothelial cells and macrophages use distinct pathways to internalize L . monocytogenes .
Our data demonstrate that a number of proteins in the endothelial shear stress-responsive pathway , including PI3K , FAK and focal adhesions , GEFH1 , and ROCK , also regulate L . monocytogenes internalization in endothelial cells; however , it is also clear from our data that the PI3K/FAK/RhoA pathway is not activated in a straightforward linear manner that starts with PI3K activity and culminates in bacterial uptake . For instance , although our data demonstrate that FAK activity is upstream of GEFH1 , implying that it is upstream of ROCK , FAK inhibition inhibits bacterial uptake but not adhesion , whereas ROCK inhibition dramatically inhibits bacterial adhesion . Likewise , although ROCK has previously been shown to directly phosphorylate and activate FHOD1[70] , formin inhibition did not affect bacterial adhesion in our assay , and formin inhibition , but not ROCK inhibition , inhibited phagocytosis-like internalization of polystyrene beads . The PI3K/FAK/Rho signaling pathway normally promotes global rearrangements in endothelial cell architecture in response to apical signaling events and mechanical deformation[62] , however many of these proteins can also act locally to modulate the chemical and mechanical environment; these local and global effects may even have opposing effects on bacterial internalization or phagocytosis-like uptake of other large objects . Here , we have identified specific proteins required for bacterial adhesion to and internalization by endothelial cells; further experiments that simultaneously combine both precise spatial and temporal control of protein activity will be necessary to dissect multiple global and local roles of these proteins during infection . We have shown that L . monocytogenes and L . innocua , a non-pathogenic relative , are internalized at comparable rates and are regulated by similar host effectors; therefore , no Listeria monocytogenes-specific effectors were necessary for bacterial internalization in HUVEC . Additionally , we have shown that microspheres are internalized at least as efficiently as bacteria , and that the presence of L . innocua did not further enhance microsphere uptake . Therefore , HUVEC exhibit constitutive phagocytic behavior that is not enhanced or inhibited by the presence of bacteria . Notably , we also show that ROCK-independent adhesion of beads appears to be distinct from ROCK-dependent adhesion of bacteria , although both uptake processes are formin-dependent to a comparable degree . Thus , while bacterial and bead adhesion are differentially regulated , our results suggest that , once adhered , the internalization mechanism is similar and formin-dependent . A hallmark of systemic listeriosis is the ability of L . monocytogenes to infect distinct cell types in distinct organs , including intestinal epithelial cells , hepatocytes , placental cytotrophoblasts , endothelial cells , macrophages and other immune cells[5]; L . monocytogenes invades these distinct cell types using both pathogen-triggered and pathogen-independent mechanisms . Intriguingly , L . monocytogenes uses distinct invasion mechanisms that capitalize on unique characteristics of these different cell types . For instance , L . monocytogenes uses the epithelial junctional protein E-cadherin to invade intestinal epithelial cells[7 , 8] and the hepatocyte growth factor receptor c-Met to invade hepatocytes[9 , 10]; such invasion requires the L . monocytogenes proteins InlA and InlB , respectively . In contrast , in this context , endothelial cells may behave more like professional phagocytes , with internalization less likely to be pathogen-specific and more likely to involve Listeria adhesion followed by co-option of a normal constitutive phagocytosis-like uptake process . Notably , the endothelial factors involved are critical endothelial regulators , including the PI3K/FAK/Rho shear stress-signaling pathway and FMNL3 , an endothelial formin that is critical for angiogenesis[53 , 71 , 72] . Although macrophages have been shown to use formins and ROCK [73 , 74] during phagocytosis of some cargo , we found that inhibition of these proteins did not inhibit macrophage-like cells from internalizing L . monocytogenes; thus , L . monocytogenes hijacks unique and distinct pathways in macrophage and endothelial infection . Given dramatically different kinetics of macrophage phagocytosis and endothelial phagocytosis-like uptake in vivo[1 , 75] , and that the PI3K/FAK/Rho signaling axis regulates endothelial architectural changes , we speculate that substantial remodeling of the endothelial cytoskeleton is required for phagocytosis-like uptake and may explain its slower kinetics . Although endothelial phagocytosis-like uptake occurs in a number of different contexts in vivo and in culture , its role in vivo is unclear . Here , we determined that the formins FMNL3 and FHOD1 likely regulate such uptake; genetic and pharmacological perturbation of these proteins can now be used to understand the role of phagocytosis-like uptake in vivo . We speculate that endothelial phagocytosis-like uptake is a surveillance strategy to remove particles from the bloodstream , particularly in cases of macrophage injury , or at sites at which macrophages have limited access . For instance , angiophagy might enhance fibrin clot clearance and restore blood flow in small diameter vessels that do not receive much immune cell traffic[1 , 2] . Phagocytosis-like uptake by endothelial cells may also recruit immune cells specifically to vulnerable sites in the vasculature to limit pathogen dissemination . Indeed , endothelial cells increase expression of pro-inflammatory cytokines and chemokines in response to L . monocytogenes[76] and Rickettssiae[77] , which comprise a group of obligate intracellular bacterial species that cause spotted fever and typhus and preferentially infect endothelial cells , likely through a direct receptor-mediated process[78 , 79] . Additionally , endothelial cells have been shown to kill internalized Rickettssiae directly in a cytokine-activated hydrogen peroxide- or nitric oxide-dependent manner[77 , 80] , and thus may contribute directly to pathogen removal . In such a setting , escape from the vacuole may be the primary pathogenic strategy of intracellular bacteria . Indeed , like L . monocytogenes , Rickettssiae species can escape the vacuole and proliferate in the endothelial cell cytoplasm[81 , 82] . Furthermore , both L . monocytogenes[12 , 19 , 31] and Rickettssiae species[81 , 82] can hijack host cell actin to move within and between cells , without exposure to the extracellular space . Thus , these pathogens may re-direct their own dissemination , rather than passively following transcytosis; this may be a mode of L . monocytogenes spread across the endothelium into the central nervous system .
S3 Table lists bacterial strains used in this study . To express fluorescent proteins in L . monocytogenes strains , plasmids were transformed into E . coli SM10 λpir by electroporation and subsequently transferred to L . monocytogenes by conjugation[83] . Constructs were stably integrated into the tRNAARG locus of the bacterial chromosome as previously described[83] . For constitutive GFP expression , plasmid pMP74 ( a gift from M . Pentecost and M . Amieva ) , in which sGFP is expressed under the Hyper-SPO1 promoter fused to the 5′ UTR of hly[84 , 85] , was incorporated into JAT745 or JAT984 to generate strains JAT1045 and JAT1046 , respectively . An identical approach was used to express a codon-optimized mTagRFP under the control of the ActA promoter ( plasmid pPL499[28] , a gift from P . Lauer ) to generate strains JAT983 and JAT985 , respectively , which express mRFP only after reaching the host cell cytoplasm . The inlAB LLOG486D strain was generated by integrating the LLOG486D mutation into JAT1084 ( a gift from M . Pentecost and M . Amieva ) , by allelic exchange[26 , 86 , 87] to generate JAT1116 . Integration was verified by sequencing the hly locus . Codon-optimized mTagRFP ( from pPL499 ) was incorporated into JAT1116 as described above to generate strain JAT1119 . HUVEC ( Lonza C2517A ) were cultured according to the manufacturer’s instructions ( EGM Bullet Kit-2 , Lonza CC-3162 ) . Infections were performed in normal growth media but lacking serum and antibiotics ( serum- and antibiotic-free media , SAFM ) . For microscopy experiments , black 96-well clear-bottom plates ( E&K Biosciences 25090 ) or glass coverslips were coated with 30μg/mL collagen type I in PBS ( Advanced Biomatrix 5005-100ML ) for 1 hour at 37°C and then washed once with PBS before cells were plated . U937 cells ( ATCC , CRL-1593 . 2 ) were grown in RPMI with 10% fetal bovine serum and gentamicin/amphotericin ( Lonza , CC-4083 ) ; for these cells SAFM consisted of RPMI without additives . DAPI ( Invitrogen D1306 ) was dissolved at 5mg/mL in dimethyl formamide and used at 1/5000 . Other drugs were dissolved in DMSO ( endotoxin-free dimethyl sulfoxide , Sigma D2650 ) at stock concentrations indicated below . Stock concentrations and sources of drugs were: 50mM LY294002 ( Sigma L9908 ) , 25mM wortmannin ( EMD Chemicals 12–338 ) , 30mM Y27632 ( EMD Chemicals 688000 ) , 10mM FAK inhibitor-14 ( FAK-14 ) ( Tocris Bioscience 3414 ) , 100mM PF573228 ( Tocris Bioscience 3239 ) . SMIFH2 ( Millipore 344092 ) solutions were freshly made with each experiment as we found that frozen stocks degraded over time . Primary antibody used for inside/outside staining was BacTrace anti-Listeria genus specific antibody ( 01-90-90 , KPL , Inc . ) . Fluorescent streptavidins used for inside/outside staining of 2 . 0μm biotinylated polystyrene beads ( Polysciences , Inc . 24172 ) were Alexa-Fluor-546-streptavidin ( Invitrogen S11225 ) and Alexa-Fluor-488-streptavidin ( Invitrogen S11223 ) . For Western blotting , rabbit monoclonal anti-Arp2 antibody ( Epitomics 5738–1 ) was used to detect Arp2 . Endothelial cells were infected as previously described[21 , 32] with the following modifications . L . monocytogenes liquid cultures were started from a plate colony , and grown overnight , spinning , at 30°C in Brain Heart Infusion ( BHI ) media ( Gibco 211059 ) supplemented with 200μg/mL streptomycin . Chloramphenicol-resistant strains were grown with 7 . 5 μg/mL chloramphenicol . Cultures were diluted in fresh media to an OD600 of 0 . 1 and returned to a spinning wheel at 30°C for 2–2 . 5 hours . Bacteria were then washed 3 times with PBS to remove any soluble factors and diluted into SAFM . HUVEC were washed once with SAFM , and bacteria were added to an MOI of 50–100 bacteria per HUVEC unless otherwise indicated . For every experiment , MOI was calculated directly by counting the colony forming units in the bacterial inoculum . To synchronize invasion , samples were spun for 10 minutes at room temperature at 500 x g prior to incubation . After thirty minutes , samples were washed four times with SAFM and , after an additional thirty minutes , media was replaced with SAFM supplemented with 20μg/mL gentamicin . Analysis was performed by flow cytometry or microscopy at 8 hours after exposure , unless otherwise indicated . Samples analyzed by microscopy were fixed for fifteen minutes in 3 . 7% formaldehyde buffered in sodium phosphate , stained with DAPI , and imaged on an ImageXpress Micro ( Molecular Devices ) using a 10X or 20X air objective; the percent of cells infected was determined as described below for the siRNA screen . For each biological replicate , 300–500 cells were analyzed . Analysis of experiments in Figs 1D , 1F and 6 was manual but the experimenter was blinded to the identity of samples during imaging and analysis . Analysis of experiments in Fig 1B and Figs 2–5 was automated as described below in Analysis of the siRNA Screen . For drug experiments ( Figs 3–6 ) , unless otherwise indicated , media was removed from cells and replaced with media containing either the drug or DMSO ( vehicle control ) , either at the time of infection or prior to infection . Cells remained in media containing the drug until 1 hour after infection , when cells were washed twice and replaced with drug-free gentamicin-containing media . HUVEC or U937 were exposed to bacteria or 6-His-LLO for 30 minutes . HUVEC were washed in PBS and then incubated in 0 . 25% trypsin-EDTA ( Invitrogen ) for 15–20 minutes to fully detach all cells; an equal amount of 6% fetal bovine serum in PBS was then added to inactivate the trypsin . U937 were in solution throughout the experiment . Propidium iodide was added to cells in solution at a final concentration of 25μM and samples were immediately analyzed on a BD LSRII Flow Cytometer ( BD Biosciences ) . Live cells were identified as described in S1A , S1B and S1C Fig For infection experiments , we determined the fraction of HUVEC that were infected as illustrated in S4 Fig For each biological replicate , 5 , 000–10 , 000 cells were analyzed . His-LLO was purified as described[24] and provided by Jennifer Robbins and Lisa Cameron . Samples were infected as described above and fixed 30 minutes after initial exposure to bacteria or beads . To quantify bacterial adhesion and internalization , inside/outside staining was performed as described previously[88] , using the BacTrace anti-Listeria genus primary antibody or fluorescent streptavidin conjugates . Samples were additionally stained with DAPI to identify HUVEC nuclei . Coverslips were mounted onto slides with VectaMount ( Vector Labs ) . Samples were imaged on a Nikon Eclipse TiE inverted fluorescence microscope equipped with a charge-coupled device ( CCD ) camera ( Andor Technologies ) using a 63X or 100X oil objective , and captured with the Micromanager[89] software package . HUVEC were identified from transmitted light images and DAPI staining . All bacteria or beads associated with individual HUVEC were counted as adherent; bacteria or beads that lacked the “outside” stain ( applied before permeabilization ) were counted as internalized . To minimize off-target effects and maximize on-target effects , siRNA pools targeting candidate genes of interest were produced by in vitro dicing as previously described[35 , 36] using purified Giardia Dicer[37] . Due to low yield for some pools in our first synthesis , we performed the synthesis twice to include all of our candidates . To avoid positional effects , the position of each siRNA pool in the final 96-well plates was randomized . For each well of a 96-well plate , 104 HUVEC suspended in SAFM were reverse-transfected with siRNAs at 20 nM final concentration using 0 . 25 μL Lipofectamine RNAiMAX ( Invitrogen 13778075 ) . The transfection mix was replaced by SAFM 8–9 hours later . Synthetic siRNAs for targeting genes of interest ( Figs 3 and 5 ) were purchased from Dharmacon ( S4 Table ) . Infections were performed approximately 72 hours after transfection . We screened each siRNA pool in 6 replicates on 3 different days for each of the 2 independent siRNA syntheses; thus , for most candidates , we collected data from 12 independent replicates . To correct for day-to-day variability in the infection itself , each plate included 10 wells of HUVEC that were not treated with siRNA and were exposed to either JAT983 or JAT985 . siRNA-treated wells were infected with JAT983 at an MOI of 50–100 . For each image , Cell Profiler[90] was used to identify nuclei and to estimate cell boundaries . Infected cells were defined using a background threshold on the images of bacteria . Foci were defined as groups of contiguous infected cells . Bacterial density in an infected cell was defined as the number of pixels in the cell above the threshold that defined the signal from bacterial fluorescence . Foci consisting of a single , unreplicated bacterium in a single cell were removed from analysis; such filtering maximized the difference between JAT983- and JAT985-infected samples . For each siRNA , in each replicate , we quantified: the fraction of HUVEC infected , the top quartile of bacterial density per infected cell , the density of foci , the top quartile of focus size , and the number of HUVEC ( used to calculate density of foci ) . We used the top quartile rather than median for the bacterial density and focus size measurements because these maximized the difference between JAT983- and JAT985-infected wells . To identify specific outliers , we used the rank-product for each metric , which corresponds to the geometric mean of the rank of each siRNA pool in each experiment , and has been used to determine outliers from microarray data[91] . Briefly , siRNA-treated wells in each replicate were ordered and assigned the rank of p/n , where p is the well’s position in the ordered list and n is the total number of siRNA-treated wells in that replicate . The rank-product for all the replicates of a given siRNA is then given by ( Πirpi/ni ) ^ ( 1/r ) , where r is the total number of replicates of that siRNA , pi is the ranking in the ith replicate , and ni is the number of RNAs in the ith replicate[91] . If all siRNA pools had the exact same effect , then each one would have a ranking that converged to 0 . 5 with increasing experimental replicates . To generate the null distribution ( for which we assume that all siRNAs gave identical effects ) , we performed identical analysis except that the names of siRNAs were randomly permutated prior to calculating the rank-product; we ranked 20 , 000 such permutation simulations to capture the probability of relatively rare events . To identify the statistical outliers in our data , we calculated the frequency of a particular siRNA’s rank among the simulations . To correct for multiple hypothesis testing ( since we screened 156 individual RNAs ) , we used the Benjamini-Hochberg Procedure to hold the false discovery rate to 0 . 05 . HUVEC were treated with control or experimental siRNA as described above . mRNA was harvested using the RNeasy Micro Kit ( Qiagen 74004 ) and cDNA was prepared using the Superscript III First-strand Synthesis SuperMix ( ThermoFisher 18080–400 ) . Genes of interest were amplified using primers specified in S5 Table . qPCR was performed using SYBR Select Master Mix ( ThermoFisher 4472908 ) on a StepOnePlus Real-Time PCR System . Normalized relative quantity ( NRQ ) and error were calculated as previously described[92] . CDH5 , ACTR2 , MYH9 , and GAPDH were used as control genes . For Western blotting , samples were treated with siRNAs as described above . After 72 hours of depletion , cells were lysed in SDS sample buffer ( 2% SDS , 10% glycerol , 0 . 02% bromophenyl blue sodium salt , 1% beta-mercaptoethanol , 5mM EDTA , 80mM Tris-HCl pH6 . 8 ) , sonicated and boiled for 10 minutes each . Samples were run on 12% SDS-PAGE gels , transferred to nitrocellulose membrane via semi-dry transfer . Total protein was evaluated by staining in Ponceau-S ( 0 . 2% Ponceau-S , 3% trichloroacetic acid , 3% sulfosalicylic acid ) . Membranes were then blocked in milk and stained with anti-Arp2 primary antibody , then horseradish peroxidase-conjugated goat anti-rabbit secondary antibody , and visualized by chemiluminescence . U937 cells were differentiated with phorbol 12-myristate 13-acetate ( PMA ) at 80nM for 36–48 hours prior to infection and were noted to be adherent at the time of infection . Infections were performed exactly as described above for endothelial cells except that U937 were infected with ActA-deficient L . monocytogenes expressing wild-type LLO ( JAT610 ) , and adherent U937 cells were infected directly from overnight liquid culture at an MOI of 80 . Infection was analysed 7 hours after infection by flow cytometry as previously described .
|
Vascular endothelial cells , which line the lumen of blood vessels , are conventionally viewed as a restrictive barrier that protects underlying tissue from blood-borne toxins and pathogens . Nonetheless , even highly restrictive endothelial cells can internalize micron-sized objects , such as blood clots , raising the question of how such phagocytosis-like uptake occurs , and whether it is mechanistically distinct from classical phagocytic pathways . We found that the pathogenic bacterium Listeria monocytogenes , which must overcome the endothelial barrier to access underlying tissue , can be taken up by primary endothelial cells ( HUVEC ) in culture . We exploited this ability to identify molecular regulators of such phagocytosis-like uptake . We found that the formin family of actin nucleators drives such uptake , whereas these proteins did not have a significant role in phagocytosis of L . monocytogenes by macrophages . Thus , our data suggest that endothelial cells and macrophages use distinct phagocytosis-like pathways to internalize L . monocytogenes . Perturbations of the regulatory proteins that we have identified here should allow for dissection of the normal physiological functions of endothelial phagocytosis-like uptake , as well as its therapeutic potential in diverse roles such as clot resolution and drug delivery .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"blood",
"cells",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"pathology",
"and",
"laboratory",
"medicine",
"evolutionary",
"biology",
"gene",
"regulation",
"pathogens",
"endothelial",
"cells",
"focal",
"adhesions",
"population",
"genetics",
"microbiology",
"immunology",
"gene",
"pool",
"epithelial",
"cells",
"developmental",
"biology",
"molecular",
"development",
"population",
"biology",
"bacterial",
"pathogens",
"small",
"interfering",
"rnas",
"contractile",
"proteins",
"actins",
"white",
"blood",
"cells",
"animal",
"cells",
"proteins",
"medical",
"microbiology",
"adhesion",
"molecules",
"gene",
"expression",
"microbial",
"pathogens",
"biological",
"tissue",
"listeria",
"monocytogenes",
"biochemistry",
"rna",
"cytoskeletal",
"proteins",
"cell",
"biology",
"nucleic",
"acids",
"anatomy",
"genetics",
"epithelium",
"biology",
"and",
"life",
"sciences",
"cellular",
"types",
"macrophages",
"non-coding",
"rna"
] |
2016
|
Endothelial Cells Use a Formin-Dependent Phagocytosis-Like Process to Internalize the Bacterium Listeria monocytogenes
|
Copper ions play an important role in ethylene receptor biogenesis and proper function . The copper transporter RESPONSIVE-TO-ANTAGONIST1 ( RAN1 ) is essential for copper ion transport in Arabidopsis thaliana . However it is still unclear how copper ions are delivered to RAN1 and how copper ions affect ethylene receptors . There is not a specific copper chelator which could be used to explore these questions . Here , by chemical genetics , we identified a novel small molecule , triplin , which could cause a triple response phenotype on dark-grown Arabidopsis seedlings through ethylene signaling pathway . ran1-1 and ran1-2 are hypersensitive to triplin . Adding copper ions in growth medium could partially restore the phenotype on plant caused by triplin . Mass spectrometry analysis showed that triplin could bind copper ion . Compared to the known chelators , triplin acts more specifically to copper ion and it suppresses the toxic effects of excess copper ions on plant root growth . We further showed that mutants of ANTIOXIDANT PROTEIN1 ( ATX1 ) are hypersensitive to tiplin , but with less sensitivity comparing with the ones of ran1-1 and ran1-2 . Our study provided genetic evidence for the first time that , copper ions necessary for ethylene receptor biogenesis and signaling are transported from ATX1 to RAN1 . Considering that triplin could chelate copper ions in Arabidopsis , and copper ions are essential for plant and animal , we believe that , triplin not only could be useful for studying copper ion transport of plants , but also could be useful for copper metabolism study in animal and human .
The phytohormone ethylene ( C2H4 ) plays important roles in plant growth and development . When exposed to ethylene gas for 3 days , dark-grown Arabidopsis seedling shows a typical triple response phenotype including a short hypocotyl and root , larger diameter hypocotyl , and an exaggerated apical hook [1 , 2] . ETHYLENE RESPONSE1 ( ETR1 ) , ETHYLENE RESPONSE2 ( ETR2 ) , ETHYLENE RESPONSE SENSOR1 ( ERS1 ) , ETHYLENE RESPONSE SENSOR2 ( ERS2 ) and ETHYLENE INSENSITIVE4 ( EIN4 ) are five ethylene receptors in Arabidopsis which function redundantly and negatively to regulate ethylene signaling and responses . The receptors signal is transferred to a downstream Raf-like protein kinase CONSTITUTIVE TRIPLE RESPONSE1 ( CTR1 ) [3 , 4] . CTR1 interacts with and phosphorylates an endoplasmic reticulum ( ER ) membrane-localized Nramp homolog ETHYLENE IN SENSITIVE 2 ( EIN2 ) . This prevents EIN2 from activating downstream components of ethylene signaling including ETHYLENE INSENSITIVE3 ( EIN3 ) and EIN3-LIKE1 ( EIL1 ) . When ethylene binds to the receptors , the phosphorylation of EIN2 by CTR1 is reduced , leading to accumulation of EIN2 and proteolytic cleavage of the cytosolic C-terminal domain of EIN2 , which enters the nucleus to initiate ethylene signaling [5–10] . Copper ions are cofactors that are required for ethylene binding to ETR1 . Ethylene insensitive mutant etr1-1 eliminates both ethylene binding and the interaction of copper ion with the receptor [11] . In Arabidopsis , the copper ions required by the ethylene receptors are transported by RAN1 , also called HEAVY METAL ATPASE 7 ( HMA7 ) , which is a Cu-transporting P-type ATPase . Two weak mutant alleles , ran1-1 and ran1-2 , were identified in a mutant screening where they responded to the ethylene receptor antagonist trans-cyclooctene ( TCO ) with a triple response phenotype which could be partially suppressed by adding copper ion to the plant growth medium . Additionally , ran1-1 and ran1-2 are hypersensitive and display a similar phenotype upon treatment with a low concentration of the copper ion chelator neocuproine . It has been shown that RAN1 is essential for the biogenesis of ethylene receptors in Arabidopsis [12–14] . A similar protein , Ccc2a , has been identified in Saccharomyces cerevisiae where it functions to transport copper ions from the copper chaperone , Atx1 , to the secretory pathway . Atx1 is a small metal homeostasis factor that protects cells against reactive oxygen toxicity caused by excess or abnormal distribution of copper ions [15–18] . A homolog of Atx1 is found in humans . This metallochaperone , HUMAN ATX1-LIKE HOMOLOG1 ( HAH1 ) , also transports copper ions and interacts with Menkes and the Wilson disease proteins [19–21] . In Arabidopsis , there are two homologs of yeast Atx1 , COPPER CHAPERONE ( CCH ) and ATX1 [22–24] . A T-DNA insertion mutant of ATX1 is specifically hypersensitive to excess copper ion as well as copper deficiency , while the over-expression of ATX1 enhances plant tolerance to excess copper ion and copper deficiency [25] . Previous yeast two-hybrid assays suggested that ATX1 and CCH△ ( without C-terminal ) interact with RAN1 and HMA5 [24 , 26] . It has been predicted that copper ions transported by RAN1 that are essential for ethylene receptors may come from ATX1 and CCH , but there is no genetic evidence to support this hypothesis [25 , 27] . To date , only the chelator neocuproine has been found to cause a plant triple response phenotype , but whether the phenotype could be suppressed by adding copper ions is unknown [14] . Neocuproine should be cautiously used as it may potentiate a cytotoxic effect of endogenous copper on cells [28] . Thus , there is no report about a specific copper ion chelator which causes a triple response phenotype specifically via a reduction in copper ions . We believe that a copper ion chelator with such a property would be useful for investigating the mechanisms of copper ion transport to the ethylene receptors . More ever , chemical genetics approaches are powerful in dealing with the function redundancy of components involved in phytohormone signaling , and have helped scientists to identify ABA receptors and novel signaling components involving in auxin and other phytohormones [29–32] . Here , we have used a chemical genetics approach to uncover a novel synthetic small molecule triplin . We found that triplin could cause a triple response phenotype in dark-grown Arabidopsis seedlings as a copper chelator . By testing the sensitivity of copper ion transport mutants to triplin , we showed that ATX1 acts upstream of RAN1 . Our genetic and biochemical results support a model where ATX1 transports copper ions to RAN1 for ethylene receptor biogenesis and signaling . Triplin does not directly act on ethylene binding to receptor . Rather , it perturbs copper ion transport involved in the interaction of RAN1 and ATX1 . In addition , triplin suppresses the toxic effects of excess copper ion on plant root growth . Thus , triplin may provide a useful chemical genetics tool to study copper ion transport in plants , and could be a lead structure for drug development for human copper disorders diseases in the future .
To identify novel small molecules that affect ethylene signaling , we used the same plant chemical genetics screening method described previously [32–34] . Arabidopsis wild type Columbia-0 ( Col-0 ) were grown and screened against a synthetic small molecule library with 12 , 000 compounds in the dark for 3 days . Those compounds that caused a triple response in seedlings were selected and their chemical genetics effects were retested . From this plant chemical screening , we acquired 14 compounds which cause a triple response in dark-grown Arabidopsis seedlings . Among these hits , the 5 strongest hits representing different structures were further assayed . Using the ethylene perception inhibitor , AgNO3 , and the ethylene insensitive mutant ein2-5 , we found that only a compound we called triplin ( 1- ( 1-morpholino-1- ( thiophen-2-yl ) propan-2-yl ) -3- ( 2- ( trifluoromethoxy ) phenyl ) thiourea ) worked through the ethylene signaling pathway ( Fig 1 and S1 Fig ) . We also examined the chemical genetic activities of 38 triplin analogs collected from the library . Our results showed 5 analogs could cause triple response phenotypes at 100 μM , and 11 analogs caused the phenotype at a higher concentration of 200 μM . The structure analysis of these active analogs indicates that the morpholino and thiourea moieties of the molecules could be important for enhancing their chemical genetics activities ( S1 Table ) . Two strategies were employed to study the mechanism of action of triplin . First , we screened ethyl methanesulfonate ( EMS ) mutagenized populations of Col-0 for triplin resistant mutants . We acquired nine triplin resistant mutants which also showed resistance to the application of the ethylene biosynthesis precursor , 1-aminocylopropane-1-carboxylic acid ( ACC ) . Among these resistant mutants , two are dominant mutants that have the same mutation as etr1-1 as revealed by DNA sequencing [1] . The other seven mutants are recessive and were genetically determined to be on the EIN2 locus by examining F1s acquired from crossings with ein2 ( S2 Fig ) . We further assayed the effects of triplin treatment on known ethylene signaling mutants . Ethylene resistant mutant etr1-1 , etr1-2 , ein2-5 and ein3 eil1 showed resistance to triplin ( Fig 1C and S3 Fig ) . Additionally , adding 500 μM AgNO3 blocked triplin’s effects on seedlings ( S6C Fig ) . Consistent with triplin affecting ethylene signaling , the ethylene-responsive gene , ERF1 , was up-regulated by triplin treatment ( Fig 1D ) . These results indicated that triplin acts on ethylene signaling . It is possible that triplin also increases ethylene biosynthesis to cause a triple response . To determine if this was occurring , we first examined how alterations in ethylene biosynthesis affect responses to triplin . We found that triplin does not increase ACSs gene expression levels . Neither the ACC synthase ( ACS ) inhibitor , aminoethoxyvinylglycine ( AVG ) nor the ethylene biosynthesis mutant cin5 affects triplin responses . Additionally , triplin treatment did not enhance ethylene biosynthesis and actually causes a decrease in ethylene production ( S4 Fig ) . Together , these results indicate that triplin treatments cause triple responses by acting on ethylene signaling but not ethylene biosynthesis . To map where triplin acts on the ethylene signaling network , we assayed how triplin treatments affect other ethylene insensitive mutants of ethylene receptors including ers1-1 , ers2-1 , etr2-1 and ein4 . Our results showed they were all resistant to triplin ( S3B Fig ) , indicating that triplin likely acts on or upstream of the ethylene receptors in the ethylene signaling network . Previous research has shown that the copper ion transporter RAN1 is involved in the biogenesis of ethylene receptors and certain ion chelators such as neocuproine can trigger a plant triple response [14] . As we expected , ran1-1 and ran1-2 mutants were hypersensitive to triplin treatments comparable with the published results for neocuproine [14] ( S10B and S10C Fig ) . Low doses of triplin ( e . g . less than 10 μM ) had no visible effects on wild type but caused a triple response in ran1-1 and ran1-2 ( Fig 2A ) . Further , introducing RAN1 into the mutants rescued a wild type response to triplin ( Fig 2B ) . However , ran1-2 etr1-1 and ran1-2 ein2-5 double mutants were both resistant to triplin treatments similar to the etr1-1 and ein2-5 single mutants ( Fig 2C ) . Together , our chemical genetics results showed that triplin is likely to act upstream of the ethylene receptors and may be involved in altering copper ion transport to the receptors . Our simple postulation is that triplin is a copper ion chelator which causes the triple response phenotype by chelating the copper ions necessary for ethylene receptor biogenesis . To determine if triplin acts as a copper ion chelator and its specificity , we examined the effects of copper ions and other ions such as Ag , Zn , Ca , Mg , Mn , Co , Ni , Na , K , Li and Mo on triplin responses in plants . Our results showed that addition of excess Cu2+ ( CuSO4 ) partially reverses the effects of triplin on plants ( Fig 2D and 2E ) . Of the other ions tested , only Ag+ ( AgNO3 ) had a similar effect ( S5 Fig and S6C Fig ) . We also examined the known copper ion chelator neocuproine and found that its effects were reversed by adding Zn2+ ( ZnSO4 ) under our assay conditions ( S10A Fig ) . To further verify that triplin acts by chelating copper ions , we tested whether triplin treatment can alleviate the toxic effects of high levels of copper ions on plant growth . Our results indicated that the growth of Arabidopsis seedlings roots treated with 50 μM CuSO4 was strongly inhibited; addition of 100 μM triplin at the same time partially reversed this inhibition of root growth ( Fig 3A and 3B ) . Higher concentration of CuSO4 causes more severe growth inhibition which can also be partially reversed by adding triplin ( S6A and S6B Fig ) . We next tested if neocuproine has a similar effect as triplin on plant grown with high levels of copper ions . To our surprise , neocuproine aggravated the copper ion toxic effects on plant growth . Another thing is that ZnSO4 could restore the phenotype caused by neocuproine . This indicated that the specificity of neocuproine is questionable ( S10A and S10C Fig ) . To better understand the effects of triplin on plants , we measured the relative copper ion content of seedlings grown with copper ions and triplin . When seedlings were grown on a normal 0 . 5xMS growth medium , adding 100 μM triplin to the medium slightly decreased the copper ion content in the seedlings from 110 μg/g to 80 μg/g . When seedlings were grown on a growth medium containing 20 μM CuSO4 , the relative copper ion content in the seedlings increased from 110 μg/g to 1 , 100 μg/g . Adding 100 μM triplin reduced the ion content from 1 , 100 μg/g to 560 μg/g ( Fig 3C ) . We tried to mix triplin with different metal ions such as Ag , Cu , Zn and Ca . Unexpectedly , we found that triplin formed a tawny turbidness with CuSO4 , and a black precipitate with AgNO3 ( S7 Fig ) . We analyzed the products formed from mixing 100 μM CuSO4 with 100 μM triplin by Matrix-Assisted Laser Desorption/ Ionization Time of Flight Mass Spectrometry ( MALDI-TOF-MS ) analysis . With this we observed a peak in the m/z of 509 . 9 , which is equal to the sum of molecular weights of triplin ( 446 . 0 g/mol ) and copper ( Fig 3D ) . Silver ions can also conjugate with triplin , but a higher concentration ( e . g . 10 mM ) of AgNO3 is needed . Conjugates of triplin with other metal ions including ZnSO4 and FeSO4 were not detected with our mass spectrographic analysis ( S8 Fig ) . Together , our results indicate that triplin is a copper chelator which can also chelate silver to a less degree . This possibly explains why both Cu2+ and Ag+ can block or partially block triplin’s effects on plant growth . In order to examine how triplin acts as a copper ion chelator to perturb ethylene signaling and cause the triple response phenotype , we first used two copper ion deficient growth medium to grow plants . One included all essential elements for plant growth except copper ions were not added . The other consisted of 0 . 5xMS growth medium supplemented with 500 μM of the copper ion chelator , bathocuproinedisulfonic acid ( BCS ) . The effects of 20 μM or 50 μM triplin on plant growth in dark were compared on these medium to control medium . The seedlings grown on both copper-deficient growth medium showed more severe triple responses in response to 20 μM triplin than seedlings grown on control medium ( Fig 4A and 4B ) . Under these same growth conditions with different triplin concentration , ein2-5 showed obvious resistance to triplin ( S9 Fig ) . This indicates that the exaggerated triple response in the absence of added copper is also dependent on ethylene signaling . These results are consistent with our earlier results indicating that triplin can chelate copper ions . However , we also noticed that simply lowering copper levels in the growth medium is not enough to cause a triple response . This led us to speculate that there may be additional actions of triplin on plants to cause the triple response . One possibility we considered was that triplin enters into plant cells to affect copper ion transport . To examine this possibility , we used RAN1 overexpression transgenic lines in the Col-0 background to test their sensitivity to triplin treatment . Our results showed that the overexpression lines are resistant to triplin treatments ( Fig 4C and S14 Fig ) . One possible explanation is that triplin acts as copper ion chelator that competes for copper with the ethylene receptors resulting in less ethylene binding to the receptors . To examine this possibility , we measured ethylene binding to membranes isolated from yeast expressing the ethylene binding domain of ETR1 in the presence and absence of triplin [14] . Triplin had no measurable effect on specific ethylene binding to the receptors ( Fig 4D ) . By contrast , ethylene binding to the receptors was reduced approximately 50% when triplin was added to the growing yeast cells expressing ETR1 ( Fig 4D ) . These results indicate that triplin is not directly affecting ethylene binding but may affect the delivery of copper ions to the receptors by affecting other copper delivery proteins . For example , triplin treatments may reduce the copper ion levels below the optimal levels needed for some copper ion chaperones . Lack of copper ions would result in a lower number of functional receptors leading to the triple response . If this is true , we should find some copper ion transport mutants that are either hypersensitive or resistant to triplin treatment . Previous studies showed RAN1 is a key copper ion transporter involved in ethylene receptor biogenesis [12–14] . However the mechanisms for how copper ions are delivered to RAN1 and how RAN1 delivers copper to the ethylene receptors are presently missing . Therefore , we took advantage of triplin to look for the protein ( s ) that deliver copper to RAN1 . We speculated that mutations in such proteins should be hypersensitive to triplin treatment . We therefore tested triplin sensitivity of T-DNA mutants of Arabidopsis that affect copper ion transport , including copt1 , copt2 , copt4 , copt5 , hma1 , hma5 , hma6 , ccs , zip2 , zip4 , cch and atx1 . This uncovered that only the copper ion chaperone mutants atx1-1 and atx1-2 are hypersensitive to triplin ( Fig 5A and 5B ) . ATX1 is a copper ion chaperone playing an essential role in copper ion homeostasis that confers plant tolerance to both copper excess and deficiency conditions in Arabidopsis [25] . atx1-1 ( SALK_026221 ) is a knock-out mutant identified previously [25] , atx1-2 ( SALK_041022 ) is another knock-out mutant identified in this research ( S11 Fig ) . Both mutants showed hypersensitivity to triplin treatments . Application of 20 μM triplin to atx1-1 and atx1-2 seedlings caused a triple response as seen with the exaggerated apical hook and shorter hypocotyl . By contrast , 20 μM triplin did not cause a triple response phenotype in wild type seedlings . Re-introducing the ATX1 gene into the atx1-1 mutants restored a wild type response to triplin ( Fig 5B ) . The hypersensitivity of atx1-1 to triplin could be partially reduced by adding copper . Furthermore , atx1-2 etr1-1 and atx1-1 ein2-5 double mutants were resistant to triplin treatment comparable with the etr1-1 and ein2-5 single mutants ( S11C Fig ) . This indicates that triplin is acting to result in lower levels of copper delivered to RAN1 , which in turn , reduces deliver of copper to the ethylene receptors . Previous work showed that the ethylene receptors are localized at ER membrane network [35] . To further examine how ATX1 is involved in RAN1 mediated copper ion transport to ethylene receptors , we compared the subcellular localization of ATX1 and RAN1 . First , we used 5-day-old light grown 35S:ATX1-GFP ( atx1-1 ) transgenic lines and observed filar and cloudy GFP signals in the cytoplasm and around the nucleus of the root meristem zone cells . In root elongation zone cells , stronger GFP signal were observed in the cytoplasm and nucleus . Follow up observations using 4' , 6-diamidino-2-phenylindole ( DAPI ) stain confirmed the nuclear signals observed in 35S:ATX1-GFP ( atx1-1 ) transgenic lines [24 , 25] ( S12A and S12B Fig ) . For the transgenic lines grown in dark , we observed strong GFP signals in the cytoplasm and nucleus of the hypocotyl cells comparable with what was observed in root elongation zone cells ( S12C Fig ) . We speculate that , in mature cells , the expression and signal of ATX1-GFP is stronger than in younger cells . We then carried out plasmolysis experiments and showed that these GFP signals observed in intact cells are not localized at the cell wall . We next scanned a thin layer of a cell and observed similar filar and cloudy GFP signals , which may be from ATX1-GFP adhered to the cell endomembrane system ( S12D Fig ) . Using a transgenic line expressing both ATX1-GFP and WAK2-mCherry , an ER marker , we observed the ATX1-GFP signals were partially co-localized with the ER markers ( S12E Fig ) . To confirm this observation , we also used a transient expression system to express ATX1-GFP and WAK2-mCherry in tobacco ( Nicotiana benthamiana ) leaf epidermal cells to observe the sub-cellular location of ATX1 . The observations in these experiments are consistent with our previous observations that ATX1 is localized to the cytoplasm and nucleus and may adhere to ER membranes ( S13A Fig ) . Using this transient expression system , we observed that GmMAN1-mcherry ( a Golgi marker ) displayed a marked punctate signal and OsREM4 . 1-mCherry ( a plasma membrane marker ) showed no filar signals , which are different from the ATX1-GFP pattern . This confirmed that ATX1 did not specifically localize to the plasma membrane and Golgi . Only WAK2-mCherry showed filar signal ( S13A Fig ) . We then examined the subcellular location of RAN1 using the transient expression system to express RAN1-GFP and WAK2-mCherry and we observed that RAN1-GFP was localized to the endomembrane and also was co-localized with the ER marker ( S13B Fig ) . Further we wondered whether ATX1 and RAN1 are co-localized in a cell . To address this question , we used the transient expression system to express RAN1-GFP and ATX1-RFP , and we observed both proteins were co-localized on the ER membrane ( Fig 5C ) . Together , our sub-cellular observations indicate both ATX1 and RAN1 can adhere to ER membranes . To explore the possibility of ATX1 and RAN1 interacting in planta , we first carried out yeast two-hybrid experiments . ATX1 was fused to a GAL4 DNA-binding domain ( BD ) and RAN1-N ( 289 amino-terminal amino acids of RAN1 ) was ligated to a GAL4 activation domain ( AD ) . Three days after these co-transformed yeast cells were grown on a synthetic dropout ( -Leu/-Trp/-His/-Ade ) medium with X-α-Gal ( 50 mg/L ) , we observed clones became blue , suggesting an interaction between ATX1 and RAN1-N in yeast cells ( Fig 5D ) . This was next confirmed by using bimolecular fluorescence complementation ( BiFC ) assays in Nicotiana benthamiana leaves . For this , 35S: cLUC-ATX1 and 35S: nLUC-RAN-N constructs were made and co-transformed into the plants , and the results showed that these proteins interact with each other ( Fig 5E ) . The co-immunoprecipitation ( Co-IP ) assays using ATX1-GFP/FLAG and RAN1-N-FLAG/GFP also showed that ATX1 interacts with RAN1 ( Fig 5F ) . These results indicate that ATX1 physically interacts with RAN1 in planta . Together , the above data supports evidences for the hypothesis that the copper ions required for ethylene receptor biogenesis and signaling are transported through ATX1 to RAN1 and finally to ethylene receptors .
The functional characterization of genes using genetic approaches in plants depends on observable phenotypes when the gene being studied is mutated . For ethylene signaling , this mutational approach has long been saturated . It is important and pressing to overcome this problem . Recently developed chemical genetics approaches using specific active small molecules provide reversible , time and strength controllable genetic perturbations for genetic research and gene function characterization [36] . In this research , we applied a high-throughput plant chemical genetics screening and uncovered triplin , a novel small molecule , which cause a triple response in dark-grown Arabidopsis seedlings . Further we demonstrated triplin is a novel copper ion chelator and it alleviates the toxic effects of high copper ion levels on root growth . Additionally , the triple response phenotype resulting from triplin treatment is reduced by adding Cu2+ and Ag+ . This is very different from another ion chelator neocuproine . Although neocuproine causes a triple response phenotype , this effect is suppressed by Zn2+ but not Cu2+ . There is some white precipitate when we mixed Zn2+ and neocuproine ( S7 Fig ) . While low concentration of copper ions could partially restore the hypersensitivity of ran1-2 to neocuproine ( S10B Fig ) . Another problem with neocuproine is that adding it to plant growth medium exaggerates the toxic effects of high levels of copper ions on plant root growth ( S10C Fig ) . So we speculate that neocuproine could cause triple response phenotype by chelating copper ions like triplin , but its specificity and safety are questionable . Other copper ion chelators such as BCS do not elicit the triple response phenotype . We believe that triplin not only chelates copper ions in the plant growth medium , but also chelates copper ions in plant cells once it enters into plant cells . We predict that this restricts the access of copper ions needed for ethylene receptor biogenesis , and this condition lead to the triple response phenotype . Our study shows that triplin is different from other characterized copper chelators and it has the potential to perturb copper ion transport and to dissect the ethylene signaling network . We still know little about the chemistry of triplin . One thing to note is triplin does not affect ethylene binding to ETR1 directly ( Fig 4D ) . This may be related to its limited solubility in water and we noticed its precipitation was enhanced by adding CuSO4 or AgNO3 ( S11C Fig ) . One possibility is its low solubility inhibits triplin to affect ethylene binding to ETR1 directly . However triplin could enter growing yeast cells expressing ETR1 when it was added to the yeast growing medium in which triplin decreased the copper level of yeast cells and affected ETR1 biosynthesis . This possibly resulted in producing abnormal ethylene receptors which could not bind ethylene properly ( Fig 4D ) . It is important to characterize triplin in term of its chemistry in future . Copper ion transport from the copper transporter RAN1 to the ethylene receptors for normal receptor biogenesis and function has been proposed , but many important details in this process are still missing [12–14] . For example , what component provides copper ion to RAN1 and where does this copper ion relay occur ? These are some important questions we need to address in copper ion transport coupled ethylene receptor biogenesis and functioning . Our results examining the effects of triplin on known ethylene-related mutants indicates that triplin acts upstream of the receptors but does not affect ethylene biosynthesis . For example , etr1-1 and ein2 are both resistant to triplin . Furthermore , a triplin hypersensitive mutant screening identified both ran1 and atx1 mutants which are posited to function upstream of the receptors to deliver copper for normal receptor biogenesis . The hypersensitivity of both the ran1 and axt1 mutants was abolished by the mutations in downstream ethylene signaling network components such as etr1-1 ( Fig 2C and S11C Fig ) . We have noticed that atx1-1 and atx1-2 are less sensitive to triplin than ran1-1 and ran1-2 ( Fig 5B ) , and the similar results were observed when treated with neocuproine ( S10D Fig ) . These results suggest ATX1 may be positioned in the network prior to RAN1 . Our ethylene binding assays indicate that triplin is not directly affecting the ethylene receptors , but appears to be affecting delivery of copper ions to the receptors . This supports the hypothesis that triplin targets copper ion transport upstream of the receptors . Although RAN1 has been proposed to play a key role in copper ion transport in ethylene receptor biogenesis and signaling , the subcellular location of RAN1 has not been determined [12 , 27] . The function of the well-known copper ion chaperone ATX1 is difficult to study by using conventional genetics approaches , since mutants such as atx1-1 do not show a phenotype related to ethylene [25] . By taking advantage of triplin , ethylene signaling mutants and manipulating the concentration of copper ions in the plant growth medium , we provided genetic evidence for the first time of the interaction between RAN1 and ATX1 , which contributes to transport copper ions to the ethylene receptors . ATX1 localizes widely in cells , indicating that it likely functions to transport copper ions to other targets in addition to RAN1 . This is consistent with the report that ATX1 is essential for copper homeostasis . For example , it may act as a copper-buffer in plant cells [24 , 25] . The other copper chaperone in Arabidopsis is CCH , but cch mutants have no ethylene related phenotypes . Additionally , we found that cch mutants had wild type responses to triplin suggesting that CCH is not important for delivery of copper ions to RAN1 and eventually , the ethylene receptors . The subcellular localization of RAN1 was not identified previously , but based on the phenotypes of its mutants it seems likely that RAN1 is active in an endomembrane system compartment , perhaps the ER [12 , 27] . Our microscopic observations using ATX1 , RAN1 , and organelle markers , as well as the results of molecule interaction assays in planta indicate that RAN1 and ATX1 are partially co-localized and interact on the ER . This supports the idea that ATX1 and RAN1 interact as part of copper transport to the ethylene receptors . Copper ions are important to human health where its disorder can result in various diseases . More and more reports show that copper disorders are closely related to human Alzheimer's disease and cancers . Therefore various copper chelators have been identified and characterized in order to cure these important human diseases [37–40] . Although a few copper ion chelators have been used in therapy , thus far their mechanisms of action are not clear [41] . One problem using these chelators is they are not very specific for copper ions [42] , and some chelators , such as neocuproine , might enhance the toxic effects of copper ions on the organism being treated as we show in this study for plant growth . To this end , triplin is a unique copper ion chelator that may provide a new lead structure for copper ion chelators related to developing drug . Thus , the model plant Arabidopsis could be another useful platform to carry out studies to uncover copper ion chelator and explore its action mechanism in vivo .
The plant chemical genetics screenings using dark-grown Arabidopsis seedlings against a structure novel and diverse synthetic chemical library of 12 , 000 small molecules from Life Chemicals Inc was performed as previously reported [32 , 33] . For all screenings , surface-sterilized Arabidopsis Col-0 seeds suspended in 0 . 1% agar were evenly distributed into 96-well plates that contained 0 . 8% agar , 0 . 3xMS salts ( Sigma-Aldrich ) , 100 μM individual chemical per well and 1% DMSO ( carrier solvent ) . Seeds were stratified for 3 days in a 4 C refrigerator , transferred to day-light for 1–4 hours then transferred in dark to grow for 3 days at 22 C in a light-tight growth cabinet . Seedlings phenotypes were recorded and imaged using a SZX16 dissecting microscope . Chemicals that caused dark-grown Col-0 seedlings triple response like phenotypes were retested . For each active chemical , its dose effect on the length of the hypocotyls or roots of assayed seedlings were measured using Image J ( NIH ) . To acquire triplin resistant mutant , we used a mutant screening strategy that is modified from [30] . Briefly 20 , 000 M2 seeds from 5 , 000 ethylmethane sulfonate ( EMS ) -mutagenized M1 Col-0 plants were surface-sterilized and grown under the same growth conditions and chemical dose as the plant chemical genetics screenings . The phenotypes of seedlings were examined under the microscope , and these with longer hypocotyl or without exaggerative apical hooks were considered as putative triplin resistant mutants , and all these mutants were then retested for their chemical genetics phenotypes in the next generation . Ethylene measurements on plants were performed as previously described [43 , 44] . Col-0 seeds were surface sterilized and planted on 0 . 5xMS solid medium . After stratification at 4 C for 3 days , plates were exposed to light for 1–4 hours and then transferred in dark to grow for 2 days at 22 C . After seed germination , every 22 seedlings per group ( n = 3 ) were transferred to a gas chromatography vial with half volume 0 . 5xMS growth medium contain 100 μM triplin , 50 μM ACC or 1%DMSO and incubated for 3 days under continual dark . Accumulated ethylene was measured by gas chromatography ( Agilent Technologies , 6890N Network GC System ) [44] . Seedlings weight is fresh weight . Three-day-old etiolated Col-0 seedlings grown on 0 . 5xMS solid medium with 20 μM CuSO4 and/or 100 μM triplin were collected . The copper ion contents in plants were then determined by inductively coupled plasma-mass spectrometry as previously described [45] . For these assays , 100 μl of the metal ions ( 100 μM or 10 mM ) and 100 μl triplin ( 100 μM or10 mM ) were mixed for MALDI-TOF-MS analysis according to [46] . A time-of–flight Axima Performance mass spectrometer ( Shimadzu , Japan ) was used . All mass spectra in this work were acquired at positive ion reflection mode . The data were controlled by a software application of MALDI-MS . The matrix is 5 mg/ml α-Cyano-4-hydroxycinnamic acid ( CHCA ) . Mass spectrum scanning range is 400-1000Da . For this , we expressed the first 128 amino acids of ETR1 containing the ethylene binding domain as a fusion protein to GST ( glutathione S-transferase ) ( ETR1 [1–128]-GST ) in Pichia pastoris as previously described [47] . Binding assays were then conducted on either intact yeast cells or isolated membranes as previously described using a radioligand binding assay [48 , 49] . Briefly , to test the effects of triplin on ethylene binding to its receptors in intact yeast cells , yeast were incubated at 30°C in the presence or absence of 100 μM triplin for 3 days . Intact yeast cells were then harvested and assayed for binding . For binding assays to membranes , cells were disrupted and membranes isolated as previously described [11] and ethylene binding determined in the presence or absence of 100 μM triplin . Saturable ethylene binding is indicated as counts per minute ( CPM ) and was calculated by subtracting the amount of radioactivity bound in the presence of excess non-radioactive ethylene from what was bound in the absence of unlabeled ethylene . We used western blots with anti-GST antibodies to ensure equal levels of ETR1 [1–128]-GST [11] . The constructs and plant transformation were performed as previously described [32] . The coding sequence ( CDS ) of ATX1 ( AT1G66240 ) and RAN1 ( AT5G44790 ) was amplified from cDNA of Col-0 by PCR using primers listed in S2 Table . Fragments were then cloned into the entry vector pDONR-zeocin by BP reactions by following the instructions of the manufacturer ( Life Technology , USA ) . Then genes were introduced to different pGWB vectors by LR reactions . The vectors were then used to transform Agrobacterium strain GV3101 , the transformed agrobacteria were finally used to transform flowering Arabidopsis plants via the floral-dip method [50] . The qRT-PCR was done as previously described [32] . For testing the expression level of ERF1 , seeds were surface sterilized and planted on 0 . 5xMS growth medium contain different chemicals in petri dish plates , after stratification at 4 C for 3 days , plates were exposed to light for 1–4 hours then transferred to dark for 3 days ( 22 C ) , and seedlings were collected . For testing the expression level of RAN1 in wild type and transgenic plants , 3-week old plant leaves were collected and total RNAs were extracted and used to synthesize the cDNAs by reverse transcription . Primers for ERF1 were made according to [43] . All primers used were listed in S2 Table . The ACTIN gene was amplified and used as an internal control . pDONR-ATX1 and pDONR-RAN1 were introduced into pGWB vectors as described in [32] . The construct combinations of ATX1:PGWB5/PGWB654 and RAN1:PGWB605 were used to transform Agrobacterium GV3101; then the Agrobacterium were infiltrated into leaves of N . benthamiana with P19 [32] . After 2-day incubation , the transformed plant leaves were observed and imaged under a confocal microscope ( Olympus FV1000 ) . For transgenic plants , the 5-day-old light grown roots or 3-day-old dark-grown hypocotyls were imaged under the confocal microscope . However photos in S12C and S12D Fig were imaged by NIKON A1R . ER , Golgi and plasma membrane associated protein markers used is AtWAK2 , GmMAN1 and OsREM4 . 1 respectively [51 , 52] . Arabidopsis WAK2-mCherry or GmMAN1-mCherry transgenic seeds are gifts from Chi-Kuang Wen lab of Shanghai Institute of Plant Physiology and Ecology , Chinese Academy of Sciences . For DAPI staining , seedlings were fixed in 5% methanol , immersed in 2 μg/ml DAPI in phosphate buffer saline ( PBS ) and viewed by fluorescence microscopy under UV light . Yeast two-hybrid experiments were performed as previously described [32] . pGADT7 ( RAN1 N terminal ) and pGBKT7 ( ATX1 ) were transformed simultaneously into the yeast strain AH109 , the colonies were transferred to a SD ( -Leu/-Trp/-His-Ade ) solid medium with X-α-gal according to the manufacturer’s protocols ( Clontech ) for 3 days . The positive clones were identified because they grew well and became blue . BiFC and Co-IP assays were performed as previously described [32 , 52] . The Split-Luciferase Complementation system was used . The construct combinations of cLuc-ATX1/nLuc-RAN1-N were used to transform Agrobacterium GV3101; then the Agrobacterium were infiltrated into leaves of N . benthamiana with P19 as previously described . After 2-day incubation , 1 mM luciferin ( Sigma ) was filtrated into the leaves and the pictures were recorded using a CCD imaging system ( Berthold , https://www . berthold . com/ ) . For the Co-IP assays , the N . benthamiana leaves were transfected with Agrobacterium containing ATX1-GFP/FLAG and RAN1-N-FLAG/GFP , and incubated for 2 days . Total proteins were extracted with extraction buffer [50 mM Tris-HCl , pH 7 . 5 , 150 mM NaCl , 5 mM MgCl2 , 1 mM EDTA , 1% Triton X-100 , and 0 . 1% protease inhibitor cocktail ( Promega ) ] for 20 minutes , and centrifuged at 4°C at 12 , 000 rpm for 20 min . The supernatant was incubated with pretreated anti-FLAG antibody coupled agarose beads ( Abmart ) for 2 hours at 4°C . Beads were washed three times with wash buffer ( 50 mM Tris-HCl , pH 7 . 5 , 150 mM NaCl , 5 mM MgCl2 , 1 mM EDTA , 1% Triton X-100 ) . The bound proteins were eluted with 2xSDS loading buffer and boiled at 100°C for 5 minutes . The eluted proteins were separated on SDS-PAGE and immunoblotted with anti-FLAG antibody ( Abmart ) and anti-GFP antibody ( Abmart ) . ATX1 ( AT1G66240 ) , RAN1 ( AT5G44790 ) , CCH ( AT3G56240 ) , ETR1 ( AT1G66340 ) , EIN2 ( AT5G03280 ) , EIN3 ( AT3G20770 ) , Triplin ( F0655-1171 )
|
Copper ions are cofactors of protein functions , and their disorder is closely related to many human diseases which drive to develop new copper chelator related drugs . In plants , copper ions are essential for ethylene receptors , but many details are unclear . Researchers need novel specific copper chelators to study these questions . Here , by using plant chemical genetics , we identified a novel chemical triplin , which could activate ethylene signaling pathway by chelating copper ions essential for ethylene receptors . Ethylene resistance mutant etr1-1 and ein2 were resistant to triplin , and copper transporter mutants ran1-1 and ran1-2 were hypersensitive to triplin . Using triplin as a molecular genetics tool , we showed copper chaperone ATX1 acts the upstream of RAN1 which transports copper ions to ethylene receptors . We provide a sample that metabolism could be studied by combining chemical genetics and known signaling pathway . Moreover , triplin could chelate copper ions effectively and specifically in Arabidopsis , and could be a useful tool to study copper ion transport in plant , and valuable for designing copper chelator related drug .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"plant",
"anatomy",
"chemical",
"compounds",
"plant",
"growth",
"and",
"development",
"ethylene",
"signaling",
"cascade",
"ethylene",
"plant",
"embryo",
"anatomy",
"brassica",
"organic",
"compounds",
"hormones",
"developmental",
"biology",
"plant",
"science",
"model",
"organisms",
"plant",
"hormones",
"experimental",
"organism",
"systems",
"seedlings",
"plants",
"fruit",
"and",
"seed",
"anatomy",
"copper",
"research",
"and",
"analysis",
"methods",
"arabidopsis",
"thaliana",
"plant",
"embryogenesis",
"plant",
"development",
"chemistry",
"biophysics",
"physics",
"biochemistry",
"plant",
"biochemistry",
"chemical",
"elements",
"signal",
"transduction",
"ion",
"transport",
"plant",
"and",
"algal",
"models",
"hypocotyl",
"organic",
"chemistry",
"cell",
"biology",
"embryogenesis",
"biology",
"and",
"life",
"sciences",
"biosynthesis",
"physical",
"sciences",
"cell",
"signaling",
"organisms",
"signaling",
"cascades"
] |
2017
|
Triplin, a small molecule, reveals copper ion transport in ethylene signaling from ATX1 to RAN1
|
Orientia tsutsugamushi , the etiologic agent of scrub typhus , is a mite-borne rickettsia transmitted by the parasitic larval stage of trombiculid mites . Approximately one-third of the world's population is at risk of infection with Orientia tsutsugamushi , emphasizing its importance in global health . In order to study scrub typhus , Orientia tsutsugamushi Karp strain has been used extensively in mouse studies with various inoculation strategies and little success in inducing disease progression similar to that of human scrub typhus . The objective of this project was to develop a disease model with pathology and target cells similar to those of severe human scrub typhus . This study reports an intravenous infection model of scrub typhus in C57BL/6 mice . This mouse strain was susceptible to intravenous challenge , and lethal infection occurred after intravenous inoculation of 1 . 25×106 focus ( FFU ) forming units . Signs of illness in lethally infected mice appeared on day 6 with death occurring ∼6 days later . Immunohistochemical staining for Orientia antigens demonstrated extensive endothelial infection , most notably in the lungs and brain . Histopathological analysis revealed cerebral perivascular , lymphohistiocytic infiltrates , focal hemorrhages , meningoencephalitis , and interstitial pneumonia . Disseminated infection of endothelial cells with Orientia in C57BL/6 mice resulted in pathology resembling that of human scrub typhus . The use of this model will allow detailed characterization of the mechanisms of immunity to and pathogenesis of O . tsutsugamushi infection .
Orientia tsutsugamushi , a gram-negative obligately intracellular coccobacillus , is the etiologic agent of scrub typhus [1] . Scrub typhus is a serious public health problem in Asia , northern Australia , and islands of the western Pacific and Indian Oceans including Korea , Japan , China , Taiwan , Indonesia , India and Thailand , and threatens one billion persons globally and causes illness in one million people each year [2] . One to two weeks after being fed upon by an infected larval Leptotrombidium mite , patients exhibit signs of infection such as an inoculation site eschar followed by lymphadenopathy , fever and rash accompanied by non-specific flu-like symptoms . Without appropriate treatment , scrub typhus can cause severe multiorgan failure with a case fatality rate of 7–15% . Doxycycline , azithromycin , rifampicin , and chloramphenicol are the antibiotics used to treat Orientia infection and are effective if begun early in the disease course [3] . However , misdiagnosis , inappropriate antibiotic treatment , and antibiotic failures have occurred , supporting the need for a vaccine [4] . Antigenic heterogeneity and immunity that wanes after infection leading to reinfections are major obstacles to vaccine development [5] , [6] . The current reemergence of scrub typhus further emphasizes the need for the development of a vaccine , which requires appropriate animal models for determining mechanisms of immunity , candidate vaccine efficacy , and correlates of immune protection . During Orientia infection , the bacteria infect endothelial cells , macrophages , cardiac myocytes , and dendritic cells [7] , [8] . Similar to most of the bacteria belonging to the family Rickettsiaceae , Orientia exhibits endothelial tropism . As such , scrub typhus is a disseminated endothelial infection that affects all organs . Primary characteristics of fatal scrub typhus pathology include diffuse interstitial pneumonia , hepatic lesions , meningoencephalitis , and coagulation disorders [7] , [9]–[11] . However , current murine models fail to reproduce the pathology of human scrub typhus . The lack of an appropriate animal model of scrub typhus has fundamentally impeded progress in this field . The currently available model , first employed more than 60 years ago , uses intraperitoneal inoculation of O . tsutsugamushi into mice , which establishes infection in the peritoneal cavity [12] , [13] . Continuous proliferation of O . tsutsugamushi in peritoneal macrophages and mesothelial cells , enlargement of spleen and liver , and severe peritonitis occur in intraperitoneally inoculated mice [14]–[16] . However , the pathology of O . tsutsugamushi in human scrub typhus consists of disseminated endothelial injury and lymphohistiocytic vasculitis as the basis for interstitial pneumonitis , hepatic damage , and encephalitis [7] , [11] , [17] , which differ considerably from the histopathologic lesions of intraperitoneally inoculated mice . Based on previous histopathological and immunohistochemical observations of scrub typhus in humans and our experience with models of rickettsioses , we hypothesized that inoculation of mice via the intravenous route would result in scrub typhus-like pathology . Our studies have determined reproducibly lethal and sublethal doses as well as the LD50 of intravenously inoculated O . tsutsugamushi in mice . We employed these doses to compare the histopathology with those of mice challenged using the classic intraperitoneal model . All parameters , including bacterial loads and histopathology of lungs , spleen , liver and brain confirmed that the intravenously inoculated mice provide a model of scrub typhus that closely resembles the human disease .
C57BL/6 ( B6 ) and C3H/HeN ( C3H ) mice were purchased from Harlan Laboratories , Houston , TX . Age- and gender-matched , 8–12 week old mice were used in all studies . Experimentally infected mice were housed in an animal biosafety level 3 facility , and all experiments and procedures were approved by the Institutional Animal Care and Use Committee ( IACUC ) of the University of Texas Medical Branch ( Protocol: 9008082 ) , Galveston in accordance with Guidelines for Biosafety in Microbiological and Biomedical Laboratories . UTMB operates to comply with the USDA Animal Welfare Act ( Public Law 89-544 ) , the Health Research Extension Act of 1985 ( Public Law 99-158 ) , the Public Health Service Policy on Humane Care and Use of Laboratory Animals , and the NAS Guide for the Care and Use of Laboratory Animals ( ISBN-13 ) . UTMB is a registered Research Facility under the Animal Welfare Act , and has a current assurance on file with the Office of Laboratory Animal Welfare , in compliance with NIH Policy . Orientia tsutsugamushi Karp strain was cultivated in Vero E6 cells or serially passaged in 8–12 week old female C57BL/6 mice ( Harlan Laboratories , Houston , TX ) . For cultivation in Vero cells , bacteria were inoculated onto confluent monolayers in T150 cell culture flasks and gently rocked for two hours at 34°C , at the end of which Dulbecco's Modified Eagles Medium ( DMEM , Gibco ) with 1% fetal bovine serum ( FBS ) and 1% HEPES buffer were added . Cells were observed for cytopathic effect , which usually occurred at 14–21 days . When areas of rounded or floating cells were observed throughout the flask , a smear was prepared , and the level of infection was assessed either by Dif-Quik ( Fisher Scientific , Kalamazoo , MI ) or immunofluorescence staining . When the flask reached 80–90% of cells infected , the cells were removed and seeded onto fresh Vero cell monolayers . This process was repeated for a total of six passages . Infected flasks were harvested by scraping , and cell suspensions were collected in Oakridge high speed centrifugation bottles and centrifuged at 22 , 000× g for 45 minutes at 4°C . The pellet was resuspended in sucrose-phosphate-glutamate ( SPG ) buffer ( 0 . 218 M sucrose , 3 . 8 mM KH2PO4 , 7 . 2 mM KH2PO4 , 4 . 9 mM monosodium L-glutamic acid , pH 7 . 0 ) and transferred to a 50 mL conical tube containing 5 mL of sterile glass beads . The conical tubes were gently vortexed at 10 sec intervals to release the intracellular bacteria and placed on ice . The tubes were then centrifuged at 700× g to pellet cell debris , and the supernatant was collected . The tubes were then centrifuged at 22 , 000× g for 45 minutes to pellet cell-free bacteria . Pellets were resuspended in SPG buffer and stored at −80°C until used . Preliminary studies were conducted using cell cultured orientiae . Animal passages were performed to rapidly produce high titered oriential stocks from infected tissues . Two groups of four 8–12 week old female C57BL/6 mice were inoculated intravenously with 1 . 25×106 focus forming units ( FFU ) of Orientia cultivated in Vero E6 cells . When the animals exhibited signs of illness , i . e . , hunched posture , lethargy , and ruffled fur , usually at six days post infection , they were euthanized , and the liver and lungs aseptically collected and placed in DMEM . Organ-specific pools were homogenized using a 7 mL glass Dounce apparatus . Homogenized samples were then rinsed with cold SPG buffer and placed in a 50 mL conical tube and centrifuged at 700× g for 10 minutes at 4°C to pellet the tissue debris . Supernatant fluid was collected and placed on ice . The tissue pellets were resuspended in 5 mL of SPG buffer , homogenized and centrifuged as above . Organ-specific supernatants were pooled on ice and then centrifuged at 22 , 000× g for 45 minutes at 4°C in a Beckman high speed centrifuge . The pellets were resuspended in 10 mL of SPG buffer , aliquoted , and stored at −80°C . It takes greater than 15 days for plaques to form in O . tsutsugamushi-infected monolayers [18 , personal observation] . Thus , in order to quantitate the number of viable bacteria in a timely manner , a focus forming assay was used [19] . Vero E6 cells in DMEM with 1% FBS and 1% HEPES were seeded onto 12-well plates and allowed to attach overnight at 37°C in a 5% CO2 atmosphere . Once the cells were confluent , serial 10-fold dilutions of oriential stocks were prepared , and 200 µL aliquots were seeded onto the confluent monolayers in triplicate . The inoculated plates were centrifuged for 5 minutes at 700× g to facilitate bacterial attachment and then incubated for two hours at 37°C in 5% CO2 . After two hours , the wells were rinsed three times with warm Dulbecco's PBS ( Cellgro , Manassas , VA ) with calcium and magnesium to remove extracellular non-viable bacteria . The wells were then overlaid with DMEM containing 1% FBS , 0 . 5% sterile methylcellulose , and 2 µg cyclohexamide and incubated at 34°C for 5 days . After 5 days , the overlay was aspirated and the monolayers gently rinsed as above . The monolayers were fixed in methanol for 30 minutes at 4°C , after which the methanol was removed and the wells rinsed as above . Wells were blocked using PBS with 1% BSA for 30 minutes at room temperature . Blocking buffer was then removed and the wells washed three times with 0 . 5% Tween-20 in PBS . An aliquot of primary polyclonal rabbit anti-O . tsutsugamushi Karp strain antibody ( 1∶500 dilution ) was added to each well and incubated at room temperature for 30 minutes . The primary antibody was removed and wells washed as above . Alexa-594 goat anti-rabbit IgG ( Invitrogen , Carlsbad CA ) diluted 1∶1 , 000 was added to each well , incubated for 30 minutes and then washed as above . Wells were examined using an inverted fluorescent microscope . Wells containing 10–100 foci of cells infected with Orientia were counted , and the concentration of focus-forming units was calculated . A large number of mouse strains has been shown to be susceptible to intraperitoneal inoculation of O . tsutsugamushi Karp strain including C57BL/6 and C3H/HeN mice [20] , but very few data exist describing susceptibility via intravenous challenge . Our laboratory has extensive experience with mouse model development for rickettsioses with these mouse strains [21]–[24] . In this study , both C57BL/6 and C3H/HeN mice strains were compared for susceptibility to Orientia infection via intravenous challenge . To determine the infectivity of the cell culture stock , intraperitoneally ( i . p . ) inoculated animals were studied in parallel with intravenously ( i . v . ) inoculated animals to ensure infectivity of the Orientia . Animals were challenged with 2 . 5×106 or 1 . 25×106 organisms and were observed daily for signs of illness for 28 days post infection ( dpi ) or until animals became moribund . Orientia tsutsugamushi Karp strain , passaged and maintained as described above , was diluted in PBS , and the bacteria were injected i . v . through the tail vein or i . p . in a volume of 200 µL . Control mice were inoculated with 200 µL of similarly prepared material from uninfected cells or tissue diluted in PBS . Animals were challenged with 1 . 25×106 , 1 . 25×105 , and 1 . 25×104 organisms either i . p . or i . v . All infected and non-infected animals were monitored for signs of illness and body weight measured daily until day 21 . Mice were necropsied at 3 , 6 , 9 , and 12 dpi or when moribund for lethally challenged animals and also at 15 dpi for sublethal i . v . challenge . Four randomly selected i . p . and i . v . inoculated animals were euthanized , and blood , brain , heart , kidney , liver , lung , lymph nodes , and spleen were collected for histopathology and blood , liver , lung , and spleen for bacterial load determination . All tissues were fixed in 10% neutral buffered formalin and embedded in paraffin . Tissue sections ( 5 µm thickness ) were stained with hematoxylin and eosin or processed for immunohistochemistry . Immunohistochemical staining was used to assess cellular distribution and intensity of Orientia infection in the organs of experimental animals . Sections were deparaffinized and rehydrated . The sections were placed on poly-L-lysine -coated slides and incubated at 70°C for 20 minutes , then rehydrated in water and treated with antigen retrieval solution . Antigen retrieval was accomplished by incubation in citrate buffer ( pH = 6 ) at 98°C for 20 minutes followed by casein endogenous IgG blocking for 15 minutes to reduce possible species cross reactivity . Endogenous alkaline phosphatase activity was quenched by incubation with Levamisole ( Sigma Aldrich , St . Louis , MO ) for 15 minutes and slides rinsed in deionized water . Nonspecific binding of antibody was blocked by incubating sections with normal goat serum and avidin blocking reagent ( Vector Laboratories , Burlingame , CA ) mixture ( 1∶10 ) for 30 minutes . Sections then were incubated for 2 hours with polyclonal rabbit anti-O . tsutsugamushi Karp strain antibody ( dilution: 1∶500 ) , followed by incubation for 30 minutes with biotinylated anti-rabbit IgG ( 1∶2000 , Vector Laboratories , Burlingame , CA ) . Signals were detected by the labeled streptavidin-biotin method with an UltraVision Alk-Phos kit ( Thermo Scientific , Waltham , MA ) . Vector Red Alkaline Phosphatase substrate ( Vector Laboratories , Burlingame , CA ) was used as chromogen , and counterstaining was performed with hematoxylin . Reagent negative controls consisted of samples in which primary antibody was replaced with normal rabbit IgG . Sections were mounted in Permount . Bacterial loads were assessed by quantitative real-time PCR [25] . DNA was extracted using a DNeasy Kit ( Qiagen , Gaithersburg , MD ) from the tissue samples , and the bacterial load at each time point and for each organ sampled was determined by quantitative real-time PCR [25] . The 47 kDa gene was amplified using the primer pair OtsuF630 ( 5′-AACTGATTTTATTCAAACTAATGCTGCT-3′ ) and OtsuR747 ( 5′-TATGCCTGAGTAAGATACGTGAATGGAATT-3′ ) primers ( IDT , Coralville , IA ) and detected with the probe OtsuPr665 ( 5′-6FAM-TGGGTAGCTTTGGTGGACCGATGTTTAATCT-TAMRA ) ( Applied Biosystems , Foster City , CA ) . Bacterial loads were normalized to total nanogram ( ng ) of DNA per µL for the same sample and expressed as the number of 47 kDa gene copies per picogram ( pg ) of DNA . At the time points that animals were euthanized , blood samples were collected in K2EDTA-coated BD microtainer tubes ( Becton Dickinson , Franklin Lakes , NJ ) and blood cell counts performed using a 950FS HemaVet apparatus ( Drew Scientific Inc . , Waterbury , CT ) that differentiates cell types by size and granularity in a 20 µL sample of whole blood . Lung tissue from lethally infected animals was collected at 6 dpi and prepared for transmission electron microscopy . For ultrastructural analysis in ultrathin sections small pieces ( ∼1 mm3 ) of tissues were fixed for at least 1 hour in a mixture of 2 . 5% formaldehyde prepared from paraformaldehyde powder , and 0 . 1% glutaraldehyde in 0 . 05 M cacodylate buffer , pH 7 . 3 , to which 0 . 03% picric acid and 0 . 03% CaCl2 were added . Then they were washed in 0 . 1 M cacodylate buffer and post-fixed in 1% OsO4 in 0 . 1 M cacodylate buffer , pH 7 . 3 , for 1 hour , washed with distilled water and stained en bloc with 2% aqueous uranyl acetate for 20 min at 60°C . The samples were dehydrated in ethanol , processed through propylene oxide and embedded in Poly/Bed 812 ( Polysciences , Warrington , PA ) . Semi-thin sections 1 µm thick were cut and stained with toluidine blue . Ultrathin sections were cut on Leica EM UC7 ultramicrotome ( Leica Microsystems , Buffalo Grove , IL ) , stained with lead citrate and examined in a Philips 201 transmission electron microscope at 60 kV .
B6 and C3H mice were compared for susceptibility , bacterial loads , and histopathology . To compare mouse strain susceptibility , i . p . inoculated animals were studied in parallel with i . v . inoculated animals . Inoculation of 103 Orientia resulted in clinical illness ( ruffled fur , hunched posture , and lethargy ) in i . p . inoculated animals at 12–15 days post infection; in contrast , signs of illness at this dose in i . v . inoculated animals were mild , i . e . , slightly ruffled fur with hunched posture but normal activity . Histopathologic examination demonstrated systemic lesions most prominently in the lungs and liver of animals inoculated i . v . with this dose of O . tsutsugamushi . Lethality was observed in the intravenous model using doses of both 2 . 5×106 and 1 . 25×106 organisms . C3H mice became moribund at 7–8 days post-inoculation ( dpi ) with either dose . The B6 mice that received 2 . 5×106 organisms became moribund at 9–11 dpi; whereas those animals receiving 1 . 25×106 organisms became moribund 10–13 dpi . Mice of both strains inoculated i . p . with either dose expired at 7–8 dpi . As both strains were susceptible all further studies were conducted with B6 mice only due to the availability of a variety of genetically modified strains on this background and the longer course similar to human scrub typhus observed in B6 mice . All mice challenged i . v . with 1 . 25×106 bacteria expired by 13 dpi ( Figure 1A ) , approximately half of the animals inoculated i . v . with 105 Orientia expired between 13 and 15 dpi ( data not shown ) , and 10% of mice challenged with 104 Orientia expired by 15 dpi ( Figure 1B ) . Intraperitoneal inoculations of all doses were uniformly lethal ( Figure 1A and B ) . Controls were monitored until 21 dpi without morbidity . Disease progression of intraperitoneally inoculated mice: At 3 dpi , there were no signs of illness except mild abdominal swelling . Overall activity was unchanged , and there was no weight change ( Figure 1C ) . Mice manifested mesenteric lymphadenopathy and mild accumulation of fibrin-containing proteinaceous fluid in the peritoneal cavity causing the lobes of the liver to adhere to one another . Portions of the gastrointestinal tract were edematous and discolored . Neither histopathologic lesions nor Orientia antigen was detected on day 3 in mice inoculated i . p . At 6 dpi , mice inoculated i . p . had begun to lose weight ( Figure 1C-red triangles ) with narrowed eyes , severely hunched posture , and swollen abdomen . The animals' activity was diminished compared to uninfected controls . Orientia antigen was detected in endothelial cells in the lungs of these animals , but with minimal cellular response . Of particular interest were the moderate accumulation of peritoneal exudate and the extensive distribution of oriential antigen in cells on the peritoneal surfaces of all abdominal organs and mild-to-severe mesothelial hyperplasia ( Figure 2 A and B ) . All animals were moribund or had expired by day 9; severe peritonitis was observed with accumulation of 2–4 mL of peritoneal exudate . Orientia antigen was detected in the lungs in association with vasculitis and interstitial pneumonia . These findings indicate that Orientia had eventually disseminated from the peritoneal cavity , but the most striking observation at this time was mesothelial hyperplasia and inflammation on the peritoneal surface of liver ( Figure 2D ) and spleen ( Figure 2E ) and the extensive Orientia infection of these cells . Proteinaceous material and infiltrating cells were also observed on the peritoneal surface of the spleen ( Figure 2E ) . The cells overlying the kidney capsule ( Figure 2C ) were also infected . Disease progression of intravenously inoculated mice: At necropsy on 3 dpi , i . v . inoculated mice had generalized lymphadenopathy but no other gross lesions . The mice had perivascular lymphohistiocytic infiltrates in the meninges , and Orientia antigen was detected in the liver and lung with associated cellular infiltrates in both organs . The kidneys were unremarkable . At 6 dpi , the animals had a slight decrease in body weight ( Figure 1C-blue boxes ) and generally appeared healthy although some animals exhibited decreased activity and slightly hunched posture . At this time point , immunohistochemistry demonstrated that systemic infection was established with most of the Orientia observed in endothelial cells of the lung ( Figure 3A and B ) , kidney ( Figure 3C ) , and liver ( Figure 3D ) . Endothelial infection was confirmed by electron microscopic analysis of lung sections from i . v infected mice ( Figure 4 ) . Cellular infiltration had increased at days 9 and 12 ( Figure 5 ) . Meningitis and cerebral perivascular infiltrates ( Figure 5A , B , C , and E ) were observed on both days , and focal cerebral hemorrhage was observed at 12 dpi ( Figure 5B ) . Pulmonary vasculitis and interstitial pneumonia ( Figure 5F and G ) became more severe as the infection progressed . Hepatic ( Figure 6A and B ) inflammatory lesions became more pronounced , and multifocal mononuclear infiltrates were numerous . Cellular infiltrates between the tubules of the kidney ( Figure 6C ) were evident at 9 days post inoculation with renal vasculitis observed at 12 dpi ( Figure 6D ) . The pathologic lesions of mice inoculated i . v . became progressively more severe through the course of infection with animals expiring on days 12–13 ( Figure 1A ) . Blood , liver , lung , and spleen were monitored for bacterial loads at each time point . At this challenge dose , i . v . and i . p . inoculated animals had detectable bacterial loads in all tissues throughout the course of disease ( Table 1 ) . The peripheral blood cell counts of i . p and i . v . inoculated animals were compared to uninfected controls and published normal ranges for B6 mice . At 3 dpi all mice showed slightly elevated WBC counts , mainly neutrophils , compared to uninfected controls , but within the normal range . At day 6 , both i . p . and i . v . inoculated animals manifested leukocytosis with lymphopenia , and i . v . inoculated animals had marginally greater elevation of WBC counts than i . p . inoculated animals . Intravenously inoculated animals had neutrophil concentrations three times greater than uninfected controls . At 9 dpi , i . v . -inoculated animals had leukocytosis , mostly neutrophilia but less than on day 6 , as well as lymphopenia . Intraperitoneally inoculated animals had lymphopenia and neutrophilia that was less severe than that of i . v . inoculated animals . At 12 days , leukocytosis persisted with neutrophil concentrations being five times greater than uninfected controls . At this time , all i . p . animals had expired ( 9 dpi ) , and all i . v . animals were moribund . Disease progression of animals inoculated i . v . with 1 . 25×104 organisms paralleled that of lethally challenged animals but with signs of illness typically appearing 2–3 days later than in lethally infected animals ( Table 2 ) . Sublethally infected animals became lethargic and developed severely hunched posture at 12–13 days . With this lower dose , 10% of animals were moribund at 13 days; the remaining animals recovered clinically between 15 and 21 days after infection . Histopathologic observations at the various time points revealed that the lesions progressed similarly to those in the lethally infected animals , but with a dose-dependent delayed onset . Hepatic lesions , mild pulmonary cellular infiltration and Orientia antigen were detected at 6 dpi; interstitial pneumonia developed between 9 and 12 dpi . On day 15 , multifocal cellular infiltrates were observed in the lungs ( Figure 7 ) , kidneys , and liver . Animals inoculated i . p . with 1 . 25×104 Orientia organisms had an incubation period two days longer than animals inoculated with the high dose before signs of illness appeared . Unlike the low dose i . v . inoculated animals , this dose administered i . p . was uniformly lethal by day 15 ( Figure 1B ) . Severe peritonitis was observed with accumulation of peritoneal exudate in excess of 2 mL . Orientia antigen was detected focally in the lungs in association with vasculitis and interstitial pneumonia on day 15 . At this time point , the animals' body weights had increased , owing to the accumulation of exudate in the peritoneal cavity . Bacterial loads of animals inoculated with Orientia were monitored at each time point . At this challenge dose , i . v . and i . p . inoculated animals had detectable bacterial loads in liver , lung , and spleen throughout the disease course ( Table 3 ) . Orientia in the blood was not detected consistently until 6 dpi for i . v . inoculated animals and 9 dpi for i . p . inoculated animals . All surviving animals began to recover weight , and signs of illness resolved ( Figure 1D ) .
Scrub typhus has been described as one of most severely neglected tropical diseases; indeed it has potentially more fatal cases annually than dengue fever [3] . It was first described in China in 84 B . C . and made its presence felt during the wars that took place in the region during the last century . Animal model development is important for understanding pathogenesis and immunity and for preclinical testing of vaccines and therapeutics . Accurate animal models for diseases are imperative to developing sound understanding of the diseases . Models that do not present similar features as the human disease may provide misleading information about the disease , further impeding the understanding of disease progression . In the case of scrub typhus , the intraperitoneally inoculated mouse model that has been used for the last 60 years results in severe peritonitis , a condition that does not occur in human scrub typhus , yielding an inappropriate model for this disease . Studies of immunity , primarily in intraperitoneally inoculated mice , were performed more than 25 years ago when many contemporary tools and concepts of immunology had not been developed . The presently developed model will enable valid determination of the mechanisms of protective immunity according to contemporary concepts of immunology in an animal system that accurately models human scrub typhus . The i . v . inoculation of O . tsutsugamushi Karp strain resulted in a hematogenously disseminated scrub typhus model that reliably produced pathology and target cell tropism similar to scrub typhus in humans . Similar to other members of the Rickettsiaceae family , Orientia predominantly infects endothelial cells after dissemination from the site of mite feeding [7] . How this occurs remains to be elucidated . The intravenously infected animals developed disseminated endothelial infection and histopathology as occurs in human scrub typhus . Both B6 and C3H mice exhibited similar disease course when challenged with Karp strain , with C3H mice being marginally more susceptible than B6 mice . Both of these mouse strains have been used in scrub typhus research as well as the study of spotted fever group rickettsioses [20]–[24] . B6 mice were chosen to fully characterize the histopathology and disease course of the i . v . model due to the abundant conditional and gene knockout strains on the B6 background for use in future studies . Although Orientia disseminated after intraperitoneal inoculation , the resulting peritonitis that occurs following this mode of infection , which does not occur in human scrub typhus , was the dominant pathological feature . The tissue bacterial loads observed during this study demonstrate that the route of inoculation is pivotal in the development of scrub typhus-like pathology . Both routes of inoculation result in bacterial dissemination ( Table 1 and 2 ) , but intravenous inoculation avoids stimulating the immune response of the peritoneal cavity and thus does not generate the lethal peritonitis observed in i . p . inoculated animals . The lower dose animals had similar bacterial distribution as the high dose animals for both routes of inoculation ( Table 2 ) , but the associated pathology also developed in an inoculation route-dependent manner . The oriential antigen at time of death in i . p . inoculated animals was observed predominantly on the peritoneal surface of the liver and spleen ( Figure 2 ) while oriential antigen in i . v . inoculated animals was predominantly in endothelial cells ( Figures 3–6 ) . The pathology of scrub typhus is characterized by multifocal cellular infiltrates around the blood vessels of all organs , particularly the brain , lungs , and liver . The central nervous system ( CNS ) is frequently involved in scrub typhus infection . Headache , nausea , vomiting , transient hearing loss , confusion , neck stiffness , delirium , and mental changes may be observed [26] . Glial nodules consisting of perivascular infiltration by lymphocytes and macrophages in the neuropil as well as perivascular hemorrhage were observed in our lethal model; those findings strongly resemble the lesions and cell tropism described in humans by Allen and Spitz ( 1945 ) and Moron et al . ( 2001 ) [7] , [11] . Respiratory involvement is common in severe scrub typhus infections . Approximately 40% of scrub typhus patients manifest cough at the time of admission [27] . Interstitial pneumonia , pulmonary edema , pleural effusions , cardiomegaly , and/or focal atelectasis are observed by chest radiography in those patients [10] , [28] . The presence of respiratory symptoms is closely linked to severity of scrub typhus [29] . Pulmonary pathology observed in humans comprises interstitial pneumonia with mononuclear cell infiltrates [7] , [11] , [30] . The intravenously infected animals developed similar lesions . Hepatomegaly and modest elevations of serum aminotransferases have been documented in humans . Those laboratory abnormalities might be associated with pathological changes in the liver similar to those described here in the intravenous mouse model of fatal scrub typhus [7] , [11] , [31] . Severe scrub typhus frequently results in acute renal failure [32] . Cellular infiltrates around the microvasculature of the kidney , particularly between the tubules , were observed by Allen and Spitz ( 1945 ) [11]; similar findings were prominent during later time points in the mouse model described here . The ideal model would involve an animal closely related to humans , i . e . , nonhuman primates ( NHP ) , and mite transmission , but both of these aspects would be difficult to obtain , both from an expense and expertise point-of-view . NHPs are expensive , and acquiring the number required to characterize the basic immunology and histopathology to validate the model would be cost prohibitive . Early studies conducted using NHPs had difficulty finding individuals that had not been exposed in nature to Orientia prior to experimentation [33]–[38] . The most recent studies to use NHPs focused on temperature and weight changes as clinical indicators of disease and did not obtain histopathologic evidence that NHPs develop pathology similar to humans; in fact , lethal infections were not achieved after i . d . inoculation [39] , [40] . Secondly , very few mite colonies exist , with most containing multiple strains of Orientia [41] , which would make the study of the immune response during infection even more challenging . Only i . d . inoculation of Orientia into NHP results in eschar-like lesions [39] , [40] . The use of mice as a model , especially C57BL/6 mice , provides many advantages including the availability of reagents and the availability of gene and conditional gene knockout strains on the B6 background . This model will allow us to study lethally and sublethally challenged animals to determine the factors that play a role in severe disease and using knockouts , adoptive transfers , or other methods to modulate the immune response to increase survival and decrease disease severity . There have been articles published recently addressing model development for scrub typhus [42] , [43] . The i . d . inoculation of Orientia by mites , and the subsequent immune response to this event , is an important step in the infection . Mite transmission resulted in variably lethal infection of the outbred mice studied , and the time to death after disease onset was 5–9 dpi which is similar to our model [42] . The published i . d . inoculation study only followed the mice for 7 dpi and only examined the dissemination of different strains of Orientia . Data were not provided to compare the entire clinical course and development of systemic pathology for these models [43] . The i . v . model aims to simulate scrub typhus once the bacteria have left the eschar and begin to systemically infect the endothelium . The histopathology of mite transmission and i . d . inoculation models has not been thoroughly characterized; thus , comparison to the i . v . model's histology is not possible at this time . The clinical signs of our mice were similar to those described in the mite transmission model , with weight loss and decreased activity preceding death [41] . Bacterial dissemination was not followed in the mite transmission model , and thus it is impossible to compare this feature to our model , but the i . d . inoculation model did show remarkably rapid dissemination from the site of inoculation into the lungs as early as 24 hours post infection [42] . In the i . v . model , orientiae were detected at every time point in the majority of animals tested with a peak on day 6 post infection . How these bacterial kinetics compare to those of the i . d . inoculation model is not known as data throughout the disease course are not available . In conclusion , the model characterized in this study closely parallels the clinical course and pathological lesions described for lethal scrub typhus in humans . Intravenous inoculation of 1 . 25×106 Orientia resulted in an acute infection that culminated in death at 12–13 dpi . Pathological progression was observed in animals euthanized at sequential time points during the course of illness . With the establishment of lethal and sublethal doses for the intravenous model of scrub typhus , it will be possible to begin elucidating , mechanistically , the host responses that result in lethal outcomes or in protective immunity . As this model was established using the C57BL/6 mouse strain , future research projects will be able to utilize the abundant gene knockout mouse strains available on this background to determine the role of specific cell types and immune components involved in scrub typhus immunity and pathogenesis . The development of this model will provide a powerful tool to characterize the immunology of scrub typhus infection and a relevant model for vaccine testing that is intended to lead to an effective vaccine that produces long lasting immunity .
|
Scrub typhus is a disease found in Southeast Asia that infects over 1 million people each year . This disease is caused by the intracellular pathogen Orientia tsutsugamushi transmitted by the bite of chigger mites . Scrub typhus is characterized by pulmonary disease and in severe cases , multiorgan system failure . The current research model utilizes an intraperitoneal route of inoculation of mice to study the host response to Orientia infection . Infection via this route results in severe peritonitis that does not occur in human scrub typhus . The development of animal models that accurately portray human disease is an important step toward understanding and managing disease . In this manuscript we describe a new mouse model that results in scrub typhus-like pathology following intravenous inoculation of mice . This model presents dose-dependent mortality with scrub typhus-like pathology that parallels human disease . Utilization of this model will provide a valuable research tool for characterizing the immune response and pathogenesis induced by O . tsutsugamushi allowing development of better treatment and an effective vaccine .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"typhus",
"medicine",
"and",
"health",
"sciences",
"liver",
"immunology",
"tropical",
"diseases",
"microbiology",
"lungs",
"bacterial",
"diseases",
"histology",
"respiratory",
"system",
"immunochemistry",
"emerging",
"infectious",
"diseases",
"population",
"modeling",
"neglected",
"tropical",
"diseases",
"animal",
"models",
"of",
"infection",
"infectious",
"diseases",
"infectious",
"disease",
"modeling",
"scrub",
"typhus",
"anatomy",
"histological",
"examination",
"biology",
"and",
"life",
"sciences",
"computational",
"biology"
] |
2014
|
A Hematogenously Disseminated Orientia tsutsugamsushi-Infected Murine Model of Scrub Typhus
|
The inner structural Gag proteins and the envelope ( Env ) glycoproteins of human immunodeficiency virus ( HIV-1 ) traffic independently to the plasma membrane , where they assemble the nascent virion . HIV-1 carries a relatively low number of glycoproteins in its membrane , and the mechanism of Env recruitment and virus incorporation is incompletely understood . We employed dual-color super-resolution microscopy visualizing Gag assembly sites and HIV-1 Env proteins in virus-producing and in Env expressing cells . Distinctive HIV-1 Gag assembly sites were readily detected and were associated with Env clusters that always extended beyond the actual Gag assembly site and often showed enrichment at the periphery and surrounding the assembly site . Formation of these Env clusters depended on the presence of other HIV-1 proteins and on the long cytoplasmic tail ( CT ) of Env . CT deletion , a matrix mutation affecting Env incorporation or Env expression in the absence of other HIV-1 proteins led to much smaller Env clusters , which were not enriched at viral assembly sites . These results show that Env is recruited to HIV-1 assembly sites in a CT-dependent manner , while Env ( ΔCT ) appears to be randomly incorporated . The observed Env accumulation surrounding Gag assemblies , with a lower density on the actual bud , could facilitate viral spread in vivo . Keeping Env molecules on the nascent virus low may be important for escape from the humoral immune response , while cell-cell contacts mediated by surrounding Env molecules could promote HIV-1 transmission through the virological synapse .
Human immunodeficiency virus type 1 ( HIV-1 ) acquires its lipid envelope by budding through the plasma membrane of an infected cell . Virus morphogenesis is directed by the viral Gag polyprotein , which is sufficient for release of virus-like particles , and also facilitates the incorporation of the viral genome and other important factors including the viral envelope ( Env ) glycoproteins ( reviewed in [1] ) . Env plays an essential role in virus replication by mediating the fusion between viral and cellular membranes during virus entry . The Env proteins are synthesized as a polyprotein precursor ( gp160 ) that is cleaved to the mature surface glycoprotein gp120 and the transmembrane glycoprotein gp41 by cellular proteases . During virus assembly at the plasma membrane , the gp120/gp41 complex is incorporated into the lipid bilayer of nascent particles as a trimer of heterodimers . Within the virion , these trimeric complexes project from the membrane surface as highly glycosylated spikes ( reviewed in [2] ) . HIV-1 exhibits a relatively low density of glycoprotein spikes on its surface compared to other enveloped viruses with only 7–14 Env complexes per virion [3] , [4] . Virus entry is initiated by gp120 binding to the viral receptor CD4 and a chemokine receptor , and completed by insertion of a fusion peptide into the host membrane and conformational changes in gp41 mediating membrane fusion ( reviewed in [2] ) . HIV-1 Env is transported to the cell surface via the secretory pathway and inserts into the lipid membrane through the gp41 transmembrane domain . However , the mechanism by which the Env glycoprotein complex is incorporated into virus particles remains incompletely understood . Genetic and biochemical evidence points to an important role of the N-terminal , membrane-apposed matrix ( MA ) domain of the viral Gag polyprotein and the unusually long ( 151 amino acids ) C-terminal tail of gp41 ( CT ) , projecting into the interior of the virion , for Env incorporation [5]–[10] . HIV-1 derivatives lacking their cognate Env proteins can incorporate heterologous viral glycoproteins , thereby adopting their distinctive entry properties , however ( ‘pseudotyping’; reviewed in [11] ) . Furthermore , deletion of the Env CT has only minor effects on virion incorporation of Env and viral infectivity in certain permissive cell lines , while strongly reducing Env incorporation and abolishing infectivity in non-permissive cells [12]–[15] . Although these results are consistent with an important role of the CT , they show that an HIV-1 specific signal is dispensible for glycoprotein incorporation . Four mutually non-exclusive models have been proposed to explain Env incorporation into retroviral particles: ( i ) random incorporation of plasma membrane proteins including Env , ( ii ) specific incorporation of Env into the virus by direct protein-protein interactions , ( iii ) co-targeting of Gag and Env to the same region of the plasma membrane and ( iv ) indirect incorporation via a bridging factor ( reviewed in [2] ) . There is experimental evidence supporting , or at least consistent with each of these models . Efficient pseudotyping by heterologous glycoproteins and the large number of cellular membrane proteins incorporated into HIV-1 ( reviewed in [16] , [17] ) indicate that there is no exclusion of non-specific proteins . Incorporation may thus be influenced by surface density of the respective protein . A direct interaction between MA and Env CT in vitro has been reported [10] and there is strong genetic evidence supporting this interaction [5]–[8] , [18] . HIV-1 assembly is believed to occur at specific , raft-like membrane lipid microdomains , and both Gag and Env have been reported to be associated with detergent-resistant membranes ( reviewed in [19] ) . Furthermore , co-expression of HIV-1 Gag with both HIV-1 Env and Ebola virus glycoproteins in the same cells showed efficient incorporation of both glycoproteins , but segregation into separate particle populations suggesting their spatial separation in virus producing cells [20] . Finally , a number of cellular proteins have been found to interact with HIV-1 Env proteins and may act as bridging factors for Env incorporation [2] . Fluorescence microscopy of HIV-1 producing cells shows patchy signals of both Gag and Env at the plasma membrane , while the majority of Env appears to reside in intracellular membrane compartments [21] . Confocal microscopy provided evidence for some colocalisation of Gag and Env at the plasma membrane [21] , but reported correlation coefficients are low [20] and the resolution of light microscopy is not sufficient to discern adjacent individual budding sites . Immunostaining of surface glycoproteins and visualization with scanning electron microscopy provided convincing evidence for a specific recruitment of Rous sarcoma virus ( RSV ) Env proteins to RSV but not to HIV-1 budding sites [22]; this study did not include HIV-1 glycoproteins , however . Studying the distribution of viral proteins at small spatial scales requires an optical resolution which is beyond the limit of light microscopy ( ∼200 nm ) . New super-resolution fluorescence microscopy techniques [23]–[25] have bypassed this resolution limit , providing spatial resolution reaching a near-molecular level . These include single-molecule localization techniques such as photoactivated localization microscopy ( PALM ) [23] and direct stochastic optical reconstruction microscopy ( dSTORM ) [26] . PALM and STORM microscopy have been employed to investigate the distribution of viral proteins upon HIV-1 cell entry [27] , [28] and to detect Gag assemblies at the plasma membrane [29]–[31] . Lehmann et al . [30] used multicolor super-resolution microscopy to investigate co-localization of the cellular restriction factor tetherin with HIV-1 budding sites . These authors also reported scattered Env distribution at Gag assembly sites , but did not further characterize Env localization [30] . Here , we performed dual-color super-resolution microscopy to analyze HIV-1 Gag and Env distribution patterns in HIV-1 producing cells . We show a CT dependent recruitment of Env to the viral budding site , with Env molecules concentrating around the Gag assemblies and in their periphery .
For detection of HIV-1 Gag in the viral context , we made use of a construct carrying the photoconvertible protein mEosFP inserted between the MA and capsid ( CA ) domains of Gag [32] . Proviral constructs carrying a gene encoding an autofluorescent protein at this position yield fluorescently labeled HIV-1 particles with wild-type morphology and infectivity upon co-transfection with equal amounts of their unlabeled counterpart [32] . Since fully assembled HIV-1 buds comprise ∼2 , 400 molecules of Gag [33] , 1 , 200 molecules of mEosFP are expected to accumulate on average at viral budding sites . In this study we employed constructs based on pCHIV [34] that encodes all HIV-1NL4-3 proteins except Nef , but is replication deficient due to deletion of the viral long terminal repeats . The ratio of Env to Gag in transfected cells is thus comparable to that found in HIV-1 infected cells . Env molecules at the cell surface were detected by indirect immunolabeling using a human monoclonal antibody against gp120 ( MAb 2G12; [35] ) . Primary antibody binding was revealed by a secondary antibody coupled to Alexa Fluor 647 followed by dSTORM imaging . Due to the higher photon yield of organic fluorophores , dSTORM imaging yields improved single-molecule localization and thus higher spatial resolution than PALM . The combination of PALM for detection of the abundant Gag molecules with a known lattice structure [1] and dSTORM for detection of the much rarer Env molecules [3] thus appeared ideally suited for the purpose of this study . Total internal fluorescence ( TIRF ) microscopy and TIRF-PALM imaging of Gag . mEosFP at the plasma membrane of transfected HeLa cells was performed for detection of HIV-1 assembly sites . At the conditions used , Gag . mEosFP was detected with a localization accuracy of ∼28 nm in PALM mode ( Figure S1 ) . TIRF microscopy revealed punctuate structures of Gag . mEosFP at the plasma membrane as described previously for Gag . eGFP [36] ( Figure 1A ) . These punctae could be clearly resolved into individual assembly sites by super-resolution imaging ( PALM; Figure 1B–D ) . Compact round assemblies with a diameter of ∼130 nm ( Figure 1D ) , closely resembling assembly sites detected by dSTORM or PALM in previous studies [29] , [30] were considered to represent HIV-1 budding structures . The PALM resolution achieved here was not sufficient to detect the semi-spherical architecture of individual Gag shells , which we have recently shown at higher resolution ( ∼18 nm ) by dSTORM imaging of Gag . SNAP assembly sites labelled with a bright synthetic fluorophore [31] . HeLa cells transfected as above were fixed and subjected to immunostaining for HIV-1 Env , followed by TIRF- and TIRF-dSTORM microscopy . Alexa Fluor 647 , coupled to the protein of interest through the primary and secondary antibody , was detected with a localization accuracy of ∼15 nm in dSTORM mode ( Figure S2 ) . TIRF microscopy showed a multi-clustered Env distribution ( Figure 2A ) , similar to previously reported results [21] . Env clusters of various sizes were observed by dSTORM imaging ( Figure 2B ) . These clusters appeared larger and less compact than the Gag . mEosFP assemblies detected in the same cells ( compare Figure 1 ) . Dual-color super-resolution microscopy was performed on HeLa cells expressing HIVmEosFP to determine the relative localization of Env with respect to viral Gag assemblies . As shown in the representative images in Figures 2C and 2D , Env clusters surrounding Gag assembly sites were often round and sometimes displayed a doughnut-like shape . A similar pattern was observed when another antibody against gp120 ( MAb b12 , [37] ) was used ( Figure S3 ) , or when a harsher fixation protocol reported to block membrane protein motility [38] was applied ( Figure S4 ) . Env clusters of similar morphology were also observed in regions lacking a detectable Gag . mEosFP signal ( Figure 2C ) , and only ∼50% of Env clusters were associated with obvious HIV-1 budding sites . Env clusters not associated with characteristic Gag assemblies may correspond to early budding sites with a low number of Gag molecules . Furthermore , co-transfection with a wt plasmid encoding unlabeled Gag was performed in our experiments , and some assembly sites may thus contain a low number of Gag . mEosFP molecules . To address this issue , HeLa cells were transfected with pCHIVmEosFP alone and subsequently processed and analyzed as above . Similar Gag assembly sites surrounded by larger Env clusters were observed , but ∼90% of Env clusters were found to be associated with Gag assemblies in this case ( Figure S3 ) . Parallel control experiments showed that omitting the primary antibody from the staining solution or analysis of cells expressing a virus derivative lacking the viral Env protein yielded only few dense Alexa Fluor 647 signals , while the Env clusters were completely lost ( Figure S5 ) . The remaining non-specific signals presumably represent aggregates of secondary antibody . Env clusters overlapping HIV-1 budding sites showed a scattered distribution and always extended beyond the area , in which Gag . mEosFP molecules were detected ( Figure 2C , D ) . Furthermore , the density of Env localization events was not enriched at the position of the Gag . mEosFP assembly compared to the directly adjacent membrane area , but often appeared lower over the actual Gag assembly site ( Figure 2C , D; see below ) . A comparative quantitative analysis of the Env signal on budding sites and released immature virions was performed to determine whether the comparatively low proportion of Env signals directly co-localizing with Gag constitutes the full complement of Env on the virion or whether peripheral Env clusters also contribute to the Env signal on the virus . Immature virions were used for this comparison , since the Gag lattice at the budding site is also immature and the Env distribution pattern on the viral surface is altered by proteolytic maturation of the inner virion structure [39] . Purified immature viral particles were fixed , stained with 2G12 as described for virus producing HeLa cells , and analyzed by dSTORM . Env signals distributed in multiple clusters were detected on the viral surface ( Figure 2E ) in agreement with recent results obtained by stimulated emission depletion microscopy [39] . Quantitative assessment revealed a similar intensity of the total Env signal on extracellular immature virions compared to the Env signal directly co-localizing with the Gag assembly site ( Figure 2F ) . Accordingly , the Env content of the virion appears to be derived from the relatively low number of Env molecules directly overlapping the Gag assembly sites , while the larger Env cluster surrounding nascent assembly sites is not incorporated into the virion . Env recruitment was also investigated in the A3 . 01 T-cell line , since T-cells represent a natural target cell of HIV-1 . A similar distribution of Gag assembly sites with overlapping Env clusters , but increased Env density at the periphery and surrounding the actual budding site was observed by super-resolution dual-color imaging of A3 . 01 cells producing HIV-1mEosFP ( Figure S6A , 6B ) . The experiments described were carried out employing fixed cells stained with complete IgG molecules , which could potentially affect distribution and detection of the molecules of interest . To exclude potential artifacts , all further experiments ( except for those with the MA ( mut ) constructs , see below ) were performed using unfixed cells ( stained at 16°C to prevent membrane protein internalization ) and immunostaining with Fab fragments . The distribution pattern for Gag assembly sites and Env molecules was largely unaltered when HeLa cells expressing HIVmEosFP were analyzed in this way ( Figure 3A ) compared to the initial protocol ( Figure 2C ) . Tightly packed clusters of Gag . mEosFP with a diameter of ∼130 nm marked HIV-1 assembly sites , which were overlayed by less dense Env clusters that commonly exhibited enrichment in the periphery and surrounding the Gag budding site ( Figure 3B ) . To investigate whether the membrane distribution of HIV-1 Env was determined by the Env protein itself or was altered by the co-expression of other viral proteins , comparative dSTORM microscopy was performed on HeLa cells expressing either HIVmEosFP as before or only HIV-1 Env in the absence of other viral proteins . Super-resolution imaging of Env alone yielded a clearly different pattern compared to Env in the viral context . The large and distinctive round or doughnut shaped Env clusters , which frequently marked HIV-1 assembly sites ( Figure 3C ) , were not seen in the absence of other viral proteins . Instead , much smaller and more dispersed clusters were observed when Env was expressed alone ( Figure 3D ) . These results demonstrate a clear influence of the other HIV-1 proteins , most likely Gag , on the membrane distribution of HIV-1 Env . To investigate whether the long Env CT and its presumed interaction with the underlying Gag lattice are important for Env distribution on the surface of HIV-1 producing cells , we performed dual-color super-resolution microscopy on HeLa cells expressing HIVmEosFPEnv ( ΔCT ) . Tightly packed clusters of Gag . mEosFP as described above were detected in this case as well and identified bona fide HIV-1 budding sites ( Figure 4A ) . Clusters of Env ( ΔCT ) appeared much smaller than observed for wild-type ( wt ) Env , however . In contrast to ( wt ) Env clusters , Env ( ΔCT ) clusters were not enriched at HIV-1 budding sites , but appeared to be more randomly distributed ( Figure 4A , B ) . Furthermore , the characteristic Env clusters observed for the wt protein ( compare Figure 3C ) were not observed in this case . Similar results were obtained in HIVmEosFP expressing A3 . 01 cells ( Figure S6 ) . Comparison of Env distribution patterns in cells expressing either HIVmEosFPEnv ( ΔCT ) ( Figure 4C ) or only Env ( ΔCT ) ( Figure 4D ) using super-resolution microscopy revealed no apparent difference with similar cluster size and distribution in both cases . Thus , the characteristic Env distribution appeared to depend on the presence of other viral proteins and on the Env CT . The most likely viral protein responsible for the observed specific Env distribution pattern in the full viral context is Gag since several mutations in its MA domain have been shown to affect Env incorporation [5]–[7] , [9] . To directly address this issue , we made use of a panel of proviral constructs carrying either the wt sequence or two point mutations ( L8S/S9R ) within the MA domain of Gag [8] , [40] ( designated here as MA ( mut ) ) in the context of Env ( wt ) or Env ( ΔCT ) . These point mutations have been shown to affect Env ( wt ) particle incorporation and thereby viral infectivity; both defects are alleviated by truncation of the Env CT [8] . Transfected HeLa cells were fixed and immunostained with antibodies against MA and Env followed by dual-color dSTORM analysis ( Figure 5 , S7 ) . Typical Gag assembly sites were readily detected in all cases and were associated with Env clusters extending beyond the assembly site for the wt construct ( Figure 5A , S7A ) . These Gag associated Env clusters were absent for MA ( wt ) in the context of Env ( ΔCT ) ( Figure 5B , S7B ) consistent with the results observed above . A similar phenotype was observed for Env ( wt ) in the context of the MA ( mut ) virus ( Figure 5C , S7C ) , indicating that the reported Env incorporation defect is due to a loss of Env recruitment to the assembly site . The combination of both mutations displayed an intermediate phenotype , with the overall Env distribution pattern ( Figure S7D ) resembling the pattern observed for the individual MA or Env mutants , but a slightly more pronounced Env-Gag co-localization revealed upon inspection of individual assembly sites ( Figure 5D ) . To describe the different Env distribution patterns at the cell membrane in an objective and quantitative manner , we performed mathematical cluster analysis using two complementary approaches . Data sets were derived from transfected HeLa or A3 . 01 cells , respectively; at least three cells per condition were included in the analysis . In a first approach , we used an image-based morphological cluster analysis ( Figure 6A ) to determine the average cluster size of Env in whole cells . The average cluster size of Env ( wt ) in virus producing cells was significantly larger than observed for Env ( wt ) expressed alone or for clusters formed in cells expressing Env ( ΔCT ) ( with or without other viral proteins ) . The distribution of cluster sizes of Env ( wt ) in HeLa cells was also analyzed by subtracting values obtained for cells expressing only Env from those obtained for HIV-1 producing cells ( Figure 6B ) . Positive values in this analysis indicate an enrichment of the respective cluster size in virus-producing cells . This analysis revealed that smaller clusters with a radius <50 nm are much more prominent in cells expressing only Env , while larger clusters with a radius of 50 to 150 nm predominate in virus-producing cells . In an independent second approach , a coordinate-based distance distribution analysis using Ripley's H-function [41] was applied ( Figure 6C , 6D and S8 ) . This approach compares the measured distribution of single-molecule localizations to a simulated random distribution , and provides information whether clustering occurs . The maximum of the H function reflects the average size of clusters . The amplitude of the H-function is a measure of the degree of clustering . The variance of the amplitude reflects the variance of the spatial organization of proteins within a cluster . Unbiased analysis of clustering confirmed the larger size of Env ( wt ) clusters in virus producing cells compared to expression of Env alone or Env ( ΔCT ) ( with or without other viral proteins ) ( Figure 6C ) . Furthermore , clusters of Env ( wt ) in HIV-1 producing cells exhibited a much smaller variation in cluster size ( Figure S8 ) and a higher degree of homogeneity of clustering ( Figure 6D ) than observed for any of the other conditions . This result shows that Env ( wt ) clusters in HIV-1 producing cells are mostly homogenous and organized and most Env ( wt ) molecules are likely to be found in a single type of arrangement . The results of both cluster analyses were similar for HeLa and A3 . 01 cells , and corroborated the results of visual inspection of the spatial distribution of HIV-1 Env . Because whole cells were analyzed in this case and no pre-selection of particular regions was made , the results of the computational analyses represent average values and provide a more general and unbiased pattern . While providing a view on the whole cell , these computational analyses were not suitable to decipher differences in the Env distribution pattern at specific localizations , e . g . HIV-1 budding sites . To obtain quantitative information on the spatial distribution of Env at such sites , it was required to analyze preselected regions . This was performed by aligning and averaging the intensity distribution of Gag . mEosFP and Env at seven HIV-1 assembly sites each from HeLa cells producing either HIVmEosFP or HIVmEosFPEnv ( ΔCT ) . This procedure should enhance features common to the respective assembly sites , while deemphasizing random distributions . The averaged images of assembly sites from cells producing HIV-1 containing Env ( wt ) or Env ( ΔCT ) , respectively , showed a very similar pattern for the respective Gag assembly sites but a dramatic difference for the Env distribution pattern ( Figure 7 ) . The full-length Env protein ( Figure 7A ) exhibited a doughnut-shaped average pattern , reflected by a bimodal distribution of Env with a local intensity minimum at the position of the peak of the Gag intensity distribution ( Figure 7A ) . In contrast , Env ( ΔCT ) ( Figure 7B ) displayed a more random distribution without any significant enrichment at or close to the actual budding site . The averaged intensity profiles recorded for cells producing the Env ( ΔCT ) virus revealed no distinctive pattern of Env and no enrichment of Env density in the region of the Gag assembly site ( Figure 7B ) , again confirming the qualitative results obtained by visual inspection of the images .
Here , we employed dual-color super-resolution microscopy to investigate the distribution of HIV-1 Env glycoproteins at the cell surface of virus-producing and Env expressing cells . A combination of PALM imaging of an autofluorescent HIV-1 Gag derivative with dSTORM imaging of Env immunostained with a fluorescent dye appeared to be the optimal strategy . In general , synthetic fluorophores have a higher photon yield than fluorescent proteins , thereby providing increased spatial resolution in single-molecule based super-resolution imaging [26] . Accordingly , dSTORM imaging of Gag assemblies had revealed spherical edge effects [31] which were obscured in previous PALM images of Gag fused to autofluorescent proteins [30] . However , the structure of the Gag lattice at HIV-1 budding sites has already been characterized by electron tomography [33] , and Gag staining with synthetic fluorophores requires cell fixation and immobilization that might affect Env protein distribution . In contrast , the cell surface protein Env can be detected on native cells by immunolabeling and subsequent detection using organic fluorophores . Thus , we applied PALM for detection of Gag , while the higher localization accuracy of dSTORM was exploited for the detailed characterization of Env surface distribution . Using this system , we have analyzed Env localization in the presence and absence of other viral proteins as well as for a variant with mutations in the MA region of Gag and determined the role of the CT for HIV-1 Env membrane distribution . The results clearly revealed a CT- and MA dependent recruitment of Env proteins to the vicinity of Gag assembly sites . Env proteins accumulated around Gag clusters , concentrating at their periphery , while the bud center displayed reduced Env density . These results are consistent with Gag-dependent Env accumulation surrounding individual sites of HIV-1 particle formation , while the observed larger extension of Env clusters compared to that of Gag assemblies suggests that other factors than direct Gag-Env interaction may contribute to Env recruitment . Comparing Env distribution on cells producing HIV-1 particles with the distribution of Env expressed in the absence of other viral components revealed a difference that was clearly recognizable even without co-detection of the HIV-1 Gag protein . A scattered membrane distribution of Env with small clusters was seen in cells expressing Env alone , while larger accumulations were detected in virus producing cells . The appearance of the larger clusters was dependent on the Env CT and was disrupted by a MA mutation that abolishes Env incorporation . Dual color analysis revealed that larger Env clusters were commonly ( ∼90% ) associated with bona fide HIV-1 assembly sites , and Env structures with the characteristic round or doughnut-like shape were absent in cells expressing Env alone . The remaining ∼10% of larger Env clusters apparently lacking Gag association could correspond to nascent HIV-1 assembly sites with a low number of Gag molecules or constitute remnants of prior assembly sites after extracellular release of the viral particle . Alternatively , they could be induced by HIV-1 mediated changes of the membrane environment leading to membrane areas conducive for Env accumulation . Unbiased computational intensity-based cluster analysis from whole-cell data confirmed that the average cluster size was largest for Env ( wt ) in the presence of other HIV-1 proteins . This was supported by subtractive distribution of cluster sizes showing that Env ( wt ) expressed alone was mainly found in small clusters ( r<50 nm ) , whereas the same protein formed clusters with a radius of 50 to 150 nm in HIV-1 producing cells . While image-based cluster analysis can be performed on whole-cell data and generates distributions of cluster sizes , it depends on the spatial resolution and pixel size of the image and requires setting intensity thresholds . Thus , we applied Ripley's H-function [41] as a complementary approach for cluster analysis . This method can only analyze regions of interest rather than whole cells and does not provide information on the distribution of cluster sizes , but it is independent of thresholds . The maximum of the H-function , which correlates with the average cluster size , again demonstrated that the largest clusters were formed by Env ( wt ) in HIV-1 producing cells . In addition , Ripley's-H function provides information on the heterogeneity of clustering through the variance of its amplitude . A very small variance was observed for Env ( wt ) in HIV-1 producing cells suggesting a single type of cluster . The sharper peak of the H-function in this case compared to the other three conditions ( Figure S8 ) further indicated more regular clustering . Thus , evaluation of the data sets by two independent types of computational cluster analysis yielded a complete and consistent picture , confirming and extending the results of visual inspection of the super-resolution images . Considering the four described models for Env incorporation , the reported data clearly argue for specific recruitment and against random incorporation for wt HIV-1 . In contrast , random incorporation appears to be likely for Env ( ΔCT ) . Averaging the Env signal at multiple HIV-1 budding sites revealed no distinctive features for the Env ( ΔCT ) virus , confirming the visual impression . Env ( ΔCT ) incorporation into HIV-1 particles may thus occur randomly and only depend on its cell surface concentration . This interpretation is consistent with its observed reduced incorporation compared to Env ( wt ) [7] , [9] and the cell type dependence of the mutant phenotype [12]–[15] . Loss of apparent accumulation of Env at viral budding sites was also observed when a RSV variant with a truncated Env CT was analyzed by scanning electron microscopy of immunostained cells [22] . Thus , CT dependent specific recruitment of Env to the assembly site may be a general feature of retroviruses rather than being determined by the long lentiviral CT . Random incorporation may also be expected for pseudotyping with other viral glycoproteins . However , Jorgenson et al . [22] reported accumulation of at least some heterologous viral glycoproteins at RSV and HIV assembly sites , arguing for specific recruitment even in the absence of the cognate Gag-Env pair . Genetic data from multiple studies support a direct interaction of the HIV-1 Env CT with the membrane-apposed MA layer [6]–[9] , [18] even though biochemical evidence for this interaction is weak . Here , we observed that a mutation in MA that had been shown to disrupt virion incorporation of Env [8] , [40] abolished formation of Env clusters at HIV-1 assembly sites . This observation is consistent with MA-dependent recruitment of Env to the viral budding site . On the other hand , Env clusters in the case of wt HIV-1 significantly extended beyond the respective Gag assembly sites . Even close inspection revealed no detectable enrichment of Gag signals surrounding bona fide budding sites , consistent with the proposition that recruitment of Gag to the nascent bud occurs mostly from a cytoplasmic pool [36] . The majority of Env signals within the respective cluster were detected in the region surrounding the bud , however . Thus , Gag density in the vicinity of HIV-1 budding sites did not appear to differ significantly from its density in other plasma membrane regions , while Env density clearly did . Furthermore , Env density within the cluster appeared lowest at the center of the viral bud , where Gag is most concentrated . These results do not rule out a direct Gag-Env interaction , but argue that an indirect mode may at least contribute to Env concentration at HIV-1 assembly sites . This function clearly depends on HIV-1 Gag since Env accumulation was not seen in the absence of other viral proteins or in the case of the MA mutant . Co-targeting of Gag and Env to pre-existing membrane microdomains with special properties could therefore also not explain our results . Conceivably , accumulation of HIV-1 Gag and possibly of other viral proteins may induce an altered membrane micro-environment which attracts Env ( wt ) molecules . Alternatively , Env may be recruited by a proteinaceous bridging factor , which directs the viral glycoproteins to the assembly site , but does not immobilize them at this position . A specific membrane environment at the viral budding site is consistent with the raft-like lipid composition of HIV-1 particles [42] , [43] and the association of budding sites with tetraspanin-enriched microdomains ( reviewed in [44] ) . The hypothesis that Gag assembly at the plasma membrane alters or induces a specific lipid environment is also consistent with results from several recent studies using fluorescence recovery after photobleaching , fluorescence resonance energy transfer or antibody co-patching of membrane proteins to reveal Gag dependent changes in marker protein mobility or distribution ( reviewed in [45] ) . Dual- or triple-color super-resolution microscopy using marker membrane proteins or lipid dyes could in the future provide direct evidence regarding the size and colocalization of such domains with viral Gag and Env clusters , and may further clarify the mechanism of Env recruitment . At first glance , the accumulation of large amounts of Env surrounding , but not co-localizing with , Gag assemblies appears surprising for viral production sites . The observed bimodal shape of the Env density profile with a minimum in the central region of the HIV-1 bud is consistent with the low density of Env trimers detected on HIV-1 virions [3] , [4] , however , and may be explained by limited compatibility of the trimeric CT with the tightly packed Gag lattice . Many cellular membrane proteins are incorporated into HIV-1 particles [16] , [17] , but efficiency dependent on the CT length has been described , suggesting steric hindrance [46] , and exclusion due to interaction of CTs with cytoplasmic factors has been observed [47] . The low number of viral glycoproteins on the particle surface has been suggested to provide a selective advantage by allowing escape of the virus from the humoral immune response [48] . Concentrating surplus Env proteins not required for infectious particle formation around viral buds , on the other hand , could play an important role in forming and maintaining the virological synapse during cell-to-cell transmission . The formation of contacts between infected and uninfected T-cells is induced by Env [49] , and viral spread through such synapses is considered to be the major mode of HIV-1 transmission in vivo ( reviewed in [50] ) . The observed large HIV-1 Env clusters may thus play an important role in viral spread and pathogenesis . Visualization of the architecture of virological synapse structures using 3D multicolor super-resolution microscopy will therefore be an important goal of future research .
HeLa cells and A3 . 01 T-cells [51] were grown at 37°C and 5% CO2 in Dulbecco's modified Eagle's medium ( DMEM; Invitrogen ) and RPMI-1640 medium , respectively . Media were supplemented with 10% fetal calf serum ( FCS ) , 100 U/ml penicillin and 100 µg/ml streptomycin . Plasmid pCHIV , expressing all HIV-1 proteins except for Nef under the control of a CMV promoter and its derivative pCHIVmEosFP have been described previously [36] . pCHIVEnv ( ΔCT ) and pCHIVmEosFP . Env ( ΔCT ) were constructed by exchanging an AgeI/XhoI fragment of the respective parental plasmid with a corresponding fragment covering the env coding region from plasmid pNL4-3CTdel144-2 [14] . The HIV-1 Env expression vector pCAGGS . NL4-3-Xba , designated here as pEnv ( wt ) , has been described previously [52] . pEnv ( ΔCT ) was kindly provided by Nikolas Herold . It was constructed based on pEnv ( wt ) by exchanging an Acc65I/XhoI restriction fragment against the corresponding fragment from an Env ( ΔCT ) expression vector kindly provided by Valerie Bosch [15] . HIV-1 proviral constructs carrying wt Gag or the MA L8S/S9R mutant defective in Env interaction [8] , [40] in combination with either Env ( wt ) or or Env ( ΔCT ) were kindly provided by F . Mammano . HeLa cells were seeded at a density of 1×104 cells in 8-well chambered cover glasses ( LabTek ) and transfected with 0 . 2 µg of plasmid/well on the following day using the FuGene HD transfection reagent ( Roche Diagnostics ) . Nucleofection of A3 . 01 cells was performed as described previously [29] . In brief: 5×106 cells were electroporated with 10 µg of each plasmid in a 0 . 4 cm cuvette ( Invitrogen ) in a volume of 500 µl of serum-free medium using a Gene Pulser Xcell ( BioRad ) . Parameters were: capacity 950 µF and 300 V . After 24 hours , cells were seeded on 8-well chambered cover glasses ( LabTek ) coated with fibronectin at a concentration of 5 µg/ml . After sedimentation , cells were fixed with 3% paraformaldehyde ( PFA ) followed by permeabilization and blocking with 2% BSA . Two staining approaches for Env were applied: fixation followed by staining , or staining followed by fixation . For the first approach , cells were fixed at 24 h post transfection ( hpt ) with 3% PFA , washed and blocked for 10 min with 2% BSA in phosphate buffered saline ( PBS ) . Harsher fixation of samples was performed using 4% PFA/0 . 2% glutaraldehyde for 30 min [38] . Cells were incubated with the monoclonal anti-gp120 antibody 2G12 ( [35] , Polymun Scientific ) for 45 min . To confirm the specificity of 2G12 staining , the anti-gp120 MAb b12 ( [37] , Polymun Scientific ) was used . The washing and blocking procedures were repeated before incubation with goat anti-human Alexa Fluor 647 secondary antibody ( Invitrogen ) . To exclude antibody induced Env clustering and fixation artifacts , we performed staining prior to fixation and used Fab fragments of the respective antibodies ( kindly provided by J . Chojnacki ) . All steps were performed at 16°C to block endocytic uptake of surface molecules . Cells were washed with PBS and samples were blocked with 2% BSA for 10 min , followed by incubation with 2G12 Fab for 40 min . Subsequently , cells were incubated with goat anti-human Alexa Fluor 647 Fab ( Dianova ) for 40 min . Stained cells were fixed with 3% PFA for 20 min and washed with PBS . Staining of the MA domain of HIV was performed using the MAb APR342 [53] ( CFAR , UK ) followed by goat-anti-mouse Alexa Fluor 532 antibody . Immature eGFP . Vpr-labelled viruses [39] , kindly provided by J . Chojnacki , were adhered to LabTek chamber slides coated with 5 µg/ml fibronectin , fixed with 3% PFA and stained with MAb 2G12 and goat anti-human Alexa Fluor 647 and subjected to dSTORM imaging . Super-resolution microscopy was performed using a custom-built microscope setup as described elsewhere [54] . Briefly , a multi-line argon–krypton laser ( Innova70C , Coherent , USA ) and a 405 nm diode laser ( Cube , Coherent , USA ) were coupled into an inverted microscope ( IX71 , Olympus , Japan ) equipped with a 63× oil immersion objective ( PlanApo 63× , NA 1 . 45 , Olympus , Japan ) suitable for total internal reflection fluorescence ( TIRF ) imaging . The excitation and emission beams were separated using appropriate dichroic mirrors and filters ( AHF , Germany ) . The fluorescence emission was detected by an EM-CCD camera ( Ixon , Andor , Ireland ) . Combined PALM and dSTORM imaging was performed sequentially , using an imaging buffer which is suitable for both photoswitching the fluorescent protein mEosFP as well as the organic fluorophore Alexa Fluor 647 or Alexa Fluor 532 [55] . Briefly , the cells were imaged in oxygen-depleted hydrocarbonate buffer ( pH 8 ) supplemented with 100 mM mercaptoethylamine ( MEA ) . First , Alexa Fluor 647 was reversibly photoswitched by irradiation with 488 nm ( photoactivation ) and 647 nm ( read-out ) . For each channel , 8 , 000 to 10 , 000 images were recorded with an integration time of 50 ms . Then mEosFP was photoactivated by irradiation with 405 nm and imaged using an excitation wavelength of 568 nm . Alternatively , Alexa Fluor 532 was photoactivated by irradiation with 514 nm . Single-molecule localization and image reconstruction was performed using the rapidSTORM software [56] . The localization accuracy of single-molecule super-resolution microscopy was evaluated experimentally as described earlier [57] using a custom written software ( Python and Scipy ) [58] . For each localized fluorophore , the distance to its nearest neighbor fluorophore in an adjacent frame was calculated . As the majority of fluorophores are detected in multiple adjacent frames , the maximum of the nearest neighbor distance distribution represents the error of localization . A prerequisite for using this approach is a statistically significant number of events ( n>4 , 000 ) . Dual-color images were recorded by adding multi-spectral beads ( Invitrogen ) to the sample and post-aligning the individual images [59] . All-distance distributions are a common tool for cluster analysis of single-molecule super-resolution data [31] , [60] , [61] . We used Ripley's K-function [41] ( 1 ) in its linearized ( 2 ) and normalized form ( 3 ) ( Ripley's H-function ) : ( 1 ) ( 2 ) ( 3 ) where r is the observation radius , n is the total number of localizations within a region of interest ( ROI ) , dij is the distance between two localizations i and j , Nr is the number of localizations around localization i within the distance r , and λ is a weighting factor correcting for the area of the ROI . We calculated Ripley's H-function for ROIs of 2×2 µm2 . In order to account for edge effects , we used a torroidal edge correction . The results were tested against a 95% confidence envelope of uniform distributions generated by Monte Carlo methods . All calculations and simulations were performed using custom software written in Matlab ( Mathworks , Natick , MA ) . Image-based cluster analysis was performed on super-resolution images by identifying cohesive regions of protein populations and determining their area . For this , a ‘quantitative’ super-resolution image with a pixel size of 10 nm was generated , in which the pixel value represents the number of localizations found at this position ( custom software written in Matlab ( Mathworks , Natick , MA ) ) . Cohesive regions in these quantitative images were identified and measured using an algorithm based on the analyze particles function from FIJI [62] . To analyze the protein distribution , we calculated average images by overlaying multiple images of individual clusters using an image pixel size of 5 nm . The individual images were aligned using the center of mass of the Gag protein distribution . Image averaging was performed using a custom software written in Matlab ( Mathworks , Natick , USA ) , intensity profiles were extracted using FIJI .
|
Newly formed HIV-1 particles assemble at the plasma membrane of virus producing cells . The inner structural protein Gag and the envelope glycoprotein Env , which are both essential components of infectious virus particles , traffic to the membrane via different pathways . Attached to the inner side of the membrane , Gag assembles into spherical particles that incorporate Env proteins in their surrounding lipid envelope . The mechanism of Env incorporation is incompletely understood , however . Here , we have exploited recently developed super-resolution fluorescence microscopy techniques that yield a near-molecular spatial resolution to analyze HIV-1 Gag and Env distribution patterns at the surface of virus producing cells . We observed recruitment of Env to the surroundings of Gag assembly sites , dependent on the presence of its cytoplasmic domain . A large proportion of Env was found in the vicinity of the Gag assembly sites rather than directly co-localizing with it . These results support an indirect mechanism of Env recruitment , presumably mediated through virus induced changes in the environment of the nascent Gag assembly . Furthermore , they suggest a role for the Env protein in HIV-1 transmission that goes beyond its well-characterized function as an entry protein on the viral surface .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"virology",
"biology",
"microbiology"
] |
2013
|
Super-Resolution Microscopy Reveals Specific Recruitment of HIV-1 Envelope Proteins to Viral Assembly Sites Dependent on the Envelope C-Terminal Tail
|
EBNA3C , one of the Epstein-Barr virus ( EBV ) -encoded latent antigens , is essential for primary B-cell transformation . Cyclin D1 , a key regulator of G1 to S phase progression , is tightly associated and aberrantly expressed in numerous human cancers . Previously , EBNA3C was shown to bind to Cyclin D1 in vitro along with Cyclin A and Cyclin E . In the present study , we provide evidence which demonstrates that EBNA3C forms a complex with Cyclin D1 in human cells . Detailed mapping experiments show that a small N-terminal region which lies between amino acids 130–160 of EBNA3C binds to two different sites of Cyclin D1- the N-terminal pRb binding domain ( residues 1–50 ) , and C-terminal domain ( residues 171–240 ) , known to regulate Cyclin D1 stability . Cyclin D1 is short-lived and ubiquitin-mediated proteasomal degradation has been targeted as a means of therapeutic intervention . Here , we show that EBNA3C stabilizes Cyclin D1 through inhibition of its poly-ubiquitination , and also increases its nuclear localization by blocking GSK3β activity . We further show that EBNA3C enhances the kinase activity of Cyclin D1/CDK6 which enables subsequent ubiquitination and degradation of pRb . EBNA3C together with Cyclin D1-CDK6 complex also efficiently nullifies the inhibitory effect of pRb on cell growth . Moreover , an sh-RNA based strategy for knock-down of both cyclin D1 and EBNA3C genes in EBV transformed lymphoblastoid cell lines ( LCLs ) shows a significant reduction in cell-growth . Based on these results , we propose that EBNA3C can stabilize as well as enhance the functional activity of Cyclin D1 thereby facilitating the G1-S transition in EBV transformed lymphoblastoid cell lines .
Epstein–Barr virus ( EBV ) is a B-lymphotropic human herpes virus that persists indefinitely in latently infected B-cells . EBV infection occurs early in life for most people and is associated with a broad spectrum of benign and malignant diseases including Burkitt's lymphoma ( BL ) , nasopharyngeal carcinoma ( NPC ) , Hodgkin's disease ( HD ) and lymphomas associated with immuno-compromised individuals , including AIDS patients and post-transplant patients receiving immune-suppressive therapy [1] . EBV infection in B-cell leads to aberrant cell division and under favorable conditions the infected B-cells will continue to proliferate indefinitely , resulting in development of immortalized lymphoblastoid cell lines ( LCLs ) [1] , [2] . One of the most noteworthy EBV-host cell interactions is the establishment of viral latency . There are three major types of latency , each having its own distinct viral-gene expression pattern [1] , [2] . Type I latency is usually noticed in BL tumors with predominant expression of EBV encoded nuclear antigen 1 ( EBNA-1 ) [1] , [2] . Type II latency is demonstrated in NPC and HD , where EBNA-1 , latent membrane protein 1 ( LMP-1 ) , LMP-2A and -2B proteins are significantly detected [1] , [2] . Type III latency , also termed as ‘growth program’ [1] , [2] is typically seen in LCLs expressing six latent nuclear proteins ( EBNA-1 , -2 , -3A , -3B , -3C , and -LP ) , three latent membrane proteins ( LMP-1 , -2A , and -2B ) , and the viral RNAs which includes the EBERs and BARTs ( 33 , 62 ) . Molecular genetics analyses have demonstrated that at least six EBV latent genes ( EBNA-1 , -2 , -3A , -3C , -LP , and LMP-1 ) are essential for in vitro immortalization [1] , [2] , indicating that a complex cascade of molecular events is required to surpass normal growth controls . One scenario which accounts for EBV-mediated B-cell immortalization is modulation of critical positive and/or negative regulators of cell-cycle progression , such as cyclins , cyclin-dependent kinases ( CDKs ) , cyclin-dependent kinase inhibitor proteins ( CDKIs ) , tumor-suppressors and apoptosis related proteins which includes p53 and pRb [3] . EBNA3C , one of the essential EBV latent antigens , has been shown to function both as a transcriptional activator and a repressor [4] , [5] , [6] . It has also been shown to interact with numerous transcription modifiers , including c-Myc [4] , prothymosin α [7] , histone deacetylases [8] , CtBP [9] , NM23-H1 [10] , DP103 [11] , SCFSkp2 [12] , p300 [13] and p53 [14] which contributes to EBV induced transformation mediated by EBNA3C . In addition , a large body of evidence indicates that EBNA3C can also deregulate the cell-cycle machinery through direct protein-protein interaction and post-translational modification of important cell-cycle regulatory proteins , including Cyclin A [15] , [16] , pRb [17] , p53 [14] , Mdm2 [18] , and Chk2 [19] . So far , studies probing EBNA3C functions provide perhaps the best link between latent EBV infection and the pRb regulated checkpoint which controls the G1-S phase transition [20] , [21] . EBNA3C was previously shown to indirectly target pRb regulated pathways [15] , [20] . EBNA3C also activates E2F-dependent promoters and can induce foci formation in colony formation assays [20] . Additionally , EBNA3C overcomes the ability of the CDK inhibitor - p16INK4A to block transformation and noticeably drives serum-starved cells through the G1-S restriction point [20] , [21] . More recently , we have shown that EBNA3C directly targets pRb and may indirectly target the pRb regulated checkpoint by associating with Cyclin A as well as Cyclin D1 known to be important in phophorylating pRb [15] , [16] . Despite this body of evidence , a clear molecular link between these molecules responsible for disrupting the G1-S phase blockage and EBNA3C is yet to be demonstrated . Cell-cycle progression is dependent on the activity of cyclins , a family of proteins whose levels oscillate in synchrony with cell-cycle progression , and its functional partner CDKs [22] . Cyclin D ( D1 , D2 and D3 ) is expressed in the mid-G1 phase in the mammalian cell-cycle [23] . Among the D-type cyclins , Cyclin D1 is the most ubiquitous and is frequently over-expressed in numerous human malignancies [24] , [25] . Cyclin D1 over-expression is often associated with increased gene expression due to gene amplification or post-translational modification [26] . Accumulation of Cyclin D1 in cancer can result in overcoming ubiquitin-mediated degradation through several distinct mechanisms [26] . Cyclin D1 , together with its catalytic partners CDK4 or CDK6 , promotes G1-S-phase transition via phosphorylation of pRb and disrupting the pRb-E2F1 repressor complex [23] . These functions of Cyclin D1 ensure efficient initiation of S phase [26] , [27] . During late G1 and S phases , Cyclin D1 is phosphorylated on Thr-286 by GSK3β , which triggers nuclear export and proteasomal degradation through E3 ubiquitin ligase , SCFFBX4-αB crystallin [26] . Thus , subversion of either of these functions may result in unrestrained cell proliferation and oncogenesis . The cyclin D1 gene is located on chromosome 11q13 , close to the bcl-1 locus , and is considered to be a proto-oncogene with evidence indicating that its derangement contributes to the development of tumors [28] . Mantle cell lymphomas have been reported to over-express Cyclin D1 due to a characteristic genetic translocation [28] . In addition , patients with tumors over-expressing Cyclin D1 have been shown to have a particularly poor prognosis [25] , [29]; however , over-expression of Cyclin D1 has been demonstrated for a vast series of human malignancies including breast cancers , esophageal cancers and pancreatic cancers [25] , [30] . Over-expression of Cyclin D1 , regardless of its gene alteration , caused abnormal cell proliferation , resulting in oncogenesis [22] , [23] , [31] . cyclin D2 , considered also as a proto-oncogene , is located on chromosome 12p13 , and unlike Cyclin D1 , Cyclin D2 has been reported to be expressed normally in B-lymphocytes [32] . Interestingly , it has been observed that ectopic over-expression of Cyclin D2 efficiently blocks cell-cycle progression [33] , suggesting an alternate role for Cyclin D2 in promoting exit from the cell-cycle and maintaining cells in a non-proliferative state . These observations suggest that D-type cyclins may have different roles depending on their levels of expression and cell type , which may also be independent of CDK activity . Reports have shown that immortalization of primary B-lymphocytes by EBV is accompanied by transcriptional activation of cyclin D2 gene but not cyclin D1 [32] , [34] . However , Cyclin D1 protein has been shown to be significantly expressed in a number of EBV positive LCLs [35] , [36] or EBV positive SCID mice lymphomas [37] . Surprisingly , these studies did not directly set out to explore the contribution of Cyclin D1 in EBV-mediated B-cell oncogenesis . A previous study from our lab showed an in vitro interaction between the EBV encoded antigen EBNA3C and Cyclin D1 [16] . The experiments described in this current study explore the consequences of this interaction in terms of EBV mediated transformation of primary B-cells as well as growth maintenance of LCLs . We now show that EBNA3C stabilizes as well as enhances the kinase activity of the Cyclin D1/CDK6 complex , and the nuclear localization of Cyclin D1 to bypass the G1 restriction point . Importantly , this study provides the first evidence to show that the essential EBV latent antigen EBNA3C targets Cyclin D1 , which is different from previous reports , and describes a potential fundamental mechanism by which EBV deregulates the mammalian cell-cycle in EBV-associated human cancers by facilitating the G1-S transition .
Myc , flag , GFP and GST tagged EBNA3C vectors have been described previously [14] , [18] . pcDNA3-HA-Ub was kindly provided by George Mosialos ( Aristotle University of Thessaloniki , Thessaloniki , Greece ) . Vectors pcDNA3-Cyclin D1 , pcDNA3-1x flag-Cyclin D2 and pcDNA3-1x flag-Cyclin D3 were provided by Alan Diehl ( University of Pennsylvania School of Medicine , Philadelphia ) and used to generate pA3F-Cyclin D by cloning PCR amplified DNA into pA3F vector [4] . GST Cyclin D1 vectors were cloned by inserting PCR amplified DNA into pGEX-2TK vector ( GE Healthcare Biosciences , Pittsburgh , PA ) . pGEX-Cyclin D1 ( 286A ) was generated by PCR using pA3F-Cyclin D1 as template . Sh-RNA vector , pGIPZ ( Open Biosystems , Inc . Huntsville , AL ) and lentiviral packaging vectors were described [38] . CDK6 cDNA cloned into pA3F vector was derived from HEK 293 cell RNA that was purified with TRIzol reagent and reverse transcribed with Superscript II ( Invitrogen , Inc . , Carlsbad , CA ) . Mouse antibodies to Cyclin D1 ( DSC-6 ) and Sp1 ( 1C6 ) , and rabbit antibody to Ub ( FL-76 ) were from Santa Cruz Biotechnology , Inc ( Santa Cruz , CA ) . Rabbit antibodies to Cyclin D2 and D3 were kindly provided by Alan Diehl ( University of Pennsylvania School of Medicine , Philadelphia ) . Mouse antibodies to flag-epitope ( M2 ) was from Sigma-Aldrich Corp . ( St . Louis , MO ) and to GAPDH was from US-Biological Corp . ( Swampscott , MA ) . Antibodies to HA-epitope ( 12CA5 ) or Myc-epitope ( 9E10 ) were prepared from cell culture supernatants as described [14] , [18] . Mouse ( A10 ) or rabbit antibody to EBNA3C were described [14] , [18] . HEK 293 , 293T and Saos-2 ( p53-/- pRb-/- ) cells were obtained from Jon Aster ( Brigham and Women's Hospital , Boston , MA , USA ) . Saos-2 and U2OS are human osteosarcoma cell line [39] . HEK 293 , HEK 293T , U2OS , and Saos-2 cells were grown in Dulbecco's modified Eagle's medium ( DMEM; HyClone , Logan , UT ) supplemented with 10% fetal bovine serum ( FBS; HyClone , Logan , UT ) , 50 U/ml penicillin ( HyClone , Logan , UT ) , 50 µg/ml streptomycin ( HyClone , Logan , UT ) and 2 mM L-glutamine ( HyClone , Logan , UT ) . BL lines BJAB , Ramos , BL41 and B95 . 8 infected BL41 ( BL41/B95 . 8 ) were kindly provided by Elliott Kieff ( Harvard Medical School , Boston , MA ) . MutuI , MutuIII were provided by Yan Yuan ( School of Dental Medicine , University of Pennsylvania , Philadelphia , PA ) . These BL lines and LCL1 and LCL2 were maintained in RPMI 1640 ( HyClone , Logan , UT ) supplemented as described above . EBNA3C expressing BJAB lines were described [14] , [18] . Unless otherwise stated all cultures were incubated at 37°C in a humidified environment supplemented with 5% CO2 . Adherent cells were transfected by electroporation with a Bio-Rad Gene Pulser II electroporator as described [14] , [18] . Peripheral blood mononuclear cells ( PBMC ) from healthy donors were obtained from University of Pennsylvania Immunology Core . As described [40] , approximately 10 million PBMC were mixed with virus supernatant in 1 ml of RPMI 1640 with 10% FBS for 4 hr at 37°C . Cells were centrifuged for 5 min at500 g , discarded the supernatant , pelleted cells and resuspended in 2 ml of complete RPMI 1640 medium in 6 well plates . EBV GFP expression visualized by fluorescence microscopy was used to quantify infection . The protein and mRNA level of the infected cells was detected after 3 days of post-infection . Transfected cells were harvested , washed with ice cold PBS and lysed in 0 . 5 ml ice cold RIPA buffer [1% Nonidet P-40 ( NP-40 ) , 10 mM Tris pH 8 . 0 , 2 mM EDTA , 150 mM NaCl , supplemented with protease inhibitors ( 1 mM phenylmethylsulphonyl fluoride ( PMSF ) , 1 µg/ml each aprotinin , pepstatin and leupeptin] . Lysates were precleared with normal mouse serum plus 30 µL of Protein A/G Sepharose ( 1 h , 4°C ) . 5% of the precleared lysate was saved for input control and the protein of interest was captured by rotating the remaining lysate with 1 µg of specific antibody overnight at 4°C . Immuno-complexes were captured with 30 µl of a 1∶1 mixture of Protein-A and Protein-G Sepharose beads , pelleted and washed 5X with ice cold RIPA buffer . For western blots , input lysates and IP complexes were boiled in laemmli buffer [41] , fractionated by SDS-PAGE and transferred to a 0 . 45 µm nitrocellulose membrane . The membranes were then probed with specific antibodies followed by incubation with appropriate infrared-tagged secondary antibodies and viewed on an Odyssey imager . Image analysis and quantification measurements were performed using the Odyssey Infrared Imaging System application software ( LiCor Inc . , Lincoln , NE ) . Escherichia coli BL21 cells were transformed with plasmids for each glutathione S-transferase ( GST ) fusion protein and protein complexes containing the tagged proteins were purified essentially as described before [14] , [18] . For in vitro binding experiments , GST fusion proteins were incubated with cell lystaes or 35S-labeled in vitro-translated protein in binding buffer ( 1x phosphate-buffered saline [PBS] , 0 . 1% NP-40 , 0 . 5 mM dithiothreitol [DTT] , 10% glycerol , supplemented with protease inhibitors ) . In vitro translation was done with the TNT T7 Quick Coupled Transcription/Translation System ( Promega Inc . , Madison , WI ) according to the manufacturer's instructions . Cells were immuno-stained as described [18] with few modifications . Briefly , U2OS cells plated on coverslips were transfected with expression vectors as indicated , using Lipofectamine 2000 ( Invitrogen , Carlsbad , CA ) according to manufacturer's protocol . After 36 h of transfection cells were fixed . B-cells were air-dried and subsequently fixed . Transiently expressed flag-tagged Cyclin D1 was detected using M2-antibody , and GFP-EBNA3C was detected by GFP fluorescence . In B-cells , endogenously expressed Cyclin D1 and EBNA3C proteins were detected using specific antibody . The slides were examined with a Fluoview FV300 confocal microscope ( Olympus Inc . , Melville , NY ) . Total RNA was isolated by using TRIzol reagent according to the instructions of the manufacturer ( Invitrogen , Inc . , Carlsbad , CA ) . cDNA was made by using a Superscript II reverse transcriptase kit ( Invitrogen , Inc . , Carlsbad , CA ) according to the instructions of the manufacturer . The primers were for cyclin D1 , 5′-TGCCCTCTGTGCCACAGATG-3′ , and 5′-TCTGGAGAGGAAGCGTGTGA-3′ , for cyclin D2 5′-TGCTCTGTGTGCCACCGACTT-3′ , and 5′-CAGCTCAGTCAGGGCATCACAA-3′ , for cyclin D3 5′-TTTGCCATGTACCCGCCATCCA-3′ and 5′-CCCGCAGGCAGTCCACTTCA-3′ , and for GAPDH 5′-TGCACCACCAACTGCTTAG-3′ and 5′-GATGCAGGGATGATGTTC-3′ . Quantitative real-time PCR analysis was done as described [18] in triplicate . 15×106 HEK 293T cells were transfected by electroporation with DNA vectors expressing a specific protein . Cells were incubated for 36 h and pretreated for an additional 6 h with 20 µM MG132 ( Enzo Life Sciences International , Inc . , Plymouth Meeting , PA ) before harvesting . Proteins were immunoprecipitated with specific antibodies and resolved by SDS-PAGE . The extent of ubiquitination of immunoprecipitated complexes were detected by HA-specific antibody ( 12CA5 ) against HA-Ub tagged proteins . 15×106 HEK 293 cells were transfected with expression plasmids . After 36 h cells were PBS washed and resuspended into hypotonic buffer [5 mM Pipes ( KOH ) pH 8 . 0 , 85 mM KCl , 0 . 5% NP-40 supplemented with protease inhibitors ) . After 10-min incubation on ice , cells were homogenized with 20 strokes in a Dounce homogenizer , nuclei were pelleted ( 2300 g for 5 min ) and the cytosolic material was collected . Nuclear pellets were PBS washed , resuspended in nuclear lysis buffer ( 50 mM Tris , pH 8 . 0 , 2 mM EDTA , 150 mM NaCl , 1% NP-40 , and protease inhibitors ) , lysed by vortexing periodically for 1 h . Soluble nuclear fraction was separated by centrifugation at 21000 g for 10 min . Total protein was measured by Bradford protein assay and 50 µg of total protein was resolved by SDS-PAGE . The efficiency of nuclear and cytoplasmic fractionation was confirmed by western blot against nuclear transcription factor Sp1 and cytoplasmic protein GAPDH . 15×106 HEK 293T cells were transfected with plasmids expressing flag-Cyclin D1 ( 5 µg ) , flag-CDK6 ( 5 µg ) and increasing amount of myc-EBNA3C ( 0 , 5 , 10 , 20 µg ) . For GSK-3β kinase assay cells were transfected with DNA vectors that express myc-tagged GSK-3β ( 10 µg ) and flag-tagged EBNA3C ( 20 µg ) . Cells were harvested and protein complexes were immunoprecipitated ( IP ) using either M2 ( for cyclin D1 ) or 9E10 ascites fluid ( for GSK-3β ) . IP complexes were then washed with buffer A ( 25 mM Tris [pH 7 . 5] , 70 mM NaCl , 10 mM MgCl2 , 1 mM EGTA , 1 mM DTT , plus protease and phosphatase inhibitors ) and incubated in 30 µl of kinase buffer B ( buffer A plus 10 mM cold ATP , and 0 . 2 µCi of [γ-32P]-ATP/µl ) supplemented with either 4 µg of histone H1 ( Upstate Biotechnology , Inc . , Lake Placid , N . Y . ) or bacterially purified GST-pRb ( residues 792-928 ) for 30 min at 30°C . The reaction was stopped by adding 2X laemmli buffer [41] and heating to 95°C for 10 min . Labeled proteins were resolved by 12% SDS-PAGE . Band quantitation was performed using the ImageQuant software ( GE Healthcare Biosciences , Pittsburgh , PA ) . Cells were transiently transfected using electroporation with plasmids as indicated in the text . After 36 hours transfection , cells were treated with 40 µg/ml cyclohexamide ( CalBiochem , Gibbstown , NJ ) and lysates were subjected to immunoblot analyses . Band intensities were quantitated using Odyssey 3 . 0 software provided by Odyssey imager ( LiCor Inc . , Lincoln , NE ) . Short-hairpin oligonucleotides directed against EBNA3C were designed ( Dharmacon Research , Chicago , IL ) . The sense strand of the EBNA3C-shRNA sequence is 5′-tcgagtgctgttgacagtgagcgaCCATATACCGCAAGGAATAtagtgaagccacagatgtaTATTCCTTGCGGTATATGGgtgcctactgcctcggaa-3′ . The sense strand of cyclin D1 sh-RNA sequence is 5′-tcgagtgctgttgacagtgagcgaCAAACAGATCATCCGCAAAtagtgaagccacagatgtaTTTGCGGATGATCTGTTTgtgcctactgcctcggaa-3′ [42] . Upper-case letters indicate 19-nucleotide ( nt ) either EBNA3C or cyclin D1 target sequences respectively and lowercase letters indicate hairpin and sequences necessary for the directional cloning into pGIPZ ( Open Biosystems , Inc . Huntsville , AL ) . Single-stranded EBNA3C and cyclin D1 oligonucleotides were first annealed and then cloned into the Xho I and Mlu I restriction sites of pGIPZ vector . The fidelity of cloned double-strand DNA was confirmed by DNA sequencing . In parallel , a commonly available control shRNA sequence ( Dharmacon Research , Chicago , IL ) : ( 5′-TCTCGCTTGGGCGAGAGTAAG-3′ ) that lacks complementary sequences in the human genome was also cloned into pGIPZ vector . Lentivirus production and transduction of EBV-transformed B-cells ( LCLs ) were essentially carried out as previously described [38] . Saos-2 ( p53-/- pRb-/- ) were transfected using Ca3 ( PO4 ) 2 method as described [38] . After 24 h transfection , cells were selected using DMEM supplemented with 1000 µg/ml G418; Invitrogen ) . After a 2-week selection , 5×106 cells were harvested , lysed in RIPA buffer and subjected for immunoblot analyses . Approximately 0 . 1×106 cells from each set of samples were plated into each well of the 6-well plates and cultured for 6 days . Viable cells from each well were counted by trypan blue exclusion method daily using a Bio-Rad TC10 Automated cell counter . For LCLs , approximately 1×106 cells were plated into each well of the 6-well plates and cultured at 37°C in complete RPMI medium . Cells were counted similarly for 20 days . Both experiments were performed in duplicate and were repeated two times . 5×106 Saos-2 ( pRb-/- ) cells were transfected as described [38] and cultured in DMEM supplemented with 1 mg/ml G418 ( Invitrogen , Inc . , Carlsbad , CA ) . After a 2-week selection , cells were fixed on the plates with 4% formaldehyde and stained with 0 . 1% crystal violet ( Sigma-Aldrich Corp . , St . Louis , MO ) . The area of the colonies ( pixels ) in each dish was calculated by Image J software ( Adobe Inc . , San Jose , CA ) . The data are shown as the average of three independent experiments . For serum starvation experiments , the culture medium was replaced with RPMI 1640 and 0 . 1% FBS for 12 h . Cells were PBS washed , fixed in cold 70% ethanol for 30 min at 20°C , PBS washed and stained 2 h in buffer containing 50 mg/ml propidium iodide , 10 mM Tris pH 7 . 5 , and 500 U/ml RNAseA in dark . PBS washed cells were analyzed for cell-cycle profile by FACS Calibur system and Cellquest software ( Becton-Dickinson Inc . , San Jose , CA ) .
In order to determine whether EBV infection alters Cyclin D expression , approximately 10×106 human resting peripheral blood mononuclear cells ( PBMC ) were infected by BAC GFP-EBV as previously described [40] for 4 h and western blot analysis was performed on samples collected 3 days after infection . The results showed that EBV infection leads to a significant induction of all three Cyclin D protein levels 3 days post-infection , with no preference for any particular D-type cyclins ( Fig . 1A ) . Similarly , western blot results of Burkitt's lymphoma ( BL ) cell line BL41 and BL41 infected with wild-type EBV strain B95 . 8 ( BL41/B95 . 8 ) also showed elevated levels of Cyclin Ds with Cyclin D1 expression more dramatically changed compared to other Cyclin Ds ( Fig . 1B ) . Since Cyclin D1 expression was induced significantly after EBV infection in both PBMC and BL cell line , we next wanted to determine if the induction was related to a specific EBV latent protein expressed during type III latency . The results showed that the levels of both Cyclin D1 and Cyclin D2 proteins were induced in type III latency BL cell line MutuIII compared to latency I expressing MutuI BL cell line ( Fig . 1C ) . These results differ with previously published observations which suggested that B-cells infected with EBV do not express Cyclin D1 [43] , [44] , [45] . However , in agreement with previously published results [32] , our real-time PCR data showed that EBV infection led to a significant increase of cyclin D2 mRNA level in LCLs ( LCL1 and LCL2 ) when compared to EBV negative BL cells ( BJAB and Ramos ) whereas , there was little or no detectable change for cyclin D1 mRNA ( Fig . 1F ) . Real-time PCR data obtained from two other matched sets of cell lines BL41 – BL41/B95 . 8 and MutuI – MutuIII also showed similar results as above ( Fig . 1G and 1H , respectively ) . These results suggest that D-type cyclins are regulated through distinctly different mechanisms in EBV infected B-cells . EBV effects on Cyclin D2 are at the level of its transcript stability whereas the effects on Cyclin D1 or D3 seem to be post-translational . To elucidate the effects of the EBV encoded essential nuclear antigen , EBNA3C on Cyclin D1 , BL lines BJAB and E3C #7 , a BJAB stably expressing EBNA3C were analyzed . The western blot results showed a significant increase in Cyclin D1 protein expression among D-type Cyclins in E3C #7 cells compared to the BJAB control cells and smaller changes in Cyclin D2 and D3 ( Fig . 1D ) . The effect of EBNA3C on Cyclin D1 steady-state levels was not due to changes in the transcription as EBNA3C expression did not alter the level of cyclin D1 mRNAs in these cells as seen above ( Fig . 1I ) . To further verify the role of EBNA3C on Cyclin D1 protein accumulation , we determined the levels of Cyclin Ds in a lymphoblastoid cell line with the EBNA3C mRNA specifically targeted by short-hairpin RNA ( Sh-E3C ) . The western blot data showed that the expression level of Cyclin D1 in the LCLs stably knocked-down for EBNA3C ( Sh-E3C ) was significantly diminished as compared to the control cell line ( Sh-Control ) ( Fig . 1E ) , however the expression levels of other Cyclin Ds was not altered ( Fig . 1E ) . These results indicate that EBNA3C can contribute to Cyclin D1 accumulation in latently infected EBV positive cells . To demonstrate that EBNA3C can stabilize Cyclin D1 protein levels , HEK 293 cells were transfected with an increasing amount of an expression construct expressing EBNA3C and tested for endogenous Cyclin D1 protein level . The results showed that EBNA3C stabilizes Cyclin D1 protein expression in a dose dependent manner ( Fig . 1J ) . We earlier determined that EBNA3C plays a critical role in modulating the ubiquitin ( Ub ) -proteasome machinery [12] , [17] , [18] . Therefore , to investigate whether the increase of Cyclin D1 levels was because of the inhibition of Ub-proteasome mediated destabilization by EBNA3C , transiently co-transfected cells were treated with the proteasome inhibitor , MG132 . The results showed that both the treatment with MG132 , and presence of EBNA3C led to a significant accumulation ( six fold ) of Cyclin D1 when compared to mock treatment or vector control ( Fig . 1K ) . Therefore the increased levels of Cyclin D1 observed in the presence of EBNA3C and MG132 is a result of stabilization of Cyclin D1 likely by EBNA3C inhibition of the Ub-proteasome degradation system . Importantly , both CDK6 and EBNA3C levels were not altered by MG132 ( Fig . 1K ) . To directly determine EBNA3C stabilization of Cyclin D1 , HEK 293 cells were transfected with flag-Cyclin D1 , flag-CDK6 , and EBNA3C expression vectors . Thirty-six hours later , cells were treated with protein synthesis inhibitor cycloheximide , and samples were collected at 0 , 1 , and 2 hours . Western blots probed with flag antibody showed that the stability of Cyclin D1 protein was significantly enhanced by EBNA3C co-expression , whereas in the absence of EBNA3C , Cyclin D1 was degraded to near completion by 2-h after addition of CHX ( Fig . 1L , grey bar ) . Cyclin D1 half life was determined to be 2 h in EBNA3C expressing cells; however , it shortened noticeably to less than 1 h when Cyclin D1 was expressed alone ( Fig . 1L , bar diagram ) . The results also indicated that both EBNA3C and CDK6 were notably stable throughout the experimental period of time and had no sign of protein degradation ( Fig . 1L , CDK6 indicated as black bar ) . Overall , the results of these experiments suggest EBNA3C can stabilize Cyclin D1 by regulating its targeted degradation likely through the Ub-proteasome degradation system . Recently we have shown that ectopic expression of EBNA3C leads to stabilization of an important cellular oncoprotein , Mdm2 by inhibiting its poly-ubiquitination [18] . The increased stability of Cyclin D1 in the presence of EBNA3C , prompted us to examine whether EBNA3C similarly inhibits poly-ubiquitination of Cyclin D1 and so enhances its stability . To explore this possibility , three cell lines were selected , the EBV negative cell line BJAB , BJAB stably expressing EBNA3C ( E3C #7 ) and an EBV positive lymphoblastoid cell line ( LCL2 ) . Immnuprecipitation using specific antibody against Cyclin D1 resulted in formation of high molecular weight species of Cyclin D1 migrating at a slower rate in BJAB cells while in BJAB cells stably expressing EBNA3C or in LCL2 significantly less of these high molecular weight bands were observed ( Fig . 2A ) . Re-probing of the same membrane with Ub specific antibody showed a similar pattern ( Fig . 2A ) . This result indicates that the activity responsible for the change in Cyclin D1 bands is present in EBV positive cells ( LCL2 ) and EBNA3C expressing cell line ( E3C #7 ) when compared to the EBV negative BJAB cells . To directly address this phenomenon , an ubiquitination experiment was set up , where HEK 293T cells were transiently co-transfected with expression constructs for HA-Ub , flag-Cyclin D1 and myc-EBNA3C and the ubiquitination of the Cyclin D1 was assessed by immunoprecipitation followed by Western blotting ( Fig . 2B ) . The result demonstrated a significant and reproducible reduction in Cyclin D1 poly-ubiquitination level in EBNA3C expressing cells ( Fig . 2B ) . Similar experiments were performed separately using two different cyclins , Cyclin A and Cyclin E to determine if this effect was specific for Cyclin D1 . However , neither Cyclin A nor Cyclin E poly-ubiquitination levels were reduced in the presence of EBNA3C ( Fig . 2C and 2D ) . To determine whether the poly-ubiquitination level of the other D-type cyclins was also affected in the presence of EBNA3C , we tested flag-tagged Cyclin D2 and D3 for ubiquitination in the absence and presence of EBNA3C . Importantly , poly-ubiquitination of both Cyclin D2 and D3 was efficiently inhibited in the presence of EBNA3C ( Fig . 2E ) . This result indicates that EBNA3C can profoundly affect the poly-ubiquitination of all Cyclin Ds and thus enhance their stability . We have shown earlier that EBNA3C interacts with Cyclin D1 in vitro along with other cyclins including Cyclin A and Cyclin E [16] . In order to determine whether EBNA3C forms a complex with Cyclin D1 in cells to enhance its stability , we performed binding assays using co-IP experiments . HEK 293T cells were co-transfected with expression constructs for myc-EBNA3C and flag-Cyclin D1 . The results showed that ectopically expressed EBNA3C associated with Cyclin D1 in cells ( Fig . 3A and 3B ) . To further determine whether this binding occurred under endogenous settings , Cyclin D1 was immunoprecipitated from EBV negative cell line , BJAB and two EBV transformed lymphoblastoid cell lines , LCL1 and LCL2 expressing EBNA3C . EBNA3C was detected by Western blot analysis using A10 , an EBNA3C specific monoclonal antibody and showed efficient co-immunoprecipitation ( Fig . 3C ) . In a separate experimental setting , Cyclin D1 was immunoprecipitated from BJAB cells and BJAB cells stably expressing EBNA3C ( E3C#10 ) . Similarly co-IP of EBNA3C was demonstrated using the A10 antibody ( Fig . 3D ) . To further corroborate the association in human cells , a GST-pulldown experiment was conducted; where bacterially expressed GST-Cyclin D1 was incubated with cell lysates prepared from either BJAB cells or BJAB cells stably expressing EBNA3C ( E3C#7 and E3C#10 ) . EBNA3C was seen to strongly associate with GST-Cyclin D1 but not with the GST control ( Fig . 3E ) . Coomassie staining of a parallel gel showed the amount of GST and GST-Cyclin D1 proteins used in the binding assay ( Fig . 3E , right panel ) . Analysis of the data from the ectopic expression system as well as cell lines endogenously expressing Cyclin D1 and EBNA3C at physiological levels strongly demonstrated an association between Cyclin D1 and EBNA3C in human cells . We have previously shown that a small N-terminal region of EBNA3C ( residues 130-160 ) binds to Cyclin D1 in vitro [16] . To map the domain of EBNA3C that interacts with Cyclin D1 , HEK 293T cells were transfected with expression constructs for flag-Cyclin D1 and either full-length EBNA3C ( residues 1-992 ) , EBNA3C residues 1-365 , EBNA3C residues 366–620 , or EBNA3C residues 621-992 . All EBNA3C expression constructs were fused in frame with a myc epitope tag at the C-terminus of the protein . As expected , the results showed that Cyclin D1 co-immunoprecipitated with full-length EBNA3C as well as with the N-terminal domain of EBNA3C ( residues 1–365 ) ( Fig 4A , left-middle panel , lanes 2 and 3 , respectively ) whereas no co-IP was detected with vector control or other truncated versions of EBNA3C ( Fig 4A , left-middle panel , lanes 1 , 4 and 5 ) . To further corroborate the binding data , an in vitro GST-pulldown experiment was performed using in vitro translated 35S-radiolabeled fragments of EBNA3C ( residues 1–100 , 1–129 , 1–159 and 1–200 ) within the N-terminal domain . In vitro precipitation experiments with bacterially expressed GST-Cyclin D1 showed strong association with residues 1–159 and 1-200 of EBNA3C ( Fig . 4B , bottom panel , lanes 3 and 4 , respectively ) , but not with EBNA3C residues 1–100 or 1–129 ( Fig . 4B , bottom panel , lanes 1 and 2 , respectively ) . All fragments of EBNA3C failed to interact with the GST control , indicating that the observed binding was specific for Cyclin D1 ( Fig . 4B , middle panel , lanes 1 to 4 ) . In an attempt to gain insights into the functionality of the association between Cyclin D1 and EBNA3C , a series of N- and C-terminal deletion mutants of Cyclin D1 ( residues 1–50 , 40–170 , 171–260 and 241–295 ) were designed according to their domain distribution [46] , [47] and tested for their ability to bind EBNA3C using in vitro binding experiments . The results of the GST-pulldown assay clearly showed that full-length Cyclin D1 , the N-terminal pRb binding region ( residues 1-50 ) and the C-terminal domain which is known to regulate Cyclin D1 stability ( residues 171–260 ) strongly associated with EBNA3C ( Fig . 4C , top panel , lanes 3 , 4 and 6 , respectively ) . However , no binding was detected with the other truncated versions of Cyclin D1 ( the CDK4/6 binding domain , residues 40–170 and the PEST domain , residues 241–295 ) or with the GST control ( Fig . 4C , top panel , lanes 2 , 5 , and 7 ) . Importantly , the C-terminal domain of Cyclin D1 ( residues 171–260 ) bound to EBNA3C with relatively higher affinity than the full-length or the N-terminal site ( Fig 4C ) . In order to determine the specificity of EBNA3C and Cyclin D1 interaction , we next performed a co-immunoprecipitation assay using all three flag-tagged D-type Cyclins . Interestingly , the results showed that EBNA3C forms complexes with all three D-type Cyclins in cells , suggesting that EBNA3C has specificity for interaction with Cyclin D1 , D2 and D3 ( Fig . 4D ) . Increased expression of Cyclin D1 has been seen in a number of cancers [25] , [30]; however , this enhanced expression is usually not sufficient to drive the oncogenic process . Emerging evidence suggests that nuclear accumulation of Cyclin D1 resulting from altered nuclear trafficking and proteolysis is critical for its oncogenic phenotype [31] . In order to determine the effect of EBNA3C on the sub-cellular localization of Cyclin D1 , asynchronously growing U2OS cells were transfected with expression vectors encoding flag-tagged Cyclin D1 and GFP-tagged EBNA3C . Localization of Cyclin D1 was determined by indirect immunofluorescence using a monoclonal antibody against the flag epitope ( Fig . 5A , panels f , h , j , l ) . While Cyclin D1 mostly localized to the cytoplasm in the absence of EBNA3C ( Fig . 5A , panels f , h ) , it was predominantly localized to the nucleus in the presence of EBNA3C ( Fig . 5A , panels j , l ) . To quantitatively compare the Cyclin D1 signals in the nuclear and cytoplasmic compartments , 10 different fields of the stained slides were examined and the bar diagram represents the mean of three independent experiments which showed that nuclear localization was increased by 20% ( Fig . 5A , bar diagram ) . To further corroborate these results showing that EBNA3C promotes nuclear localization of Cyclin D1 , the sub-cellular localization of endogenous Cyclin D1 was determined in three different cell lines – EBV negative BL cell line BJAB , BJAB cells stably expressing EBNA3C ( E3C# 7 ) and an EBV transformed B-cell line LCL2 , using a specific antibody against cyclin D1 . As anticipated , the results showed that cyclin D1 was predominantly localized in the nucleus of both EBNA3C positive BJAB cells ( Fig . 5B , panels f , g ) and EBV positive cells LCL2 ( Fig . 5B , panels j , k ) , but was almost exclusively cytoplasmic in the EBV negative BJAB cells with no EBNA3C expressed ( Fig . 5B , panels b , c ) . Based on immuno-fluorescence studies , we observed that Cyclin D1 localization was mainly restricted to the cytoplasmic fraction of asynchronously growing cells . However , on expression of EBNA3C the localization of Cyclin D1 was predominantly nuclear . To further support these data , transiently transfected HEK 293 cells were subjected to sub-cellular fractionation and fractionated proteins were analyzed by immunoblot analysis . The result showed that flag-tagged Cyclin D1 alone was detected approximately 50% in both cytoplasmic and nuclear fractions , respectively ( Fig . 6A , compare lanes 1 and 4 ) . However , when co-transfected with EBNA3C , flag-Cyclin D1 was detected predominantly within the nuclear fraction ( Fig . 6A , compare lanes 3 and 6 ) , with an approximately 50% increase compared to flag-Cyclin D1 alone ( Fig . 6A , compare lanes 1 and 3 ) . EBNA3C was detected completely within nuclear fraction ( Fig . 6A , lanes 2 and 3 ) . The efficiency of cytoplasmic and nuclear fractionation was confirmed by localization of nuclear transcription factor Sp1 and cytoplasmic protein GAPDH ( Fig . 6A ) . These observations strongly suggested that the apparent nuclear trans-localization of Cyclin D1 mediated by EBNA3C , as determined by indirect immuno-fluorescence microscopy or sub-cellular fractionation assay may be due to deregulation of the critical regulatory kinase GSK-3β , a negative regulator of Cyclin D1 nuclear retention and protein stability [31] . We thus decided to examine whether EBNA3C can nullify the effect of GSK-3β on Cyclin D1 function . GSK-3β can direct the nuclear export of Cyclin D1 via a CRM1-dependent pathway [31] . To examine whether EBNA3C can block Cyclin D1 nuclear export , we tested the ability of EBNA3C to override GSK-3β triggered Cyclin D1 nuclear export . To test this possibility , HEK 293 cells were transfected with expression vectors encoding flag-tagged Cyclin D1 , with or without GSK-3β and myc-tagged EBNA3C . Fractionated cell lysates were analyzed by western blot to clarify flag-tagged Cyclin D1 localization . As expected , Cyclin D1 was primarily present in the cytoplasmic fraction both in the absence and presence of GSK-3β ( Fig . 6B , lanes 1 and 4 ) . In contrast , Cyclin D1 was largely detected within the nuclear fraction when co-expressed with EBNA3C ( Fig . 6B , lane 3 ) . Interestingly , even in the presence of GSK-3β nuclear fractionation of Cyclin D1 was greatly increased when co-expressed with EBNA3C compared with the vector control ( Fig . 6B , compare lanes 1 and 3 ) . GSK-3β has been shown to phosphorylate Cyclin D1 on Thr-286 in vitro [31] , and is postulated to be a major regulator of protein levels and intracellular distribution of Cyclin D1 [31] . To establish a plausible explanation for the inhibitory effects of EBNA3C on GSK-3β dependent Cyclin D1 subcellular localization , we first asked whether EBNA3C can form a complex with GSK-3β to negatively modulate its activity and to also determine whether the kinase activity of GSK-3β is inhibited in the presence of EBNA3C . To this end , we co-expressed myc-tagged GSK-3β and flag-tagged EBNA3C and assessed their interaction through co-immunoprecipitation experiment . The results showed that indeed EBNA3C can form a complex with GSK-3β ( Fig . 6C , compare lanes 3 and 4 ) . Next , an in vitro kinase assay was conducted where GSK-3β was immuno-precipitated in the absence and presence of EBNA3C , and tested for its ability to phosphorylate recombinant GST-Cyclin D1 proteins ( wild-type and T286A mutant Cyclin D1 ) . The results showed that the phosphorylation level of wild-type GST-Cyclin D1 by GSK-3β was reduced by more than 4 fold in the presence of EBNA3C ( Fig . 6D , compare lanes 1 and 2 ) . As expected , no phosphorylation bands were observed in case of mutant GST-Cyclin D1 ( T286A ) indicating the specificity of this experiment ( Fig . 6D , lanes 3 and 4 ) . Parallel blots showed the protein expression levels in whole cell-lysate ( Fig . 6D ) , and the amount of purified GST-Cyclin D1 used in this experiment ( Fig . 6D ) . These results indicated that EBNA3C may regulate Cyclin D1 sub-cellular localization probably by blocking the function of GSK-3β . To address the functional consequences as a result of the association of Cyclin D1 and EBNA3C , we tested the activity of Cyclin D1/CDK6 complexes for the ability to phosphorylate histone H1 or recombinant GST-pRb ( residues 792-928 ) . HEK 293T cells were transiently transfected with increasing amounts of a myc-tagged EBNA3C expression construct . Flag-tagged Cyclin D1/CDK6 immunoprecipitated complexes were assayed for in vitro kinase activity as determined by histone H1 or GST-pRb phosphorylation ( Fig . 7A and B , respectively ) . The results showed that Cyclin D1-dependent kinase activity increased in a dose-responsive manner with increased expression of EBNA3C ( Fig . 7A and B ) . Phosphorimager analysis revealed 1 . 6-times more P32-Histone H1 and 2 . 3-times more P32-GST-pRb ( Fig . 7A and B ) . Parallel blots showed the expressed protein levels ( Fig . 7A and B , top two panels ) and the amount of substrates ( histone H1 or GST-pRb ) used in this study ( Fig . 7A and B ) . Cyclin D1/CDK4/6 complexes are rate-limiting for G1 progression by contributing to the sequential phosphorylation of pRb , and thereby canceling the growth-suppressive function of pRb , thus facilitating entry into S-phase [26] , [27] . Previously we have shown that EBNA3C facilitates pRb degradation by enhancing its poly-ubiquitination through recruitment of the SCFSkp2 E3 ligase activity [17] . To test whether EBNA3C coupled with Cyclin D1/CDK6 complex regulates pRb stabilization , a stability assay was performed using cycloheximide ( CHX ) treated Saos-2 ( pRb-/- p53-/- ) cells co-transfected with plasmids expressing myc-tagged pRb , flag-tagged Cyclin D1 , flag-tagged CDK6 , and EBNA3C ( Fig . 7C ) . The results clearly showed that independent expression of either Cyclin D1/CDK6 complex or EBNA3C reduced pRb expression levels ( Fig . 7C [upper panel] , compare lanes 1-9 ) . Surprisingly , when both EBNA3C and Cyclin D1/CDK6 complex were expressed together , little or no pRb was detected ( Fig . 7C [lower panel] , lanes 1-3 ) , indicating that EBNA3C can also facilitate pRb degradation in cooperation with Cyclin D1/CDK6 either through stabilization of Cyclin D1 ( Fig . 7C [lower panel] , compare lanes 4–9 ) or increasing kinase activity of Cyclin D1/CDK6 complex . In order to probe whether EBNA3C enhances pRb poly-ubiquitination in a Cyclin D1-dependent manner for degradation , we performed an in vivo ubiquitination assay . HEK 293T cells were co-transfected with expression constructs for myc-tagged pRb , HA-tagged Ub , flag-tagged Cyclin D1 and CDK6 and untagged EBNA3C as indicated ( Fig . 7D ) . pRb was immunoprecipitated with myc antibody , and ubiquitinated-pRb was detected by probing blots with HA antibody . In agreement with the previous result , poly-ubiquitination of pRb was significantly enhanced in the presence of EBNA3C alone ( Fig . 7D , compare lanes 3 and 4 ) and slightly further elevated in the presence of both EBNA3C and Cyclin D1/CDK6 complex ( Fig . 7D , compare lanes 4 and 6 ) indicating that EBNA3C together with Cyclin D1/CDK6 is important for inducing pRb poly-ubiquitination and its subsequent degradation . To determine the effect of EBNA3C and Cyclin D1/CDK6 complex on pRb mediated cell growth suppression , an osteosarcoma cell line , Saos2 , was transfected with the expression plasmids for myc-tagged pRb , flag-tagged Cyclin D1 , flag-tagged CDK6 and EBNA3C as indicated in the figure ( Fig . 8A–D ) . Cells were additionally transfected with a GFP expression vector . The cells were selected with G418 for 2 weeks and the proliferation rate of the selected cells was measured by an automated cell counter for 6 days ( Fig . 8 ) . Dead cells were excluded using Trypan Blue staining . The rationale for choosing Saos2 as recipient cells was that cell growth suppression and morphological changes can be observed in Saos2 cells that express pRb de novo [48] . The results showed that EBNA3C together with Cyclin D1/CDK6 complex effectively reduced the growth suppressive effect of pRb . The cell-proliferation rate of cells expressing pRb either with EBNA3C or Cyclin D1/CDK6 complex was 1 . 5-2 fold higher than pRb alone ( Fig . 8A ) . However , interestingly EBNA3C together with Cyclin D1/CDK6 complex significantly enhanced the cell proliferation rate , which was approximately either 6 fold higher than pRb alone or 3 fold higher than pRb when co-expressed with either EBNA3C or Cyclin D1/CDK6 complex ( Fig . 8A ) . To check the expression levels of these proteins , the selected cells were subjected to western blot analysis ( Fig . 8B ) . The results showed that the pRb expression levels were significantly reduced in EBNA3C or Cyclin D1/CDK6 expressing samples , whereas no changes of expression were observed for other proteins ( Fig . 8B ) . GAPDH was used as an internal loading control and expression of GFP indicated equivalent amount of total protein lysate prepared from selected cells ( Fig . 8B ) . In order to corroborate the previous experiment , we next performed a colony formation assay , where cells were similarly transfected with different combinations of expression constructs as stated above . After selection of the transfected cells with G418 similarly as stated above for 2 weeks , the number of antibiotic-resistant colonies was counted ( Fig . 8C–D ) . In agreement with the previous experiment , the results showed that co-expression of both EBNA3C and Cyclin D1/CDK6 proteins with pRb in Saos-2 cells resulted in an increase in the number of colonies compared to pRb alone ( Fig . 8C , compare panels 1–3 and Fig . 8D , bar diagram ) . However , interestingly EBNA3C together with Cyclin D1/CDK6 complex markedly increased the antibiotic-resistant colonies ( Fig . 8C , compare panels 1-4 and Fig . 8D , bar diagram ) . Overall , these results indicate that EBNA3C can utilize the function of Cyclin D1/CDK6 to neutralize the growth inhibitory effect of pRb . In the context of the above-described results , we hypothesized that EBNA3C exploits Cyclin D1/CDK6 to promote LCL proliferation by inactivating pRb . To address this , LCLs were stably transduced with lentiviruses that express short hairpin RNA to silence either EBNA3C ( Sh-E3C ) or cyclin D1 ( Sh-CyD1 ) . The Sh-Control RNA is not complementary to human genome sequences . Stable transduction was verified by GFP expression ( Fig . 9A ) . The expression levels of knocked down genes among these cells were then detected by Western blot analysis ( Fig . 9B ) . The results showed that the level of EBNA3C or Cyclin D1 was knocked down by sh-RNA whereas LCL1 transduced with sh-Control had levels similar to LCL1 ( Fig . 9B ) . The results also showed that pRb expression levels were slightly increased in both EBNA3C and Cyclin D1 knocked down samples , whereas there were no alterations observed for other Cyclin D expression levels ( Fig . 9B ) . In order to determine whether both EBNA3C and Cyclin D1 are critical to maintain the proliferation of EBV transformed cells , a proliferation analysis was done ( Fig . 9C ) . The results showed that the proliferation rate of both wild-type LCL1 and LCL1 infected with the lentivirus control sh-RNA ( Sh-Control ) expressing physiological level of both EBNA3C and Cyclin D1 was significantly higher than that of LCLs with Sh-E3C and Sh-CyD1 ( Fig . 9C ) . In agreement with the previously published results [20] , [49] , we also observed that the proliferation rate of LCLs containing Sh-E3C with reduced levels of EBNA3C expression was approximately 3 fold slower than that of control cell-lines ( Fig . 9C ) . Interestingly , the proliferation rate of LCLs with Sh-CyD1 was 50% higher than LCLs with Sh-E3C and only about 1 . 5 fold lower than that of control . This suggests that other D-type cyclins might be involved in LCL growth , particularly Cyclin D2 which was shown earlier to be associated in EBV mediated lymphomagenesis and probably transcriptionally up-regulated by one of the other major EBV latent antigen LMP1 [45] . However , it is clear from repeated analyses that cyclin D1 knock-down correlates with an increase in doubling time . The results support the notion that EBNA3C and cyclin D1 are critical for driving the growth of EBV transformed cells . It has been shown earlier that both EBV positive cells and cells stably expressing EBNA3C can bypass G1/S phase checkpoint caused by serum starvation [20] , [35] . Cell-cycle profiles of cells cultured in medium with 0 . 1% FBS were analyzed by flow cytometry ( Fig . 10 ) . Analyses of serum-starved , EBV negative cell lines BJAB and DG75 and LCLs sh-E3C and sh-CyD1 revealed an increased percentage of cells at the G0/G1 phase of the cell cycle ( Fig . 10A , B ) and decreased percentage of cells in the G2/M phases ( Fig . 10A , C ) . Fig . 10B and 10C represents the difference in both G0/G1 and G2/M phases due to serum starvation , respectively . However , under the same culture conditions , the EBV-positive LCLs - LCL1 , LCL2 , LCL1-with Sh-control and BJAB-cells stably expressing EBNA3C ( E3C# 7 and E3C# 10 ) continued through the cell-cycle without being arrested at any particular phase ( Fig . 10A histograms , B and C ) . Furthermore , the results also indicated that upon knockdown of both EBNA3C and Cyclin D1 , LCLs underwent a substantial degree of apoptosis ( Sub G0 ) in response to serum starvation , similar to EBV negative cell lines , BJAB and DG75 ( Fig . 10A ) . However , there was no sign of apoptosis observed either in BJAB cells stably expressing EBNA3C or wild-type LCLs ( Fig . 10A ) . Altogether , this experiment demonstrated that EBNA3C and Cyclin D1 positively contribute to cell growth in EBV transformed cells and are critical for overriding the G1 block as a result of serum starvation .
The cyclin D1 gene amplification has been observed in cancers of the breast , head and neck or larynx [50] , [51] , [52] . Chromosomal rearrangement is another cause of Cyclin D1 over-expression associated with centrocytic lymphomas [53] , parathyroid adenomas [54] and mantle cell lymphoma [28] . The obvious association of Cyclin D1 with cancer has led the investigators to uncover its oncogenic properties . In fact , Cyclin D1 was shown to cooperate with the Ras oncoprotein for cell transformation [55] . Earlier reports have suggested that immortalization of primary B-lymphocytes by EBV is accompanied by transcriptional activation of the cyclin D2 gene but not cyclin D1 [43] , [44] , [45] . However , a number of studies showed noticeable changes in Cyclin D1 protein levels in both EBV positive LCLs [35] and EBV positive SCID mice lymphomas [37] . Despite the controversy regarding the Cyclin D1 expression in EBV positive B-lymphoma cells , it is clear that in order to deregulate the entire mammalian cell-cycle , EBNA3C manipulates G1 restriction point through disruption of Cyclin/CDK-pRb-E2F pathway [20] . Cyclin D1 is over-expressed in a variety of human cancers that do not exhibit cyclin D1 gene amplification or structural abnormalities of the cyclin D1 locus , which suggests that increased Cyclin D1 stability is a potential mechanism . Mutations of cyclin D1 at T286 and P287 have been found in human tumors [24] and alter Cyclin D1 nuclear localization as well as stability . Our data showed that both EBV infection in primary B-cells and EBV persistence in cancer cell lines increased protein stability . However the cyclin D1 mRNA level was unchanged . Similar to virus infection , EBNA3C gene expression increased Cyclin D1 levels without altering mRNA levels . It is important to determine if these effects also occur in vivo . The results presented here also demonstrated that the expression of Cyclin D2 and D3 were up-regulated in quiescent cells infected with EBV probably through distinctly different mechanisms . EBV infection or its transforming protein latent membrane protein 1 ( LMP1 ) up-regulates Cyclin D2 expression in primary B-lymphocytes and Burkitt's lymphoma cells [45] . None of the studies have shown an important role for Cyclin D3 in EBV-mediated cell transformation . Studies have suggested that the D-type cyclins may have non-overlapping functions at specific steps in B-cell differentiation [32] , and that the expression of different D-type cyclins may be influenced by EBV infection through distinctive pathways . Thus , a potential mechanism which involves the contribution of Cyclin D1 in EBV-mediated B-cell transformation is the constitutive induction of these key cell-cycle regulators which leads to pRb hyper-phosphorylation and uncontrolled cell proliferation . Several lines of evidence suggest that Cyclin D1 is targeted by the E3 ligase , SCFFBX4-αB crystallin for degradation [26] . Elevated expression of FBX4 and αB crystallin is also found to trigger the destruction of wild-type Cyclin D1 , but not the phosphorylation-deficient Cyclin D1 mutant , D1T286A [26] . Thus , impairment of SCFFBX4-αB crystallin function may also account for Cyclin D1 overexpression . Data from the ubiquitination assay showed that EBNA3C efficiently inhibits Cyclin D1 poly-ubiquitination , which led us to speculate that EBNA3C may interact with this particular E3 ligase and inhibit its ability to ubiquitinate Cyclin D1 . The SCFSkp2 ligase has also been shown to be involved in the degradation of Cyclin D1 [56] , [57] , [58] , and knockdown of Skp2 molecule promoted marked accumulation of Cyclin D1 [57] . EBNA3C interacts with SCF components to regulate the stability of p27KIP1 and pRb [12] , [17] . It is likely EBNA3C inhibition of SCFSkp2 reduces Cyclin D1 ubiquitination . EBNA3C may be a deubiquitinase or associate with one to regulate the stability of Mdm2 [18] and likely Cyclin D1 . Since the expression level of Cyclin D1 is related to the potential for malignancy and the prognosis of a variety of cancers [30] , [31] , revealing the mechanisms governing the ubiquitin-proteasome mediated degradation of Cyclin D1 is of importance in designing therapeutic interventions . Conceivably , this approach could amplify the therapeutic window using Cyclin D1 as a target and enhance the efficacy of conventional drugs against EBV mediated oncogenesis . We have shown earlier that EBNA3C can interact with Cyclin D1 using an in vitro GST-pulldown experiment [16] . Here , we examined the molecular association between EBNA3C and Cyclin D1 complexes to obtain a more in-depth understanding of the different domains of EBNA3C that modulate the activity of Cyclin D1 which will lead to further understanding the basic mechanism by which EBV regulates the mammalian cell-cycle . EBNA3C associates with Cyclin D1 via the same N-terminal domain , residues 130-190 , that has been shown to bind many critical cell-cycle regulators [18] including other Cyclins - A and E [16] . In addition , a recent genetic study using recombinant EBV expressing conditionally active EBNA3C showed that deletion of this particular domain could not support cell proliferation of EBV transformed LCLs , signifying the importance of this domain within EBNA3C [49] . The association of EBNA3C with different Cyclins suggests is perhaps cell-cycle dependent and one may replace another depending on the stage in the cell-cycle , which ultimately leads to aberrant cell proliferation in EBV transformed cells . The previously published data and the data herein were generated using asynchronously growing cells; therefore it would be interesting to further elucidate these interactions in a cell-cycle dependent manner . However , using chemical synchronization is likely to distort the true activities underlying EBNA3C function with Cyclin complexes . Nevertheless , we will be undertaking this line of experimentation in the near future . To promote G1-S phase transition , nuclear localization of Cyclin D1 is critical and it occurs either via decreased proteolysis in cytoplasm which facilitates nuclear import or through inhibition of GSK-3β function which triggers nuclear export via phosphorylation at T286 [27] , [59] . Immunofluorescent studies showed that EBNA3C expression enforces nuclear localization of Cyclin D1 . Increased stability and nuclear accumulation of Cyclin D1 in the presence of EBNA3C prompted us to examine whether EBNA3C can also negatively regulate GSK-3β function linked to the stability of Cyclin D1 . Indeed , our data show that EBNA3C forms a complex with GSK-3β and significantly reduces its kinase activity toward Cyclin D1 , thereby enhancing the nuclear retention of Cyclin D1 . Altogether , these data suggest that either by increasing nuclear import by blocking the poly-ubiquitination level of Cyclin D1 or inhibiting nuclear export of Cyclin D1 via inhibiting the kinase activity of its negative regulator GSK-3β , EBNA3C would likely ensure the efficient nuclear accumulation of Cyclin D1 during G1-phase . However , we cannot eliminate the possibility that EBNA3C may also facilitate Cyclin D1 nuclear accumulation through additional strategies . Cyclin D1 is central to the coordination of the cell-cycle progression at the G1 to S phase transition by integrating the control of pRb phosphorylation with the transcriptional activity of E2F [60] . Cyclin D1 in association with its binding partner , CDK4 or 6 phosphorylates pRb to facilitate S phase entry [60] . Previously we have shown that EBNA3C enhances the kinase activity of Cyclin A/CDK2 complex [15] and recruits an E3 ligase SCFSkp2 to degrade pRb [17] . Similarly , here we show that by an in vitro kinase assay EBNA3C increases the activity of Cyclin D1/CDK6 complex toward both Histone H1 and a truncated mutant of pRb . Moreover , EBNA3C in conjunction with Cyclin D1/CDK6 complex increases pRb poly-ubiquitination and thereby enhances its degradation process . In addition , we show EBNA3C coupled with Cyclin D1/CDK6 complex significantly abolishes the growth suppressive function of pRb in Saos-2 cells . Studies using serum starved conditions have shown that both EBV and its essential nuclear antigen , EBNA3C can bypass G1 restriction point probably through disruption of Cyclin/CDK-pRb-E2F pathway [21] , [36] . LMP1 has also been shown to be associated with resistance to G1 arrest during serum starvation [36] . Taking advantage of these findings , together with the use of sh-RNA mediated gene knockdown strategies , we have generated knockdown lymphoblastoid cell-lines targeting both EBNA3C and cyclin D1 transcripts and assayed for cellular proliferation to carefully determine the plausible role of both of these viral and cellular oncoproteins . Indeed , our results show that both EBNA3C and Cyclin D1 are critical for efficient proliferation of EBV infected B-cells . Moreover , the results point out that upon knockdown of these gene products , cells undergo significant apoptosis , probably through induction of the activities of the tumor suppressors – p53 and pRb . Earlier results [14] and the data herein adequately show that EBNA3C critically regulates the growth suppressive properties of both p53 and pRb . Overall , we have shown in this report that the essential EBV latent antigen , EBNA3C physically interacts with and stabilizes Cyclin D1 by blocking nuclear export or inhibiting the poly-ubiquitination . In addition , EBNA3C alters pRb phosphorylation as well as stability by enhancing Cyclin D1/CDK6 kinase activity , thereby nullifying pRb mediated growth suppressive activity ( Fig . 11 ) . Furthermore , knockdown of both EBNA3C and Cyclin D1 expression by lentivirus-delivered sh-RNA demonstrated that both EBNA3C and Cyclin D1 play a critical role in cell proliferation in EBV transformed cells . Thus , the present study provides an insight into the mechanisms linked to the development of EBV-associated B-cell lymphomas through the enhancement of a major cell-cycle component , Cyclin D1 , which is known to orchestrate the activities of a vast range of cellular networks that are important in the development of human cancers .
|
Epstein-Barr virus ( EBV ) , a ubiquitous human herpesvirus , is linked to the development of multiple cancers , including lymphomas and epithelial carcinomas . EBNA3C , one of its essential latent antigens encoded by EBV , is expressed in EBV-associated lymphomas and contributes to aberrant cell growth after EBV infection . Cyclin D1 over-expression is associated with numerous cancers and is crucial for the transition from G1 to S phase in the mammalian cell-cycle . This study demonstrates that EBNA3C can enhance the functional activity of the Cyclin D1/CDK6 complex which in turn facilitates the G1 to S transition by neutralizing the growth inhibitory effects of pRb . Thus , manipulation of Cyclin D1 functions by EBNA3C provides a favorable environment to promote malignant transformation of EBV infected B-cells .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"infectious",
"diseases/viral",
"infections",
"virology/viruses",
"and",
"cancer",
"oncology/hematological",
"malignancies",
"microbiology/cellular",
"microbiology",
"and",
"pathogenesis"
] |
2011
|
Epstein-Barr Virus Nuclear Antigen 3C Facilitates G1-S Transition by Stabilizing and Enhancing the Function of Cyclin D1
|
The repair of DNA double-strand breaks must be accurate to avoid genomic rearrangements that can lead to cell death and disease . This can be accomplished by promoting homologous recombination between correctly aligned sister chromosomes . Here , using a unique system for generating a site-specific DNA double-strand break in one copy of two replicating Escherichia coli sister chromosomes , we analyse the intermediates of sister-sister double-strand break repair . Using two-dimensional agarose gel electrophoresis , we show that when double-strand breaks are formed in the absence of RuvAB , 4-way DNA ( Holliday ) junctions are accumulated in a RecG-dependent manner , arguing against the long-standing view that the redundancy of RuvAB and RecG is in the resolution of Holliday junctions . Using pulsed-field gel electrophoresis , we explain the redundancy by showing that branch migration catalysed by RuvAB and RecG is required for stabilising the intermediates of repair as , when branch migration cannot take place , repair is aborted and DNA is lost at the break locus . We demonstrate that in the repair of correctly aligned sister chromosomes , an unstable early intermediate is stabilised by branch migration . This reliance on branch migration may have evolved to help promote recombination between correctly aligned sister chromosomes to prevent genomic rearrangements .
Homologous recombination ( HR ) is a mechanism of DNA double-strand break repair ( DSBR ) that is conserved from bacteria to humans [1] . It involves resection of the broken DNA ends to generate single-stranded DNA overhangs , coated in a recombinase , which search the genome for homologous sequences and catalyse a reaction termed strand-invasion [2] . The product of strand-invasion is a joint molecule ( JM ) , containing multiple DNA duplexes and frequently comprised of D-loops and Holliday junctions ( HJs ) , also referred to as 3-way and 4-way DNA junctions , respectively . From the JM , DNA synthesis is established to restore the genetic information lost as a result of the break . Once synthesis is complete , the JM is resolved to generate the recombinant products of repair . When strand-invasion occurs between DNA sequences that are not fully homologous , such as between regions of repetitive DNA located on the same or different chromosomes , gross chromosomal rearrangements can occur . In higher organisms , where repetitive sequences are known to make up a substantial proportion of the genome , gross chromosomal rearrangements are associated with cancer [3] , [4] , [5] . This suggests that mechanisms exist for ensuring the correct pairing of sister chromosomes during HR . In order to gain further insight into the mechanism of HR , it is necessary to be able to detect different intermediates of repair as they are formed in live cells . To achieve this , it is desirable to work with a system for generating a site-specific DNA double-strand break ( DSB ) that can be efficiently repaired by HR with an unbroken sister chromosome . Such a system was described in 2008 in Escherichia coli [6] . This system uses an inducible hairpin endonuclease , SbcCD , to cleave a DNA hairpin that forms on the lagging-strand template following replication of a 246 bp interrupted palindrome that has been inserted into the chromosomal lacZ gene ( Figure S1 ) . Despite the fact that E . coli has a single origin of chromosomal DNA replication , this cleavage reaction generates a two-ended DSB at lacZ ( Figure 1A ) implying that cleavage occurs post-replication [6] . We distinguish the two sides of the break as origin-proximal ( OP ) and origin-distal ( OD ) , also labelled OP and OD in all relevant figures ( Figure S1 ) . The DSB was shown to be efficiently repaired by RecBCD-mediated HR ( Figure 1B ) [6] . In order to accumulate intermediates of repair generated by this system , it is necessary to prevent their resolution . In E . coli , the proteins RuvABC and RecG have been implicated in resolving intermediates of HR . HJs are branch migrated by RuvAB and resolved via cleavage mediated by RuvC [7] , [8] , [9] , [10] . Due to a strong synergistic effect of mutations in the ruv and recG genes in the efficiency of conjugational recombination , P1 transduction and survival following exposure to ionizing radiation and ISceI-mediated DSBs , a functional overlap of these proteins has been proposed , suggesting that RecG may also be implicated in resolving HJs [11] , [12] . Throughout this paper we use the term resolution in its general sense of converting a molecule containing HJs to one without ( i . e . resolution can be by branch migration , DNA replication , or cleavage ) . In support of a role of RecG in resolution , in vitro experiments have shown that both RuvABC and RecG process the same synthetic DNA junctions [13] , [14] . Additionally , in vivo suppression of ruv mutations , by expression of the cryptic HJ resolvase RusA , also requires RecG [15] . Furthermore , the Mycobacterium tuberculosis RecG homologue , MtRecG , was shown to process similar branched DNA junctions in vitro [16] . However , it is important to note that many different roles for RecG have been proposed in the literature . Early work has shown that RecG antagonises RecA-mediated strand-exchange [17] , [18] . This was puzzling given that RecG promotes recombination and led to the proposal that RecG might facilitate RecA-mediated strand-exchange from a 3′ invading substrate while antagonising strand-exchange from a 5′ invading substrate [17] , [19] . In subsequent work , it has been argued that RecG catalyses replication fork reversal following UV irradiation [20] and prevents over-replication caused by replication fork collision , by converting 3′ to 5′ single-strand flaps [21] , [22] , [23] , [24] . Whether or not these proposed activities relate to the synergy of recG and ruv mutations has not been clear , and the diverse consequences of a single recG mutation , as well as the ability of the purified protein to process many different substrates , have generated a complex picture of RecG's biological role . Using the palindrome-based system for inducing DSBR between sister chromosomes , we analyse the intermediates of repair accumulated in the absence of the ruv and recG genes to elucidate their function during DSBR and gain further insight into the precise mechanism of repair . We show that RuvABC is the main HJ branch migration and resolution complex in E . coli and that RecG is required for the formation of HJs , by converting 3-way DNA junctions ( D-loops ) to 4-way DNA junctions ( HJs ) . We go on to show that in the absence of both RuvAB and RecG , DNA is lost at the breakpoint due to an inability of a ΔruvAB ΔrecG mutant to catalyse branch migration . We conclude that branch migration , catalysed by either RuvAB or RecG , is essential for stabilising intermediates of DSBR by promoting the conversion of 3-way DNA junctions into 4-way DNA junctions , a conclusion that can explain the synergistic behaviour of ruv and recG mutants . We propose that this mechanism for stabilising intermediates favours DSBR reactions that occur between correctly aligned sister chromosomes , thus serving as a mechanism for ensuring correct pairing of sisters and , in turn , accurate repair of DSBs .
ruv and recG mutants have been shown to be sensitive to DNA damage and this sensitivity is exacerbated in ruv recG double mutants [11] , [12] . In accordance with these studies , DNA damage induced by SbcCD-mediated cleavage of a palindrome caused a loss of viability in single ΔruvAB or ΔrecG mutants that was severely exacerbated in the double ΔruvAB ΔrecG mutant ( Figure 2 and Figure S2 ) . Presumably , the decrease in viability is a consequence of the accumulation of toxic DNA repair intermediates that would normally be processed by these proteins . To detect these hypothetical repair intermediates and determine their structures , constructs containing three repeats of the crossover hotspot instigator , Chi ( χ ) , were integrated 1 . 5 kb either side of the palindrome in order to enrich for recombination intermediates in close proximity of the DSB ( Figure 1C ) . Subsequently , DNA from strains containing these constructs was isolated , digested with restriction endonucleases , and separated by two-dimensional ( 2D ) agarose gel electrophoresis; a useful technique for distinguishing between 3-way DNA junctions and 4-way DNA junctions ( Figure 3A ) . Three fragments surrounding the DSB were detected using radioactive probes ( Figure 3B ) . All membranes were exposed for the same amount of time and intermediates were quantified relative to linear DNA ( Figure 3C and S3 ) . As shown in Figures 3CII , an increase in intermediates was detected in ΔruvAB , ΔrecG and ΔruvAB ΔrecG mutants specifically in conditions in which DSBs were induced ( DSB+ ) relative to a very low background of spontaneous intermediates detected in the absence of induced breaks ( DSB− ) ( Figure S3 ) . A ΔruvAB mutant , accumulated a significant amount of 4-way junctions , presumably HJs , when DSBs were induced ( Figure 3CI; red arrows and 3CIII ) . This was not the case when DSBs were induced in either ΔrecG or ΔruvAB ΔrecG mutants ( Figure 3CI and 3CIII ) . As 4-way junctions accumulated in a ΔruvAB mutant but not in a ΔruvAB ΔrecG mutant , this suggests that RecG cannot simply be required for the resolution of 4-way junctions and must be required for their formation; presumably by catalysing the conversion of 3-way to 4-way junctions , an activity that has been reported for RecG in vitro [19] , [20] , [25] , [26] . It is interesting to note that the analysis of the ΔruvAB mutant reveals the existence of preferred configurations of branched DNA , which are seen as spots on the 2D gels ( Figure 3CI ) . The placement of these spots is reproducible suggesting that they reflect DNA structures that accumulate in preference to others . Further work is required to determine what these structures are and how they are formed . Spots on the 4-way junction spike may reflect asymmetrically placed single HJs or double HJs and spots on the 3-way junction arcs may reflect positions of preferential single-strand invasion or pausing of DNA synthesis . However , these 3-way junction spots do not simply correlate with the expected positions of single-strand invasion predicted by the positions of Chi ( χ ) sites . As 3-way junctions are expected to form early in the reaction via strand invasion , as well as later during re-synthesis of the broken DNA , further work is required to understand their provenance . 2D agarose gel electrophoresis is only suitable for analysing small chromosomal fragments ( 2–7 Kb ) . In order to determine whether intermediates of repair could be located across larger regions of the chromosome , pulsed-field gel electrophoresis ( PFGE ) was used as it allows the separation of big fragments of DNA . Additionally , branched DNA does not run into a pulsed-field gel ( PFG ) , but remains trapped in the wells , and this allows it to be separated from its linear counterpart [27] . Plugs containing chromosomal DNA were digested to release three fragments surrounding the DSB ( yagV , lacZ , and araJ ) ( Figure 4A ) . The total amount of DNA detected in these fragments ( the sum of the signal from the gel and the well ) was normalised to a control fragment , of a similar size , located on the opposite side of the chromosome ( cysN ) to account for differences in loading between samples . Additionally , the proportion of DNA that was retained in the wells of the gels was also measured as this DNA included the branched intermediates of repair ( Figure 4B–E ) . In conditions of no DSBs ( lanes 1 , 2 and 3 for each probe ) , little DNA , of all the fragments probed , was retained in the wells ( Figure 4B–E ) . A similar result was obtained when DSBs were induced in a recombination proficient strain ( Figure 4B; lane 4 for each probe ) . Upon inducing DSBs in a ΔruvAB mutant , a large proportion of the lacZ fragment , containing the DSB , was detected in the well of the gel whereas little of the yagV and araJ fragments appeared to contain branched DNA ( Figure 4C ) . In a ΔrecG mutant , DSB induction resulted in a small amount of branched DNA in all three fragments ( Figure 4D ) . Unexpectedly , analysis of the DNA extracted from a ΔruvAB ΔrecG double mutant showed that when DSBs were induced , a significant amount of the DNA at the breakpoint ( lacZ fragment ) was lost ( Figure 4E ) . It should be noted here that this result explained the low yield of DNA in the 2D gel analysis of the ΔruvAB ΔrecG double mutant . The reader should be aware that the DNA species obtained from the ΔruvAB ΔrecG mutant visualised using 2D gel electrophoresis ( Figure 3 ) , represent the minority of molecules recovered when DSBs were induced in that background . The lacZ probe lies between the palindrome and the OP 1 . 5 Kb 3x χ array , in a region of DNA predicted to be degraded pre-RecBCD-mediated loading of RecA and strand-invasion . Therefore , loss of DNA in this region may suggest an inability of this mutant to initiate DNA synthesis associated with repair . However , a significant loss of DNA was also detected in the OD araJ fragment , which lies beyond the OD 1 . 5 Kb 3x χ array . This profile suggests that the loss of DNA observed may not be due to an inability to re-establish DNA synthesis , but due to an inability to form repair intermediates close to the DSB . Interestingly , in the OP yagV fragment , there was no loss of DNA but a dramatic accumulation of branched DNA . 2D agarose gel electrophoresis confirmed this accumulation of intermediates but revealed that there was still no bias towards the accumulation of either 3-way or 4-way DNA junctions when DSBs were induced , as was seen with the same mutant in the DNA remaining at the locus of the breakpoint ( Figure 3C and Figure 5 ) . A ΔruvAB ΔrecG mutant , shown to lose DNA at the site of a DSB , is both unable to branch migrate and resolve HJs . In order to determine which of these activities is required to prevent the loss of DNA observed , a ΔruvAB ΔrecG mutant was compared to a ΔruvC ΔrecG mutant . A ΔruvC ΔrecG mutant still retains RuvAB and should therefore be able to catalyse branch migration . However , RuvAB cannot resolve HJs in the absence of RuvC , so HJs should remain unresolved in this background . The ability of RuvAB to catalyse branch migration in the absence of RuvC was confirmed by PFGE ( Figure 6 ) . A significant amount of branched DNA was accumulated in the wells of the PFGs in ΔruvAB and ΔruvC mutants ( Figure 6C ) , consistent with the hypothesis that HJs are only resolved when all components of the RuvABC complex are present . However , the branched DNA accumulated in a ΔruvAB mutant was located within the lacZ fragment containing the DSB , while in a ΔruvC mutant , branched DNA was detected in all three fragments surrounding the break . This is indicative of RuvAB-mediated branch migration being active in the absence of RuvC . Once this was verified , PFGE was used to check whether a ΔruvC ΔrecG mutant lost DNA in response to DSBs and to compare this to DNA loss in a ΔruvAB ΔrecG strain ( Figure 7 ) . In order to detect DNA located OP of the DSB and beyond the point of initial RecBCD-mediated loading of RecA and strand-invasion , a new probe , codB , that binds 8 . 5 Kb OP to the 3x χ array , was designed ( Figure 7A ) . Between the breakpoint and the codB probe , as well as the 1 . 5 Kb 3x χ array , there is an endogenous χ site located 5 Kb from the breakpoint , in the cynX gene . Assuming a 20%–35% probability of χ site recognition , these four χ sites should be responsible for between 59% and 82% of strand-invasion events [28] , [29] , [30] . As shown in Figure 7 , DNA hybridising to the codB probe was lost in a ΔruvAB ΔrecG mutant when DSBs were induced , consistent with the hypothesis that intermediates of repair are not stable in this background . Interestingly , this loss did not occur in a ΔruvC ΔrecG mutant . These results imply that the loss of DNA observed in a ΔruvAB ΔrecG mutant is due to an inability to branch migrate intermediates of repair , rather than an inability to resolve HJs , and this results in the destabilisation of repair intermediates .
Due to a synergistic effect of mutations in the ruv and recG genes , it had originally been argued that these proteins may provide alternative pathways for resolving HJs . We have corroborated the observation that mutations in both ruvAB and recG result in enhanced sensitivity to DSBs compared to the respective single mutations when DSBs are induced by SbcCD-mediated cleavage of a palindrome ( Figure 2 and Figure S2 ) . However , analysis by 2D agarose gel electrophoresis of the DNA at the DSB has confirmed that this enhanced sensitivity was not accompanied by an accumulation of HJs ( 4-way DNA junctions ) ( Figure 3C ) . This result argues against the view that RuvABC and RecG are simply redundant because they provide alternative pathways to resolve HJs . 4-way DNA junctions were indeed accumulated close to the DSB in a ΔruvAB mutant , consistent with a role of RuvAB in processing HJs ( Figure 3C ) . However , these 4-way junctions were not accumulated in proximity to the DSB in a ΔruvAB ΔrecG mutant , arguing that RecG is required for their formation . The use of PFGE for studying intermediates of DSBR revealed why 4-way DNA junctions were not accumulated close to the DSB in a ΔruvAB ΔrecG mutant . In the absence of both RuvAB and RecG , DNA was lost at the site of the DSB . This was accompanied by an accumulation of branched DNA over 30 Kb away from the breakpoint ( Figure 4 and 5 ) . For the DNA in the lacZ locus to be lost , and for intermediates of repair to be present in the yagV fragment , the OP DNA end must be processed , by RecBCD , from the lacZ fragment to the yagV fragment . This is surprising as RecBCD will encounter eight endogenous χ sites ( as well as the OP 3x χ array ) in the region of the chromosome between the DSB and the yagV fragment and should induce RecA-mediated strand-invasion as a result [31] . This suggests that in a ΔruvAB ΔrecG mutant background , the products of RecA-mediated strand-invasion are not stable , which allows RecBCD to process a region of the chromosome that would not be processed in a wild type context . χ sequences around the E . coli chromosome are distributed asymmetrically to limit DNA end processing by RecBCD on the OP side of a DSB [32] . The asymmetry detected for OP accumulation of branched DNA and OD loss of DNA in a ΔruvAB ΔrecG mutant reflects this asymmetry of endogenous χ sequences , strengthening the hypothesis that the degradation is mediated by RecBCD . There are eight endogenous χ sites between the break and the OP yagV fragment that itself contains two χ sites and only one endogenous χ site between the break and the OD araJ fragment that contains no χ sites . We conclude that in a ΔruvAB ΔrecG mutant the products of strand-invasion are transient and non-productive for repair due to an inability to branch-migrate 3-way junctions and form 4-way junctions . This leads to the disruption of the 3-way junctions and the formation of a new DNA end for RecBCD to process . When the next χ site is recognised , a new event of strand-invasion is initiated , which is once again disrupted by a lack of branch migration activity . Over time , the broken chromosome is degraded . We propose that in ruvABC+ recG+ cells , when sister chromosomes are correctly aligned , branch migration is facilitated and this stabilises intermediates of repair by promoting the formation of 4-way DNA junctions . This favours the accurate repair of DSBs . This interpretation is supported by the observation that the frequency of ectopic recombination is increased in recG mutant strains in a chromosomal direct repeat deletion assay [33] , [34] , [35] and in recombination between chromosomal and plasmid homologies [36] . In the direct repeat assay , this is the case unless the replicative helicase is compromised [33] , [35] . The redundancy we observe in the stabilisation of JMs can explain the synergistic defect caused by ruv and recG mutations and this no longer necessitates the previously proposed redundancy in HJ resolution . However , redundancy at this stage cannot be excluded . Furthermore , if RuvABC and RecG do not provide alternative pathways for the resolution of HJs , such pathways must nevertheless exist otherwise recG and ruv mutations would be epistatic . This has led us to consider again the evidence that recG and ruv provide two pathways for HJ resolution . The strongest evidence in favour of this hypothesis is the observation that suppressors of the UV sensitivity of ruv mutations cause activation of the cryptic HJ resolvase , RusA , and this suppression requires RecG [15] . The simplest interpretation of this result is that the branch migration activity of RecG translocates HJs to positions where they are cleaved by RusA . However , RusA is not expressed in the absence of the activating mutation , rus , and no HJ resolvases other than RusA and RuvC have been discovered in E . coli [37] . Furthermore the requirement for recG in the suppression of ruv by rus can now simply be explained by the destabilisation of JMs that we observe in a recG ruvAB double mutant . If JMs are not formed , then they cannot be resolved by RusA . This leaves the question of whether there exists a pathway to resolve HJs that is an alternative to cleavage by RuvABC . The genetics argue that this is so . Ruv mutants are only modestly recombination defective but recG ruv double mutants are as defective as recA . This is synergy , not epistasis , arguing that the presence of RuvABC or RecG can provide alternative ways of successfully catalysing recombination . If synergy is explained by redundancy of RuvAB and RecG at the stage of JM formation and RuvABC provides a way to resolve HJs then there must also be a way to resolve HJs in the absence of RuvABC . What is this route ? The observation that HJ resolution in the absence of RuvABC leads to substantial yields of chromosome dimers [11] , [27] demonstrates clearly that this pathway can generate crossover products and excludes models such as double HJ dissolution by branch migration that would produce only non-crossovers . It has been suggested that new rounds of DNA replication initiated at the chromosomal origin can sometimes pass through HJs and generate the resolved chromosomes [27] . To explain the synergy of recG and ruv , given the assumption that the activities were redundant for HJ resolution , it was suggested that RecG might facilitate this reaction . However , the results presented here open up the possibility that the replication forks that manage to pass through HJs may do so without the help of RecG . It is clear from our work that HJs accumulate in a ruvAB mutant , implying that they persist long enough to be detected and the data shown in Figure 6 argue that JMs are not resolved before they can be branch migrated by RuvAB . These data are not well explained by an immediate role of RecG in HJ resolution but are compatible with a delay of resolution in the absence of RuvABC as predicted if resolution is mediated by the next round of DNA replication initiated at the chromosomal origin . Many functions have been proposed for RecG , including the resolution of Holliday junctions [11] , [12] , replication fork reversal following UV irradiation [20] , conversion of 3′ flaps to 5′ flaps in the termination of replication [21] , [22] , [23] , [24] , destabilisation of RecA promoted strand exchange [17] , [18] and stabilisation RecA-promoted strand exchange [17] , [19] . Our results clearly demonstrate the importance of the role of RecG , as an alternative to RuvAB , in stabilising RecA-promoted strand exchange in DSBR . Many models for the repair of DNA DSBs have been proposed over the years and these are reviewed in detail by Pâques and Haber [38] . Some of the models predict the formation of 4-way DNA junctions , from 3-way DNA junctions , and some do not . Most models for the repair of two-ended DSBs in Saccharomyces cerevisiae implicate invasion of one DNA end followed by DNA synthesis that uncovers a region of homology to induce an event known as second-end capture . This can be processed to generate a double HJ intermediate that has been detected in vivo in meiotic and mitotic cells [39] , [40] , an intermediate that may be resolved by branch migration or HJ cleavage ( Figure 8 – HJ resolution ) . Alternatively , the invading strands can be ejected and re-annealed , prior to the completion of the double-HJ structure , in a reaction known as synthesis-dependent stand-annealing ( Figure 8B – SDSA ) , a mechanism that has the advantage of not generating crossover outcomes . If strand-invasion were to occur at short regions of homology , such as repetitive elements , rather than at correctly aligned sister chromatids or homologous chromosomes , second-end capture may be disfavoured . If it does occur , and resection proceeds beyond the region of homology , resolution by SDSA would minimise genome instability by ensuring non-crossover outcomes [41] . In S . cerevisiae , during the repair of a two-ended DSB in which second-end capture is prevented , the invading end can be repaired by break-induced replication ( BIR ) ( see [42] for a recent review ) . BIR has been shown to involve multiple rounds of strand-invasion in the initial phase of the reaction , consistent with repair-intermediate instability [43] . Furthermore , BIR is mutagenic consistent with a D-loop migration mechanism in which short-lived mismatches are not corrected but , instead , are copied in a conservative mode of DNA replication [44] , [45] , [46] ( Figure 8C ) . These observations suggest that second-end capture plays an important role in promoting accurate repair of two-ended DSBs . Indeed , second-end capture prevents BIR and promotes gene conversion through the operation of a recombination execution checkpoint ( REC ) that senses the proximity and orientation of the two recombining ends before DNA synthesis is initiated . When such ends are sensed , as is the case with a two-ended DSB , accurate repair is ensured and the outcome is directed towards gene conversion [47] . In contrast to DSBR in eukaryotes , in E . coli , DSBR involves extensive DNA degradation followed by the re-establishment of replication forks via the PriA-DnaB pathway of replisome loading [2] , [48] , [49] . This is understood to result in the formation of converging replication forks that restore the DNA between the two recombining ends ( Figure 1B ) . Within this model of DSBR , the stabilisation of intermediates by second-end capture should not be possible . We suggest that branch migration is an alternative to second-end capture for stabilising an intermediate that can be then converted to a 4-way DNA junction . The stabilisation of recombination intermediates by branch migration , which we have observed , is expected to work equally well for two-ended and one-ended DSBs . On the other hand , the stabilisation of intermediates determined in some way by second-end capture , by definition , cannot operate at one-ended DSBs . These types of DSBs do arise endogenously from replication forks that run into replication fork barriers , single-stranded DNA nicks or gaps , and from cleavage of reversed forks , and are thought to be the most common type of break encountered by all cells [50] , [51] , [52] . As second-end capture cannot be implicated as a mechanism for stabilising the intermediates generated from the repair of one-ended DSBs , this raises the intriguing question of how they can be stabilised in eukaryotic cells . The repair of one-ended sister chromatid breaks is distinguished from inter-chromatid BIR by the requirement of Rad51 , Rad52 , Rad54 and Rad59 [53] but little is known about the pathway of repair including how early intermediates are stabilised . One possibility is that some one-ended breaks await the formation of a second end produced by the firing of a replication origin situated on the other side of the causative lesion ( i . e . a two-ended break is generated from the sum of two one-ended breaks occurring one on each side of the same inducing lesion ( such as a persistent single-strand gap ) ) . The mechanism discovered here presents a solution adopted by E . coli that is expected to work equally well at one-ended and two-ended breaks . Repair of a DSB by HR with a sister chromosome has evolved to be accurate , despite the fact that genomes contain regions of repetitive sequence that could act as substrates for incorrect pairing . Here we show that the E . coli proteins RuvAB and RecG do not simply provide alternative pathways for the resolution of HJs , as previously suggested , but play redundant roles in stabilising recombination intermediates between sister chromosomes .
All strains used are listed in the supporting information . See Table S1 for a list of strains , Table S2 for plasmids used in the construction of the strains , Table S3 for oligonucleotides used in the construction of the plasmids and protocols S1 and S2 for methods used in the construction of the strains and plasmids . Overnight cultures grown in 5 ml L-broth were diluted to an optical density ( OD600nm ) of 0 . 02 and grow at 37°C with agitation to an OD600nm of 0 . 2 . The PBAD-sbcDC construct was induced by adding 0 . 2% arabinose . If PBAD-sbcDC was to be repressed as well as induced , the culture ( OD600nm of 0 . 2 ) was split in two and either 0 . 5% glucose or 0 . 2% arabinose was added . Cultures were put back at 37°C to grow for 60 minutes . Cells were harvested at 4°C and washed 2X in TEN buffer ( 50 mM Tris , 50 mM EDTA , 100 mM NaCl , pH 8 . 0 ) . Cells were re-suspended in TEN buffer to an OD600nm of 80 ( for 2D agarose gel electrophoresis ) or an OD600nm of 4 ( for PFGE ) and mixed with an equal volume of 0 . 8% ( for 2D agarose gel electrophoresis ) or 2% ( for PFGE ) low melting point agarose ( Invitrogen ) prepared in TEN buffer equilibrated to 50°C . The agarose/cell mix was poured into plug moulds ( BioRad ) and allowed to set . Plugs were treated in NDS solution ( 0 . 5 M EDTA , 10 mM Tris , 0 . 55 M NaOH , 36 . 8 mM lauroyl sarcosine; pH 8 . 0 ) supplemented with 1 mg/ml of proteinase K ( Roche ) and put at 37°C overnight . Fresh NDS + proteinase K was added for a second overnight and plugs were stored at 4°C in fresh NDS . To digest , a plug was washed in 1X restriction buffer for 6 hours , replacing the buffer every hour . The plug was placed in fresh 1X restriction buffer , supplemented with the restriction enzyme and incubated at 37°C overnight with rocking . A plug digested with a restriction enzyme was run in the first dimension in 1X TBE ( 89 mM Tris-borate , 2 mM EDTA ) on a 0 . 4% ( w/v ) agarose gel and run at 1 V/cm for 26 hours at 4°C . The lane was sliced out , rotated 90° , and set in the second dimension agarose ( 1% in 1X TBE supplemented with 0 . 3 µg/ml ethidium bromide ) . The second dimension was run at 6 V/cm for 10 hours at 4°C . The DNA was transferred to a positively charged nylon membrane by Southern blotting and cross-linked using UV-light . A plug digested with a restriction enzyme was run on a 1% ultra high gel strength agarose ( AquaPor ) prepared in 0 . 5X TBE and run on a CHEF-DR II PFGE ( BioRad ) at 6 V/cm for 10 hours at 4°C . Switch time was set to 5–30 seconds with an inclusion angle of 120° . The DNA was transferred to a positively charged nylon membrane by Southern blotting and cross-linked using UV-light . DNA was detected using 32P α-dATP incorporated ( using Stratagene Prime-It II random primer labelling kit ) into a PCR fragment . Probes were hybridised to membranes overnight at 65°C in 10 ml of Church-Gilbert buffer ( 7% SDS , 0 . 5 M NaH2PO4 , 1 mM EDTA , 1% BSA ) . Membranes were washed at 60°C in 2X SSC ( 1X SSC: 0 . 15 M NaCl , 0 . 015 M Na-citrate ) , supplemented with 0 . 1% SDS , for 15 minutes and then 0 . 5X SSC , supplemented with 0 . 1% SDS , for 30 minutes . Labelled membranes were exposed to GE healthcare storage phosphor screens and scanned using a Molecular Dynamics Storm 860 phosphor imager scanner . Images were quantified using GE healthcare ImageQuant TL . See Table S3 for the oligonucleotides used in the generation of the probes .
|
Genetic recombination is critically important for the repair of DNA double-strand breaks and is the only repair mechanism available to the bacterium Escherichia coli . Repair requires that the appropriate location on an unbroken sister chromosome is recognised as a repair template , and this can be accomplished by a system that detects the presence of extensive DNA sequence identity . We show here that the two known branch migration activities of the cell , RuvAB and RecG , provide alternative mechanisms for stabilising early recombination intermediates . In their absence , broken DNA is extensively degraded at the site of the break consistent with abortion of recombination . It has previously been proposed that RuvABC and RecG can substitute for each other in the resolution of four-way Holliday junctions , whereas we show that they play a synergistic role in the formations of these junctions . Our results demonstrate that branch migration provides a mechanism capable of stabilising recombination intermediates when extensive DNA sequence homology is available , a reaction that may contribute to ensuring that repair occurs at an appropriate location on a sister chromosome .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"biochemistry",
"cell",
"biology",
"nucleic",
"acids",
"biology",
"and",
"life",
"sciences",
"dna",
"dna",
"recombination",
"molecular",
"cell",
"biology"
] |
2014
|
Branch Migration Prevents DNA Loss during Double-Strand Break Repair
|
The pathogenic fungus Cryptococcus neoformans exhibits morphological changes in cell size during lung infection , producing both typical size 5 to 7 μm cells and large titan cells ( > 10 μm and up to 100 μm ) . We found and optimized in vitro conditions that produce titan cells in order to identify the ancestry of titan cells , the environmental determinants , and the key gene regulators of titan cell formation . Titan cells generated in vitro harbor the main characteristics of titan cells produced in vivo including their large cell size ( >10 μm ) , polyploidy with a single nucleus , large vacuole , dense capsule , and thick cell wall . Here we show titan cells derived from the enlargement of progenitor cells in the population independent of yeast growth rate . Change in the incubation medium , hypoxia , nutrient starvation and low pH were the main factors that trigger titan cell formation , while quorum sensing factors like the initial inoculum concentration , pantothenic acid , and the quorum sensing peptide Qsp1p also impacted titan cell formation . Inhibition of ergosterol , protein and nucleic acid biosynthesis altered titan cell formation , as did serum , phospholipids and anti-capsular antibodies in our settings . We explored genetic factors important for titan cell formation using three approaches . Using H99-derivative strains with natural genetic differences , we showed that titan cell formation was dependent on LMP1 and SGF29 genes . By screening a gene deletion collection , we also confirmed that GPR4/5-RIM101 , and CAC1 genes were required to generate titan cells and that the PKR1 , TSP2 , USV101 genes negatively regulated titan cell formation . Furthermore , analysis of spontaneous Pkr1 loss-of-function clinical isolates confirmed the important role of the Pkr1 protein as a negative regulator of titan cell formation . Through development of a standardized and robust in vitro assay , our results provide new insights into titan cell biogenesis with the identification of multiple important factors/pathways .
The ubiquitous environmental yeast Cryptococcus neoformans is a basidiomycetous yeast that has been estimated to cause over 200 , 000 new cases of meningoencephalitis with greater than 180 , 000 deaths per year worldwide [1] , occurring mostly in immunocompromised individuals with acquired immunodeficiency syndrome ( AIDS ) [2] . The natural history of most of the cases of this invasive fungal infection proceeds through 3 stages: ( i ) primary infection via inhalation of desiccated yeasts or basidiospores , with development of sub-clinical pneumonia and spontaneous resolution via granuloma formation; ( ii ) latency of dormant yeast cells , as demonstrated epidemiologically [3] and biologically [4] and ( iii ) reactivation and dissemination upon immunosuppression , with meningoencephalitis as the most severe clinical presentation of disease [5] . From the environment to interactions with hosts , the yeasts experience drastic changes that reflect a capacity to rapidly adapt and survive in host tissues and cause disease [4 , 6–9] . In hosts , C . neoformans is exposed to various stresses including high temperature , nutrient deprivation , low pH , hypoxia and high levels of free radicals [10] . In response to the host environment , morphological changes are required to survive and cause disease [11] . Specifically , C . neoformans alters its morphology and produces enlarged cells referred to as giant or “titan cells” [12 , 13] . This phenomenon has been observed in animal and insect models of cryptococcosis , as well as in human lung and brain infections [12–16] . Titan cells have increased cell body size , ranging from 10 μm up to 100 μm in diameter [12 , 13 , 17–21] as compared to the 5–7 μm size of typical cells . Studies exploring titan cell biology have revealed that these are: ( i ) uninucleate polyploid cells [12 , 13 , 20]; ( ii ) possess a large single vacuole; ( iii ) are surrounded by a thick cell wall [22]; and ( iv ) have a dense and highly crosslinked capsule [13 , 17 , 22] . Titan cells also exhibit increased resistance to various stresses including phagocytosis [21] , oxidative and nitrosative stress [20 , 21] , and resistance to the antifungal drug fluconazole [20] . Importantly , titan cell production also enhances dissemination , survival and virulence in a mouse model of infection [19] . Titan cell formation is known to be regulated by the G-protein coupled receptors Gpr5 and Ste3a , that signal through the Gα subunit protein Gpa1 to trigger the cyclic adenosine monophosphate / protein kinase A ( cAMP/PKA ) signaling pathway [15 , 20 , 23–26] . The cAMP/PKA pathway is critical for regulation of other virulence factors in C . neoformans , including capsule formation [23 , 24] , notably through its action on the ubiquitin-proteasome pathway [27] . Pka1 is known to be negatively regulated by the protein Pkr1 , and pkr1Δ mutant strains exhibit enlarged capsule [23] . Further studies show that titan cells formation is increased by high PKA1 expression or low Pkr1 activity and is decreased by low PKA1 expression [6] . Downstream of the PKA pathway , Rim101 , a major transcription factor that again controls production of many virulence factors , is also necessary for titan cell production [18] . To date , studies of titan cell formation have been hindered by an inability to consistently and reproducibly generate large quantities of titan cells in vitro . Although several methods have been reported for inducing large cells in vitro , there have been persistent problems in easily and consistently implementing these protocols across laboratories [17] , presumably because the variables that contribute to titan cell inducing conditions are not well understood . In this study , we identified robust in vitro conditions that generate enlarged cells with many of the in vivo titan cell characteristics and used this protocol to explore environmental and genetic factors involved in titan cell formation . The genetic determinants of titan cell formation have been investigated through a genotype-phenotype correlation study in H99-derivative laboratory reference strains , through analysis of deletion and complementation in reference strains , and analysis of genetic defects in clinical isolates using whole genome data and complementation .
While growth in minimal medium using standard growth conditions had no effect on cell size , we identified growth conditions that stimulated the production of enlarged yeast cells and optimized this experimental protocol , referred to as our in vitro protocol , using the reference strain H99O ( S1 Fig ) . Observation of these in vitro-generated large cells by microscopy shows many characteristics of titan cells including increased cell body size ( diameter >10 μm ) , refractive cell wall , large central vacuole , and peripheral cell cytoplasm distribution , similar to in vivo titan cells ( Fig 1A ) . Our in vitro protocol proved to be reproducible with H99O reference strains generating titan cells in three different laboratories throughout the world , although some variability in the overall proportion of titan cells generated was observed ( Fig 1B ) . Specifically , the proportion of titan cells was 39 . 4% [interquartile range ( IQR ) , [34 . 1–40 . 7] in Lab 1 , 21 . 0% [12 . 5–26 . 8] in Lab 2 and 29 . 9% [23 . 1–48 . 5] in Lab 3 . The distribution of yeast cell body size from the in vitro protocol varied from 3 . 7 to 16 . 3 μm ( median 10 . 2 [8 . 5–11 . 5] ) , whereas in vivo it varies from 3 . 6 to 41 . 8 μm ( median 14 . 8 [11 . 2–18 . 45] ) with 84% of the yeasts classified as titan cells ( Fig 1C ) . Titan cells differ from typical cells in various characteristics including capsule size budding rate , DNA content , cell wall and capsule structure , and the extent of melanization [12 , 13 , 20 , 22] . Comparison of in vitro titan cells ( TC ) to typical cells ( tC ) showed a significant increase in capsule size ( median 4 . 8 μm in titan cells vs 2 . 7 μm in typical cells , p<0 . 001 ) similar to that observed in vivo ( median 10 . 5 μm in titan cells vs 8 . 0 in typical cells , p<0 . 001 , Fig 1D ) . The capsule thickness of in vivo titan cells was increased compared to the in vitro titan cells . The budding rate of in vitro titan cells was also significantly increased ( median 82 . 5 m per bud ) compared to typical cells ( median 89 . 0 m per bud ) ( p = 0 . 018 ) , whereas it was similar for titan and typical cells produced in vivo . Interestingly , overall budding rate was faster in vivo than in vitro cells ( 68 and 89 m per bud , p<0 . 001 ) ) . The budding rate of both in vivo and in vitro tutan cells and typical cells were faster than cells grown in stationary phase ( 111 . 5 and 124 . 5 m per bud in minimal medium ( MM ) and YPD , respectively ) ( Fig 1E , S1 Movie , S2 Movie ) . To analyze DNA content , yeasts obtained at the end of the in vitro protocol were stained with propidium iodide ( PI ) and DNA content of titan ( TC , FSC/SSChigh ) and typical ( tC , FSC/SSClow ) cells was compared to haploid ( H99O ) and diploid ( AD7-77 ) strains grown in Sabouraud medium . DNA content was higher in titan ( FSC/SSChigh ) than typical ( FSC/SSSlow ) cells ( Fig 1F ) , with an increase in the proportion of polyploid cells in the titan cell population as observed by a PI fluorescence greater than the diploid control strain ( red arrow , Fig 1G ) . In contrast , the typical cells had the same PI fluorescence pattern as the haploid H99O cells . Similarly , a reverse gating strategy based on PI intensity shows yeasts with the highest PI intensity were large titan cells ( red arrows , S2 Fig ) . Calcofluor white ( CFW ) staining was used to analyze cell wall chitin content . After multispectral imaging flow cytometry—gating on the titan and typical cell populations under both in vitro and in vivo conditions ( Fig 2A and 2B ) —CFW fluorescence intensity ( Fig 2C and 2D ) showed significantly increased fluorescence of titan cells compared to typical cells in vitro ( 322539 ± 3072 vs 123062 ± 20727 , p<0 . 0001 , Fig 2C ) and in vivo ( 144909 ± 38487 vs 27622 ± 7412 , p<0 . 0001 , Fig 2D ) . Cell sorting based on CFW staining and fluorescence microscopy allowed us to validate that cells exhibiting the higher CFW intensity were titan cells ( FSC/SSChigh , S3B Fig ) . Measurements of chitin content using fluorescence microscopy also showed significant increases in titan cells compared to typical cells ( fluorescence intensity/pixel/cell 87 . 9 [71 . 7–107 . 7] vs 66 . 5 [51 . 7–79 . 4] , respectively , p<0 . 0001 , S4A Fig ) . Chitin levels were also assessed by N-acetylglucosamine ( Gluc-NAc ) content . Gluc-NAc levels were higher in titan cells ( 156 . 3 mM/g [153 . 7–203 . 7] ) than typical cells ( 97 . 7 [87 . 2–119 . 3] ) , ( p<0 . 001 , S4B Fig ) . Furthermore , titan cells exhibited pronounced melanization ( S4C Fig ) , as measured by blackness on the pictures ( S4D Fig ) , compared to typical cells ( S4D Fig ) , with a median of the ( max—mean grey intensity per pixel ) of 20065 [18785–21887] in titan vs 13067 [9660–15998] in typical cells ( p<0 . 0001 ) . Finally , capsule structure was also investigated based on the binding pattern of monoclonal antibodies specific for capsular polysaccharides [28] using multispectral imaging flow cytometry ( Fig 2E and 2F , S5A Fig ) and immunofluorescence ( S5B Fig ) . Based on the fluorescence pattern of the 2D10 antibody [29] , the algorithm modulation and bright details intensity R7 allowed us to discriminate the distribution of the capsule staining in titan ( TC ) and typical ( tC ) cells and showed with almost no overlap between both population in vitro ( Fig 2E ) and in vivo populations ( Fig 2F ) . Variability of staining was observed for E1 ( IgG1 ) and 13F1 ( IgM ) antibodies ( S5A Fig ) . No pronounced differences between the capsule structures of titan and typical cells was observed visually during immunofluorescence staining ( S5B Fig ) . To examine temporal changes in cell size induced by our protocol , we measured yeast cell sizes at 0 , 4 , 8 , 16 , 24 and 120 h using automated analysis . This automated analysis correlated with manual size measurements ( Interclass correlation = 0 . 99 ) and titan/typical cell classification ( Kappa test = 0 . 81±0 . 07 ) . The median cell size increased during the first 24 h of incubation , starting at 5 . 7 μm [5 . 4–6 . 0] and increasing to 9 . 7 μm [8 . 4–11 . 0] ( Fig 3A ) . The first titan cells were observed at 8 h with a progressive increase in the proportion of titan cells overtime , reaching a plateau by 24 h ( Fig 3B ) . Temporal changes in cell morphology were determined using light microscopy and live cell imaging . Cells with the large vacuole characteristic of titan cells appeared between 4 and 8 h ( Fig 3C , white arrow ) . Live cell imaging over 12 h ( Fig 3D , S3 Movie ) showed that: ( i ) titan cells swelled from the progenitor typical sized cells and ( ii ) titan cells divided to produce typical sized daughter cells . CFW staining is known to transfer only partially to daughter cells upon division resulting in lower fluorescence in daughter cells while remaining at a high level in mother cells [4 , 30] . Pulsed CFW staining was used to further monitor the ancestry of titan and typical cell populations over time using flow cytometry ( Fig 3E ) . The initial CFW stained population ( 0 h ) consisted of typical cells ( FSClow ) with high CFW fluorescence intensity ( black density lines ) . At 24 h , two populations were observed . The titan cell population ( FSChigh ) had high CFW fluorescence intensity ( black arrow ) , indicating these cells were generated by the swelling of typical sized cells in the original culture . The second population consisted of typical cells ( white arrow , FSClow ) with low calcofluor fluorescence , consistent with newly formed daughter cells ( Fig 3E , right panel ) . Combined , these data show that the titan cells derived from the initial inoculated cells and daughter cells are typical sized . We tested several parameters affecting steps 2 and 3 of our protocol described in S1 Fig and identified parameters that significantly influenced titan cell generation , as measured by cell size distribution and proportion of titan cells . The first parameter we tested was the growth medium and transition between different growth media ( Fig 4A ) . Initial culture in YPD ( step 2 ) then transfer to MM ( step 3 ) resulted in the highest median cell size at 9 . 1 μm [6 . 9–11 . 1] . Initial culture MM followed by transfer to MM produced fewer titan cells , 25 . 5% ( 118/463 ) vs 39 . 5% ( 182/461 ) , although the titan cells tended to be larger with cell body diameters over 20 μm . Light exposure during step 3 also increased both median cell size and proportion of titan cells ( Fig 4B ) . In the light , median cell size was 9 . 4 μm [7 . 3–11 . 4] vs 8 . 4 μm [7 . 1–9 . 9] in the dark , with a titan cell proportion of 41 . 6% ( 983/2362 ) vs 24 . 3% ( 1072/4399 ) , respectively ( Fig 4B ) . Incubation temperature at 30°C at step 3 increased cell size distribution compared to 37°C ( 9 . 4 μm [7 . 3–11 . 4] vs 7 . 0 μm [6 . 2–8 . 2] ) as well as the proportion of titan cells ( 41 . 6% ( 983/2362 ) vs 8 . 1% ( 297/3652 ) ) ( Fig 4C ) . The pH of the minimal medium at step 3 also influenced cell size ( Fig 4D ) , with pH = 5 . 5 producing significantly larger cells and proportion of titan cells ( 9 . 1 μm [6 . 9–11 . 2] and 38 . 6% titan cells ) , compared to either lower pH ( pH = 4: median 5 . 1 μm [4 . 4–5 . 8] ( 0% ) ) or higher pH ( pH = 7: 8 . 2 μm [7 . 2–9 . 4] ( 16 . 4% ) , or pH = 8 . 5: 6 . 9μm [5 . 9–7 . 9] ( 0 . 7% ) ) ( Fig 4D ) . Finally , hypoxia at step 3 also increased median cell size compared to normoxia ( 7 . 5 μm [5 . 9–9 . 7] ) , with chemically induced hypoxia yielding higher median cell sizes compared to physically induced hypoxia] ( 10 . 1 μm [7 . 8–12 . 5] vs 8 . 9 μm [7 . 3–10 . 9 , p<0 . 0001 ) ( Fig 4E ) . The proportion of titan cells in normoxia ( 14 . 5% ( 732/5050 ) was lower than in chemically induced hypoxia or physically induced hypoxia ( 63 . 0% ( 732/1161 ) and 38 . 6% ( 1264/3004 ) , respectively ) ( p<0 . 0001 ) . We then tested hosts factors that could interact in vivo with yeast cells in the lung such as anticapsular antibodies , serum and phosphatidylcholine . Both serum and phosphatidylcholine have already been implicated in titan cell formation [13 , 18 , 31] . Co-incubation at step 3 with monoclonal antibodies that bind to different epitopes of the capsule inhibited titan cell generation , with a decreased cell size of 7 . 2 μm [6 . 1–8 . 3] for E1 mAb , and 6 . 8 μm [5 . 9–7 . 7] for 18B7 mAb compared to the untreated control ( 8 . 9 μm [7 . 1–10 . 6] ) ( Fig 5A ) , and a significantly smaller proportion of titan cells ( 3 . 2% ( 44/1360 ) with 18B7 , 5 . 3% ( 67/1273 ) with E1 compared to 33 . 1% ( 327/987 ) for the untreated control ) . The addition of fetal calf serum ( FCS ) significantly decreased median cell size ( 7 . 3 μm [6 . 4–8 . 1] vs ( 9 . 1 μm [7 . 1–11 . 1] , Fig 5B ) and the proportion of titan cells ( 2 . 5% ( 99/3911 ) vs 38 . 2% ( 1234/3228 ) , p<0 . 0001 ) compared to control , as did the addition of phosphatidylcholine ( PC ) ( 8 . 0 μm [7 . 0–9 . 0] vs 9 . 0 μm [7 . 1–11 . 2] for the median cell size , ( Fig 5C ) and the proportion of titan cells ( 14 . 7% ( 344/2340 ) vs 38 . 2% ( 1077/2820 ) , p<0 . 0001 ) . To understand if an alteration in yeast metabolism induced by ergosterol , protein or nucleic acids inhibition affected titan cell formation , we tested the effect of co-incubation of fluconazole ( inhibitor of ergosterol synthesis ) and flucytosine ( inhibitor of nucleic acids formation and transcription ) and cycloheximide ( translation inhibitor ) at step 3 . Fluconazole ( FLC ) exposure resulted in significantly smaller median cell sizes compared to the drug-free control ( 7 . 1 μm [6 . 2–8 . 3] ) at 1 mg/L , 6 . 8 μm [5 . 9–7 . 8] at 2 mg/L , and 6 . 5 μm [5 . 6–7 . 2] at 4 mg/L vs 9 . 4 μm [7 . 5–11 . 3] in the control , Fig 5D ) and a significantly smaller proportion of titan cells ( 5 . 9% ( 124/2073 ) at 1 mg/L , 2 . 9% ( 55/1919 ) at 2 mg/L and 1 . 0% ( 19/1877 ) at 4 mg/L vs 40 . 9% ( 877/2146 ) in the control , p<0 . 0001 , Fig 5D ) . Flucytosine exposure significantly decreased the mean yeast cell size at all concentrations tested ( 6 . 5 μm [5 . 9–6 . 9] at 1 mg/L , 6 . 5 μm [6 . 1–6 . 9] at 2 . 5 mg/L , and 6 . 4 μm [5 . 9–6 . 8] at 5 mg/L compared to control ( 8 . 3 μm [6 . 8–10 . 3] , Fig 5E ) with no titan cells observed upon flucytosine exposure . Cycloheximide exposure at 0 . 1 μg/mL also significantly decreased the mean cell size from 9 . 0 μm [6 . 3–12 . 7] to 6 . 1 μm [5 . 5–6 . 9] ( p<0 . 0001 ) and the proportion of titan cells from 43 . 3% ( 797/1840 ) to 0 . 1% ( 20/1663 ) ( Fig 5F ) . Of note , the viability of the cells recovered at step 4 after 5 d of drug exposure was unchanged for fluconazole but reduced for flucytosine and cycloheximide ( p<0 . 0001 compared to unexposed , S6 Fig ) . We also tested if iterative subcultures with or without the presence of active molecules ( CFW or fluconazole ) affected titan cell formation , assuming that the cell wall and the global metabolism of the sub-cultured progeny would be impaired in the presence of high concentrations of the cell wall toxic drug ( CFW ) or fluconazole , respectively . We analyzed the impact of repeated sub-culture of the cells prior to step 1 on titan cell production ( S7 Fig ) . Sub-culture on Sabouraud agar eight times ( 8 Sub ) spanning a one-month period significantly decreased the median cell size ( 8 . 60 μm [7 . 02–10 . 07] ) compared to the initial culture ( 0 Sub ) ( 9 . 17 [6 . 99–10 . 90] ) ( p<0 . 0001 ) . Addition of CFW to induce cell wall stress during sub-culture ( 8Sub+CFW ) significantly decreased the median cell size ( 8 . 03 μm [6 . 82–9 . 46] ) compared to the 8Sub control . Thus , iterative exposure to fluconazole during sub-culture significantly increased the median yeast cell size to 10 . 15 μm [8 . 04–13 . 23] ( 8Sub+FLC ) compared to the 8Sub control ( p<0 . 001 , S7 Fig ) . Previous studies in a murine pulmonary infection model showed that inoculum concentration can impact titan cell production [12 , 13] . To explore this phenomenon further , we examined cell size changes in response to different initial concentrations of cells at step 3 ( Fig 6A ) . Initial cell concentrations significantly impacted the median cell size of the yeast population ( p<0 . 0001 ) , with the highest median cell size observed at 106 cells/mL ( 9 . 2 [7 . 3–11 . 1] ) compared to 105 cells/mL ( 6 . 3 [5 . 1–8 . 3] ) , 104 cells/mL ( 6 . 2 [5 . 1–7 . 9] ) and 107 cells/mL ( 5 . 9 [5 . 2–6 . 5] ) ( Fig 6A ) . Similarly , the proportion of titan cells was significantly higher at 106 cells/mL ( 37 . 6% ( 896/2382 ) compared to 10 . 3% ( 279/2716 ) at 104 cells/mL , 14 . 3% ( 346/2420 ) at 105 cells/mL and 0% at 107 cells/mL ( 0/2177 ) , p<0 . 0001 . Previous study reported that pantothenic acid ( PA vitamin B5 ) is involved in quorum sensing and growth rate in C . neoformans [32] . The addition of PA had no effect on median cell size ( 8 . 35 μm [6 . 9–10 . 2] and 8 . 3 μm [6 . 4–10 . 8] , p = 0 . 8011 , Fig 6B ) , but significantly increased the proportion of titan cells ( 37 . 9% ( 1435/3785 ) vs 26 . 9% ( 983/3650 ) ) . In specifically implemented experimental settings , the proportion of titan cells was influenced by the concentration of PA with a significant increase in titan cells at 0 . 125 μM ( 56 . 5 [50 . 6–61 . 1] and 12 . 5 μM ( 47 . 6 [35 . 6–50 . 3] ) ( Fig 6D ) . In parallel , analysis of the growth curves of the yeast showed a significant increase in the doubling time ( slope ) at ≥0 . 125 μM of PA ( Fig 6E ) , suggesting a lack of correlation between titan cell formation and growth rate because titan cell formation was completely inhibited at 1250 μM of PA while the doubling time increased . Recent studies in C . neoformans also implicate the role of the small Qsp1 peptide in quorum sensing [33] . Addition of Qsp1 peptide significantly decreased median cell size from 9 . 1 μm [7 . 1–11 . 2] to 8 . 5 μm [6 . 9–10 . 1] ( Fig 6C ) and titan cell proportion from 38 . 4% ( 1075/2798 ) to 26 . 6% ( 915/3439 ) , p<0 . 0001 . Addition of Qsp1 peptide inhibited the formation of titan cells in H99O ( Fig 6C ) and KN99α ( Fig 6F ) . In the qsp1Δ , pqp1Δ and opt1Δ deletion mutants that cannot produce or import a functional Qsp1 peptide [33] , titan cell generation was increased compared to KN99α , confirming the negative regulation of Qsp1 peptide in titan cell formation ( Fig 6F ) . When qsp1Δ , pqp1Δ were complemented with Qsp1 but not with scrambled Qsp1 peptides , titan cell formation was similar ( increased titan cell formation ) to that of the mutant alone . The complementation of the opt1Δ deletion mutant with Qsp1 or scrambled Qsp1 did not rescue the parental phenotype suggesting that the import of Qsp1 is crucial for its action on the yeast cells ( Fig 6F ) . Previous whole genome sequencing studies identified single-nucleotide polymorphisms ( SNPs ) and insertions/deletions ( indels ) between H99-derived strains recovered from various laboratories ( Table 1 ) [34] . To determine whether any of these SNPs or indels affected titan cell generation , we tested the H99S , H99W , H99 CMO18 , H99L , KN99α strains . H99O produced significantly more titan cells than the other H99-derived strains , p<0 . 0001 ( Fig 7A , S8A and S9 Figs ) . These H99 derivative strains were also tested for titan cell formation in the lungs of infected mice ( S8A and S9 Figs ) . As with in vitro titan cell production , all the H99 derivative strains showed lower levels of titan cell formation in vivo when compared to H99O ( p<0 . 0001 ) , with the exception of KN99α that had equivalent titan cell production to H99O ( Fig 7A , S8A and S9 Figs ) . Two genes , LMP1 and SGF29 , are dramatically affected by SNPs/indels in the H99 derivatives; LMP1 has a frameshift deletion ( H99W and H99 CMO18 ) and SGF29 is deleted ( KN99α and H99L ) [35] . To determine if these genes are involved in titan cell production , we analyzed lmp1Δ and sgf29Δ deletion mutants for in vitro and in vivo titan cell formation ( Fig 7 and S8 and S9 Figs , respectively ) . In vitro , the sgf29Δ mutant in the H99O background had half the titan cell formation of the H99O wild-type strain [8 . 1% ( 49/600 ) to 4 . 2% ( 25/600 ) , p<0 . 0001] . The sgf29Δ mutant in the hypervirulent H99S also manifested no titan cells generation , as did an lmp1Δ mutant in this background . Complementation of LMP1 and SGF29 in this H99S mutant restored titan cells generation to that found in the parental strain; 1 . 2% ( 13/935 ) and 1 . 4% ( 7/600 ) , respectively vs H99S 1 . 6% ( 25/1540 ) ( Fig 7B ) . Importantly , the same trend was observed for in vivo titan cell formation . In vivo , the lmp1Δ H99S mutant produced only 3 . 5% ( 21/600 ) titan cells compared to 14% for H99S ( 84/600 ) , p<0 . 0001 , and this decrease in titan cell production was restored in the lmp1Δ:LMP1 H99S strains ( 9 . 5% ( 57/600 ) ) ( S8B Fig ) . The sgf29Δ mutant in the H99O background reduced titan cell formation in vivo from 18 . 8% ( 113/600 ) to 9 . 3% ( 37/400 ) ( p<0 . 0001 ) . In H99S , complementation of Sgf29 ( sgf29Δ:SGF29 ) in H99S restored titan cell generation to wild-type H99S levels from 5% ( 30/600 ) to 19 . 8% ( 237/1200 ) ( Fig 7B , S8B Fig ) ( p<0 . 0001 ) . SREBP is a gene involved in response to hypoxia , so we tested titan cell formation in the sre1Δ mutant . The proportion of titan cells was significantly decreased in the sre1Δ mutant at 5 . 1% ( 53/920 ) compared to KN99α [14% ( 337/2358 ) ] ( p<0 . 0001 ) ( S10 Fig ) . The signal transduction pathway Gpr/PKA/Rim101 regulates titan cell formation in vivo [18] . Briefly , the G-protein coupled receptor 5 and Ste3a pheromone receptor signal through Gpa1 to trigger the cAMP/PKA signaling cascade , ultimately activating the Rim101 transcription factor . This pathway regulates virulence factors such as capsule or melanin [36 , 37] . To determine if this same pathway was critical for titan cell generation in vitro , we examined cell enlargement in the gpr4Δ , gpr5Δ , gpr4Δ/gpr5Δ , rim101Δ , and cac1Δ mutants and their complemented strains in both the H99 and KN99α genetic backgrounds ( Fig 7C and 7D , S9 Fig ) . In the H99O genetic background , Rim101 function was similar to that observed in vivo , with little titan cell formation in the rim101Δ mutant ( 1 . 9% ( 51/2600 ) ) and full restoration of titan cell production in the complemented strain ( 9 . 2% ( 239/2600 ) ( p<0 . 0001 ) ( Fig 7C ) . In KN99α , the rim101Δ , gpr4Δ/gpr5Δ , and cac1Δ mutants had no titan cell formation , but surprisingly both of the single gpr4Δ and gpr5Δ mutants also lacked titan cell formation ( Fig 7D ) . This is in contrast to in vivo where titan cell production was rescued by GPR5 alone [18] . Taken together , these data suggest that signaling through both Gpr4 and Gpr5 via the cAMP/PKA pathway to Rim101 is required for titan cell production in vitro . To determine whether the Gpr/PKA/Rim101 pathway can impact titan cell formation in clinical isolates , we also screened a total of 56 clinical isolates for their ability to produce titan cells . Two isolates ( AD2-06a and AD2-02a ) produced a more titan cells relative to H99O ( ratio of clinical strain/H99O of 2 . 6±0 . 3 and 1 . 4±0 . 3 , respectively ) . Three additional isolates ( AD4-37a , AD1-95a , AD4-43a ) produced fewer titan cells than H99O ( ratio = 0 . 4±0 . 3 , 0 . 2±0 . 1 , 0 . 1±0 . 0 , respectively ) . Titan cell production in the other clinical isolates was close to zero ( ratio between 0 . 1 and 0 . 01% for five , less than 0 . 01% for six , and no titan cells at all for the remaining 39 isolates ) ( Fig 8A ) . The complete genome sequence was obtained for 41 of the 68 screened clinical isolates and a phylogenetic tree of these strains and the H99O reference strain shows high genetic diversity including VNI , VNII , VNBII isolates ( Fig 8B , Table 2 ) . Compared to H99O , the high titan cells generating strain AD2-06a harbored 31 , 229 SNPs and was closely related to AD3-55a ( 31 , 171 SNPs ) and AD3-41a ( 28 , 599 SNPs ) , which were both unable to produce titan cells . A total of 19 genes , including CNAG_00570 ( PKR1 ) , were disrupted in AD2-06a and not in AD3-55a and AD3-41a ( S1 Table ) . Of note , no common genetic mutation , insertion or deletion was observed in the 3 strains that produced a significant ( >1% ) proportion of titan cells ( H99O , AD2-06a , AD1-95a ) as compared to all the other examined isolates . Duplication of chromosomal regions were observed in the French sequenced clinical isolates ( S2 Table ) with AD2-06a harboring a large duplication of chromosome 9 ( Chr9 , region 465–665kb ) . To assess whether the Chr9 duplicated region could be responsible for titan cell formation , we explored a larger collection of C . neoformans strains with complete genome sequences [38] and discovered five additional clinical isolates ( Ug2459 , CCTP20 , FFV14 , WM-148 , WM-626 ) harboring partial duplications on chr9 duplication ( S11 Fig ) . We analyzed titan cell formation in these 5 isolates , but only the Ug2459 strain generated titan cells in vitro and only at a low proportion ( S11 Fig ) , suggesting that genes located within in the AD2-06a Chr9 duplication are not involved by themselves in titan cell formation . AD2-06a was isolated from the initial cerebrospinal fluid ( CSF ) sample of an HIV-infected patient at baseline ( diagnosis of cryptococcosis ) . Another isolate , AD2-07 , was recovered from the CSF of the same patient after 13 d of amphotericin B treatment . Thus , AD2-06a and AD2-07 are closely related with only 137 different SNPs and 40 indels . AD3-55a is in the same clade and also very similar with 370 different SNPs and 64 indels , although recovered from another patient and in another place ( Fig 8B and 8C ) . By contrast , these three isolates differ from H99O by 29 , 000 SNP and 3 , 386 indels . AD2-07 and AD3-55a were both unable to produce titan cells ( Fig 8A , 8B and 8E ) , allowing a more fine-scale analysis of SNPs linked to titan cell formation in these closely related strains ( Fig 8C ) . Specifically , AD2-07 produced 1 . 0% ( 31/3047 ) titan cells compared to 39 . 1% , ( 639/1633 ) in AD2-06a and 15 . 2% ( 304/2001 ) in H99O ( p<0 . 0001 ) . The median cell size of AD2-07 was significantly decreased compared to AD2-06a and H99O ( 5 . 9 μm [5 . 2–6 . 6] for AD2-07 , 8 . 5 μm [7 . 0–13 . 0] for AD2-06a , and 7 . 7 μm [6 . 5–9 . 2] for H99O , p<0 . 0001 , S12 Fig ) . Comparison of the AD2-06a , AD2-07 , and AD3-55a genomes identified four genes with loss-of-function mutations in AD2-06a but not in AD2-07 or AD3-55a: CNAG_00570 ( PKR1 ) ( Fig 8D ) , CNAG_07475 ( hypothetical protein ) , CNAG_01240 ( hypothetical protein ) and CNAG_05335 ( hypothetical protein ) . More precisely , AD2-06a had a frameshift mutation at glycine 125 in the CNAG_00570 ( PKR1 ) leading to a truncated protein of 138 amino acids ( Fig 8D ) . Pkr1 is the cAMP-dependent protein kinase regulatory subunit that interacts with Pka to regulate the phosphorylation activity of Pka . To further explore PKR1 in clinical isolates , we analyzed additional , previously sequenced , clinical isolates ( S3 Table ) that harbored mutations leading to Pkr1 truncation ( Bt156 , Bt58 , 8–1 , Bt77 , Bt117 , Ug2462 ) for titan cell formation [39] . Specifically , a frameshift mutation at amino acid 14 introducing a premature stop codon at position 96 for Bt156 , as well as stop codons introduced at positions 130 for AD2-06a , 258 for Bt58 , 302 for 8–1 , 439 for Bt77 , 441 for Bt117 , and 445 for Ug2462 were observed ( Fig 8D , Table 3 ) . We hypothesized that the strains with highly impacted/truncated Pkr1 protein would produce more titan cells , similar to the AD2-06a isolate . AD2-06a and Bt156 , the strains with the largest truncation , had high levels of titan cell formation with a ratio of 2 . 8±1 . 0 for AD2-06a and 1 . 3±0 . 6 for Bt156 compared to H99O ( Fig 8E ) , and titan cell proportions and median cell sizes of 39 . 1% ( 639/1633 ) ( median 8 . 5μm [7 . 0–13 . 0] ) and 17 . 9% ( 469/2614 ) ( median 7 . 7μm [6 . 4–9 . 4] ) , respectively ( S12 Fig ) . For the other strains , the ratio , proportion of titan cells , and median size were decreased compared to H99O: 0 . 2±0 . 0 , 3 . 6% ( 89/2464 ) and 6 . 9 μm [6 . 0–7 . 8] for 8–1 strain; 0 . 2±0 . 1 , 2 . 9% ( 110/3834 ) and 6 . 5 μm [5 . 8–7 . 1] for Bt77; 0 . 1±0 . 0 0 . 9% ( 27/3081 ) and 6 . 4 μm [5 . 7–7 . 2] for Bt117; and 0 . 1±0 . 1 , 1 . 5% ( 54/3628 ) 5 . 4μm [4 . 7–6 . 1] for Ug2462 , p<0 . 0001 ) ( Fig 8E , S12 Fig ) . To directly test whether the truncated Pkr1 protein impacted titan cell production in strain AD2-06a , the functional KN99α allele of the PKR1 gene was introduced into the strain ( AD2-06a:PKR1 ) and titan cell formation analyzed . Titan cell production was significantly decreased by complementation ( p<0 . 0001 ) with 44 . 9% ( 574/1279 ) for AD2-06a:PKR1 as compared to 64 . 6% ( 700/1083 ) for AD2-06a and 23 . 2% ( 555/2363 ) for H99O ( Fig 8F ) . To further explore the function of PKR1 in titan cell formation , we also tested the ability of pkr1Δ in a KN99α background to generate titan cells compared to KN99α , and found that pkr1Δ produced more titan cells with a ratio of 4 . 9±1 . 6 compared to the parental strain KN99α ( Fig 9 ) . The pkr1Δ median cell size ( 8 . 1 μm [6 . 7–9 . 5] ) exceeded that of KN99α ( 6 . 4 μm [5 . 5–7 . 2] ) p<0 . 0001 ) with a significant increase in the proportion of titan cells at 28 . 5% ( 695/2436 ) for pkr1Δ vs 4 . 6% ( 121/2614 ) for KN99α ( p<0 . 0001 , Fig 9A ) . We also analyzed titan cell production in two additional independent pkr1Δ and complemented pkr1Δ:PKR1 strain in the H99 background . The pkr1Δ-1 and complemented pkr1Δ:PKR1-1 gave ratios of 2 . 9±1 . 5 and 1 . 8±0 . 7 , respectively , while the pkr1Δ-2 and complemented pkr1Δ:PKR1-2 ratio were 1 . 9±0 . 8 and 1 . 4±0 . 4 , respectively , compared to the H99 parental strains ( 1 . 0±0 . 5 ) ( Fig 9B ) . In both strains , complementation significantly reduced the proportion of titan cells generated ( p<0 . 0001 ) from 29 . 8% ( 685/2493 ) for pkr1Δ-1 to 18 . 8% ( 593/3217 ) for pkr1Δ:PKR1-1; and from 19 . 9% ( 422/2364 ) for pkr1Δ-2 to 13 . 9% ( 359/2588 ) for pkr1Δ:PKR1-2 with H99 at 10 . 3% ( 357/3674 ) . We also tested the role of PKA1 and PKR1 using the galactose-inducible and glucose-repressible versions of PKA1 and PKR1 mutants [6] . In these mutants , when incubated in galactose minimal medium ( Fig 9C ) , the genes are turned on whereas when incubated in glucose minimal medium the genes are turned off ( Fig 9D ) . In galactose minimal medium , PGAL7::PKA1 and PGAL7::PKR1 had titan cell production ratios of 5 . 6±1 . 1 and 0 . 6±0 . 5 compared to H99 , respectively . The proportion of titan cells was significantly increased upon PKA1 induction [46 . 9% ( 794/1724 ) ] and reduced by PKR1 induction [4 . 7% ( 222/5031 ) ] compared with H99 in galactose minimal medium at 14 . 8% ( 402/3108 ) ( p<0 . 0001 , Fig 9C ) . In glucose minimal medium , PGAL7::PKA1 and PGAL7::PKR1 had titan cell production ratios of 0 . 0±0 . 0 and 2 . 4±1 . 0 , respectively . The proportion of titan cells was significantly decreased upon PKA1 repression [0 . 2% ( 6/4156 ) ] and increased by PKR1 repression [25 . 4% ( 527/2425 ) ] compared with H99 at 10 . 3% ( 357/3674 ) ( p<0 . 0001 , Fig 9D ) . At least 2 other genes ( TSP2 , USV101 ) have been linked to titan cells formation in vivo although their role is less clear , so we directly explored their phenotypes in our in vitro protocol . To examine the role of tetraspanin 2 ( Tsp2 ) in titan cell formation in vitro , we analyzed three independent tetraspanin 2 deletion mutants ( tsp2Δ-1 , tsp2Δ-2 and tsp2Δ-3 ) and two complemented mutants ( tsp2Δ-1:TSP2-1 and tsp2Δ-1:TSP2-2 ) for their titan cell formation compared to the wild-type KN99α . Titan cell production was significantly increased in the deletion mutants ( 23 . 9% ( 518/2161 ) for tsp2Δ-1 , 55 . 7% ( 990/1778 ) for tsp2Δ-2 , 44 . 3% ( 714/1610 ) tsp2Δ-3 , 4 . 6% ( 119/2576 ) ) compared to the wild-type and complemented strains ( tsp2Δ-1:TSP2-1 , 5 . 8% ( 228/3877 ) for tsp2Δ-1:TSP2-2 and 4 . 9% ( 173/3472 ) for KN99α ) ( p<0 . 0001 ) ( Fig 9E ) . Similarly , usv101Δ median cell size ( 10 . 6 μm [8 . 7–12 . 7] ) was higher than the parental strain KN99α ( 7 . 6μm [6 . 6–8 . 8] ) and the proportion of titan cells was 62 . 7% ( 1260/2008 ) for usv101Δ vs 10 . 5% ( 238/2262 ) for KN99α ( p<0 . 0001 ) ( S9 Fig ) . We finally selected additional sequenced clinical isolates that harbored mutations leading to Usv101 or Cac1 truncation . Bt88 showed a truncation of Usv101 and an increased titan cells formation ratio of 0 . 6±0 . 4 whereas the other isolates belonging to the VNBII lineage harboring a CAC1 mutation ( Bt133 and Bt31 , Bt40 , Bt89 and Bt105 ) did not show increased titan cells formation ( ratio 0 . 0±0 . 0 ) ( S13 Fig ) .
We identified and validated a new protocol allowing robust generation of titan cells in vitro . This protocol was discovered serendipitously while testing conditions that could induce dormancy in C . neoformans [4] . We observed yeast cell enlargement under defined growth conditions , then optimized those conditions for titan cell production . It is important to note that the utility of other published protocols to generate titan cells in vitro are hindered by issues with inter-laboratory reproducibility [12 , 31] . To establish the inter-laboratory transferability of our protocol , we independently tested it in two other laboratories ( K . Nielsen and A . Casadevall ) and observed that the protocol produced similar results in all laboratories , although slight variations of the materials and equipment used produced subtle variability in the percentage of titan cells generated in the 3 labs . Interestingly , titan cell production in vitro was also optimized by Ballou et al . 2017 and Zaragoza et al . 2017 , using a different set of growth conditions [40 , 41] . Exploration of the similarities and differences between these protocols will likely identify the critical environmental conditions that trigger titan cell production in vivo . Titan cells obtained in vitro exhibited many of the characteristics of in vivo titan cells recovered from the lungs of infected mice [17] . Similar to previous work on in vivo titan cells , we defined the in vitro titan cells as having a cell body size > 10 μm and typical cells with a cell body size ≤ 10 μm [12 , 13 , 18 , 21] . Titan cells generated with our in vitro protocol were also polyploids , as previously shown in vivo [12 , 13] . Melanization was increased in in vitro titan compared to typical cells . Capsule size was slightly increased in the in vitro titan cells compared to typical cells , but this difference was lower than previously shown in vivo [12 , 13] . We also demonstrated that , regardless of the capsule size differences between in vitro and in vivo titan cells , the structure of the in vitro titan cell capsule was different to that in typical cells , a phenomenon also observed in vivo [13 , 22] . The cell wall was thicker in in vitro titan cells compared to typical cells , as previously analyzed in vivo [22 , 31 , 42] . The increased chitin in the titan cell wall results in a detrimental immune response that exacerbates disease[42] . These findings suggest a fundamental difference in titan and typical surface structure that may contribute to reduced titan cell phagocytosis [12 , 13 , 21] . These cell surface differences also underscore the complex regulation of these major virulence factors and shows intricate adaptation of C . neoformans to both in vitro and in vivo conditions . Titan cell generation in vitro allowed a detailed kinetic analysis that revealed titan cells are formed between 4 and 8 h . Using calcofluor white staining to follow cell fate and cell division [4 , 30] , we showed that titan cells were exclusively derived from cells present in the initial inoculum that evolved progressively toward the titan cell phenotype . In contrast , typical cells were a mixture of cells from the initial inoculum and new cell replication . Titan cell division produced typical sized haploid cells , as shown previously for in vivo titan cells [12 , 13] and confirmed in our study . We published in vivo data that validate this observation [4]: using yeasts recovered from the lung of mice at one week after inoculation of C . neoformans stained with calcofluor and multispectral imaging flow cytometry , we showed that the CalcofluorHigh population was associated with yeast cells harboring high cell size parameters compatible with titan cells [4] . In vitro , titan cell generation coincided with the appearance of a large vacuole in the yeast cells at 4 and 8 h of incubation . Recent evidence places vacuoles at the center of networks enabling nutrient resources to be degraded , sorted and redistributed [43] . As the vacuole volume occupies much of the total volume of the titan cell body , one can imagine that cell-cycle regulation could be impacted , ultimately leading to polyploidy [20] . The fact that our in vitro protocol consistently produced titan cells also allowed us to test factors that influenced their appearance . In terms of environmental factors , we showed that titan cell production was influenced by pre-culture medium , initial pH , light exposure , temperature , type of medium and hypoxia . A metabolic switch between YPD ( rich medium ) pre-culture and minimal medium ( poor medium ) incubation was a key factor to induce titan cell generation . This switch is a stress that induces many metabolic modifications and has been studied extensively in Saccharomyces cerevisiae [44] . Hypoxia is another stress factor encountered by human pathogenic fungi during infection [45] , and a strong signal for titan cell production in vitro . Oxygen levels in healthy human tissues are 20–70 mmHg ( 2 . 5–9% O2 ) , but can be less than 10 mmHg ( ∼1% O2 ) in hypoxic or inflamed tissues or inside granulomas [46] . We know from previous work on pulmonary aspergillosis that hypoxia has been observed in infected lungs of mice [47] . Titan cells have been reported in human pulmonary cryptococcosis and well-studied in murine pulmonary infection following inhalation [12–13 , 15–21] , although they are also observed in mouse lungs after intravenous inoculation of animals [4] . These observations lead us to hypothesize that low oxygen levels in the lungs could be a signal for titan cell formation . A major transcriptional regulator of the fungal hypoxia response is the sterol regulatory element-binding protein ( Srebp ) [48] . Deletion mutants of the SREBP gene ( sre1Δ ) in C . neoformans display defects in adaptation to hypoxia , ergosterol synthesis , susceptibility to triazole antifungal drugs and cause a reduction of virulence [48] . Importantly , the sre1Δ mutant also showed defects in titan cell production in vitro , highlighting the role of hypoxia in titan cell production . Interestingly , quorum sensing is also involved in titan cell production . Indeed , the initial concentration of yeasts in minimal medium dramatically impacted titan cell generation , with 106 cells/mL being the optimal cell concentration to generate titan cells in vitro . No titan cell formation was observed with a starting concentration of 107 cells/mL , likely due to rapid consumption of the nutrients preventing metabolic modifications needed to generate titan cells . Alternatively , addition of the quorum Qsp1 peptide [33 , 49] to wild type cultures ( already producing Qsp1 ) decreased titan cell production , suggesting that Qsp1 negatively regulates titan cell production . Cleavage and internalization of Qsp1 were critical because the qsp1Δ , pqp1Δ , opt1Δ mutants all showed increased titan cell production . Pantothenic acid ( vitamin B5 ) has also been implicated in quorum sensing in C . neoformans [32] . Addition of pantothenic acid dramatically increased titan cell formation at concentrations between 0 . 125 and 12 . 5 μM . These results demonstrate that intercellular communication is important for titan cell formation and that this sensing process involves cell-cell communication instead of simple nutrient sensing . Induction of titan cells due to the presence of host factors such as temperature , addition of lipids , presence of serum , and antibodies was also tested . The presence of E1 [50] and 18B7 [51] anti-capsular IgG antibodies decreased titan cell production . This decrease could be related to changes in yeast metabolism induced directly by anti-capsular antibodies , as shown previously [52] , and provides a new mechanism by which antibodies could alter the course of infection . One could imagine that specific anti-capsular antibodies may not reach cryptococcal cells in the alveolar space at sufficient concentration to impair enlargement . The presence of surfactant protein-D , considered as an opsonin in the lung , could impair antibody fixation [53] , thus inhibiting the inhibitory effect of IgG antibodies and allowing titan cell formation in the lung . Interestingly , addition of serum ( 5% FCS ) or L-α-phosphatidylcholine decreased titan cell production , which is different from other protocols for titan cells generation [40 , 41] . This difference may be because our protocol induces titan cells through parallel or independent pathways to those triggered by serum or lipids . Overall , these results imply the existence of numerous triggers for titan cell formation mediated through independent signaling pathways . How these pathways ultimately interact , both positively and negatively , to regulate titan cell production still needs to be explored . Titan cell production was inhibited by addition of fluconazole , flucytosine and cycloheximide—even at concentrations that or below the MIC of the drug . Fluconazole is known to inhibit the 14-alpha-demethylase ( Erg11 ) involved in ergosterol synthesis , leading to plasma membrane instability and accumulation of toxic precursors [54] . Flucytosine is a base analogue leading to inhibition of DNA replication and protein synthesis [55] . Cycloheximide is known to impact protein synthesis through inhibition of translation [27] . Thus , titan cell production likely involves an active process requiring protein and nucleic acid production , as well as plasma membrane integrity ( normal ergosterol quantity ) . Conversely , serial passage in the presence of fluconazole increased titan cell production , suggesting compensatory changes in response to fluconazole also impacted titan cell production . These data have profound implications for in vivo titan cell production , as prolonged drug therapy could prevent or enhance titan cell formation . In previous studies , exposure of titan cells to fluconazole selected for aneuploidy and drug resistance in the daughter cells [20] . In contrast , our studies show exposure to cell-wall stress , induced by serial passage on CFW agar , decreased titan cell production . In these sub-culture experiments , we did not investigate subsequent genomic or metabolic changes that arise under these stress conditions . Serial sub-culture could have induced genetic rearrangements ( aneuploidy , SNPs , indels ) or epigenetic variation that altered titan cell production . Our protocol is easy to implement for study of the molecular and genetic mechanisms underlying titan cell generation . Our in vitro assay allowed us to identify host , environmental and yeast factors that impact titan cell production . By taking advantage of strains harboring genetic differences and clinically relevant genetic truncations , we were able to assess genetic factors modulating titan cell production . However , these studies also highlight that variability in titan cell formation cannot be completely explained by the acquisition of genetic events , with H99-derivative strains showing diversity in titan cell production that does not fully correlate with genetic modifications . The observation that titan cell formation in KN99α differs in vitro ( lower than H99O ) and in vivo ( equivalent to H99O ) highlights this issue and suggests further strain adaptation that are yet to be characterized . We uncovered new genes involved as positive or negative regulators of titan cell production . Sgf29 , is a component of the SAGA complex that binds H3K4me2/3 and recruits histone deacetylases in S . cerevisae [56] . The LMP1 gene is known to be involved in virulence in a mouse model and in mating [34] . We showed here that both genes are positive regulators of titan cell formation , although their mechanism of action remains unclear . We also showed in vitro the critical role of the Gpr/PKA/Rim101 pathway in titan cell formation , previously characterized in vivo [18] . Gpr5 signals through Gpa1 to trigger the PKA pathway that activates the transcription factor Rim101 [18 , 36] . In addition , we identified three genes that are negative regulators of titan cell formation , including PKR1 ( known to act as a regulatory subunit in the PKA pathway [23] , TSP2 that encodes a glucose repressor of laccase in C . neoformans [57] , and USV101 that is a pleiotropic transcription factor in C . neoformans known to regulate capsule formation and pathogenesis [58] . In Saccharomyces cerevisiae , the Pka1/Pkr1 complex is a heterotetramer with 2 catalytic subunits and 2 regulatory units . This complex is dissociated in the presence of cAMP [59] . Moreover , the architecture of the functional domains of Pkr1 include one interaction domain/dimerization at amino acids 2 to 40 and two cAMP binding domains at amino acids 219 to 351 and 353 to 473 , based on INTERPRO data ( Fig 8D ) . Consequently , the cAMP binding domain on Pkr1 is critical for the dissociation of the PKA1/PKR1 complex . Analysis of differences in titan cell production , combined with complete genome sequencing , allowed us to identify naturally occurring mutations in the PKR1 gene that impact titan cell production . Both PKR1 mutations have a stop codon ( Gly125fs for AD2-06 and Asp14fs for Bt156 ) that reduces the protein length . Interestingly , AD2-06a was the incident clinical isolate and a recurrent isolate recovered after 13 days of amphotericin treatment ( AD2-07 ) did not harbor this PKR1 mutation . In addition , a PKR1 mutation leading to intron retention was found in a relapse isolate [60] and shown to be associated with less virulence than the incident isolate . Whether the virulence differences observed with the relapse isolates are linked to titan cell formation needs to be further investigated . We identified TSP2 as a negative regulator of titan cell formation based on deletion mutants and complemented strains . TSP2 is known to interact with the cAMP/PKA pathway—tsp2Δ mutant strains phenotype are reversed by the addition of cAMP [57] . These data suggest that TSP2 inhibits the cAMP pathway and reinforces the major role of cAMP in titan cell formation . No natural TSP2 mutants were observed in our collection of clinical isolates . Interestingly , all isolates from the VNBII have a mutation in the CAC1 gene leading to the functional defect of the Cac1 protein . Out of the six VNBII isolates ( Table 3 ) , only Bt88 that harbored an additional functional abolition of Usv101 was able to produce as much titan cells as H99O did . Therefore , in Bt88 , titan cell formation resulted in the equilibrium between the abolition of CAC1 ( positive regulator ) and USV101 ( negative regulator ) . Altogether , these results show proof of concept that our in vitro protocol can be used to identify and characterize genes required for titan cell production . Our preliminary analysis only identified a handful of genes involved in titan cell production , but it is likely that many more are involved in generation of this complex cell morphology . Our data provide new insights into the genesis of titan cells and the environmental , host and genetic factors that influence their production . Finally , our data show that this in vitro protocol can be used to reproducibly generate titan cells that have similar characteristics to titan cells generated in vivo . The conditions identified for titan cell formation provide a robust system that could be invaluable to dissect the molecular mechanisms that underlie titan cell formation and allow the identification of naturally occurring mutations that regulate titan cell formation . These studies will enhance our understanding of the impact and mechanisms of yeast morphological changes on pathobiology .
Mice ( purchased from Jackson Laboratories , Bar Harbor , ME ) were handled in accordance with guidelines defined by the University of Minnesota Animal Care and Use Committee ( IACUC ) , under approved protocol numbers 1010A91133 and 130830852 , and in accordance with the protocols approved by JHSPH IACUC protocol M015H134 . All animal experiments were performed in concordance with the Animal Welfare Act , United States federal law , and NIH guidelines . The strains and clinical isolates of C . neoformans used in the study are listed in S4 Table . The study was started with H99 strain called H99O that was kindly provided by J . Heitman ( Duke University , Durham , NC ) in the late 90’s . The reference strain KN99α and strains from the Madhani collection were provided from Kirsten Nielsen's lab and the Fungal Genetic Stock Center [61] , respectively . C . neoformans strains were grown in liquid Yeast Peptone Dextrose ( YPD , 1% yeast extract ( BD Difco , Le Pont de Claix , France ) 2% peptone ( BD Difco ) , 2% D-glucose ( Sigma , Saint Louis , Minnesota , USA ) ) and in minimal medium ( MM , 15mM D-glucose ( Sigma ) , 10 mM MgSO4 ( Sigma ) , 29 . 4mM KH2PO4 ( Sigma ) , 13mM Glycine ( Sigma ) , 3 . 0 μM Thiamine ( Sigma ) , [32] ) . Minimum inhibitory concentration ( MIC ) of H99O for fluconazole ( FLC ) and flucytosine ( 5FC ) ( both purchased from lsachim , Shimadzu Group Company , Illkirch-Graffenstaden , France ) were determined by the EUCAST method and were 8 and 4 mg/L , respectively . C . neoformans strain from stock cultures stored in 20% glycerol at -80°C was cultured on Sabouraud agar plate at room temperature ( step 1 ) . After 2 to 5 d of culture , approximately 107 cells were suspended in 10 mL YPD in a T25cm3 flask and cultured 22 h at 30°C , 150 rpm with lateral shaking until stationary phase ( final concentration = 2x108cells/mL ) ( step 2 ) . Then , one mL of the suspension was washed twice with MM . The cell concentration was adjusted to 106 cells/mL in MM and the suspension was incubated in a 1 . 5 mL tube ( Eppendorf ) with the cap closed , at 30°C , 800 rpm for 5 d using an Eppendorf Thermomixer® ( Hamburg , Germany ) ( step 3 ) . Cell size was determined as described below . Cells with body size >10 μm were considered as titan cells as described [12] . Results are expressed as median cell size [interquartile range , IQR] or as median [IQR] of the proportion of titan cells in a given condition for H99 or as a ratio compared to the proportion of titan cells obtained with the H99O in experiments involving other strains ( clinical isolates , other H99 strains and mutants ) . In specific experiments , 104 cells/mL were incubated in 100 well plate ( Fischer Scientific ) and incubated at 30°C with agitation in the Bioscreen apparatus ( Fischer Scientific ) . In specific experiment using PGAL7 inducible mutants in H99 , MM with galactose at 15mM ( galactose MM ) was used in parallel to MM containing glucose . Yeasts were observed after India ink staining and capsule thickness was determined as the size of the thickness in pixel of the white area surrounding the cell wall imaged with an Olympus AX 70 microscope and analyzed using the ImageJ software available at https://imagej . nih . gov/ij/ and the Multi_measures plugin . Multispectral flow cytometry was used to quantify chitin content after calcofluor white staining ( CFW , fluorescent brightener 28 , 0 . 0001 μg/mL CFW in PBS ) and capsule structure after immunostaining of three anti-capsular antibodies ( E1 IgG1 monoclonal antibody [50] , both 2D10 [29] and 13F1 [29] IgM monoclonal antibody 30 m at 10 μg/mL ) and then incubation with FITC coupled anti-IgG or -IgM secondary antibodies ( 15 m at 1:1000 concentration in PBS ) . The antibody 18B7 has not been used for this specific experiment because it produces aggregation that prevented ImagestreamX testing . Pictures were taken in flow and analyzed using various existing algorithms . We used ImageStreamX with the INSPIRE software ( Amnis Corporation ) . Cell suspensions were adjusted to 107 in 200 μL and 10 , 000 cells were recorded at 40-fold magnification in 3 different channels including the bright field channel ( BF ) and 2 fluorescence channels ( channel 1: 430-505nm [CALCO]; channel 2: 470-560nm [Anticapsular antibodies] ) . Data analysis was performed using the IDEAS software ( Amnis Corporation ) after fluorescence compensation procedures . The first step consists in the definition of a mask that delineates the relevant pixels in each picture . Then , 54 algorithms ( calculations made for each event within a defined mask ) are available to analyze size , texture , location , shape or signal strength . Using basic algorithms , unfocused events , yeasts aggregates were excluded [4] . First titan cells ( TC ) and typical cells ( tC ) were selected based on a dot plot Area/Diameter . We decided to avoid overlap between populations and select well separated population based on their size after control using the bar added on the picture of the yeasts ( see Fig 2A and 2B ) . For chitin content , the calcofluor intensity histogram using the Intensity of algorithm in channel 01 have been generated for TC and tC ( see Fig 2C and 2D ) . For capsule structure , the algorithms dedicated to structure analysis were tested and Modulation and Bright details intensity R7 algorithms in channel 2 have been found to separate the capsule structure of titan cells from typical cells populations . For each population of interest , the geometric mean was calculated using the IDEAS software . Additional experiments using fluorescence microscopy for chitin content measurement was performed after calcofluor white staining ( CFW , fluorescent brightener 28 ) adapted from [42] . Briefly , 107 C . neoformans cells in 10 mL MM were washed once and 500 μL of 3 . 7% formaldehyde in PBS was added . Cells were incubated at room temperature for 30–40 m , inverting the tube every 5 m . Samples were washed twice in PBS , cell concentration was adjusted to 106 cell/mL . The supernatant was removed and 1 mL of 0 . 0001 μg/mL CFW in PBS was added , and incubated 5 m at 25°C . Cells were then washed twice in PBS . Results were expressed as median [IQR] of the mean fluorescence intensity/pixel/cell after picture analysis as described below . Capsule immunofluorescence ( IF ) of titan cells was done by incubating approximately 5x106 cells/mL with 10μg/mL of murine-derived monoclonal antibodies to the cryptococcal capsule ( IgG1 18B7 , IgG1 E1 , IgM 12A1 , IgM 2D10 [29 , 50] in blocking solution ( 1% bovine serum albumin in PBS ) . Cells and mAb mixtures were done in 1 . 5 mL microcentrifuge tubes at 37°C for 1 h under continuous mixing . Next , cells were washed three times with PBS by centrifugation ( 5 , 000 rpm for 5 m at room temperature ) and incubated for 1 h at 37°C with 5 μg/mL fluorescently labelled secondary-mAbs , goat anti-mouse IgG1-FITCs or IgM-TRITCs ( Southern Biotech ) in blocking solution and 1 μg/mL of Uvitex2b ( Polysciences , Warrington , PA ) solution to visualize the fungal cell wall . Cells were washed three times with PBS by centrifugation , mounted in glass coverslips and imaged with an Olympus AX 70 microscope equipped with blue , green and red fluorescent filters using 40x and/or oil immersion 100x objectives . Capsule immunofluorescence of titan cells preparations performed in two independent experiments gave consistent results . Titan cells melanization was induced following step 3 . Cells were washed once with minimal medium , suspended in 1mL of minimal medium supplemented with 1mM of L-DOPA ( Sigma D9628 ) , transferred to a 5mL Erlenmeyer flask ( for normal oxygenation ) and incubated at 30°C under continuous mixing at 200 rpm for 3 d . Since melanin is resistant to acid hydrolysis , a spherical melanin “ghost” remains following incubation of black cells with 12N HCl for 1 h at 100°C ( a reduced and modified version of the procedure in [62] . Acid-resistant melanin “ghosts” were washed three times in PBS by centrifugation and visualized using light microscopy . Melanization was measured using imageJ in Icy software by manually circling each cell and measuring the mean gray intensity / pixel /cell . Blackness was calculated as the maximum mean grey intensity minus the mean grey intensity / pixel /cell . Increasing melanin content will result in higher blackness . H99 cells were grown in MM for 48 h in vitro . Cells were centrifuged 2 m at 14 , 000 rpm and were then washed twice with sterile water . These cells were exposed to γ-radiation to remove layers of the capsule polysaccharide [13 , 20] . Cells resuspended in sterile water were transferred to a 24-well flat-bottom plate and irradiated for 45 m: dose 560 Gy ( 56 , 000 rad ) . Titan cells and tyical cells were separated [20] . Washed irradiated cells were filtered using CellMicroSieves ( BioDesign Inc . of New York , Carmel , NY ) with a 10 μm pore size . The CellMicroSieves were rinsed with PBS to remove typical cells from the filter . To recover the titan cells population , the CellMicroSieves were inverted and the membrane was washed with PBS . The TCs population was concentrated by centrifugation at 12 , 000 g for 1 m . To recover the typical cells population , the filter flow-through was concentrated by centrifugation at 12 , 000 g for 1 m . Cellular chitin quantification was adapted from [20 , 42 , 63] . Purified in vitro titan cells and typical cells were collected by centrifugation at 14 , 000 rpm for 2 m and the media were removed . Dry weights were measured following 2–3 d of evaporation at 37°C . Dried pellets were extracted with 1 mL 6% KOH at 80°C for 90 m . Samples were centrifuged at 14 , 000 rpm for 20 m . Each pellet was suspended in 1 mL PBS and spun again . Each pellet was suspended in 0 . 2 mL of McIlvaine’s Buffer ( 0 . 2 M Na2HPO4 , 0 . 1 M citric acid , pH 6 . 0 ) . Five μL of purified Streptomyces griseus chitinase ( 5 mg/mL in PBS ) was added to hydrolyze chitin to Gluc-NAc and incubated for 3 d at 37°C . Chitinase-treated samples were spun at 14 , 000 rpm for 1 m , each 10 μL of sample supernatant was combined with 10 μL 0 . 27 M sodium borate , pH 9 . 0 . Samples were heated to 99 . 9°C for 10 m . Upon cooling to room temperature , 100 μl of DMAB solution ( Ehrlich’s reagent , 10 g p-dimethylaminobenzaldehyde in 12 . 5 mL concentrated HCl , and 87 . 5 mL glacial acetic acid ) was added , followed by 20 m incubation at 37°C . Hundred μL was transferred to 96-well plates , and absorbance at 585 nm was recorded . Standard curves were prepared from stocks of 0 . 2 to 2 . 0 mM of Gluc-NAc ( Sigma , Saint Louis , Missouri , USA ) . The amount of Gluc-NAc was calculated as mmol/g cells ( dry weight ) . Results are expressed as median [IQR] . A 96-well microtiter plate was filled with 200 μL of a 106/mL cell suspension in PBS and centrifuged 5 m at 4000 rpm . The pellet was suspended in 150 μL of ethanol 70% and incubated in the dark 1 h at 4°C . After discarding the supernatant , a 50 μL mix composed of 44 μL NS ( 0 . 01M Tris HCL pH 7 . 2 , 1 mM EDTA , 1mM CaCl2 , 0 . 25 M Sucrose , 2 . 12 mM MgCl2 , 0 . 1 mM ZnCl2 ) , 5μL RNase A at 10mg/mL and 1 . 25μL PI at 0 . 5mg/mL was added in each well as described [64] . After a 30 m incubation at 30°C in the dark , the plate was sonicated 1 m and each sample diluted at 1:40 in 50mM Tris HCl . The fluorescence intensity was measured using the Guava easyCyte 12HT Benchtop Flow Cytometer ( Guava , MERCK , Kenilworth , New Jersey ) . Selection of singlets by gating allowed ( i ) determination of PI intensity on channel YelB ( 583/26 ) in FSC/SSChigh ( TC ) and FSC/SSClow ( tC ) ; ( ii ) determination of the FSC/SSC distribution in PIhigh and PIlow population . FlowJo software v . 10 was used to analyze the data . The graphs of the number of yeasts were normalized to the mode to depict the data in terms of '% of max' . The % of max denotes the number of cells in each bin ( the numerical ranges for the parameter on the x axis ) divided by the number of cells in the bin that contains the largest number of cells . Knowing that CFW staining does not alter C . neoformans viability and that daughter cells harbored lower CFW signal due to partial cell wall transmission from mother to daughter cells [4 , 30] , we analyzed cell size and CFW fluorescence intensity of the progenies of titan cells and typical cells following the in vitro protocol on 106 cells of H99O pre-stained with CFW ( channel BluV 448/50 using Guava ) . Budding rates were determined after yeasts ( 105 cells composed of titan cells and typical cells ) previously incubated using our protocol or in vivo ( see below ) were directly deposited in a 35 mm sterile culture dish in minimal medium without agitation and incubated at 30°C . Pictures were taken every 2 or 5 m by phase microscopy using the Axiovert 200M inverted microscope with 40x or 20x objectives ( Carl Zeiss MicroImaging , NY ) , used in conjunction with an AxiocamMR camera . Cell size evolution over time was assessed for strain AD2-06a by dynamic imaging ( Nikon Biostation ) . Briefly , 35 mm sterile culture dish ( Hi-Q4 , Nikon ) were coated for 1 m with E1 antibody at 2 mg/L in order to provide anchor for the capsule . Yeasts ( 105 cells ) were added in 1 mL MM and incubated at 32°C for 18 h . Series of 221 images were taken by phase-contrast microscopy every 5 m at ×100 magnification . Merging was done using ImageJ software in Icy Software . To analyze the various factors that could impact titan cells generation , we modified the various steps of our in vitro protocol . For step1 , stress was produced by 8 subcultures ( twice a week for one month ) on agar medium or on agar supplemented with CFW ( 20mg/L ) or with fluconazole ( 32mg/L ) . For step 2 , the pH of MM ( normally at 5 . 5 ) was set at 4 , 7 or 8 . 5 without buffering . For step 3 , initial cell concentration ( from 104 to 107cells/mL ) was tested . Hypoxia was generated physically by closing the cap of the Eppendorf® tube during 5 d or chemically upon incubation in MM supplemented with 1 nM CoCl2 , cap closed , as already described [65] . The production of titan cells was also assessed in the presence of various reagents added at step 3: ( 1 ) Qsp1 peptide ( NFGAPGGAYPW , [33] ) ( Biomatik , Cambridge , Canada ) was resuspended at 10mM in water and stored at -80°C until use at 10 μM final with the scrambled peptide ( AYAPWFGNPG ) as a control; ( 2 ) pantothenic acid purchased from Sigma ( Saint-Louis , Missouri , USA ) used at 125 μM; ( 3 ) monoclonal anti-capsular antibodies E1 [50] and 18B7 [51] used at a final concentration of 166 μg/mL in MM [66]; ( 4 ) decomplemented fetal calf serum ( FCS , Invitrogen , Carlsbad , CA , USA ) at 5% in MM; ( 5 ) L-α-Phosphatidylcholine from egg yolk ( Sigma , Saint-Louis , Missouri , USA ) was extemporaneously reconstituted at 5 mM in MM; ( 6 ) antifungal drugs ( fluconazole and flucytosine ) were tested at the concentrations close to the MIC ( 2-fold dilutions ) with the diluent ( DMSO or water ) as control . Results are expressed as median [IQR] . Growth in the presence or in the absence of antifungal drugs was evaluated by enumeration of yeast cells concentration at step 4 of our protocol using the Guava cytometer , starting from 106 cells inoculated at step 1 . C . neoformans strains were cultured overnight at 30°C in YPD broth medium ( BD , Hercules , Canada ) . Yeast cells were collected by centrifugation , washed with phosphate buffered saline PBS and resuspended in sterile saline . For titan cells analysis in vivo , groups of 6- to 8-week-old C57BL/6J mice ( Jackson Labs , Bar Harbor , Maine ) were anesthetized by 5% isofurane inhalation for 1–5 m , infected intranasally with 2 × 105 cells in a 40 μL volume and sacrificed at D6 . In these experiments , 84% of titan cells were obtained . For mutant screening , groups of 6- to 8-week-old C57BL/6J mice ( Jackson Labs , Bar Harbor , Maine ) were anesthetized by intraperitoneal pentobarbital injection and infected intranasally with 5 × 106 cells in a 50 μL volume . Infected mice were sacrificed by CO2 inhalation at 3 d post-infection . The lungs were harvested , homogenized , and then resuspended in 10 mL PBS supplemented with collagenase ( 1 mg/mL ) [20] . Cell homogenates were incubated for 1 hour at 37°C with agitation , and washed several times with double distilled water . The C . neoformans cells were fixed with 3 . 7% formaldehyde for 40 m , washed 3 times with sterile PBS , and then resuspended in sterile PBS . The proportion of titan cells and typical cells were determined by microscopy . Data presented were from 3 mice per strain , except for strains sgf29Δ in H99O , H99S that had 2 mice per strain . We PCR amplified the PKR1 and TSP2 genes using the primer KN99α DNA as substrate and the following primers ( PKR1F: AAGCTTggaatgaagatgaaattagtacgtg; PKR1R: ACTAGTgtccatcattgctgtaacttggttg; TSP2F: GAGCTCaactccgatgatcatggactcgg; TSP2R: GAGCTCtgcccaagagactagagtgtaacc ) . The 2559 bp TSP2 and the 2000 bp PKR1 amplicons were cloned in the pGEMT easy vector ( Clontech ) and sequenced . The pNE609 and pNE610 plasmids were then constructed by cloning the PKR1 and TSP2 DNA fragments into the pSDMA57 plasmid [67] using the SpeI/HindIII and SacI cloning sites , respectively . To create transformants , the plasmids pSDMA57 containing PKR1 amplicon was linearized with BaeI and biolistically transformed into AD2-06a and Bt156 clinical strains . To complement the tsp2Δ mutant , pSDMA57 plasmid containing TSP2 gene was linearized with BaeI and biolistically transformed into the tsp2Δ mutant strain . All transformants were selected on YPD supplemented with neomycin . Genomic DNA was purified from the transformants and PCR was used to check the presence of PKR1 and TSP2 genes in the transformed strains . PCR reactions contained 1 μl gDNA , 2 . 5 μl of each of the 10 mM primer stocks ( PKR1 forward , PKR1 reverse , TSP2 forward , TSP2 reverse ) 5 μl Taq buffer , 4 μl dNTPs , 0 . 25 μl ExTaq polymerase ( New England Biolabs , USA ) and 34 . 75 μl sterile water . The cycling parameters were 35 cycles of 94°C for 20 seconds , 54°C for 20 seconds and 72°C for 90 seconds . Products were visualized using electrophoresis with 0 . 8% TAE agarose gel . To differentiate between random integration , single insertion , and tandem insertion into the safe haven , we performed a similar PCR as above using primers UQ1768 , UQ2962 , UQ2963 , and UQ3348 as previously described [67] . Genomic DNA was adapted for Illumina sequencing using Nextera reagents . Libraries were sequenced on an Illumina HiSeq to generate 101 base reads . most data was previously described [39 , 68] and one additional isolate was newly sequenced for this study ( AD2-07 ) ( NCBI SRA accession SRR5989089 ) . Reads were aligned to the C . neoformans H99 assembly ( GenBank accession GCA_000149245 . 2 [34] using BWA-MEM version 0 . 7 . 12 [69] . Variants were then identified using GATK version 3 . 4 [70] , where indels were locally realigned , haplotypeCaller was invoked in GVCF mode with ploidy = 1 , and genotypeGVCFs was used to predict variants in each strain . All VCFs were then combined and sites were filtered using variantFiltration with QD < 2 . 0 , FS > 60 . 0 , and MQ < 40 . 0 . Individual genotypes were then filtered if the minimum genotype quality < 50 , percent alternate allele < 0 . 8 , or depth < 5 . Variants were then functionally annotated with SnpEff version 4 . 2 [71] . For phylogenetic analysis , the 535 , 968 sites with an unambiguous SNP in at least one isolate and with ambiguity in at most 10% of isolates were concatenated; insertions or deletions at these sites were treated as ambiguous to maintain the alignment . Phylogenetic trees were estimated using RAxML version 8 . 2 . 4 [72] under the GTRCAT model in rapid bootstrapping mode . For determination of Pkr1 architecture domains , the INTERPRO tool was used ( http://www . ebi . ac . uk/interpro/protein/J9VH50 ) . To increase the number of events analyzed in each condition tested/each parameter analyzed ( cell size , capsule size and chitin content ) , pictures of 3–5 fields were taken with an AxioCam MRm camera ( Carl Zeiss , Oberkochen ) at x40 on interferential contrast microscope ( DMLB2 microscope; Leica , Oberkochen ) . Image were then analyzed ( for cell size and chitin content ) using Icy software v . 1 . 9 . 2 . 1 . [73] ( icy . bioimageanalysis . org ) and a specific plugin ( HK-Means plugin ( http://icy . bioimageanalysis . org/plugin/HK-Means ) that allows analysis of multiple structures from a bright field . Preliminary experiments were done to compare results obtained with Icy to "manual" measurements by analyzing about 200 cells on the same pictures for 3 independent experiments . In subsequent experiments , results were pooled for a given condition from 2 to 3 independent experiments after good reproducibility was assessed . Statistical analysis was performed with STATA® software ( College Station , Texas , v13 . 0 ) . To validate the cell size determination using the Icy software , the intraclass correlation coefficient was calculated . The ability of the automated method to classify the C . neoformans cells as titan cells or typical cells compared to visual measurement was evaluated using the Kappa test [74] . To compare titan cells generation in the various conditions , non-parametric tests were performed using the Kruskal-Wallis test for multiple comparisons or Mann Whitney test as required . GraphPad Prism software ( v . 6 ) was used to generate graphs . All sequence data from this study have been submitted to NCBI BioProject ( https://www . ncbi . nlm . nih . gov/bioproject ) under accession number PRJNA174567 . The AD2-07 sequence is available in the NCBI SRA under the accession number SRR5989089 ( https://www . ncbi . nlm . nih . gov/sra/SRR5989089/ ) )
|
Cryptococcus neoformans is a yeast that is capable of morphological change upon interaction with the host . Particularly , in the lungs of infected mice , a subpopulation of yeast enlarges , producing cells up to 100 μm in cell body diameter–referred to as titan cells . Along with their large size , the titan cells have other unique characteristics such as thickened cell wall , dense capsule , polyploidization , large vacuole with peripheral nucleus and cellular organelles . The generation of a large number of such cells outside the lungs of mice has been described but was not reproducible nor standardized . Here we report standardized , reproducible , robust conditions for generation of titan cells and explored the environmental and genetic factors underlying the genesis of these cells . We showed that titan cells were generated upon stresses such as change in the incubation medium , nutrient deprivation , hypoxia and low pH . Using collections of well characterized reference strains and clinical isolates , we validated with our model that the cAMP/PKA/Rim101 pathway is a major genetic determinant of titan cell formation . This study opens the way for a more comprehensive picture of the ontology of morphological changes in Cryptococcus neoformans and its impact on pathobiology of this deadly pathogen .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"methods"
] |
[
"flow",
"cytometry",
"cell",
"physiology",
"chitin",
"cell",
"walls",
"cryptococcus",
"neoformans",
"medicine",
"and",
"health",
"sciences",
"cryptococcus",
"yeast",
"infections",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"microbiology",
"cell",
"formation",
"fungi",
"materials",
"science",
"fungal",
"diseases",
"macromolecules",
"fungal",
"pathogens",
"cellular",
"structures",
"and",
"organelles",
"materials",
"by",
"structure",
"research",
"and",
"analysis",
"methods",
"polymers",
"specimen",
"preparation",
"and",
"treatment",
"staining",
"mycology",
"polymer",
"chemistry",
"infectious",
"diseases",
"spectrum",
"analysis",
"techniques",
"medical",
"microbiology",
"microbial",
"pathogens",
"chemistry",
"spectrophotometry",
"yeast",
"cytophotometry",
"cell",
"staining",
"eukaryota",
"cell",
"biology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"organisms"
] |
2018
|
Titan cells formation in Cryptococcus neoformans is finely tuned by environmental conditions and modulated by positive and negative genetic regulators
|
Schistosomiasis affects more than 200 million individuals worldwide , with a further 650 million living at risk of infection , constituting a severe health problem in developing countries . Even though an effective treatment exists , it does not prevent re-infection , and the development of an effective vaccine still remains the most desirable means of control for this disease . Herein , we report the cloning and characterization of a S . mansoni Stomatin-like protein 2 ( SmStoLP-2 ) . In silico analysis predicts three putative sites for palmitoylation ( Cys11 , Cys61 and Cys330 ) , which could contribute to protein membrane association; and a putative mitochondrial targeting sequence , similar to that described for human Stomatin-like protein 2 ( HuSLP-2 ) . The protein was detected by Western blot with comparable levels in all stages across the parasite life cycle . Fractionation by differential centrifugation of schistosome tegument suggested that SmStoLP-2 displays a dual targeting to the tegument membranes and mitochondria; additionally , immunolocalization experiments confirm its localization in the tegument of the adult worms and , more importantly , in 7-day-old schistosomula . Analysis of the antibody isotype profile to rSmStoLP-2 in the sera of patients living in endemic areas for schistosomiasis revealed that IgG1 , IgG2 , IgG3 and IgA antibodies were predominant in sera of individuals resistant to reinfection as compared to those susceptible . Next , immunization of mice with rSmStoLP-2 engendered a 30%–32% reduction in adult worm burden . Protective immunity in mice was associated with specific anti-rSmStoLP-2 IgG1 and IgG2a antibodies and elevated production of IFN-γ and TNF-α , while no IL-4 production was detected , suggesting a Th1-predominant immune response . Data presented here demonstrate that SmStoLP-2 is a novel tegument protein located in the host-parasite interface . It is recognized by different subclasses of antibodies in patients resistant and susceptible to reinfection and , based on the data from murine studies , shows protective potential against schistosomiasis . These results indicate that SmStoLP-2 could be useful in a combination vaccine .
Schistosomiasis is an important parasitic disease , caused by trematode worms of the genus Schistosoma; it affects approximately 200 million of individuals primarily in developing countries and an estimated additional 500 to 600 million are at risk . The digenetic blood fluke , Schistosoma mansoni , is one of the major causative agents [1] . Parasite eggs are trapped in the liver and intestine , where they induce granuloma formation and fibrosis , the main cause of morbidity and mortality in schistosomiasis [2] . Chemotherapy is an important control strategy against this parasitic disease [3]; however , it has not reduced the endemicity [4] and rapid reinfection demands frequent treatment [5] . Therefore , it is considered that an effective vaccine combined with chemotherapy would be an efficient control mechanism [6] . Until October 2003 , schistosome research suffered from limited genomic information; this situation has changed significantly with the simultaneous publication of the S . mansoni and S . japonicum transcriptomes [7] , [8] . These initiatives , together with the advent of entire S . mansoni genome sequencing , all boosted by advances in bioinformatics , have markedly changed the schistosome vaccine research field . Simultaneously with the publication of the transcriptome data , and its scrutiny for genes with functions that would indicate their surface exposure to allow interaction with the host immune system , a series of novel genes were suggested as potential vaccine candidates based on their functional classification by Gene Ontology [8] . One of these , stomatin , was assigned a role in lipid raft formation or receptor binding by Gene Ontology categorization . Actually , it is most similar to the sub-family Stomatin Like Protein 2 ( SLP-2 ) , of which the best characterized gene is the human ortholog [9] . The protein was also proposed as a schistosome drug target [8] , since the human ortholog was described as interacting with anti-malarial drugs , participating in the transfer of the drug Mefloquine to the intracellular parasite via a pathway used for the uptake of exogenous phospholipids [10] . The SLP-2 was first identified in humans ( HuSLP-2 ) ; it presents , like other stomatins ( e . g . Stomatin , SLP-1 and SLP-3 ) , a central Stomatin , Prohibitin , Flotillin , HflK/C ( SPFH ) domain that may mediate interactions with plasma and mitochondrial membranes [11]–[13] . HuSLP-2 is the first member of this family that lacks an N-terminal hydrophobic domain , displaying a mitochondrial targeting sequence in this region . Additionally , a palmitoylation centered on Cys29 could not be identified [9] , [11] . The function of stomatins , including SLP-2 , remains undetermined . In erythrocytes , it may link stomatin or other integral membrane proteins to the peripheral cytoskeleton , playing a role in the regulation of ion channel conductance or in the organization of sphingolipids and cholesterol-rich lipid rafts [9] . More recently , this gene has been investigated as a novel cancer-related gene over-expressed in certain kinds of human tumours [14] , [15] , and in the assembly of mechanosensation receptors [16]–[22] . Moreover , it has been proposed to function as a link between synapse-polarized mitochondria and T-cell receptor ( TCR ) signalosomes , contributing to modulate TCR signalling and T cell activation [23] , [24] . In this work , we describe and characterize a novel S . mansoni stomatin like protein 2 ( SmStoLP-2 ) . Data obtained here establishes that SmStoLP-2 is present in the tegument of adult worms and schistosomula . In addition , we evaluated the reactivity of rSmStoLP-2 antigen against the sera from individuals living in endemic areas for schistosomiasis in Brazil , showing that the groups resistant and susceptible to reinfection showed different antibody profiles . We subsequently demonstrated the ability of anti-rSmStoLP-2 serum to inhibit penetration and migration of cercariae in vivo . Lastly , immunization of mice with rSmStoLP-2 induced a Th1-type of immune response and a significant reduction in worm burden upon challenge with cercariae .
Schistosoma mansoni adult worms ( BH strain ) were obtained by perfusion of mice , 7–8 weeks after infection . Eggs , miracidia , cercariae , and schistosomula were obtained as previously described [8] . Cercaria number and viability were determined using a light microscope prior to infection . This study was conducted according to the principles expressed in the Declaration of Helsinki . The study was approved by the Institutional Review Board of Fundação Oswaldo Cruz ( 0083/99-CEP/FIOCRUZ ) . All patients or their legal guardians provided written informed consent for the collection of samples and subsequent analysis . All animals were handled in strict accordance with good animal practice as defined by Animals Use Ethics Committee of UFMG ( Universidade Federal de Minas Gerais , Brazil ) and Instituto Butantan ( São Paulo , Brazil ) , and the study was conducted adhering to the institution's guidelines for animal husbandry . Peripheral blood was obtained from individuals with different genetic background living in three endemic areas for schistosomiasis ( ‘Melquiades’ , ‘Caatinga do Moura’ and ‘Côrrego do Onça’ , all in the state of Minas Gerais , Brazil ) . These individuals were classified in four groups according to their infection status and the selection of subjects was performed based only on the criteria for inclusion and exclusion of each group independent of previous knowledge of immune responses for each individual . Non-infected ( NI ) individuals are healthy people from non-endemic areas without any parasite infection or contact with contaminated water . One group was shown to be stool-negative after treatment with praziquantel and was classified as resistant to S . mansoni reinfection ( RR ) [25] . The water contact exposure was determined using previously described methods [26]–[28] , objectively evaluated by observers and studied population had at least one contact daily . Individuals classified as susceptible to S . mansoni reinfection ( SR ) were shown to be stool-positive following treatment with praziquantel ( 40 mg/kg ) ( at 1 and 5 months ) . The sera from RR and SR groups were obtained six months after praziquantel treatment and these individuals were examined for S . mansoni infections using the Kato–Katz technique before treatment and one , 6 and 12 months after treatment to check for reinfection rates [25] . Individuals grouped as infected ( INF ) showed stool-positive examination and no treatment history ( never received anti-helminthic treatment , as determined by survey ) . These infected patients had infection levels that varied from 48 to 224 epg ( egg counts per gram of feces ) . For each time point , three independent ( consecutive days ) stool samples were taken and two slides were prepared from each sample . These patients or their legal guardians gave informed consent after explanation of the protocol that had been previously approved by the Ethical Committee of Fundação Oswaldo Cruz . Details regarding sex and age of the individuals included in this study are described in Table S1 . Total RNA was isolated from adult worms ( 1 g ) using TRIzol reagent ( Invitrogen ) , followed by mRNA purification with oligo ( dT ) -cellulose columns according to the manufacturer's instructions ( Amersham Biosciences ) . The SuperScript™ plasmid system for cDNA synthesis and cloning ( Invitrogen ) was used for cDNA library construction following the manufacturer's protocol . The cDNA fragments were directionally ligated into the SalI/NotI cloning sites of the pSPORT1 vector and transformed into competent Escherichia coli DH5α . Specific oligonucleotides were designed using the EST assembly partial sequence from the São Paulo Transcriptome data ( SmAE 606856 . 1 , http://bioinfo . iq . usp . br/schisto6/ ) together with an EST from TIGR ( BF936634 ) . The 5′ and 3′ oligonucleotides , CACCATGATTCGTAGTATCATTGG and CTATTCTTGTTTATCGCTATC , were used in a PCR reaction to amplify the complete open reading frame of SmStoLP-2 from a cDNA library made from adult worms . The PCR reaction was performed using Platinum Pfx enzyme ( Invitrogen ) , and initiated with one cycle of 5 min at 94°C , followed by 30 cycles of 30 s at 94°C , 1 min at 55°C , and 3 min at 68°C . PCR products were purified from agarose gel electrophoresis , cloned into pENTR/D TOPO cloning vector ( Invitrogen ) , and sequenced to confirm its identity . Blast and PSI-Blast searches against the non-redundant protein sequence database , using SmStoLP-2 as a query , were used to identify orthologs of SmStoLP-2 . Additionally , we searched the S . mansoni genome ( GeneDB , http://www . genedb . org/genedb/smansoni/ ) for proteins with Pfam SPFH/Band 7 domains . Post-translational modification prediction: the signal peptide prediction was performed using the SignalP 3 . 0 server ( http://www . cbs . dtu . dk/services/SignalP/ ) , transmembrane helices were analyzed by TMHMM version 2 . 0 ( http://www . cbs . dtu . dk/services/TMHMM-2 . 0/ ) , palmitoylation sites were predicted by CSS-Palm ( http://csspalm . biocuckoo . org/1 . 0/index . php ) [29] , and mitochondrial targeting sequence as predicted by the MitoProt program ( http://ihg2 . helmholtz-muenchen . de/ihg/mitoprot . html ) . Molecular weight ( MW ) and isoelectric point ( pI ) were calculated with the Compute pI/Mw tool ( http://www . expasy . ch/tools/pi_tool . html ) . For phylogenetic analyses , alignments of protein sequences were performed using the ClustalX 1 . 83 software . The tree was constructed using Clustal with the Neighbour Joining method , excluding positions with gaps . The numbers represent the confidence of the branches assigned by bootstrap ( in 1000 samplings ) . The TreeView program ( http://taxonomy . zoology . gla . ac . uk/rod/treeview . html ) was used to visualize the tree . To produce a recombinant SmStoLP-2 , the full-length cDNA sequence was directionally cloned by recombination into pDEST17 ( to produce a protein that contains an N-terminal hexahistidine tag ) and transformed into BL21 ( DE3 ) ( Invitrogen ) . For protein expression , the transformed cells were grown in 600 ml LB plus ampicilin ( OD600 = 0 . 6 ) . Isopropyl-β-D-thiogalactopyranoside ( IPTG ) was added to the culture to a final concentration of 1 mM , and cells were incubated for 3–4 h at 37°C . Cells were harvested by centrifugation and resuspended in 50 ml of lysis buffer ( 50 mM sodium phosphate pH 8 . 3 , 0 . 3 M NaCl ) . The cell suspension was passed twice ( 1500 psi ) through a French press and the crude homogenate was centrifuged at 20 , 000×g for 40 min . The pelleted inclusion bodies were washed twice with wash buffer ( lysis buffer , 2% Triton X-100 , 2 M urea ) and finally resuspended in solubilization buffer ( lysis buffer , 5 mM beta-mercaptoethanol , 20 mM imidazole , 8 M urea ) . The recombinant protein was refolded from the inclusion bodies by diluting 100-fold into equilibration buffer ( solubilization buffer without urea ) . The recombinant protein was then purified by metal affinity chromatography using the Akta Prime system ( Amersham Biosciences ) under native conditions . Briefly , the sample was loaded onto a Ni2+-NTA column ( 5 ml bed volume ) pre-equilibrated with the same buffer . The column was washed with 10 bed volumes of the equilibration buffer and then eluted with 20–500 mM imidazole linear gradient . The main peak was pooled and the protein purity of fractions was assessed using sodium dodecyl sulfate polyacrylamide gel electrophoresis ( SDS-PAGE ) . Further , the elution buffer was exchanged with Phosphate Buffer Saline pH 7 . 4 ( PBS ) before use of this protein . Polyclonal rat serum was produced against preparations of recombinant SmStoLP-2 . Rodents were inoculated four times , at 21-day intervals with 100 µg of protein mixed with TiterMax adjuvant ( CytRx Corporation; first dose ) or PBS ( in subsequent doses ) . Fifteen days after the last inoculation , rodents were exsanguinated . The sera were used at a dilution of 1∶10 , 000 ( v∶v ) in Western blots and 1∶100 in indirect immunofluorescence assays . CD measurements were carried out on a Jasco J-810 Spectropolarimeter at 20°C equipped with a Peltier unit for temperature control . Far-UV CD spectrum was acquired using a 1 mm path length cell at 0 . 5 nm intervals over the wavelength range from 190 to 260 nm . Five scans were averaged for each sample and subtracted from the blank average spectra . The protein concentration was kept at 10 µM in 10 mM sodium phosphate buffer pH 7 . 4 . Total parasite extracts from eggs , miracidia , cercariae , 10-day old schistosomula and adult worms of S . mansoni were prepared in 40 mM Tris , pH 7 . 4 , 2% SDS plus protease inhibitor cocktail ( Sigma ) through sonication ( 4 cycles of 2 min , with pulses of 0 . 75 s , 40% amplitude ) . The samples were centrifuged at 20 , 000×g for 30 min at 4°C and the supernatant was quantified and used for assays . The soluble fraction of adult worms and schistosomula was obtained in a similar way , with the exception of 2% SDS in the sonication buffer . After centrifugation at 20 , 000×g for 30 min at 4°C , the supernatant was recovered , and the insoluble pellet was sonicated in the presence of 2% SDS , which after centrifugation at 20 , 000×g for 30 min at 4°C , originated the so-called insoluble fraction . Their protein concentrations were determined with a RC DC Protein Assay Kit ( Bio-Rad , CA , USA ) . Samples of purified rSmStoLP-2 and extracts ( 20 µg ) were submitted to SDS-PAGE . The gel was electroblotted onto a PVDF membrane , which was blocked with 0 . 02 M Tris ( pH 7 . 5 ) and 0 . 3% Tween 20 containing 5% dry milk for 16 h at 4°C . Subsequently , the membrane was incubated in a 1∶10 , 000 dilution with primary antibody in blocking buffer plus 150 mM NaCl for 3 h at room temperature . After three washes using Tris 10 mM ( pH 7 . 5 ) , the membrane was incubated in a 1∶4000 dilution with secondary goat anti-rat IgG conjugated to horseradish peroxidase ( HRP ) ( Pierce ) for 1 h and after three washes using Tris 10 mM ( pH 7 . 5 ) , the membrane was treated with ECL plus ( GE ) reagent according to manufacturer's instructions . The sample used in this experiment was kindly provided by Dr . Simon Braschi ( University of York , England , UK ) . Briefly , the tegument was removed by a freeze/thaw method and surface membranes enriched by sucrose-gradient centrifugation as previously described [30] , [31] , generating a gradient pellet . Proteins were sequentially extracted from the gradient pellet using a three-step process with reagents of increasing solubilizing power as follows: Extract 1 ( soluble proteins ) : 40 mM Tris , pH 7 . 4; Extract 2 ( non-covalent , but firmly bound proteins ) : 5 M urea ( BDH , VWR International , Dorset , UK ) , 2 M thiourea ( BDH ) in 40 mM Tris , pH 7 . 4 ( Extraction Buffer 2; EB2 ) ; Extract 3 ( GPI-anchored and single membrane spanners ) : EB2 plus 4% CHAPS ( Sigma ) and 2% N-decyl-N , N-dimethyl-3-ammonio-1-propane sulphate ( SB 3–10; Sigma ) , pH 7 . 4; Final pellet ( multispanning membrane proteins ) : solubilized with 40 mM Tris , pH 7 . 4 plus 2% SDS . Mitochondrial enriched fraction was prepared from adult worm tegument in isotonic mitochondrial buffer ( MB ) ( 210 mM mannitol , 70 mM sucrose , 1 mM EDTA , 10 mM HEPES pH 7 . 5 ) supplemented with complete protease inhibitor cocktail ( Sigma ) . The tegument membranes were obtained by centrifugation at 100×g for 30 min at 4°C . The resulting supernatant was centrifuged at 10 , 000×g for 10 min at 4°C to purify the mitochondrial fraction . The resulting pellets were ressuspended in 40 mM Tris , pH 7 . 4 plus 2% SDS . The protein concentration was estimated by the method of Lowry with a RC DC Protein Assay Kit ( Bio-Rad , CA , USA ) . Anti-Mitofusin-1 antibody ( Mfn1 ( H-65 ) ( Santa Cruz Biotechnology ) ( 1∶200 dilution ) was used as a mitochondrial tracker in Western blot experiments followed by incubation with secondary goat anti-rabbit IgG conjugated to HRP ( Sigma ) . Freshly perfused worms were embedded in OCT medium in a pre-cooled beaker of isopentene , frozen in liquid N2 . Eight-micrometer cryostat adult worm sections were adhered to silanized glass slides ( DakoCytomation ) and fixed in acetone for 30 min at −20°C before blocking with PNT ( PBS 1x , 10% Naive rabbit serum and 0 . 1% Tween 20 ) for 4 h at room temperature . They were then incubated with anti-rSmStoLP-2 antisera diluted 1∶100 in PNT for 2 h at room temperature . After washing five times with PBS 0 . 1% Tween 20 , pH 7 . 4 ( PBS-T ) , an Alexa Fluor 488 conjugated anti-rat IgG 1∶200 ( v∶v ) , 20 mM DAPI ( 4′ , 6-diamidino-2-phenylindole dihydrochloride , Molecular Probes ) to visualize nuclei , and 0 . 1 µg/ml phalloidin rhodamine ( Molecular Probes ) to stain actin microfilaments , were added to PNT solution , and the samples incubated for 1 h at room temperature . Sections were washed five times with PBS-T , and then mounted in Fluorescent Mounting Medium ( DakoCytomation ) . In order to label the whole parasite , 7-day cultured schistosomula were fixed in 4% paraformaldehyde in phosphate-buffered saline ( PBS ) for 1 h on ice , washed with PBS and kept at 4°C until use , the hybridisation conditions were the same used for adult worm sections . Rat pre-immune sera were used as negative control . Images were acquired in a Zeiss LSM 510 Meta confocal system , attached to a Zeiss Axiovert 100 microscope using a LD-Achroplan 20x/0 . 4 or C-Apochromatic 63x/1 . 2 water immersion objectives with differential interference contrast . Sera of schistosomiasis patients living in endemic areas in Brazil were tested by ELISA as previously described [32]–[34] to measure the levels of immunoglobulin isoytpes to rSmStoLP-2 protein . For this assay , 96 well flat-bottom microtiter plates ( Nunc ) were coated overnight at 4°C with 100 µl/well of rSmStoLP-2 at a concentration of 5 µg/ml in 0 . 1 M carbonate bicarbonate buffer ( pH 9 . 6 ) . The plates were then blocked with 10% bovine fetal serum in PBS ( pH 7 . 4 ) for 2 h at room temperature . Subsequently , the plates were washed three times with PBS plus 0 . 05% Tween-20 ( PBS-T ) . Serum samples diluted 1∶50 ( IgG ) and 1∶40 ( IgA ) in PBS-T ( 100 µl/well ) were added in duplicate and the plates incubated for 1 h at room temperature . Peroxidase-labelled anti-human IgG and anti-human IgA ( Sigma ) was added at dilutions of 1∶10 , 000 and 1∶1000 ( 100 µl/well ) , respectively . After 1 h at 37°C , the plates were washed and orthophenyl-diaminobenzidine plus 0 . 05% hydrogen peroxide in phosphate citrate buffer ( pH 5 ) was added ( 100 µl/well ) . This mixture was than incubated for 15 min at room temperature , and the reaction was stopped by addition of 5% H2SO4 ( 50 µl/well ) . Absorbance was read at 492 nm using a microplate reader ( Bio-Rad , Hercules . CA , USA ) . To measure IgG subclasses , the previous protocol was modified . The serum dilution was changed based on the isotype to be detected . Serum samples diluted 1∶30 ( IgG1 ) , 1∶5 ( IgG2 or IgG4 ) and 1∶80 ( IgG3 ) were added to the plates and incubated for 2 h at 37°C , as previously described [32] . After washing , peroxidase labelled mouse anti-human antibody was added in each well at concentrations of 1∶1000 ( IgG1 , IgG3 , IgG2 or IgG4 ) , and the plates were incubated for 16 h at 4°C . The subsequent steps were identical to those described for the other isotypes . Cercariae penetration and migration inhibition assays were adapted from a previously described method for Necator americanus [35] . Briefly , six aliquots of 100 µl , each containing 100 cercariae of S . mansoni , were incubated for 1 h at 37°C with 50 µl of sera from rats immunized with rSmStoLP-2 formulated with TiterMax . Sera from rats injected with saline was used as control . To eliminate any effect of complement , the sera were previously heated for 30 min at 56°C . After the 1h incubation period , 800 µl of pond water was added to the samples and the whole volume ( 950 µl ) was then applied to the shaven abdomen of six anaesthetized mice ( 90 mg/Kg of Ketamine and 10 mg/Kg of Xylazine ) . The cercariae were allowed to penetrate by the ring method for 30 min at room temperature . Non-penetrating parasites were evaluated by counting those that remained on the surface of the skin , which were collected by removing the remaining liquid with a pipette and washing the skin twice with 1 ml of PBS . The mean value was considered the percentage of larval penetration inhibition by the antiserum . Six weeks after percutaneous penetration , 6 mice per group were sacrificed with a lethal dose of Ketamine/Xylazine solution . Perfusion fluid ( Saline solution , 500 units/L of heparin ) was pumped into the aorta artery , and perfused worms were collected from the hepatic portal vein . Adult male and female worms were counted using a stereomicroscope . Five to six week-old female C57BL/6 from the Universidade Federal de Minas Gerais ( UFMG ) animal facility , were supplied with food and water ad libitum . Groups of C57BL/6 mice were lightly anaesthetized ( with 45 mg/kg of Ketamine and 10 mg/kg of Xylazine ) and injected subcutaneously in the nape of the neck with 3 doses , at 15-day intervals , of 25 µg of protein mixed with Freund's Complete Adjuvant ( Sigma; first dose ) or Freund's Incomplete Adjuvant ( in subsequent doses ) . In the control group , PBS with Freund's adjuvant was administered using the same immunization protocol . Challenge infections were performed 2 weeks following the final immunization . Mice were anaesthetized with 90 mg/Kg of Ketamine and 10 mg/Kg of Xylazine and exposed percutaneously to 100 cercariae by the ring method on their shaven abdomens . Six weeks after percutaneous challenge infections , 10 mice per group were sacrificed and perfused as described in the cercariae inhibition of penetration assay . The protection was calculated by comparing the number of worms recovered from each vaccinated group with its respective control group , in two independent experiments . The livers were collected from the same animals fixed in 10% paraformaldehyde , processed for paraffin embedding and histopathological sections performed using microtome at 6–7 µm and stained in a slide with hematoxilin-eosin ( HE ) . The number of granulomas was obtained from the liver sections using 10× objective in a microscope . The area from each liver section was calculated using capture in scanner followed by analysis in the KS300 software connected to a Carl Zeiss image analyzer , and the number of granulomas calculated by the area of the liver . Mice were bled from the retro orbital plexus and ELISA was performed to confirm the titer of specific anti-rSmStoLP-2 IgG , IgG1 and IgG2a in the serum of immunized animals . Briefly , 96 well flat-bottom microtiter plates ( Nunc ) were coated overnight at 4°C with 100 µl/well of rSmStoLP-2 at a concentration of 5 µg/ml in a 0 . 1 M carbonate bicarbonate buffer ( pH 9 . 6 ) . The plate was then blocked with bovine fetal serum 10% in PBS for 2 h at room temperature . Further , the plates were washed three times with PBS plus 0 . 05% Tween-20 ( PBS-T ) . One hundred microliters of each serum diluted 1∶100 in PBS-T was added per well and incubated for 1 h at room temperature . Plate-bound antibody was detected by peroxidase-conjugated anti-mouse IgG , IgG1 and IgG2a ( Southern Biotechnology ) diluted in PBST 1∶10 , 000 , 1∶5000 and 1∶2000 , respectively . After 1 h at 37°C , the plate was washed and orthophenyl-diaminobenzidine plus 0 . 05% hydrogen peroxide in phosphate citrate buffer ( pH 5 ) was added ( 100 µl/well ) . This mixture was then incubated for 30 min at room temperature , and the reaction was stopped by addition of 5% H2SO4 ( 50 µl/well ) . Absorbance was read at 492 nm using a microplate reader ( Bio-Rad , Hercules , CA , USA ) . Animals that received PBS with Freund's adjuvant were used as negative control . Cytokine experiments were performed using splenocyte cultures from individual mice immunized with rSmStoLP-2 plus CFA/IFA ( n = 5 for each group ) . Splenocytes were isolated from macerated spleens of individual mice 10 days after the third immunization and washed twice with sterile PBS . The cells were adjusted to 1×106 cells per well in RPMI 1640 medium ( Gibco , CA , USA ) supplemented with 10% FBS , 100 U/ml penicillin G sodium , 100 µg/ml streptomycin sulfate , 250 ng/ml amphotericin B . Splenocytes were maintained in culture with medium alone or stimulated with rSmStoLP-2 ( 25 µg/ml ) or concanavalin A ( ConA ) ( 5µg/ml ) as previously described [36] . The 96-well plates ( Nunc ) were maintained in an incubator at 37°C with 5% CO2 . For cytokine assays , polymyxin B ( 30 µg/ml ) was added to the cultures and this treatment completely abrogated the cytokine response to LPS , as previously described [37] . Culture supernatants were collected after 48 h of rSmStoLP-2 stimulation for IL-4 and TNF-α analysis and 72 h of rSmStoLP-2 stimulation for IL-10 and IFN-γ . The assays for measurement of IL-4 , IL-10 , IFN-γ and TNF-α were performed using the Duoset ELISA kit ( R&D Diagnostic ) according to the manufacturer's recommendations . Student's t-test was used and the two-tailed p-value was calculated to compare experimental and control groups on challenge infections , antibody profiles and cytokine assays in mice . For the human humoral response against rSmStoLP-2 , the Kruskal–Wallis test was used to evaluate the significance of the results of all groups compared to the non-infected ( NI ) . The Mann–Whitney test was used to evaluate the significance of antibody measurements obtained between the groups resistant to S . mansoni reinfection ( RR ) versus the groups susceptible to reinfection ( SR ) .
The full-length sequence of the S . mansoni cDNA encoding Stomatin like protein-2 was obtained by PCR from an adult worm cDNA library with specific oligonucleotides . The resulting full-length cDNA ( GenBank accession EU531730 ) displays an ORF of 1077 bp , encoding a protein of 358 amino acids with a predicted molecular mass of approximately 39 . 5 kDa and an isoelectric point of 5 . 83 . BlastP comparisons of the deduced S . mansoni protein sequence to GenBank showed that the best match ( E-value = 5×10−95 ) was to Danio rerio hypothetical protein , with 58% identity and 81% similarity over 355 amino acids . The next best match was against an unknown S . japonicum protein ( probably an incomplete SLP-2 ) . This was followed by several other SLP-2 proteins including human ( 58% of identity ) , therefore we designated this gene as SmStoLP-2 ( since there was already another gene named as Sm-SLP-2 , although not related to Stomatin like proteins [38] ) . SmStoLP-2 contains the stomatin signature sequence ( residues 31–189 ) ( outlined by a dashed box in Figure 1 ) , and is recognized as part of the Pfam SPFH/Band 7 family with an E-value of 1 . 2×10−75 . Additionally , searching the S . mansoni genome ( GeneDB ) for proteins with Pfam SPFH/Band 7 domains , we found putative orthologues of H . sapiens stomatin ( Band 7 ) , and C . elegans Mec-2 ( Figure 1 ) . We identified a further five schistosome stomatin-related genes ( data not shown ) . Human stomatin ( Band 7 ) may associate with membranes via a hairpin loop ( continuous box ) with both the N- and C- termini facing the cytoplasm ( Figure 1 ) . This domain is conserved among several members of the SPFH/Band 7 superfamily , such as C . elegans MEC-2 and S . mansoni Stomatin and Mec-2 , but is absent in SLP-2 members ( Figure 1 ) . We further identify in all SLP-2 sequences putative signal peptides ( ranging from 16 to 32 amino acids at the N-terminal region ) , which were predicted to be a mitochondrial targeting sequence ( dashed - dotted box ) . Some previously recognized stomatin family members , such as human stomatin , have a consensus sequence for palmitoylation centered on Cys29 and Cys86 [39] , which apparently increase the affinity of stomatin for the membrane . Further examining the distribution of potential post-translational modifications of SmStoLP-2 , we found three putative sites for lipid modification ( palmitoylation ) centered on Cys11 , Cys61 and Cys330 ( Figure 1 , underlined ) . Surprisingly , these palmitoylation sites were not detected in any other analysed member of the SLP-2 subfamily , except for the S . japonicum ortholog ( data not shown ) . Phylogenetic analysis of the SFPH superfamily confirmed that SmStoLP-2 is a member of the stomatin family , grouping it in a branch with other SLP-2s ( Figure S1 ) . Like other stomatins [13] , [40] , [41] , it is distantly related to flotillin , prohibitin , and HflK/C . Two putative flotillin and two putative prohibitins genes were identified in the S . mansoni genome . As expected , and probably due to its prokaryotic origin , no orthologues of HflK/C were found ( Figure S1 ) . E . coli transformed with pDEST17-SmStoLP-2 showed a band at 45 kDa when induced with IPTG , which is slightly higher than the expected molecular mass for rSmStoLP-2 ( Figure 2A ) . The bacteria were lysed by a French Press and separated into soluble and insoluble fractions . The insoluble fraction ( inclusion bodies ) was shown to contain the majority of the recombinant protein ( Figure 2B , lanes 1 , 2 ) . The inclusion bodies were extracted with 8 M urea , refolded by dilution and purified by affinity chromatography on nickel-charged columns through an imidazole linear gradient from 20 to 500 mM ( Figure 2B , lanes 3–8 ) . The fractions were pooled and dialyzed to remove imidazole , yielding 8 . 0 mg of rSmStoLP-2/L culture . Circular Dichroism spectra indicated that the rSmStoLP-2 contains a regular secondary structure , although the proportions of secondary structure elements ( α-helix and β-sheet ) were not calculated ( data not shown ) . Extracts were prepared from cercariae , schistosomula , adult worms , eggs and miracidia stages of S . mansoni and subjected to immunoblotting with rat anti-rSmStoLP-2 serum , showing comparable levels of expression in all stages across the parasite life cycle . Native SmStoLP-2 observed in schistosome extracts , migrates with a molecular mass higher than that predicted , which was comparable to rSmStoLP-2 ( ∼49 kDa ) ( Figure 2C ) . It is not known whether an additional smaller band ( ∼47 kDa ) could be a product of post-translational modification , alternative initiation , protein degradation or alternative mRNA splicing . Extracts from schistosomula and adult worms were separated into soluble and insoluble fractions and Western blot analysis revealed SmStoLP-2 to be present in the insoluble fractions in both stages ( Figure 2D ) . A higher molecular mass band can be seen in the insoluble fraction of schistosomula , similar to that observed in the recombinant protein ( P ) . The two most prominent bands correspond closely to the monomeric and dimeric forms of the protein at 49 and 98 kDa . To further characterize the distribution of SmStoLP-2 in S . mansoni tegument , differential extractions of tegument membrane proteins were analysed . Western blot using anti-rSmStoLP-2 serum , revealed that SmStoLP-2 was recovered in the first extraction fraction solubilized with urea/thiourea ( Figure 2E ) , suggesting SmStoLP-2 to be firmly bound , although non-covalently , to the tegument membranes . On the other hand , SmStoLP-2 displays a mitochondrial signal sequence , which , if functional , may target it to the tegumental mitochondria . To address this issue , we isolated the tegument and performed a differential fractionation , separating the membrane and mitochondrial fractions . The anti-rSmStoLP-2 antibody recognized the protein in both fractions ( Figure 2F ) . Mitochondrial enrichment was ascertained using a Mitofusin-1 antibody , which only detected this protein in the mitochondria-enriched fraction . Immunolocalization studies using rat serum raised against rSmStoLP-2 revealed through confocal fluorescence microscopy , that SmStoLP-2 is mainly expressed in the tegument of the adult S . mansoni male and female worms and seems to be expressed at lower levels in the muscle cells of male worms ( Figure 3A and E ) . In an attempt to localize SmStoLP-2 in relation to the cytoskeletal tegument components , we used phalloidin-rhodamine as an actin marker . As shown in Figure 3P , there was some overlap staining on the muscle layers of adult male worms , revealed by the yellow signal . In contrast , the green band in the tegument , which corresponds to the main location of SmStoLP-2 , did not seem to be co-localized with actin ( Figure 3P ) . Additionally , the protein in male adult worms appears to be located more basally in the tegument , but it is interesting to note that the green band also seems to be running around and outside of their dorsal tubercles ( Figure 3P ) . Intact schistosomula stained with anti-rSmStoLP-2 and phalloidin-rhodamine suggested SmStoLP-2 to be external to the muscle layers in the tegument , as revealed by the green band running around and externally to the red band ( Figure 3I–K , and R–T ) ; additionally , the phalloidin-rhodamine internal labelling confirms that the parasites were well permeabilized ( Figure 3J and M ) . No staining was observed in male and female sections or intact schistosomula incubated with naive rat serum ( Figure 3C , G and L ) . Preliminary experiments indicate that , also in cercariae , SmStoLP-2 would be located in the evolving tegumental layer . We evaluated by ELISA the specific reactivity of anti-SmStoLP-2 antibodies in sera of individuals with different status of resistance and susceptibility to S . mansoni reinfection . The sera of schistosomiasis patients , with the exception of the group susceptible to reinfection ( SR ) , had significant levels of total anti-SmStoLP-2 IgG as compared to the non-infected group ( Figure 4A ) . Furthermore , individuals from the group resistant to reinfection ( RR ) had increased levels of anti-SmStoLP-2 IgG when compared to individuals susceptible to reinfection ( SR ) . Regarding IgA , statistically significant levels of antibodies to rSmStoLP-2 were observed in the INF and RR groups when compared to the NI group ( Figure 4B ) . Once more , the RR group produced more anti-SmStoLP-2 IgA as compared to the SR individuals . The IgG subclass profile of schistosomiasis patients was characterized predominantly by IgG1 , IgG2 and IgG3 antibody responses to rSmStoLP-2 . Individuals resistant to reinfection ( RR ) displayed at least a 2-fold higher level of IgG1 , IgG2 and IgG3 anti-SmStoLP-2 antibodies as compared to those susceptible to reinfection ( SR ) ; these isotypes were also significantly higher when compared to the NI group ( Figure 4C ) . Concerning IgG4 , this antibody isotype was not detected in any of the groups studied . In an attempt to check if the anti-rSmStoLP-2 antibodies could impair penetration of cercariae and their survival afterwards , we performed a skin penetration inhibition assay . As shown in Figure S2A , the rat anti-rSmStoLP-2 serum inhibited cercarial skin penetration by 77% , as compared with 40% inhibition by serum from rats that received saline only ( p = 0 . 002 ) . Six weeks after the infection , we assessed the parasite load in the infected mice; data revealed that only 12% of the penetrating parasites matured to adult worms in the group in which cercariae were incubated with anti-rSmStoLP-2 , while 42% matured in the group incubated with control serum ( Figure S2B ) . In a typical infection in the murine model , usually the maturation rate is around 35–40% [42] . C57BL/6 mice were immunized with 3 doses of rSmStoLP-2 formulated with Freund's adjuvant and sera were analyzed by ELISA at 15 , 30 , 45 , 60 , 75 and 90 days for production of anti-SmStoLP-2 antibodies . Significant titers of specific anti-rSmStoLP-2 IgG antibodies were detected at all time points , showing a plateau after the third dose ( data not shown ) . To determine the IgG isotype profile induced by immunization , specific IgG1 and IgG2a to rSmStoLP-2 were also analyzed . The levels of specific IgG1 and IgG2a and the IgG1/IgG2a ratio indicate that until the second dose there is a predominant Th2 response and after the third immunization there occurs a drift towards a more balanced or Th1-modulated immune response ( Table 1 ) . In order to investigate the cytokine profile induced by the rSmStoLP-2 immunization regimen described above , we isolated splenocytes 10 days after the third immunization . Cytokine production ( IFN-γ , TNF-α , IL-4 and IL-10 ) was measured in the culture supernatants from in vitro rSmStoLP-2-stimulated spleen cells of immunized mice . Statistically significant levels of IFN-γ , signature of Th1-type immune response , and TNF-α , a proinflammatory cytokine , were produced in the stimulated splenocytes from the rSmStoLP-2-immunized group as compared with the control ( Figure 5A and B ) . Additionally , high levels of the modulatory cytokine , IL-10 , were also observed ( Figure 5C ) , and no secretion of IL-4 , a Th2 cytokine , was detected ( data not shown ) . These results indicate that immunization of mice with rSmStoLP-2 formulated with Freund's adjuvant induces a Th1-predominant immune response , with increased levels of IFN-γ , TNF-α and IL-10 and non-detectable levels of IL-4 secretion . In order to determine the protective potential of rSmStoLP-2 , immunized mice were challenged with 100 cercariae . The worms were recovered by perfusion 6 weeks after challenge and results were expressed as the “mean worm burden” ( mean ± S . D . ) and are summarized in Figure 6 . The animals immunized with rSmStoLP-2 in Freund's adjuvant showed 30 and 32% reduction in worm burden against challenge infection in two independent experiments when compared to the control group . Analysis of egg counts in the liver did not show a statistically significant reduction in oviposition .
In this report , we have identified SmStoLP-2 as a member of the stomatin super family , displaying several properties shared with other SLP-2 proteins and some unique features . The widespread distribution of the ‘conserved’ SPFH domain across life kingdoms has been taken as an indication of its ancient origin , suggesting the common ancestry and functional homology of all SPFH proteins [13] . However , in a recent review , it has been proposed that SPFH grouping has little phylogenetic support , probably due to convergent evolution of its members [41] . Independently of its origin , our phylogenetic analysis of the deduced SmStoLP-2 protein has grouped it together with human SLP-2 , and SLP-2 from Danio rerio and Xenopus tropicalis , and at some distance from SLP-1 , stomatin ( band 7 ) and SLP-3 . We can highlight the following primary sequence features: 1 ) SmStoLP-2 lacks an N-terminal hydrophobic domain , similar to other SLP-2 members [9]; 2 ) SmStoLP-2 and SjStoLP-2 are the unique members of SLP-2 family , which show putative sites for palmitoylation , a property that could enhance the hydrophobicity of proteins and contribute to their membrane association; 3 ) additionally , SmStoLP-2 , like all SLP-2 proteins , seems to have a mitochondrial targeting peptide in the N-terminal region . The recovery of SmStoLP-2 in the insoluble fraction of parasite extracts suggests that the protein could be membrane-associated . Furthermore , solubilization of SmStoLP-2 from tegument membranes after the treatment with the chaotropic agents , urea and thiourea , indicates that it should be non-covalently bound to the tegument . It is interesting to note that SmStoLP-2 was not identified on Braschi's proteomic study [30] , which could be explained based on the differences in sensitivity of the two methods used to detected the protein ( mass spectrometry and Western blot ) . Moreover , albeit not with a quantitative analysis , the fact that SmStoLP-2 protein is present in the free-living ( freshwater ) cercariae and miracidia , as well as in the egg stage , suggests that the protein has other function ( s ) , not exclusively associated with the tegument , which are common to both free-living and parasitic stages . It has been recently proposed that HuSLP-2 may interact with actin ( a cytoskeletal constituent ) , vav ( a small GTPase that regulates cytoskeleton reorganization ) and Nck ( an adaptor protein that links transmembrane and scaffolding molecules to the cytoskeleton ) [23] . The confocal immunofluorescence images of the parasites confirm the SmStoLP-2 tegument localization and suggest some weak co-localization with actin , only on muscle layers of adult male worms . In addition , SmStoLP-2 seems to be located externally to the muscle layers in the 7-day-old schistosomulum . Data from the HuSLP-2 suggests that there are at least two cellular pools of this protein: one associated with the plasma membrane and the other with mitochondria [9] , [11] , [24] . It is important to note that the tegumental cytoplasmic layer lying under the surface membranes , contains small mitochondria [43] . However , confocal immunofluorescence microscopy does not have sufficient resolution to address this question with confidence . Our results on the differential fractionation of tegument extracts analyzed by Western blot addressed this issue and strongly suggested that SmStoLP-2 also displays a dual targeting , one associated to the tegument membrane and one to the mitochondria . As a consequence of the studies in the attenuated cercaria vaccine model , the schistosomula is believed to be the target of protective immunity [42] . Given the tegument localization of SmStoLP-2 in the schistosomula suggested by our immunolocalization results , this molecule should be accessible as an immune target . In individuals putatively resistant to reinfection ( RR ) , the antibody response mounted against SmStoLP-2 , consisted mainly of the cytophilic antibodies IgG1 and IgG3 , which have opsonization properties , cell dependent cytotoxicity , and the ability to activate the classical complement pathway , functions which could be involved in the resistance to S . mansoni reinfection . Elevated levels of IgG1 , IgG2 and IgG3 have been linked to the human resistant status for several vaccine candidates , such as , Sm23 , Sm28 , Sm14-FABP , Sm29 and TSP-2 [32] , [44] , [45] . Concerning IgA levels , investigators have associated the increased levels of this isotype with resistance to reinfection stimulated by Sm28GST antigen [46] , [47] , which parallels our results , where high levels of IgA antibodies to rSmStoLP-2 were observed in patients which are resistant to reinfection ( RR ) . Although no function has been ascribed for SmStoLP-2 , our finding that anti-rSmStoLP-2 antibodies inhibits cercariae skin penetration and migration , suggests that the molecule may have an important role in larval host entry and in migration through the tissues until the lungs before reaching the portal hepatic system . This finding provides further support for testing SmStoLP-2 as a vaccine candidate against murine schistosomiasis . In the murine model , rSmStoLP-2 induced high levels of anti-rSmStoLP-2 IgG after the second immunization and showed a reduced IgG1/IgG2a ratio at 45 days after the first immunization . Additionally , we confirmed by cytokine analysis that rSmStoLP-2 immunization elicited a Th1-predominant type of immune response characterized by production of high levels of IFN-γ and no detection of IL-4 . To determine if rSmStoLP-2 conferred protection against S . mansoni infection , immunized mice were challenged with cercaria and worm burden analyzed . Immunization with rSmStoLP-2 induced a 30-32% worm burden reduction . A primary obstacle in the research and development of a schistosomiasis vaccine is a lack of understanding of what type of immune response should be induced . In the irradiated cercariae vaccination model , protection can be either based on a Th1 , a Th2 , or a mixed Th1/Th2 immune response [48] . However , in the case of recombinant proteins , Th1 inducing antigens have been described to induce protection against infection in the mouse model [36] , [49]–[53] . The role of IFN-γ in the protective immunity to schistosomiasis is well described in mice exposed to the irradiated vaccine and there is compelling evidence that immune elimination of challenge parasites occurs in the lungs . Since IFN-γ is likely to be required to activate pulmonary macrophages which may mediate the protective response [54] , we could expect that similar mechanisms would be involved in rSmStoLP-2 protective immunity . Recently , HuSLP-2 has been proposed as a potentially useful target for immunotherapy in humans , since it modulates effector T cell responses [24] . Thus , down-regulation of HuSLP-2 expression could be valuable in the course of autoimmune disease treatment , since it decreases T cell reactivity; alternatively , enhancement of HuSLP-2 expression could be explored in vaccine development , since it would increase T cell responsiveness [24] . Given that orthologs often , but not always , have the same function , it is still unclear what is the function of SmStoLP-2; it could have an immunomodulatory role like its human ortholog [24] or could acquire a different function on the parasite-host interface , like providing structural scaffolding for the tegument or supporting the traffic of vesicles to the surface plasma membrane [9] , organizing the peripheral cytoskeleton and assembly of multichain receptors , such as ion channels [11] , [55] , or even mechanosensation receptors [18] , [20] . Further investigations of SmStoLP-2 will be valuable in understanding how the tegument functions as the parasite-host interface . A critical issue in vaccine design is the use of an appropriate adjuvant and/or delivery system to induce the suitable immune response . Experiments are underway investigating rSmStoLP-2 with different adjuvant formulations , which would be suitable for use in humans . In conclusion , our study showed that SmStoLP-2 is a novel tegument protein , being recognized by different subclasses of antibodies in patients resistant and susceptible to reinfection , and in the light of data obtained from murine studies , protective properties against schistosomiasis were revealed . We believe that an ideal vaccine may require the combination of quite a few antigen targets to induce an effective protection against the parasite , and SmStoLP-2 could contribute to reach this goal .
|
Schistosomiasis is a parasitic disease causing serious chronic morbidity in tropical countries . Together with the publication of the transcriptome database , a series of new vaccine candidates were proposed based on their functional classification . However , the prediction of vaccine candidates from sequence information or even by proteomics or microarrays data is somewhat speculative and there remains the considerable task of functional analysis of each new gene/protein . In this study , we present the characterization of one of these molecules , a stomatin like protein 2 ( SmStoLP-2 ) . Sequence analysis predicts signals that could contribute to protein membrane association and mitochondrial targeting , which was confirmed by differential extractions of schistosome tegument membranes and mitochondria . Additionally , confocal microscope analysis showed SmStoLP-2 present in the tegument of 7-day-old schistosomula and adult worms . Studies in patients living in endemic areas for schistosomiasis revealed high levels of IgG1 , IgG2 , IgG3 and IgA anti-SmStoLP-2 antibodies in individuals resistant to reinfection . Recombinant SmStoLP-2 protein , when used as vaccine , induced significant levels of protection in mice . This reduction in worm burden was associated with a typical Th1-type immune response . These results indicate that SmStoLP-2 could be useful in association with other antigens for the composition of a vaccine against schistosomiasis .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"biotechnology",
"infectious",
"diseases/helminth",
"infections",
"molecular",
"biology",
"immunology"
] |
2010
|
Schistosoma mansoni Stomatin Like Protein-2 Is Located in the Tegument and Induces Partial Protection against Challenge Infection
|
APE1 is a multifunctional protein with a DNA base excision repair function in its C-terminal domain and a redox activity in its N-terminal domain . The redox function of APE1 converts certain transcription factors from inactive oxidized to active reduced forms . Given that among the APE1-regulated transcription factors many are critical for KSHV replication and pathogenesis , we investigated whether inhibition of APE1 redox function blocks KSHV replication and Kaposi’s sarcoma ( KS ) phenotypes . With an shRNA-mediated silencing approach and a known APE-1 redox inhibitor , we demonstrated that APE1 redox function is indeed required for KSHV replication as well as KSHV-induced angiogenesis , validating APE1 as a therapeutic target for KSHV-associated diseases . A ligand-based virtual screening yielded a small molecular compound , C10 , which is proven to bind to APE1 . C10 exhibits low cytotoxicity but efficiently inhibits KSHV lytic replication ( EC50 of 0 . 16 μM and selective index of 165 ) and KSHV-mediated pathogenic phenotypes including cytokine production , angiogenesis and cell invasion , demonstrating its potential to become an effective drug for treatment of KS .
Kaposi’s sarcoma-associated herpesvirus ( KSHV ) , also termed human herpesvirus type 8 ( HHV8 ) , is a member of the γ-herpesviridae subfamily . This virus has been proven to be the etiological agent of Kaposi’s sarcoma ( KS ) [1] . Almost 100% of KS lesions , regardless of their source or clinic subtype ( i . e . , classic , AIDS-associated , African endemic , and post transplant KS ) , are infected with KSHV . KS is the most common malignancy associated with HIV-infection . About 20% of AIDS patients develop KS with most of them ( 60% ) manifesting with oral lesions [2] . Additionally , KSHV is also associated with two lymphoproliferative diseases , namely primary effusion lymphoma ( PEL ) [3] and multicentric Castleman’s disease ( MCD ) [4] . Currently there is no definitive cure for KS and other KSHV-associate diseases . The KSHV life cycle consists of two phases , latent and lytic [5] . When KSHV infects a target cell , it establishes latent infection by default and expresses a few latent genes to maintain latent infection . Establishment of latency is a viral strategy to avoid host immune surveillance and fuse symbiotically with the host for persistent infection . Spontaneous lytic replication occurs in a small portion ( 1–2% ) of infected cells , releasing infectious virions to infect fresh cells in order to sustain the population of latently infected cells , that otherwise would be quickly lost by segregation of latent viral episomes as spindle cells divide [6] . Furthermore , KSHV lytic replication is also crucial for efficient dissemination from its long-term reservoir to the sites of disease and providing paracrine regulation for KS development [7–10] . Therefore , both latent and lytic cycles of KSHV are important for viral pathogenicity . KSHV lytic replication may serve as a therapeutic target for treatment of KS due to its etiological role in KS . Currently classic cancer therapies are generally used to treat KS patients , which include surgical excision and radiation therapy for patients with a few lesions in a limited area and chemotherapy for patients with extensive or recurrent KS [2] . The chemotherapeutics that have been approved by the FDA include liposomal anthracycline products [11] ( liposomal doxorubicin or liposomal daunorubicin ) [12] , paclitaxel and interferon-alpha [13–15] . However , these therapeutic agents do not target the etiological virus and the tumor response to any chemotherapeutic regimen is only transient . For AIDS-KS , HAART regimens are associated with regression in the size and number of existing KS lesions[16–19] . However , despite its dramatic decrease in frequency since the advent of HAART , KS remains the most common AIDS-associated cancer . In addition , there is an emergence of a new HAART-associated syndrome . In a subset of HIV-seropositive individuals , starting HAART in the setting of advanced HIV infection results in a paradoxical clinical worsening of the existing infection or the appearance of a new condition including KS in a process known as immune reconstitution inflammatory syndrome or IRIS[20 , 21] . Since IRIS-KS is the result of responses by a recovered immune system to KS-causing pathogen ( i . e . KSHV ) , the treatment of KSHV-seropositive , HIV-positive patients with a combination of antiretroviral ( HAART ) and anti-KSHV chemotherapeutics is expected to yield positive results . However , there is currently no available drug effectively targeting KSHV . Apurinic/apyrimidinic endonuclease 1 ( APE1 ) , also termed Redox factor-1 ( Ref-1 ) , is a multifunctional protein . Its C-domain carries a DNA base excision function ( acting as apurinic/apyrimidinic endonuclease ) and its N-domain has a redox activity controlling gene expression for cell survival pathways . The redox function of APE1 converts its substrate proteins from oxidized inactive form to reduced active form and affected transcription factors include AP-1 [22] , NF-κB [23] , Egr-1 [24] , HIF-1α [25] , p53 [26] , Pax protein [27] and COX-2 [28] . The DNA base excision repair function of APE1 has been extensively explored as a therapeutic and chemopreventive target [29] . It was demonstrated that the redox function of this protein is also associated with many malignancies , such as pancreatic cancer [30] , ovarian cancer [31] , and glioblastoma [32] . Indeed the inhibition of APE1 redox function decreases cell proliferation , prevents the angiogenesis progress , and blocks the differentiation of endothelial precursor cells [33] . As many of the transcription factors under the regulation of the APE1 redox function are known to be involved in KSHV lytic replication and other pathogenic phenotypes such as neo-angiogenesis , endothelial cell differentiation and cell invasion , we asked if APE1 redox function is critically required for KSHV lytic replication and tumorigenesis and if the enzyme can serve as an effective target for drug development for KSHV-associated diseases . In this study , we have validated APE1 as an effective target for blocking KSHV lytic replication and some pathogenic phenotypes . A small molecular compound was identified to be a novel inhibitor of APE1 redox function through a ligand-based virtual screening using an in-house three-dimensional ( 3D ) molecular superimposing algorithm , WEGA and shown to be effective in inhibition of KSHV lytic replication , virus-mediated endothelial differentiation , neo-angiogenesis and cell invasion .
During the switch from latency to lytic replication , KSHV uses several signaling pathways to activate the viral Replication Transcription Activator ( RTA ) to initiate viral lytic replication cascade . It was reported that the MEK/ERK , JNK and P38 multiple mitogen-activated protein kinase pathways are necessary and sufficient for activating the RTA promoter through activation of the transcription factor AP-1 [34 , 35] . The DNA binding activity of AP-1 to its target DNA was found to be dependent on the redox activity of APE1 [22] , suggesting that APE1 might be a regulatory factor for KSHV lytic replication . To explore this hypothesis , we first determined the role of APE1 in KSHV lytic replication through shRNA-mediated knock down of APE1 expression in cells and examining the effects of APE1 silencing on KSHV lytic replication . An shRNA against human APE1 was introduced into BCBL-1 cells by lentiviral transduction . In the transduced cells , the APE1 shRNA effectively down-regulated APE1 expression by 80% in comparison to the cells transduced with control shRNA ( Fig 1A ) . Then the cells were treated with TPA to induce KSHV reactivation . The viral DNA replication was examined in these shRNA-transduced cells by measuring intracellular viral genomic DNA at different time points up to 120 hour post-induction using real-time PCR . The effect of APE1 knock down on progeny virion production was also determined . The culture media of shRNA-transduced cells were collected up to 5 days post-induction and the amounts of virion particles in the media were determined by quantifying encapsidated viral DNA with real-time PCR . Results showed that with APE1 knock down , both KSHV lytic DNA replication and virion production were significantly reduced ( Fig 1B and 1C ) . The expression of immediate-early gene RTA in shRNA-transduced BCBL-1 was determined by qRT-PCR and the result showed a significant reduction of RTA transcript with APE1 knock down ( Fig 1D ) . The importance of the redox function of APE1 to KSHV lytic replication was further examined using a pharmacological inhibitor , namely E3330 , which has been proven to be a redox function inhibitor of APE1 [36] . BCBL-1 cells were treated with TPA to induce lytic viral replication . Three hours after the induction , the cells were exposed to E3330 of various doses . Forty-eight hours post-induction , total DNA was extracted and intracellular viral genomic DNA was determined by real-time PCR . The half maximal inhibitory concentration of E3330 ( IC50 for viral DNA replication ) was calculated to be 17 . 7 μM ( Fig 1E ) . The effect of E3330 on progeny virion production was also determined . Five days post-induction , the E3330-treated cell culture media were collected and virion particles were determined by quantifying encapsidated viral DNA in the media . The half maximal antiviral effective concentration ( EC50 ) was calculated from extracellular virion doses response curves to be 8 . 5 μM ( Fig 1E ) . Cytotoxicity of E3330 on BCBL-1 cells was assessed by Trypan blue exclusion that provided a homogeneous method for estimating both the numbers of viable and nonviable cells present in culture of each treatment . The half maximal cytotoxic concentration ( CC50 ) was determined to be 112 μM , yielding a selectivity index ( CC50 / EC50 ) of 13 ( Fig 1E ) . Taken together , both studies with shRNA knock down and the pharmacological inhibitor validates that APE1 is essential for KSHV lytic replication and can serve as an effective target for antiviral agents against KSHV infection . KS is an angiogenic and invasive tumor and abnormal neovascular channel that fills with red blood cells is a pathological feature of KS . Although the nature and cellular origin of KS cells remains contentious , KSHV infection of both endothelial cells and mesenchymal stem cells ( MSCs ) confer the cells with certain KS features including angiogenic , invasive and transformation phenotypes [37 , 38] . When human MSCs from periodontal ligament ( PDLSC ) were infected with KSHV , increased angiogenesis activity was shown in an in vitro Matrigel tubulogenesis assay ( Fig 2B ) . In addition , KSHV-mediated angiogenesis has been demonstrated to involve paracrine regulation , and the leading hypothesis for the mechanism is that KSHV proteins , including LANA and vIRF-3 , interact with and up-regulate HIF-1α , leading to increased VEGF-A expression [39–42] . KSHV induction of VEGF-A is responsible for the induction of angiogenesis , as well as proliferation of KS spindle cells . In addition , IL-6 , IL-8 , and other cytokines also play important roles in KSHV-associated angiogenesis . IL-6 and IL-8 are regulated by MAPK pathway through the transcription factor AP-1 in KSHV-infected cells [43 , 44] . We found that the conditioned media of KSHV-infected PDLSC could increase the angiogenesis activity of PDLSCs , indicating that KSHV can induce angiogenesis of oral MSCs in a paracrine manner ( Fig 2C ) . Since both AP-1 and HIF-1α are known to be substrates of APE1 [45] and the redox activity of APE1 is reported to be associated with angiogenesis in cancer [28] , we asked whether APE1 is crucial for KSHV-induced angiogenesis in PDLSCs . To this end , APE1 shRNA-transduced PDLSCs were infected with GFP-carrying recombinant KSHV ( rKSHV . 219 ) at an MOI ( multiplicity of infection ) of 20 ( viral genome equivalent ) and infection rates were from 72 to 87% on the basis of GFP expression . The effect of APE1 silencing on angiogenesis capacity of MSCs was assessed through a Matrigel tubulogenesis assay . The KSHV-PDLSCs transduced with control shRNA were able to form capillary-like tubules on Matrigel that represents the later stage of angiogenesis . By contrast , the KSHV-PDLSCs expressing APE1 shRNA showed reduced tubulogenesis activity ( Fig 2D ) . Furthermore , paracrine-induced angiogenesis was also APE1-dependent as the conditioned medium of APE1 shRNA expressing cells showed lower capability to induce paracrine-associated angiogenesis ( Fig 2E ) . To confirm that APE1 is indeed required for KSHV-induced angiogenesis of PDLSCs through regulating the transcription and secretion of angiogenic growth factors , we examined the effect of APE1 silencing on the transcription and production of VEGF-A , IL-6 and IL-8 of KSHV-infected PDLSCs . Total RNAs were extracted from KSHV-PDLSCs transduced with control or APE1 shRNAs and subjected to quantitative real-time PCR using probes for VEGF-A , IL-6 and IL-8 . The levels of APE1 and LANA mRNAs were also evaluated as references . The results showed that along with reduced expression of APE1 , the levels of VEGF-A , IL-6 and IL-8 mRNAs were significantly decreased in APE1 knock down cells ( Fig 2F ) . In parallel , the culture media of KSHV-PDLSCs with APE1 or control shRNA were collected and the levels of secreted VEGF-A , IL-6 and IL-8 were measured by ELISA . Results demonstrated dramatic reduction of these angiogenic factors in the culture media of APE1 shRNA-expression cells ( Fig 2G ) . Therefore , we conclude that APE1 is a critical regulator for paracrine regulation of angiogenic growth of KSHV-infected KS progenitor cells . APE1 is an important regulator of AP-1 , HIF-1α , and many other transcription factors , of which many are known to be involved in KSHV replication and pathogenic processes . Thus , APE1 represents a promising target for antiviral against KSHV and associated diseases . In order to identify potential inhibitors against APE1 , a ligand-based virtual screening was employed using an in-house three-dimensional ( 3D ) molecular superimposing algorithm ( WEGA ) . Using E3330 as a query molecule , compounds within our natural compound repository , Guangdong Small Molecule Tangible Library [46] were screened on the basis of 3D shape and pharmacophore features ( Fig 3A ) . Conformational ensembles ( maximum size of 250 ) were generated for each compound in the database through the CAESAR algorithm in Discovery Studio 3 . 5 . Following shape-based alignment , the similarities between the query and target molecules were calculated by WEGA . Among the highly scored compounds exhibiting similar shape and pharmacophore features to E3330 were C10 ( Fig 3B ) , which was chosen for further characterization . Differential scanning fluorimetry ( DSF ) and circular dichroism ( CD ) were employed to test whether C10 is capable of interacting with APE1 . DSF allows the identification of a compound binding to its target protein through the observed thermal shift ( ΔTm ) of the protein [47–49] . APE1 was incubated with C10 at 37°C for 30 min and then the fluorescence dye was added . The melting curves with increasing temperature were obtained from a real-time instrument LightCycler 480 . Results showed that C10 shifted the Tm of APE1 by 1 . 3°C ( Fig 3C ) , indicating the direct binding of C10 to APE1 . As a reference , E3330 shifted the Tm of APE1 by 1 . 0°C . CD is an assay using polarized light to measure the melting temperature changes influenced by compound binding . The result also confirmed that C10 destabilizes the structure of APE1 and so does E3330 ( Fig 3D ) . The binding affinities of C10 and E3330 to APE1 were determined by using surface plasmon resonance ( SPR ) and Kd’s of interaction of C10 and E3330 with APE1 were measured to be 189 nM and 1 . 63 μM , respectively , indicating a higher binding affinity of C10 to APE1 than that of E3330 ( Fig 3E ) . Since the ability of AP-1 binding to its DNA binding sites has been shown to require redox regulation by APE1 , electrophoresis mobility shift assay ( EMSA ) was employed to investigate the ability of C10 to inhibit the redox function of APE1 . Although the double-stranded DNA fragment containing AP-1 binding motif could be shifted when incubated with a nuclear extract , purified AP-1 proteins ( c-jun/c-fos heterodimer ) failed to bind to the DNA fragment ( Fig 4A ) . In accordance with redox regulation of APE1 , the binding of AP-1 heterodimer to the double-stranded AP-1 oligonucleotides occurred only in the presence of reduced form of APE1 ( Fig 4A ) . Untreated APE1 ( oxidized form ) showed almost undetectable AP-1 DNA binding activity , confirming that AP-1 DNA binding activity is dependent on redox function of APE1 in this system . To test whether C10 can inhibit APE1 redox activity and , as a consequence , block AP-1 DNA binding activity , the EMSA experiment was performed with increasing concentrations of C10 . As shown in Fig 4B , a dose-dependent inhibitory response by C10 was established , yielding an IC50 of 17 . 0 μM . By comparison , a relatively weaker effect on APE1-dependent AP-1 binding was observed for E3330 ( IC50 = 36 . 0 μM , Fig 4C ) . To examine whether C10 also affects the C-terminal function of APE1 , an in vitro assay for APE1 DNA excision activity was established . A 40-bp DNA with an U-G base-pair was prepared by annealing of two synthetic oligonucleotides . Uracil-DNA glycosylase ( UDG ) recognizes the U-G base pair and removes uracil base to product a single base lesion . APE1 is able to recognize the absent site and its exonuclease activity shortens the 40-bp DNA to 18 bp ( Fig 4D ) . As shown in Fig 4E , the APE1 exonuclease activity was unaffected in the presence of up to 100 μM of C10 . By contrast , Myrecetin , a known APE1 exonuclease function inhibitor and served as a positive control , showed a dose-dependent inhibitory response ( Fig 4F ) . Taken together , the results demonstrated that C10 is an inhibitor specific to the APE1 redox function . As KSHV lytic replication is dependent on the redox function of APE1 , we wondered if the new APE1 inhibitor C10 effectively inhibits KSHV lytic replication . BCBL-1 cells were induced into lytic replication by TPA . Three hour post-induction , the cells were exposed to C10 in a wide range of concentrations . Intracellular viral genomic DNA was determined at 48 hour post-induction . Encapsidated viral DNA was measured for released virions in the media 5 days post-induction . The IC50 and EC50 values of C10 were determined from the dose-response curves of the intracellular DNA and extracellular virion to be 4 . 6 and 0 . 16 μM , respectively ( Fig 5A ) . The cytotoxity of C10 was assessed for BCBL-1 using trypan blue exclusion method for cell viability at 48h and CC50 values was determined to be 26 . 4 μM , leading to a selectivity index ( SI = CC50/EC50 ) of 165 . In addition , the cytotoxicity of C10 to primary lymphocytes was also assessed with human peripheral blood mononuclear cells ( PBMCs ) after 48 hour treatment and data were shown in Fig 5B . The cytotoxicities of C10 to different types of cells , including PDLSC , iSLK . 219 , HUVEC and HEK293T , were also examined and shown in Supporting Information ( S4 Fig ) . KSHV lytic replication is controlled by the viral transcription activator RTA . Since RTA is known to be regulated by AP-1 signal transduction , inhibition of APE1 dependent AP-1 DNA binding would therefore disrupt this signaling . Thus , functional inhibition of APE1 redox activity by C10 is speculated to result in the decreased expression of RTA expression . To test this , we employed promoter-reporter assays to study the effects of C10 on the activation of the AP-1 promoter as well as the RTA promoter by AP-1 ( c-Jun/c-fos ) . 293T cells were transfected with an AP-1 promoter-luciferase reporter plasmid or an RTA-promoter-luciferase reporter , respectively . Six hours late , C10 was added into the culture media in varying concentrations . TPA was then added 24h post-transfection to activate AP-1 signaling . The activation of the AP-1 promoter as well as the RTA promoter in the absence and presence of C10 were measured through the luciferase activities 36 hours post-induction . C10 was able to block TPA-induced AP-1 promoter ( Fig 5C ) and RTA promoter activities in a dose-dependent manner ( Fig 5D ) . These results indicate that AP-1 redox inhibitor C10 blocks KSHV lytic replication through inhibiting TPA-induced AP-1 pathway and RTA expression . Interestingly , C10 was also found to be able to block auto-regulation of the RTA promoter activity by RTA itself ( Fig 5E ) . This result suggests that APE1 redox function is also required in some event downstream of RTA such as activation of RTA or associated cellular protein ( s ) . To examine if C10 is able to block KSHV lytic replication induced by ectopically expressed RTA , we employed iSLK . 219 cell line in that KSHV lytic replication is under control of doxycycline-inducible RTA [50 , 51] . iSLK . 219 was treated with doxycycline ( DOX ) for 3 hours , then C10 in a wide range of concentration was added to the culture medium . In this system , C10 was found to be able to inhibit viral lytic DNA replication with an IC50 value of 5 . 8 μM and virion production with an EC50 value of 3 . 2 μM , respectively ( Supporting Information , S1 Fig ) . KS is a vascular tumor and pathological neoangiogenesis is a hallmark of the cancer . As KSHV-mediated angiogenesis or angiogenic growth factor production is dependent on APE1 ( Fig 2 ) , possibly through maintenance of redox status of AP-1 and HIF-1α and their DNA binding activities , we investigated whether C10 could be an effective inhibitor of KSHV-elaborated neoangiogenesis . In a Matrigel tubulogenesis assay , the ability of KSHV-infected oral MSCs to form capillary-like tubules was examined after treatment with C10 . As shown in Fig 6A , the tubulogenesis of KSHV-PDLSC was inhibited by C10 in a dose dependent manner . E3330 was also included in the assay as a reference and showed inhibitory activity at higher doses ( Fig 6A ) . In addition , C10 and E3330 were capable of inhibiting paracrine-mediated angiogenesis with conditioned media from KSHV-infected PDLSCs ( Fig 6B ) . We also examined C10 for its effect on the angiogenic capability of KSHV-infected HUVECs and their conditioned media . Similar inhibitory effects of C10 on angiogenesis of endothelial cells were observed in comparison to that of PDLSCs ( Supporting Information , S2 Fig ) . Next , we employed a murine Matrigel plug assay to examine the ability of C10 to inhibit angiogenesis ex vivo . KSHV-infected and mock-infected PDLSCs were mixed with Matrigel ( without any cytokines ) and implanted into C57BL/6 mice . In day 7 , the matrigel plugs were stripped and examined for new blood vessel formation . As shown in Fig 6C , KSHV-infected PDLSCs exhibited increased blood vessel formation on the Matrigel plug . When C10 in varying concentrations was mixed with Matrigel , KSHV-induced angiogenesis was dramatically inhibited in a dose dependent manner ( Fig 6C ) . The effect of C10 on KSHV-mediated angiogenesis in the ex vivo assay was quantitated by measuring the hemoglobin content after homogenizing the Matrigel plugs using Drabkin’s reagent [52] . The result , shown in Fig 6C ( right panel ) , is consistent with that of in vitro Matrigel tubule formation assay . KS has been considered as a cytokine disease as KS lesions over-produce cytokines , especially angiogenic growth factors , that contribute to the major pathogenic features of KS [53 , 54] . In the early stage , KS is not a real sarcoma but an angiohyperplastic-inflammatory lesion mediated by inflammatory cytokines and angiogenic factors [55] . The finding that the paracrine-regulated angiogenesis requires the redox function of APE1 ( Fig 2 ) prompted us to further investigate if blockade of APE1 redox function by C10 could reduce cytokine and angiogenic factor production . KSHV-infected MSCs were shown to produce a wide array of cytokines that resembles KS cells . KSHV-infected PDLSCs were treated with either C10 or E3330 for 24 hours . After changing media to remove C10 and E3330 from the media , cells were continued to culture for 24 hours . Then the culture media were collected and examined for the expression of VEGF-A and IL-6 by ELISA . The results showed that both C10 ( Fig 7A ) and E3330 ( Fig 7B ) were able to reduce the levels of these angiogenic growth factors . Similar results were also obtained with HUVECs where KSHV-induced productions of VEGF-A , IL-6 and IL-8 were inhibited by C10 ( Supporting Information , S3 Fig ) . Cell invasion is a feature of KS and KSHV infection of human primary endothelial cells has been reported to promote cell migration and invasion [56 , 57] . Using a Matrigel-transwell assay , we investigated the invasive property of KSHV-infected PDLSCs . Cells were stimulated for chemostasis in a-MEM containing 10% FBS and seeded in a Matrigel on the upper chamber of a Transwell containing a-MEM free of FBS . The cells that migrated through the membrane were visualized by crystal violet staining . KSHV-infected PDLSC showed increased migration and invasion activity ( Fig 8A ) . In addition , PDLSCs in the KSHV-PDLSC conditioned medium also had elevated migration and invasion ( Fig 8B ) , suggesting the KSHV-mediated cell migration is paracrine regulated , possibly through VEGF [57] . shRNA-mediated knock down of APE1 led to a significant decrease in cell invasion ability in both KSHV-infected PDLSCs and the use of the conditioned media ( Fig 8C and 8D ) . Consistently , C10 was also able to block cell migration and invasion of KSHV-PDLSC ( Fig 8E ) . Taken together , APE1 is required for KSHV-mediated cell migration and invasion and APE1 redox inhibitor C10 can efficiently block this phenotype . C10 was identified through a ligand-based virtual screen using E3330 as a template molecule and proven to be a new inhibitor of APE1 redox function . C10 blocks the redox function of APE1 and shows no demonstrable effect on the base excision endonuclease activity ( Fig 4 ) . The two functions of APE1 are independent and separable in the primary amino acid sequence . The cysteine at position 65 ( Cys65 ) is critical for the redox function [58] . However , APE1 crystal structure showed that Cys65 is deeply buried inside the protein [59 , 60] , raising a question how C10 interacts with APE1 and gains access to the redox active site . In the early study on E3330 , it was surprisingly found that E3330 binds at the C-terminal domain , distant from Cys65 in a NMR study [61] . At high concentration , E3330 interacts with both regions of the N-terminus ( amino acids 68−74 ) and C-terminus ( amino acids 266−273 ) as mapped by hydrogen/deuterium exchange mass spectrometry [36] . The results suggest that a locally unfold state could be the redox active state of APE1 . To understand how C10 binds to APE1 and blocks redox function , molecular docking and molecular dynamics ( MD ) simulation were performed on the ligand-receptor complex system . First , the time-dependent root mean square deviation ( RMSD ) values of the complex backbone and ligand trajectories were acquired for 40 ns MD simulation . When C10 was docked at the C-terminal pocket , MD simulations produced stable RMSD curves , which fluctuated around 1 . 3 Å ( Fig 9A ) , suggesting that C10 fits well in the C-terminal pocket . However , if C10 was docked into the N-terminal domain , the RMSD curves fluctuated 2 to 4Å ( Fig 9B ) , indicating that C10 was easily dropped out from the N-terminal pocket . These data suggest that C10 prefers to bind with the C-terminal rather than the N-terminal domain of APE1 . This is consistent with the crystal structure , which shows that the redox active site of APE1 is buried inside the protein . Then , we docked C10 into the C-terminal pocket and performed MD simulations to explore the binding pose of C10 in the pocket . The binding mode derived from MD simulation results indicated that C10’s aromatic groups prefer to form σ–π interactions with Trp280 and its carbonyl group trends to form hydrogen bonds with Arg177 side-chain ( Fig 9C ) . Now the question is how C10 , through interacting with those amino acids in the C-terminal , affects APE-1 redox activity . Toward this end , we compared the RMSD curves of APE1 and APE1-C-C10 backbone trajectories in 40 ns MD simulation . The RMSD curve for APE1 backbone was stable , while the RMSD curves for APE1-C-C10 backbone show higher fluctuations ( ranging from 1 to 1 . 5 Å ) in comparison with APE1 alone with RMSD within a range from 0 . 5 to 1Å ( Fig 9D ) . This result informs that C10 binds with the C-terminal pocket and results in conformation transitions of APE-1 . It is observed from the MD simulation that in the absence of C10 , Cys65 is hidden behind an α-helix in APE1 ( Fig 9E ) , while binding of C10 into the C-terminal pocket lifts up the α-helix domain , therefore resulting in Cys-65 exposed to the solvent ( Fig 9F ) . This conformational change ( or called locally unfolded ) allows C10 to form a covalent bond with Cys65 , as revealed in a covalent docking ( Fig 9G ) , which leads to the blockage the formation of the disulfide bond between Cys-65 and Cys-95 and inhibition of the redox activity of APE1 .
In the current study , we validated APE1 as an effective target for blocking KSHV replication and treating KS and other KSHV-associated malignancies . This study revealed crucial roles of APE1 in KSHV lytic replication , as well as in the development of KS pathogenic features such as cytokine production , angiogenesis and cell invasion . Furthermore , a novel compound was identified that binds to APE1 and blocks its redox function , thereby exhibiting antiviral activity against KSHV and KS pathogenic feature development . The salient features and implication of these findings are as follows . ( 1 ) APE1 is a multifunctional protein with the DNA base excision activity in its C-terminal region and redox activity in its N-terminal domain . These two functions are completely independent in their actions as the mutation of Cys65 abolishes the redox function but does not affect the DNA repair function , whereas the mutation of His309 , which is required for the DNA repair function , does not affect the redox function[62] . Both functions of APE1 have been intensively explored for therapeutic purposes . Selective inhibitors against APE1 base excision repair activity have been exploring to be used in combination with DNA-interactive anticancer drug to enhance the efficacy of chemotherapy [63] . The APE1 redox regulation of a class of transcription factors affecting cancer survival and growth makes the protein attractive to be used as a target for cancer therapeutic strategy [64] . APE1 is overexpressed in human pancreatic cancer . The specific APE1 redox inhibitor E3330 is reported to cause tumor growth inhibition in cell lines and pancreatic cancer xenograft model in mice , demonstrating the potential of APE1 inhibitors in pancreatic cancer treatment [30] . In the current study , we , for the first time , demonstrated that APE1 plays crucial roles in KSHV lytic replication as well as other critical aspects of KS development through its redox function and validated APE1 as an effective target for halting viral replication and treatment of KS . Among the transcription factors whose transcriptional activities are dependent on APE1 redox activity , many are known to participate in KSHV life cycle and pathogenicity . ( i ) KSHV reactivation is initiated through mitogen-activated protein kinase ( MAPK ) /extracellular signal-regulated kinase ( ERK ) pathway or Ras/Raf/MEK/ERK/Ets-1 pathway that lead to activation of RTA by AP-1[34 , 35 , 65] . Through redox regulation , APE1 determines if AP-1 is able to activate the RTA promoter . Therefore , APE1 redox function is absolutely required for KSHV reactivation and lytic replication . ( ii ) KS is considered a cytokine disease and KS cell growth and angiogenesis is dependent on autocrine and paracrine cytokine regulation . Pathogenic roles of HIF-1α , AP-1 and COX-2 in KSHV-mediated overexpression of a variety of cytokines have been established [35 , 41 , 66] . APE1 is known to regulate HIF-1a , AP-1 and COX-2 and , as a consequence , controlling cytokine production[22 , 25 , 28] . Our results are consistent with the previous reports and confirm the roles of APE1 in KSHV-induced cytokine production , validating APE1 as an effective target for halting KSHV replication and inhibiting development of KSHV-associated malignancies . The critical roles of APE1 in KSHV-associated cytokine production through redox regulation of DNA binding activities of the transcription factors including HIF-1α , COX-2 and AP-1 in turn explain the pharmacological mechanism underlying inhibition of KSHV-mediated cytokine production , angiogenesis and cell invasive growth by APE1 inhibitor C10 . ( 2 ) In addition to the well-characterized target proteins of APE1 mentioned above , our study suggests that APE1 may be involved in other steps in KSHV life cycle or pathogenicity by regulating other APE1 substrate proteins that have not yet been elucidated . ( i ) KSHV reactivation is regulated through a positive feedback mechanism [67] . RTA auto-regulates its own expression at the transcriptional level by using cellular RBP-Jκ notch signaling pathway[68 , 69] . We found that the APE1 redox function is required for RTA auto-regulation ( Fig 5E ) , suggesting that some protein involving the auto-regulation , such as RBP-Jκ or RTA , might be a substrate of APE1 and under the redox regulation by APE1 . ( ii ) The EC50 of C10 for blocking virion particle release is 29-fold lower than IC50 for inhibiting intracellular viral DNA replication ( Fig 5A ) , suggesting APE1 has an additional role in a certain step between viral DNA replication and virion assembly ( e . g . regulation of late gene expression ) , which can also be affected by C10 . Further investigation is warranted for the roles of APE1 in these steps of viral life cycle , which may lead to identification of new substrates of APE1 and expand our understanding on the role of APE1 in biology . ( 3 ) Since KSHV is heavily dependent on APE1-regulated cell proliferation and survival pathways for viral replication and pathogenic feature development , APE1 has been validated to be an excellent target for anti-KSHV and KS therapy . Its inhibitors are able to affect not one but many steps towards KSHV replication and KSHV-associated malignancy development . Multi-target drugs have raised considerable interest in the last decade due to their advantages in the treatment of complex diseases[70] . Although C10 is a specific inhibitor of APE1 and not an authentic multi-target compound in a narrow sense , it practically impacts multiple proteins and affects multiple biological processes . Therefore , C10 has a greater potential than a compound that affects a single biological reaction in becoming an effective drug to treat complex diseases such as KS and other cancers . However , the multiple targeting property of a compound often raises an issue of cytotoxicity . Therefore we have closely paid attention to the cytotoxicity issue with C10 . In cell-based assays , C10 showed relatively low cytotoxicity with an acceptable selectivity index ( e . g . , SI = 165 for inhibition of KSHV virion production in B cells ) . E3330 was also reported to exhibit low cytotoxicity [71] . It could be explained by that the transcription factors under APE1 redox regulation are mainly over-expressed in malignant or virally infected cells so that APE1 redox inhibitors affect cancer cells and viral replication in a greater extent than normal somatic cells[72 , 73] . This notion is supported by the fact that APE1 redox function is not evolutionally conserved but only present in mammals . In contrast , the base excision repair function of APE1 is conserved during phylogeny , which is an ortholog of E . coli Xth ( exonuclease III ) [74] . Nevertheless , the safety and cytotoxicity of C10 is an important issue that warrants further evaluation in future pre-clinical and clinical research . ( 4 ) APE1 is associated with the progression of various human diseases[62] . The redox function of the enzyme regulates several pathways relevant to cell proliferation and cancer survival [64] . Dysregulation of APE1 has been reported to be associated with cancer where alteration of APE1 was found in both expression and subcellular localization levels [30 , 32] . Inhibition of APE1 redox function decreases cell proliferation and migration of cancer cells , and blocks the differentiation of endothelial precursor cells and angiogenesis [33] . As a novel and effective inhibitor of APE1 redox function , C10 has potential to become an anti-cancer agent to treat cancers and other angiogenesis-related human diseases . In addition to cancer , APE1 was shown to be associated with non-malignant angiogenesis-related disease . Aged macular degeneration ( AMD ) is the leading cause of severe vision loss in the elderly and neovascularization in the retina is a key pathology in AMD and other ocular diseases[75] . Currently the most advanced treatment for AMD is to block retinal neoangiogenesis using anti-VEGF agents , but the responses to the therapy are usually transient and disease recurrence occurs[76] . Inhibition of APE1 redox function appears to be a better strategy as APE1 inhibitors with multi-targeting property not only blocks VEGF production , but also reduces inflammatory reaction that also contributes to AMD development . It has been shown that inhibition of APE1 redox activity by E3330 blocks retinal angiogenesis in vitro and in an animal model[77 , 78] . As an effective APE1 redox inhibitor , we see potential value of C10 in treatment of AMD and other angiogenesis-related diseases . Taken together , revelation of the multiple-roles of APE-1 redox function in KSHV lytic replication and viral pathogenic feature development ( cytokine production , angiogenesis and migration ) and the discovery of a novel APE1 redox inhibitor inform a new strategy toward controlling KSHV infection and treating KSHV-associated diseases .
This research was approved by the Animal Ethics Review Board of Sun Yet-sen University and Medical Ethics Review Board of Sun Yet-sen University . We have an existing Institutional Animal Care and Use Committee approval for the animal use ( Approval No . 2015–041 , Animal Ethics Review Board of Sun Yet-sen University ) . The experiment was carried out strictly following the Guidance suggestion of caring laboratory animals , published by the Ministry of Science and Technology of the People's Republic of China . The human sample collection and the use of PBMCs and PDLSCs in our research were approved the Medical Ethics Review Board of Sun Yet-sen University ( Approval No . 2015–028 ) . Written informed consent was provided by study participants . BCBL-1 cells ( obtained from the National Institutes of Health AIDS Research and Reference Reagent Program ) , an effusion lymphoma cell line that latently infected with KSHV , were grown in RPMI 1640 medium supplemented with 10% heat-inactivated fetal bovine serum ( FBS ) and penicillin-streptomycin ( 50 units/ml ) and 100 mg/ml amphotericin B sodium deoxycholate . Peripheral blood mononuclear cells ( PBMCs ) were isolated from whole blood of a healthy donor ( obtained from the Zhongshan School of Medicine , Sun Yat-Sen University ) and cultured in RPMI1640 medium supplemented with 10% FBS and antibiotics . Human embryonic kidney HEK293T cells were purchased from American Type Culture Collection ( ATCC ) and cultured in Dulbecco's modified Eagle's medium ( DMEM ) supplemented with 10% FBS and antibiotics . Periodontal ligament stem cells ( PDLSCs ) ( obtained from the Hospital of Stomatology of Sun Yat-sen University ) were isolated from the periodontal ligament tissues ( n = 5 referred to 5 independent cultures from different individuals ) and maintained in alpha minimal essential medium ( aMEM , GIBCO Life Technologies ) containing 10% FBS , 200 mM L-glutamine and antibiotics . An shRNA lentiviral vector targeting the 3’-UTR site of APE1 mRNA ( Clone ID: NM_080649 . 1-1305s1c1 ) was purchased from Sigma-Aldrich . Lentiviral particles were prepared by transfection HEK293T cells with pLKO . 1-shAPE17958 ( or pLKO . 1-shControl ) , psPAX2 , pMD2 . G plasmids in the ratio of 4:3:1 . Media containing lentiviruses were harvested at 48 h and 72 h and used to transduce BCBL-1 cells and PDLSCs . Transduced BCBL-1 cells were selected under 2 μg/ml puromycin for a week , while KSHV-PDLSCs was transduced transiently and cells were used 72 h after transduction . BCBL-1 cells were treated with TPA to induce KSHV lytic replication . Three hour post-induction , APE1 inhibitors in a wide range of concentration were added into the culture medium . Total DNA was purified using a DNasey kit ( Magen ) 48h post-induction . KSHV genomic DNA copy number was quantified by real time PCR on a Roche LightCycler 480 instrument using primers specified for LANA and normalized to GAPDH . The half-maximal inhibitory concentration ( IC50 ) values of compounds were determined from a dose-response curve of KSHV DNA content values from TPA-induced and chemical-treated cells . The viral DNA contents with those of uninduced cells subtracted were divided by those of the control cells with no drug treatment and then represented on the y-axes of dose-response curves: y-axis value = ( TPAX − no TPAX ) / ( TPA0 − no TPA0 ) , where X is any concentration of the drug and 0 represents nondrug treatment . The IC50 on viral DNA synthesis for each compound was calculated using GraphPad Prism software . Extracellular virion numbers were estimated by determining encapsidated viral genomic DNA . Five days postinduction with TPA , BCBL-1 culture media were collected and virion particles were cleared by passing through 0 . 45-μm filters . Virions were pelleted from the medium supernatant . The virion preparations were treated with Turbo DNase I ( Takara ) at 37°C for 1 hour followed by proteinase K digestion . Encapsidated viral DNA was extracted with phenol-chloroform . Extracted DNA was precipitated with ice-cold ethanol , and the final DNA pellet was dissolved in TE buffer . The KSHV genomic DNA in virions was measured by real-time PCR with primers directed to LANA ( ORF 73 ) as described above . Virion DNA copy numbers were calculated from a standard curve established using rKSHV . 219 . KSHV virion numbers were presented as the copy numbers of viral genomic DNA per milliliter of culture supernatant . BCBL-1 and PBMCs were treated with chemical inhibitors in a wide range of concentration for 2 and 5 days . The viability of cells was assessed by counting Trypan blue-stained cells using a Countstar instrument . The half-maximal cytotoxic concentration ( CC50 ) was calculated from dose-response curves with Graph-Pad Prism software . The cytotoxicity of chemical inhibitors to cells was also assessed in cell metabolic level using MTT assay . PDLSCs were plated in 96-well plates at 5000 cells/well and treated with inhibitors in various concentrations for 48 h . The medium was replaced with DMSO . Absorbance of MTT was measured at 570nm . The structure of C10 [ ( 1E , 6E ) -1 , 7-bis ( 5-methylfuran-2-yl ) hepta-1-6-diene-3 , 5–dione] was acquired from a tangible compound repository , the Guangdong Small Molecule Tangible Library [46] . The compound C10 used in this study was synthetic , prepared by using the procedure illustrated in Supporting Information ( S5 Fig ) . Boric anhydride ( 0 . 42 g , 6 mmol ) and acetylacetone ( 3 . 00 g , 30 mmol ) were suspended in 10 mL EtOAc , and the mixture was stirred for 3 h at 70°C . After removing the solvent , the resultant solid was washed with hexane . Then 20 ml of EtOAc , 5-methylfuran-2-carbaldehyde ( 6 . 00 g , 54 mmol ) , and tributyl borate ( 1 . 38 g , 6 mmol ) were added and the mixture was stirred for 24 h . Butylamine ( 21 . 90 mg , 0 . 3 mmol ) dissolved in EtOAc was added and the mixture reacted at 70°C for 24 h . Then 1 N HCl was used to adjust the pH to 5 . The product was extracted from the water layer with EtOAc . The C10 was purified by recrystalization from EtOAc to yield 3 . 75 g red crystal , giving a percent yield of 90 . The purity of C10 was determined by HPLC equipped with ZORBAX SB-C18 column ( 4 . 6x250mm , 5μm particle size ) and a UV/VIS detector setting of λ = 254nm and 460nm . Compound was eluted with CH3OH solvent system and assayed by HPLC , which confirmed the purity of the compound to be ≥99% ( Supporting Information , S5B Fig ) . 1H-NMR and 13C-NMR spectra data were recorded on a Bruker AvanceIII spectrometer at 400 MHz using TMS as reference ( Bruker Company , USA ) . Detailed results can be found in Supporting Information ( S6 Fig ) . APE1 protein was expressed and purified from E . coli and used to study the interaction between APE1 with C10 . AP-1 proteins ( c-jun and c-fos ) were purified for APE1 enzymatic assays . APE1 and c-fos cDNAs were cloned into the pET-28a vector with a hexahistidine ( 6xHis ) tagged at the N-terminus . c-jun was cloned into pGEX-4T-1 vector with a GST tag . The Roseta E . coli transformed with each of the plasmids , were grown in LB media and induced with 1mM isopropylb-D–thiogalactoside ( IPTG ) when the culture reached the density of 0 . 6 OD . The bacterial culture grew at 37°C for 4 h for APE1 and at 25°C for 4 h for c-jun and c-fos . Pelleted cells were resuspended and sonicated in lysis buffer containing PMSF with ( for GST-tag ) or without DTT ( for His-tag ) . His-tagged APE1 and c-fos were purified with Ni2+-NTA-resin and eluted with buffer containing imidazole . GST-tagged c-jun was purified using glutathione beads and eluted with reduced glutathione . Protein concentrations were determined using the BCA protein assay kit ( Thermo Scientific ) . Purified APE1 protein ( Supporting Information , S7 Fig ) was diluted to 2 μM and incubated with C10 or E3330 in different concentrations at 37°C for 30 min . Fluorescence dye ( SYPRO orange , Invitrogen ) was added to a final concentration of 5x to the protein solution . Scanning was performed in a Lightcycler 480 instrument and run the temperature scan from 25–95°C min-1 . Melting curve was plotted by the software LightCycler 480 Software release 1 . 5 . 0 SP4 ( Roche ) . Melting temperature ( Tm ) of each concentration was obtained by LightCycler 480 Protein Melting software ( Roche ) . Far-UV CD spectra were acquired on a Chirascan spectropolarimeter ( Applied Photophysics Ltd . , Leatherhead , UK ) in a 225−260 nm interval . APE1 ( 5 μM ) with C10 ( 10 μM ) or E3330 ( 500 μM ) were performed in in 10 mM phosphate buffer ( pH 7 . 4 ) , using a 0 . 1 cm path-length cuvette . Thermal denaturation profiles were obtained by measuring the temperature dependence of the signal at 230 nm in the range of 20−90°C with a resolution of 0 . 5°C and a 1 . 0 nm bandwidth . A temperature controller was used to set up the temperature of the sample; the heating rate was 1°C min-1 . Data were collected at 0 . 2 nm resolution with a 20 nm/min scan speed and a 4s response and were reported as the unfolded fraction versus temperature . Interactions of APE1 with C10 and E3330 were analyzed by using a ProteOn XPR36 SPR instrument ( Bio-Rad , Hercules , CA ) . APE1 protein in 20 nM of 10 mM sodium acetate buffer , pH 5 . 5 was immobilized using amine coupling on the EDAC/Sulfo-NHS-activated surface of GLH biosensor chip channel L1 ( Bio-Rad ) . The surface was blocked with 1M ethanolamine . The final immobilization level for APE1 was approximately 12 , 000 RU . Channel L2 ( reference channel ) of GLH biosensor chip was also activated with EDAC/Sulfo-NHS and blocked with 1M ethanolamine . Compounds in phosphate buffered saline ( PBS ) , pH 7 . 4 , containing 0 . 005% Tween-20 ( PBST ) , were injected at 20 ml/min for 240s at concentrations of 10–3 . 125×10−4 μM ( 1:2 dilutions ) . Following the compound injection , the chip surface was regenerated with 40s pulses of 0 . 85% H3PO4 and running buffer . All experiments were performed at 25°C . In each of the kinetic studies , the interactions of C10 in six concentrations with APE1 and reference channel ( L2 ) were monitored in parallel . The compound concentration data collected were reference-subtracted using ProtedOn Manager 2 . 0 . Each set of sensorgrams was globally analyzed using the 1:1 Langmuir binding model to obtain the kinetic rate constants ( Kon and Koff ) . Global kinetic rate constants ( ka and kd ) were derived for each reaction , and the equilibrium dissociation constant , KD , was calculated using the equation KD = kd / ka . An EMSA-based assay was adapted to measure APE1 redox activity following the APE-1-dependent AP-1 DNA binding ability [36 , 79] . APE1 was reduced by incubating in 0 . 25 mM DTT over night at final concentration of 1 μM . Reduced APE1 ( 2 μL , the final concentration was 0 . 1 μM ) were incubated with purified c-jun/c-fos ( 1:1 ratio ) in EMSA reaction buffer ( 10 mM Tris-HCl pH 7 . 5 , 50 mM NaCl , 5mM MgCl2 ) for 20 min at 37°C . cy5 . 5-labelled double-stranded DNA was added and incubated for another 20 min . The samples were resolved on 5% nondenaturing polyacrylamide gels ( TBE-PAGE ) at 4°C , 100V for 1 h and then scanned with an Odyssey imager ( LI-COR ) . Antibodies against c-jun and c-fos ( Abclone ) were included in the EMSA for supershift of specific band . A 40-bp double-stranded DNA with an U-G mismatch pair was prepared by annealing the following two oligonucleotides: Cy5 . 5–5’-GTAAAACGACGGCCAGTGUATTCGAGCTCGGTACCCGGGG-3’ and 5’-CCCCGGGTACCGAGCTCGAATGCACTGGCCGTCGTTTTAC-3’ . The substrate DNA was incubated with Uracil-DNA glycosylase ( UDG ) at 37°C for 10 min . Uracil-DNA glycosyase removes the uracil-base from the U-G base-pair to produce an abasic site . Then APE1 was added with its buffer and incubated for another 30 min at 37°C . APE1 recognizes the abasic-site and cleave the DNA at the abasic site leading to release of the cy5 . 5- fluorophore labeled 18-nucleotide fragment , which can be resolved from the uncleaved 40-nucleotide fragment ( intact substrate ) on 20% denaturing polyacrylamide gels ( TBE-Urea PAGE ) and visualized with an Odyssey imager ( LI-COR ) . The promoter sequence of the RTA promoter ( 3kb ) was cloned into pGL3-basic vector ( Promega ) to generate pRTA-luc . The promoter-reporter plasmids pAP-1-luc was provided by Dr . Ersheng Kuang at Sun Yat-sen University . Subconfluent 293T cells grown in 48-well plates were co-transfected with 50 ng of pAP-1-luc or pRTA-luc and 5ng of pRL-TK by using lipofectamine 2000 reagent ( Life Technologies ) . Twenty-four hours after transfection , cells were treated with 12-O-Tetradecanoyl-phorbol-13-acetate ( TPA ) . The pRL-TK plasmid expresses Renilla luciferase and was used as an internal control . For RTA auto-regulation assay , 293T cells were co-transfected with pRTA-luc , pRL-TK and pCR3 . 1-ORF50 . Thirty-six hour post-induction , the luciferase assay was performed with Promega's Dual-luciferase assay kit . Each sample was duplicated and each experiment was repeated at least three times . iSLK . 219 cells ( kindly provided by Ke Lan lab of Institut Pasteur of Shanghai , Chinese Academy of Sciences , Shanghai , China ) , carrying rKSHV . 219 [51 , 80] , were induced for lytic replication by 1 μg/ml doxycycline and 1 mM sodium butyrate for 5 days . The culture media were filtered through a 0 . 45-m filter and centrifuged at 100000 g for 1 h . The pellet was resuspended in 1/100 volume of 1X PBS and stored at -80°C until use . PDLSCs were seeded at 2x105 cells per well in 6-well plates . Cells were infected with KSHV in the presence of polybrene ( 5 μg/ml ) at an MOI = 20 ( viral genome copy equivalent ) . After two hours , the inoculum was removed and replaced with fresh culture medium . Cells were incubated under 5% CO2 at 37°C . Forty eight-well plates were coated with Matrigel ( 100μl/well ) and incubated at 37°C for 1h to allow gelation to occur and avoid bubble . PDLSC or KSHV-PDLSC were resuspended in 200 μl a-MEM without FBS and placed on the top of the gel . The cells was incubated at 37°C with 5% CO2 for 8 h , and images of tube formation were captured using a ZEISS fluorescence microscope . The quantification of the tube was using the software ImageJ to measure the total length of tube in the image . The average value was used for the histogram . Female C57BL/6 mice ( 4 to 6 weeks old ) were obtained from the Sun Yat-sen University Animal Center . Five animals were used per treatment group . 5-10x106 cells in 100–200 μl medium were prepared and thoroughly mixed with 500 μl matrigels to total volume of 600–800 μl , and subsequently implanted to mice by inguinal injection . After even days , the mice were killed and matrigel plugs were removed . The Matrigel plugs were photographed and subsequently dissolved in 1 ml Dispase reagent for 16 h at 37°C . After removal of debris by centrifugation , hemoglobin was qualified using Drabkin's reagent ( Sigma-Aldrich ) . Cell invasion assays were carried out in 24-well transwell units ( millipore ) . Briefly , polycarbonate filters with 8-μm pores were coated with 60 μl of matrigel-gel . PDLSCs or KSHV-PDLSCs ( 15 , 000 cells ) in serum-free media were placed in the upper wells and the lower chambers were filled with 10% FBS medium . After 24 h incubation the cells that had passed through the filter were stained with crystal violet . The number of migrated cells was counted from multiple randomly selected microscopic visual fields using ImageJ software . Photographs were obtained and independent experiments were performed in triplicate . Levels of VEGF-A , IL-6 and IL-8 in cell culture supernatants were determined using specific ELISA kits for human VEGF-A ( eBioscience ) , IL-6 ( BD ) and IL-8 ( Xin Bosheng , China ) according to the manufacturer instructions . Virtual molecular docking of C10 with APE1 protein was executed using Autodock suite 4 . 2 . 6 [81] . APE1 protein was downloaded from Protein Data Bank ( PDBID: 4QHE ) . Structure files of ligands were prepared for molecular docking by defining the number of torsion angles , addition of hydrogen atoms and conversion into software specific file format ( pdbqt ) . Similarly , APE1 protein was also prepared by removing bad contacts , addition of hydrogen atoms , removal of needless water molecules , and conversion of file format into pdbqt . 5 Å near residues 68−74 and 266−273 on APE1 were defined as E3330 binding site over APE1 [36 , 61] . First , a prepared ligand was virtually docked against APE1 protein blindly . Then , the ligand was docked at the defined binding site using Lamarckian Genetic Algorithm of Autodock 4 . 2 . The GPU accelerated Amber Molecular Dynamics suite with Amber ff99SB force field was employed for the all atoms explicit MD simulations of receptor-ligand complexes ( http://ambermd . org/#Amber12 ) . The receptor-ligand complexes were solvated with TIP3P water model in a cubic periodic boundary box to generate required systems for MD simulations and systems were neutralized using appropriate number of counterions . The distance between box wall and the complex was set to greater than 10Å to avoid direct interaction with its own periodic image . Neutralized system was then minimized , heated up to 300 K temperature and cooled down to equilibrated until the pressure and energies of systems were stabilized . Finally , the equilibrated systems were used to run 40 ns MD simulations . The average structures were calculated from the equilibrated stage of the MD trajectories ( from 35 to 40 ns ) and subsequently optimized with steepest descents for 200 steps . The minimized average structure was then used for the covalent docking . DOCKovalent [82] is a covalent adaptation of DOCK3 . 6 [83 , 84] . Given a pre-generated set of ligand conformation and a covalent attachment point , it exhaustively samples ligand conformations around the covalent bond and selects the lowest energy pose using a physics-based energy function . For the docking reported in this work , Cys65 residue was defined as attachment site . The bond angle ( C-ligand covalent attachment point-rest of ligand ) of 109 . 5 ± 10° , also in 2 . 5° increments . Scoring was as described [82] using a physics-based energy function which uses pre-calculated van der Waals electrostatics ( calculated with DELPHI ) , and solvent-excluded de-solvation [83] grids . The receptor is kept fixed throughout the docking simulation . All data were analyzed by two-tailed Student’s t-test and one-way ANOVA in GraphPad Prism , followed by comparisons performed using Bonferroni method . p values < 0 . 05 were considered significant ( *P<0 . 05 , **P<0 . 01 and ***P<0 . 001 ) .
|
As a major AIDS-associated malignancy , Kaposi’s sarcoma ( KS ) is caused by Kaposi’s sarcoma-associated herpesvirus ( KSHV ) . Currently there is no definitive cure for KS . In this study , we identified a cellular protein , namely APE1 , as an effective therapeutic target for blocking KSHV replication and inhibiting the development of KS phenotypes . We showed that the redox function of APE1 is absolutely required for KSHV replication , virally induced cytokine secretion and angiogenesis . Blockade of APE1 expression or inhibition of APE1 redox activity led to inhibition of KSHV replication and reduction of cytokine release and angiogenesis . Furthermore , we identified a novel small molecular compound , C10 , which exhibited specific inhibitory activity on APE1 redox function and was demonstrated to efficiently inhibit KSHV replication and paracrine-mediated KS phenotypes such as angiogenesis and cell invasion . As a potent inhibitor of APE1 redox , C10 not only has value in development of a novel therapeutics for KS , but also may be used in therapies for other human diseases such as leukemia , pancreatic cancer and macular degeneration .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"cytokines",
"pathology",
"and",
"laboratory",
"medicine",
"cardiovascular",
"physiology",
"pathogens",
"immunology",
"microbiology",
"cancers",
"and",
"neoplasms",
"viral",
"structure",
"angiogenesis",
"viruses",
"oncology",
"developmental",
"biology",
"dna",
"replication",
"dna",
"viruses",
"molecular",
"development",
"dna",
"herpesviruses",
"oxidation-reduction",
"reactions",
"medical",
"microbiology",
"microbial",
"pathogens",
"chemistry",
"kaposi's",
"sarcoma-associated",
"herpesvirus",
"kaposi",
"sarcoma",
"viral",
"replication",
"sarcomas",
"virions",
"immune",
"system",
"biochemistry",
"electrochemistry",
"nucleic",
"acids",
"virology",
"physiology",
"viral",
"pathogens",
"genetics",
"biology",
"and",
"life",
"sciences",
"chemical",
"reactions",
"physical",
"sciences",
"organisms"
] |
2017
|
An APE1 inhibitor reveals critical roles of the redox function of APE1 in KSHV replication and pathogenic phenotypes
|
The growing availability of high-quality genomic annotation has increased the potential for mechanistic insights when the specific variants driving common genome-wide association signals are accurately localized . A range of fine-mapping strategies have been advocated , and specific successes reported , but the overall performance of such approaches , in the face of the extensive linkage disequilibrium that characterizes the human genome , is not well understood . Using simulations based on sequence data from the 1000 Genomes Project , we quantify the extent to which fine-mapping , here conducted using an approximate Bayesian approach , can be expected to lead to useful improvements in causal variant localization . We show that resolution is highly variable between loci , and that performance is severely degraded as the statistical power to detect association is reduced . We confirm that , where causal variants are shared between ancestry groups , further improvements in performance can be obtained in a trans-ethnic fine-mapping design . Finally , using empirical data from a recently published genome-wide association study for ankylosing spondylitis , we provide empirical confirmation of the behaviour of the approximate Bayesian approach and demonstrate that seven of twenty-six loci can be fine-mapped to fewer than ten variants .
As a result of linkage disequilibrium ( LD ) , loci identified by common variant genome-wide association analyses are often large and may contain hundreds of variants . Characterization of the specific variants driving these signals has clear benefits , most obviously through highlighting the specific causal molecular mechanisms , which may provide insight into complex disease pathophysiology . Progress in this endeavor has been hampered by incomplete coverage of human sequence variation ( so potential causal variants are missing from genome-wide association study [GWAS] and imputation data sets ) , and by lingering uncertainties concerning the frequency spectrum of the underlying risk-alleles [1–3] . However , whole-genome sequence data ( e . g . from the 1000 Genomes Project ) now offer near complete coverage across the common and low frequency allele ranges at least ( minor allele frequency [MAF] >1% ) in multiple ancestry groups [4] . At the same time , evidence is accumulating—from large-scale GWAS [3 , 5] , fine-mapping [5–7] , re-sequencing [2 , 8] and trans-ethnic studies [9 , 10]–that most ( though not all ) GWAS signals are driven by causal variants that are themselves common , with evidence supporting alternative rare variant association models ( e . g . synthetic association [1] ) restricted to a few loci ( e . g . NOD2/CARD15 in inflammatory bowel disease [8] ) . These advances make it possible to perform systematic fine-mapping studies that interrogate the vast majority of potential causal variants at GWAS loci . Indeed , several recent studies have reported successful refinement of GWAS loci [6 , 7] . These efforts often include a trans-ethnic component that seeks to leverage ancestral differences in LD patterns between common SNPs to aid causal variant localization [9–11] . However , substantive questions related to the reliability , precision and performance of GWAS fine-mapping efforts remain unanswered . In this study we combined simulation-based and empirical approaches to address these questions .
To quantify fine-mapping precision , we simulated ten association studies of 1 , 000 cases and 1 , 000 controls in 1Mb regions from the CEU haplotypes present in the 1000 Genomes Phase 1 dataset [4] . In each region we assigned a single , additive effect , causal variant of varying odds ratio ( OR ) and risk allele frequency ( RAF ) . For each combination of OR and RAF , we simulated 1000 random “replicate” regions . On these data , we performed case-control association testing under an additive model , followed by meta-analysis of all ten studies for each region ( such that the total sample size examined was 20 , 000 ) . On the basis of the meta-analysis summary statistics for each of the variants , we generated approximate Bayes’ factors ( ABFs ) –describing the evidence in favor of association for a given variant–using the method proposed by Wakefield [12] . From these ABFs , the posterior probability of each variant to be causal was derived , and used to assemble “credible sets” that contained all variants with a cumulative posterior probability of a causal variant exceeding a chosen threshold ( here , 95% or 99% ) for each of the 1000 replicate regions [6] . The haplotypes used for the simulations are derived from whole-genome sequencing data for the 1000 Genomes project [4] , and as such represent a “gold standard” scenario with nearly the entire low-frequency and common variant spectrum covered . Although whole-genome sequence data are likely to offer the most accurate framework for fine-mapping , currently available whole-genome sequencing datasets are relatively small and lack power to discriminate between highly-correlated variants . As a result , most fine-mapping studies rely on the much larger datasets available from the meta-analysis of GWAS and/or custom array datasets after imputation . To represent such designs , we filtered these whole-genome data to those present on the Illumina HumanOmni2 . 5 BeadChip array , corresponding to genotype data equivalent to those from dense GWAS or a custom fine mapping array such as the Metabochip [13] . For this “GWAS” scenario , these variants were subsequently imputed up to the 1000 Genomes Phase 1 , all ancestries , release 3 panel . Additionally , we created a “GWAS with failure” scenario–representing a GWAS with a genotype missingness rate per cohort–where , for each of the twenty case or control cohorts ( each with n = 1000 ) in an analysis , we removed all genotypes from 5% of the variants ( after downsampling ) . This equates therefore to an ~0 . 25% missingness rate per variant across each set of 20 , 000 samples . From the simulated data , we first sought to establish whether the credible sets generated from these simulations by the approximate Bayesian approach had appropriate coverage of the simulated causal variant under a model of association . In the gold standard scenario ( i . e . unfiltered 1000 Genomes data ) , where the causal variant would always be present in the sample , we estimated the median coverage of the causal variant at 98 . 5% and 99 . 8% , for the 95% and 99% credible sets , respectively ( S1 Fig ) . In the downsampled GWAS scenario , coverage of the causal variant was broadly comparable to that under the gold standard – 97 . 8% for the 95% credible set and 99 . 6% for the 99% credible set . The introduction of a random 5% variant drop-out per study population in the GWAS with failure scenario could result in loss from analysis , in some of the datasets , of the causal variant , or of variants that are influential with respect to the imputation scaffold . ABFs were rescaled by maximum effective sample size to compensate for genotype missingness . Results showed that the introduction of random missingness primarily affected simulations with little power to detect genome-wide significant association . At all other simulated OR and RAF combinations , missing genotype data led to only modest attrition of the causal variant coverage ( S1 Fig ) . Having shown that the credible set performance was well calibrated , we next asked how often the causal variant was identical to the most strongly associated ( lead ) SNP . In situations where the two show near complete overlap , more sophisticated fine-mapping strategies would be unnecessary . In the gold standard scenario where the causal variant had the highest OR ( 1 . 5 ) and RAF ( 50% ) , 79% of the simulations accurately identified the causal variant as the most strongly associated ( Fig 1 ) . However , when the risk allele was less common ( RAF 5% ) and the phenotypic effect more in line with that observed at more typical GWAS loci ( OR 1 . 1 ) , the probability of accurately identifying the causal variant by focusing on the most associated variant dropped to just 2 . 4% ( Fig 1 ) . All subsequently presented results are based on the GWAS scenario , as this most closely resembles the majority of contemporary fine mapping studies . In this scenario , the numbers ranged from 70% ( RAF 50% and OR 1 . 5 ) to 2 . 4% ( RAF 5% and OR 1 . 1 ) . These results confirm that focusing solely on the most significant variant is insufficient for identification of the causal variant at most common variant GWAS loci , so more elaborate approaches to fine-mapping–such as construction of credible sets–are needed . Next , we evaluated the properties of credible sets for fine-mapping over a range of association models . When the association signal was strong—e . g . RAF 50% and OR 1 . 5—the median 95% credible set contained only a single variant ( and two in the median 99% set ) . Ninety percent of all replicates generated credible sets containing <11 ( 95% ) and <14 variants ( 99% ) ( Fig 2A and 2B ) . In the simulations with lesser power to detect association ( RAF 10% , OR 1 . 2–58 . 5% power ) , the median set contained 15 variants at 95% and 24 variants at the 99% posterior probability cut-off , but the range was far larger ( the 90th centiles now were 96 and 183 variants–Fig 2A and 2B ) . Overall , the size of the credible set had a strong log-negative correlation with the power to detect genome-wide association ( r2 = 0 . 61 and r2 = 0 . 68 for the 95% and 99% credible sets–S2 Fig ) ; at lower power , fine-mapping ability was mostly attenuated . We then asked how many of the loci could be successfully fine-mapped . For this purpose , we deemed resolution of the credible set to fewer than ten variants to constitute success , on the grounds that this is a tractable number that would encourage researchers to undertake detailed functional evaluation . The success-rate for the 95% credible sets ranged from 88 . 4% for RAF 50% variants ( OR = 1 . 5 ) to 36 . 9% for RAF 10% ( OR = 1 . 2 ) variants in simulations with >50% power . The equivalent figures for the 99% credible sets were 85 . 5% and 28 . 4% respectively . In the least well performing simulation setting overall ( RAF 5% at OR = 1 . 1 ) , 95 . 3% of the replicates did not fine-map at all , with the credible sets containing > 500 variants within the 1Mb region . However , the results showed that the approximate Bayesian fine-mapping approach provided both good coverage of the causal variant and was often able to successfully refine loci to credible sets of a size that would support exhaustive functional follow-up . To establish the impact of different fine-mapping strategies and causal variant scenarios , we compared the approximate Bayesian fine-mapping approach to a more simplistic strategy of retaining all variants with r2 to the lead variant exceeding some pre-defined threshold ( 0 . 5 , 0 . 8 or 0 . 9 ) . We found that , in simulations with great power to detect genome-wide significant association , the approximate Bayesian approach resulted in smaller credible sets than sets based on r2 thresholds ( e . g . 0 . 8-fold reduction of r2 >0 . 9 variant set containing fewer than 10 variants compared to 95% credible sets at RAF 50% and OR 1 . 5 –Fig 3A ) . At replicates with less strong association signal , stringent r2 thresholds ( 0 . 9 , and , to a lesser extent , 0 . 8 ) often resulted in smaller fine-mapped variant sets than those generated by the approximate Bayesian approach ( Fig 3A ) . However , the smaller sets for the r2 approach at lower OR and RAF were accompanied by a substantial increase in the false-positive rate ( e . g . 24% of r2 >0 . 9 sets with fewer than 10 variants do not actually contain the causal variant at RAF 10% and OR 1 . 2 –Fig 3A ) . This increase in false-positives was not seen using approximate Bayesian credible sets ( Fig 3A ) . Trans-ethnic datasets are widely promoted as a useful adjunct to fine-mapping approaches . We therefore explored the value of adding samples from a distinct ancestral group , focusing on individuals of African descent . The reduced LD typically observed in African-descent populations should support improved fine-mapping , at least at those loci where the same causal variants are segregating across ancestries . To investigate this , we took the same design of 10 studies , each of 1000 cases and 1000 controls , but generated some to all of these studies ( 1 , 3 , 5 or 10 ) using the Yoruba ( YRI ) component of 1000 Genomes . The simulated causal variant was kept identical to that in the CEU simulations , and with the same effect size . However , the RAF for the causal variant in the YRI haplotypes diverged , and was typically lower than that observed in CEU . We focused these simulations on the less extreme GWAS scenarios–causal variant RAFs of 10% , 20% at OR = 1 . 2 ( defined by CEU data ) –and saw a marked improvement in fine-mapping resolution compared to the equivalent CEU-only study . For example , in an analysis of 5 CEU and 5 YRI samples , the probability that the lead GWAS variant was also the causal variant increased 1 . 7-fold ( 27% to 46%; S3 Fig ) for RAF of 10% and 1 . 3-fold ( 45% to 60%; S3 Fig ) for RAF of 20% . The increased fine-mapping resolution was also reflected in the credible sets: the number of successfully refined loci ( < 10 variants ) increased on average by 1 . 5-fold ( RAF 10% ) and 1 . 3-fold ( RAF 20% ) in the 50% YRI design compared to a CEU-only study ( Fig 3B ) . Failure to see incremental improvements in fine-mapping resolution as YRI proportion increased ( the results for 50% YRI are better than for 100% YRI ) could reflect the advantage in signal localization possible from analyses that can benefit from the divergent LD patterns across the two ethnicities . However , it might also be due to the causal variant RAF being derived from the CEU population: as demonstrated in the broad evaluation of the properties of approximate Bayesian fine-mapping above , ABFs are dependent on power , and substantially lower risk allele frequencies in the YRI population result in loss of association signal . Overall , these results confirm the utility of multi-ethnic design for fine-mapping purposes , at least when the assumption of a shared causal allele is realized . With the emergence of genome-wide functional annotations , such as those generated by the ENCODE , GENCODE and NIH Roadmap Epigenomics projects , these data can now be used to inform fine-mapping [14] . We considered whether including such prior information in the approximate Bayesian framework would further improve its performance . For illustration purposes , we chose to set an elevated prior on exonic variants , but any other functional annotation for which there is evidence of disproportionate functional impact , such as islet enhancers in type 2 diabetes [15] , could also have been used . We simulated one thousand 1Mb random regions containing a causal coding variant of OR 1 . 2 and RAF 10% as described ( see Methods ) , but in these replicates the causal variant was limited to exonic sequence . This was followed by fine-mapping with or without a functional prior of ten-fold greater probability on coding variants–roughly corresponding to the fold over-representation of coding variants observed in GWAS studies [16] . The weighted approach resulted in a 1 . 4-fold reduction in the 95% , and 1 . 5-fold reduction in the 99% credible set size compared to non-weighted sets derived from the same data ( S4A Fig ) . In line with this , there was an increase in the percentage of successfully fine-mapped loci in the weighted scenario , with 35 . 8% of the weighted 95% credible sets containing fewer than ten variants , compared to 26 . 9% for the non-weighted sets ( S4B Fig ) . For the 99% credible sets , the rates of successful fine-mapping were 25 . 4% and 17 . 0% respectively ( S4B Fig ) . Using a model-based approach , it has been shown that , when causal variation was limited to small genomic regions ( 100 loci of 10kb each ) containing a large amount of the total trait variance ( 25% ) [16] , functional priors increased fine-mapping performance . Our results show that similar improvements in fine-mapping results can be obtained using the approximate Bayesian approach at loci containing variants which explain much less of the total heritability . All the above simulations assume a single causal variant at each locus . To determine how the approximate Bayesian approach would deal with the presence of multiple signals at a locus , unbeknown to the investigator , we simulated 1000 random regions containing two distinct causal variants of equal RAF ( 10% ) and effect size ( OR 1 . 2 ) . As might be expected , the size of the 95% credible sets showed a 1 . 3-fold increase ( median 15 versus 19 variants with a single versus two causal variants present ) , whereas the 99% credible sets increased by 1 . 5-fold ( 24 versus 35 variants–S3C Fig ) . This was accompanied by a reduction in the number of successfully refined loci ( 1 . 3-fold; S3D Fig ) . However , this still meant that at 30 . 5% of replicate loci at least one causal variant was included in a 95% credible set containing fewer than ten variants–only in 5 . 2% were both causal variants present in such a set . In practice , a preliminary round of conditional analysis would allow such instances of multiple association signals at the same locus to be detected , and fine-mapping efficiency could then be maximized by considering each of the component signals separately . Finally , we used an empirical data set to investigate if we could recapitulate the effectiveness of the fine-mapping observed in the simulations . We applied the approximate Bayesian approach ( as described in the Methods ) to data for twenty six non-MHC loci genotyped at high density ( using the Illumina Immunochip with subsequent 1000 Genomes European ancestry imputation ) for 9 , 049 cases and 13 , 607 controls of European origin from a published association study for ankylosing spondylitis ( AS ) [5] . The mean RAF 32% ( range 5–48% ) and OR 1 . 17 ( 1 . 11–1 . 65 ) of the reported lead variants at these loci were broadly in line with the simulated datasets . Approximate Bayesian fine-mapping resulted in median 95% and 99% credible sets containing 20 ( 1–295 ) and 38 ( 2–1 , 113 ) variants , respectively ( S1 Table ) . In line with the simulations , the success of fine-mapping across these 26 loci was correlated with the effect size and RAF at each locus ( and hence with the power to detect the locus in association testing–Fig 4A ) . Several of the loci that lie most outside the interquartile range of observed 95% credible set sizes in the simulations show extensive LD: index variants at the IL27 , NPEPPS and SH2B3 loci all are in LD ( r2>0 . 1 ) with a proportion of variants in the 1Mb interval that compare to the top 10 quantile of simulated loci . Seven of the twenty six loci ( 27% ) were reduced to fewer than ten 95% credible variants ( S1 Table ) . Overall , twenty four of the twenty six credible sets derived from our European-only analysis included the lead variant reported in the larger multi-ethnic meta-analysis from Cortes et al . [5] . One example of our approximate Bayesian approach revealing strong functional candidates can be found at the FCGR2A locus . The 99% credible set for this locus , which is associated not only with AS , but with a range of other autoimmune conditions [5 , 17] , contained only six variants spanning 11kb of the promoter region and first 3 exons of FCGR2A ( Figs 4B and S5 ) . One of these six variants within the credible set was a missense variant ( rs1801274 ) in FCGR2A , which encodes an immunoglobulin Fc receptor gene found on the surface of many immune response cells . In conclusion , simulated and empirical data analyses demonstrate that fine-mapping represents an effective strategy for causal variant localization , at a subset of loci at least . The simulations also emphasize that trans-ethnic study designs can improve resolution further . It is still true that the vast majority of GWAS data is generated from individuals of European descent , and even trans-ethnic studies such as Cortes et al . generally contains only a modest non-European contribution ( ~14% of the total sample size ) . Therefore , empirical assessment of the full value of trans-ethnic over ethnic-specific fine-mapping will ultimately depend on the generation of large scale non-European GWAS data . The simulations clearly show that fine-mapping is sample size dependent , so the utility of this approach is dictated by the availability of large-scale data . Fine-mapping resolution is most precise at strong association signals , such that for a variant with allelic OR of 1 . 2 and RAF of 50% , fine-mapping in 10 , 000 cases and 10 , 000 controls of European origin can be expected to reduce the credible set to fewer than ten variants at around 60% of loci . For less strong signals , resolution is typically less good , but , even here , the generation of well-calibrated credible sets facilitates integration with genomic annotation data . Indeed , one advantage of the approximate Bayesian framework used here is that prior information from external data resources such as ENCODE [18] , can readily be included to upweight highly-annotated variants . This can help prioritize specific signals and aid the selection of appropriate functional assays for follow-up . In addition , fine-mapping strategies utilizing ABFs are computationally inexpensive and can be applied to publicly available association summary statistics from GWAS data without the need to access individual level genotypes . This should allow rapid deployment of fine-mapping analyses to existing GWAS data sets , and encourage efforts to convert common variant GWAS data into an improved understanding of disease biology .
We used sequence data from 85 CEU individuals from 1000 Genomes Phase 1 , release 3[4] ( single nucleotide variants only ) , to simulate 1Mb autosomal genomic regions with HAPGEN2 [19]–an algorithm which resamples known haplotypes and thus maintains experimentally-derived LD patterns . When evaluating the effect of multi-ethnic study design on the fine-mapping resolution , one , three , five or all of the ten studies were simulated based on haplotypes from 88 YRI rather than 85 CEU individuals from 1000 Genomes Phase 1 , release 3 . One thousand non-overlapping 1Mb regions were randomly selected from the mappable human genome reference . In the center of each region , causality was assigned to a single causal variant of specified risk allele frequency ( RAF; 50% , 20% , 10% or 5% ) with an additive phenotypic effect ( OR 1 . 5 , 1 . 2 , 1 . 1 and 1 . 0 , the last corresponding to the null model of no association ) . In the evaluation of the impact on fine-mapping of multiple causal variants in a region , 1000 regions containing two distinct variants ( CEU r2<0 . 05 ) with a MAF~10% in the central 750kb of the 1Mb interval were selected . Haplotypes were simulated using HAPGEN2 with , with each of the two variants assigned an additive phenotypic effect ( OR = 1 . 2 ) . For each RAF/OR parameter setting , genotypes were generated for ten “studies” of 1 , 000 cases and 1 , 000 controls for each of the one thousand “replicate” genomic regions . Case-control association analysis was performed by logistic regression under an additive model , implemented in SNPTEST , and summary statistics from the ten studies aggregated by fixed-effects meta-analysis using GWAMA [20] . We performed these analyses using an unfiltered set of 1000 Genome-derived genotypes , equivalent to an association study performed on genome-wide sequence data ( “gold standard” scenario ) . To represent a more typical fine-mapping scenario , variants from each study were downsampled to the content of an Illumina HumanOmni2 . 5 BeadChip array ( “GWAS” scenario ) . This is also a similar density to that achieved within established GWAS regions using recent custom arrays such as Metabochip [13] or Immunochip [21] . To capture the effects of genotyping failure , downsampling was performed with and without 5% random variant failure per cohort ( n = 1000 cases or controls ) in each simulated GWAS ( “GWAS with failure” scenario ) . These GWAS genotypes were then imputed ( using IMPUTE2 [22] ) up to the 1000G Phase 1 , all ancestries release 3 panel before analysis as above . Only well-imputed variants ( INFO score >0 . 4 ) with a MAF > = 1% were included in further analyses . While privacy concerns mean that individual-level genotype data for GWAS are generally unavailable to researchers , large amounts of summary statistics are available online . Summary data do not allow full specification of disease and null models , but the information contained within can be used in a Bayesian framework by instead approximating Bayes’ factors as proposed by Wakefield in 2007 [12] . This approximation assumes that the likelihood distribution for association is summarized by the regression parameters with a prior for association centered on 0 ( which corresponds to the null of no association ) and variance dependent on W , which describes the strength of association conditional on its existence . The value for W is set at 0 . 4 , which equates to a 95% belief that the relative risk corresponding to departure from the null model is less than 1 . 5 . As demonstrated by Wakefield [12] , this results in an equation for the approximate Bayes Factor ( ABF ) given as ABF=11−rexp ( −Z22r ) where Z is the Z-statistic describing the strength of association derived from the regression and r a shrinkage factor r=WV+W defining the ratio of prior variance to total variance . The resulting ABF is dependent the effect size through Z and the power of the study through the variance of the effect V . ABFs for variants with missing genotype data were corrected by rescaling variance to the maximum observed effective sample size ( provided sample size was within 30% of the maximum observed sample size ) . Posterior probabilities for association were calculated based on the ABFs all variants in each of the simulated regions . For assessing the relation between fine-mapping performance and power to detect genome-wide association , we used Quanto v1 . 2 . 4 ( available from http://biostats . usc . edu/Quanto . html ) . The software was run in “Gene only” setting , assuming a disease prevalence of 0 . 55% in line with that of ankylosing spondylitis . The power for each combination of OR and RAF , for both simulations and empirical data , was calculated individually .
|
Over the last few years , several approaches for fine-mapping genome-wide association studies ( GWAS ) loci have been proposed and used to localize potential causal variants . However , the performance of these types of tests is often poorly characterized . In this study , we used extensive simulations to show that statistical fine-mapping can indeed accurately reduce the number of likely causal variants at common GWAS loci . These approaches can be further improved by changes in study design , such as the inclusion of multiple ethnic groups in the study population . Finally , we demonstrate the utility of this type of approach on a recently published genome-wide association study for ankylosing spondylitis , where we could fine-map seven of the twenty-six loci to a number of variants ( n = 10 ) which is tractable for follow-up in a laboratory setting .
|
[
"Abstract",
"Introduction",
"Results",
"Methods"
] |
[] |
2015
|
Evaluating the Performance of Fine-Mapping Strategies at Common Variant GWAS Loci
|
Hypomethylating agents reactivate tumor suppressor genes that are epigenetically silenced in cancer . Inevitably these genes are resilenced , leading to drug resistance . Using the MLH1 tumor suppressor gene as a model , we showed that decitabine-induced re-expression was dependent upon demethylation and eviction of promoter nucleosomes . Following decitabine withdrawal , MLH1 was rapidly resilenced despite persistent promoter demethylation . Single molecule analysis at multiple time points showed that gene resilencing was initiated by nucleosome reassembly on demethylated DNA and only then was followed by remethylation and stable silencing . Taken together , these data establish the importance of nucleosome positioning in mediating resilencing of drug-induced gene reactivation and suggest a role for therapeutic targeting of nucleosome assembly as a mechanism to overcome drug resistance .
The DNA hypomethylating agents decitabine ( 5-aza-2′deoxycytidine ) and azacitidine ( 5-azacytidine ) are established therapies for myeloid malignancies and show promise in treating solid tumors [1] . These drugs are cytidine analogs that covalently trap the DNA methyltransferase I ( DNMT1 [NP_001124295] ) protein onto DNA , targeting the enzyme for proteasome degradation . The resulting depletion of DNMT1 leads to passive demethylation in dividing cells . The observed effects of low dose decitabine on cell growth , differentiation [2] and enhanced immunological responses to tumor-associated antigens [3] are thought to be due to the re-expression of critical genes silenced by aberrant promoter hypermethylation . Sustained gene re-expression has been associated with clinical response [4] , [5] , supporting the view that it is critical to the therapeutic mechanism of action of these drugs . Clinically relevant low doses of decitabine and azacitidine can lead to sustained changes in gene expression that are associated with reduced tumorigenicity in mice bearing transplanted tumor xenografts [6] . Like all anti-cancer therapies , resistance to hypomethylating agents ultimately develops , and without alternative therapies patients succumb quickly to their disease [7] . A variety of mechanisms have been proposed to explain resistance , including insufficient drug uptake by membrane transporters , deficiency of the enzyme required for drug activation ( deoxycytidine kinase ) , or increased drug metabolism through deamination by cytidine deaminase [8] . Although these mechanisms may explain the resistance of some cell lines in vitro [8] , a recent study showed they do not explain acquired resistance in patients [9] . Numerous in vitro studies show that gene re-expression following decitabine treatment is transient [10] , [11] . The silencing of genes initially re-expressed by decitabine treatment is termed gene resilencing [12] , [13] . If sustained gene re-expression correlates with clinical response then gene resilencing is likely to play a role in the development of drug resistance . Therefore , understanding the mechanistic basis of gene resilencing is a prerequisite for the development of therapies that cause sustained gene re-expression and prolonged clinical response . The rapid onset of gene resilencing is unlikely to be explained by DNA remethylation , which in vitro studies have shown occurs gradually over several weeks [11] , [13] . Also drug resistance occurs despite persistence of DNA hypomethylation [9] . This suggests other epigenetic mechanisms , such as nucleosome positioning or histone modifications , are responsible for driving the resilencing of genes . We reasoned that promoter nucleosome positioning could explain gene resilencing independently of DNA remethylation . To address this hypothesis we mapped the temporal onset of epigenetic changes at the MLH1 [NM_000249] gene promoter following exposure of RKO cells to decitabine . We used this model because the MLH1 gene is biallelically methylated and silent in this cell line and activation of the promoter with decitabine has been shown to involve nucleosome eviction [14] .
Optimization experiments showed that 72 hours of decitabine treatment at a concentration of 2 . 5 µM led to near maximal re-expression of MLH1 and depletion of DNMT1 protein in RKO cells ( Figure 1A , B ) . Using these conditions we profiled global methylation changes at baseline ( day 0 ) , as well as throughout decitabine treatment ( days 1–3 ) and during recovery ( days 4–45 , Figure 1C ) . As expected , we observed a decrease in global methylation from 3 . 7%±0 . 2 in untreated RKO cells to 0 . 8%±0 . 1 at day 4 . Global methylation levels remained low up to day 8 and recovered gradually to near baseline levels of 3 . 2%±0 . 2 by day 45 ( Figure 1D ) . In SW620 cells , global remethylation was also observed although it occurred more slowly than in RKO cells ( Figure 1E ) . Prior to treatment , MLH1 mRNA levels in RKO cells were undetectable ( Figure 2A ) . Methylation analysis of the regions indicated in Figure 2B showed 95% methylation by pyrosequencing , whilst allelic bisulfite sequencing showed hypermethylation on 100% ( 25/25 ) of promoter molecules ( Figure 2A , C and D ) . Two days after withdrawal of decitabine ( day 5 ) MLH1 mRNA reached maximal levels ( Figure 2A ) . This coincided with a near maximal decrease in promoter methylation , which dropped to 46 . 3% at day 5 ( Figure 1D ) , with 31 . 3% ( 10/32 ) molecules showing complete or near complete demethylation ( defined as no more than two methylated CpG dinucleotides per molecule; Figure 2C ) . Having demonstrated that short-term exposure to decitabine had re-expressed MLH1 we then determined the temporal onset of MLH1 resilencing in RKO cells by profiling MLH1 mRNA levels up to day 45 . The initial stages of resilencing ( defined as the point when mRNA levels begin to decrease after decitabine withdrawal ) started at day 6 , and by day 17 expression levels were only 17 . 5% of the level of maximal MLH1 expression observed at day 5 ( Figure 2A ) . However , throughout this period ( up to day 8 ) methylation levels remained low . For example , by day 8 methylation levels were very similar to day 5 at 51 . 3% , with 35 . 6% ( 16/45 ) demethylated promoter molecules despite mRNA levels declining by over 50% ( Figure 2A , C , D ) . These results show that promoter remethylation does not precede or coincide with MLH1 resilencing . Interestingly , when remethylation did occur ( between days 8 and 17 ) , we found preferential methylation of five CpG sites immediately upstream of the MLH1 transcription start site ( TSS; Figure 2C ) that are critical to the regulation of expression [15] . In SW620 cells we also observed fluctuations in MLH1 expression levels ( between 1 . 5-fold and 0 . 4-fold relative to baseline at day 0 ) during and after decitabine exposure , though these changes were not related to DNA methylation ( Figure S1 ) . We next profiled histone modification changes using native ChIP , focusing on the ‘active’ histone marks H3K9ac and H3K4me3 ( Figure 2E and Figure S1C ) . For control experiments we measured the levels of these histone modifications at the GAPDH [NM_002046] promoter ( Figure S1C ) . Prior to treatment with decitabine the levels of H3K9ac and H3K4me3 at the MLH1 promoter were either very low or undetectable ( Figure 2E ) . Levels of H3K9ac and H3K4me3 increased and decreased with a very similar trend to the re-expression and resilencing of MLH1 . By contrast , in SW620 cells the levels of H3K4me3 and H3K9ac at the MLH1 promoter were comparable with levels at the GAPDH promoter as expected due to the high levels of MLH1 expression in this cell line ( Figure S1C ) . We also found the levels of the ‘repressive’ histone modification H3K27me3 remained very low or undetectable at the MLH1 promoter throughout treatment in RKO cells ( data not shown ) . As a control for H3K27me3 , we used primers specific for the MYOD1 [NM_002478] promoter , which is enriched for this histone modification . Taken together these data confirm that MLH1 re-expression is accompanied by promoter DNA demethylation and acquisition of the active histone marks H3K9ac and H3K4me3 . These data also demonstrate that MLH1 re-expression is tightly linked to DNA demethylation whereas the early stages of resilencing are independent of DNA remethylation . Having demonstrated the relationship between MLH1 expression , promoter demethylation and active histone modifications , we sought to determine how nucleosome levels across the promoter change during and after decitabine exposure . We firstly determined nucleosome levels and positioning in untreated RKO cells . This was done using MNase digestion coupled with qPCR ( MNase-qPCR ) at nine regions across the MLH1 promoter ( Regions I–IX; Figure 3A ) , as well as with Nucleosome Occupancy and Methylome Sequencing ( NOMe-Seq; Region N1 in Figure 3A ) . This showed dense nucleosome occupancy across the MLH1 promoter and precisely mapped the positions of two nucleosomes , one at the MLH1 TSS and one within exon 1 ( Figure 3A , B ) . This was evident from a small region of DNA accessible to the GpC methyltransferase M . CviPI ( asterix , Figure 3A ) and MNase ( Region V , Figure 3B ) indicating a region of linker DNA between two adjacent and precisely positioned nucleosomes . Furthermore , MNase-qPCR detected strong nucleosome positioning at Regions III and VI flanking this site of MNase and M . CviPI accessibility ( Figure 3B ) . Decitabine-induced MLH1 re-expression was associated with the eviction of these nucleosomes ( Figure 3C–E ) , as well as nucleosomes from all regions across the MLH1 promoter ( Figure S2A–G ) . By day 3 ( final day of decitabine exposure ) nucleosome levels at Regions III and VI were lowest at 25 . 4% and 22 . 6% respectively , relative to levels in untreated cells . These results confirm that in addition to DNA demethylation , MLH1 re-expression is dependent upon eviction of promoter nucleosomes , as described in previous reports [14] . In SW620 cells ( normally expressing MLH1 ) we found much lower levels of nucleosome occupancy at Regions III and VI . Interestingly , decitabine treatment of SW620 cells also resulted in nucleosome eviction from these regions ( Figure S2H ) , which may explain the initial increase in MLH1 expression seen in SW620 cells following decitabine exposure ( Figure S1A ) . Next , we determined how nucleosome levels across the MLH1 promoter change during the initial stages of resilencing and compared this with the MLH1 expression and promoter methylation data described above . At day 5 , when MLH1 re-expression was maximal , nucleosome levels across the promoter remained low ( Figure 3C–E ) . Strikingly , by day 7 ( just 4 days after decitabine withdrawal ) we found that nucleosomes had reoccupied the MLH1 TSS and exon 1 and that this coincided with the decline of gene expression ( Figure 3C–E and Figure S2 ) . Restoration of nucleosome occupancy occurred despite the MLH1 promoter remaining hypomethylated ( Figure 3F ) suggesting that the reassembly of nucleosomes at the MLH1 TSS might initiate gene resilencing , and that this precedes DNA remethylation . Given our observation that nucleosome deposition occurred in the absence of remethylation it was important to determine nucleosome occupancy and DNA methylation on individual molecules . In doing so , we sought to determine whether nucleosome reassembly occurs on demethylated promoter molecules which would indicate that this was an initiating event in gene resilencing and a prerequisite for remethylation . We designed a second NOMe-Seq assay across the MLH1 TSS Region N2 , Figure 3A ) . This assay was designed to preferentially amplify unmethylated DNA , allowing us to determine nucleosome occupancy on DNA molecules that had been demethylated by decitabine treatment . Using this assay , we profiled nucleosome occupancy in untreated SW620 cells . We found that 91% ( 21/23 ) of molecules were nucleosome depleted across the TSS ( Figure 3G ) and accordingly , MLH1 was highly expressed ( Figure S1A ) . Next we analyzed decitabine treated RKO cells at day 5 when MLH1 re-expression was maximal . This revealed nucleosome eviction from the TSS on a proportion ( 9/22 ) of demethylated promoter molecules ( Figure 3H ) confirming that re-expression was associated with the eviction of nucleosomes , but also suggesting that some demethylated promoter molecules remain nucleosome bound . However , by day 7 , when resilencing of gene expression had begun ( Figure 3E ) , NOMe-Seq showed that all demethylated molecules assayed were nucleosome occupied across the TSS ( Figure 3I ) . This clearly shows that nucleosomes reassemble onto demethylated promoter molecules and that nucleosome occupancy rather than DNA methylation is associated with reduced gene expression .
This study shows that MLH1 resilencing is initiated by a rapid restructuring of chromatin architecture , characterized by the reassembly of nucleosomes at the TSS . This restructuring occurs prior to DNA remethylation , suggesting that gene resilencing following exposure to decitabine is controlled by a hierarchy of epigenetic events . Whilst previous studies have described decitabine-induced gene reactivation in detail , in this study we specifically focused on the molecular events associated with gene resilencing . This is technically challenging as it requires a model system to determine the temporal relationship between gene expression and promoter epigenetic changes , ideally at daily intervals . It also requires sampling of large numbers of cells to measure each variable at each time point . This renders such experiments impossible using material from patients receiving decitabine . Furthermore , variability in the molecular drivers between patients would make the identification of a model gene a major obstacle . To overcome these difficulties , we chose to track the resilencing of the MLH1 gene in RKO colorectal carcinoma cells following decitabine exposure . MLH1 is an archetypal gene inactivated by hypermethylation and loss of expression plays a pivotal role in the development of colorectal and other cancers [16] . This gene has been extensively epigenetically characterized using a variety of MLH1-specific assays [14] , [17] . Furthermore , RKO cells show biallelic hypermethylation of the MLH1 promoter allowing us to examine epigenetic changes on a homogeneous population of silent MLH1 promoter molecules at baseline . By measuring MLH1 expression levels at daily intervals we were able to precisely pinpoint the initiation of resilencing . This then allowed us to demonstrate that resilencing began when the MLH1 promoter remained maximally demethylated , which confirms previous reports that the gradual rate of global and site-specific DNA remethylation cannot explain the swiftness of gene resilencing [11] , [13] . Instead , we found that MLH1 resilencing was tightly linked to nucleosome position and histone modifications . In addition to the loss of H3K4me3 and H3K9ac , we found that nucleosome levels rapidly recovered following decitabine exposure , that nucleosome reassembly coincided with the decline of MLH1 expression , and that nucleosomes reoccupied the TSS of demethylated molecules . Our data suggest that the reassembly of nucleosomes at the TSS is a prerequisite for remethylation and that this is an important factor in determining the future epigenetic state of the reactivated MLH1 promoter . A limitation of our study is that these data relate to one cell line and one promoter ( MLH1 ) , potentially impacting on the generalizability of our findings . However , our proposition that nucleosomes initiate gene resilencing after drug exposure is supported by two recent studies , the first describing differentiation of NCCIT cells and the second describing the silencing of a GFP transgene . The study of You et al . showed that differentiation of NCCIT cells was associated with nucleosome assembly at the shared NANOG/OCT4 enhancer and that this resulted in the loss of expression of these genes [18] . This study also showed that hypermethylation of the NANOG promoter and enhancer followed nucleosome assembly and gene silencing . In the second study , Si et al . tracked the resilencing of a GFP transgene driven by a cytomegalovirus ( CMV ) promoter and showed that histone H3 density increased at the CMV promoter five days after withdrawal of decitabine [13] . By combining our findings with those of previous studies we have constructed a model describing MLH1 resilencing following decitabine exposure ( Figure 4 ) . Prior to treatment , MLH1 is silent and the promoter is methylated and occupied by nucleosomes ( Figure 4A , F and G ) . Decitabine-induced re-expression is associated with demethylation and nucleosome eviction from the TSS ( Figure 4B , F , and G ) as shown in this study and by others [14] . Nucleosome eviction from the TSS is associated with increased H3K9ac and H3K4me3 at remaining promoter nucleosomes . We found that trimethylation of H3K4 was tightly linked to MLH1 expression levels , which agrees with a previous report that H3K4me3 is required for anchoring the TFIID transcription factor subunit of the RNApolII complex [19] . The initial stages of resilencing are associated with the loss of H3K9ac and H3K4me3 , which most likely coincides with the loss of RNApolII from the promoter and nucleosome reassembly at the TSS ( Figure 4C ) . Note that gene resilencing occurs when nucleosomes re-enter the promoter on demethylated molecules ( Figure 4C , F and G ) . Finally , gradual remethylation of the MLH1 promoter over several weeks consolidates the silenced state ( Figure 4D , E and G ) . We found that remethylation occurred preferentially at five CpG sites that overlap with the site of nucleosome reassembly ( Figure 4D ) . The preferential remethylation of these five CpG sites may be explained by a previous report describing the recruitment of DNMT3L-DNMT3A/B complexes to nucleosomes that were unmethylated at lysine 4 of histone H3 [20] . The driving force behind nucleosome reassembly is at present unclear but it may be dependent on the surrounding chromatin context . For example at bivalent promoters , which are characterized by the presence of H3K4me2/3 and H3K27me3 , complete and rapid epigenetic repression might be reinstated due to the persistence of H3K27me3 [21] . In our study , we found that H3K27me3 remained very low or undetectable throughout treatment indicating this is unlikely to trigger resilencing at the MLH1 locus . It is possible however that other histone modifications could be driving gene resilencing prior to remethylation . Interestingly , the persistence of the RNApolII complex at the promoter of TMS1 is a critical factor in the long-term stability of decitabine-induced reactivation [10] . We propose that at reactivated promoters , equilibrium exists between the binding of the RNApolII complex and the reassembly of nucleosomes . Reduced binding of the RNApolII complex would invite nucleosome reassembly at the TSS , preventing further binding of the RNApolII complex and ultimately leading to promoter remethylation . This hypothesis is consistent with previous reports showing that continued binding of the RNApolII complex protects against de novo methylation [22] . A key question is whether gene resilencing is a result of active chromatin re-modeling or clonal expansion of cells that did not respond to decitabine treatment . Although we did not measure cell death in treated cells , we consider two components of our data strongly support our conclusion that resilencing of MLH1 is an active process . Firstly , nucleosome levels at the MLH1 promoter recover before DNA methylation levels . This stepwise recovery in chromatin structure argues against passive resilencing , which would be associated with simultaneous re-emergence of repressive chromatin features ( methylated and nucleosome occupied DNA ) . Secondly , single-molecule analysis of the MLH1 promoter shows that nucleosomes reassemble onto demethylated molecules , and since MLH1 is biallelically hypermethylated prior to treatment , this shows that we are measuring changes within cells that were demethylated by decitabine exposure . Our finding that nucleosome occlusion of demethylated promoters can initiate gene resilencing has clear implications for the development of epigenetic therapies . Furthermore , our study raises the possibility that measurement of nucleosome occupancy at the TSS of critical genes may provide a more informative marker of emerging drug resistance than the measurement of promoter methylation .
The colorectal cancer cell lines SW620 ( MLH1 promoter unmethylated and gene expressed ) and RKO were maintained in DMEM media supplemented with 25 mM glucose , 10% ( v/v ) FBS , 100 units penicillin , 100 µg/mL streptomycin and 2 mM glutamate ( Life Technologies ) and grown at 37°C in 5% CO2 . Cells were treated every 24 hours for a period of 72 hours by replacing media supplemented with the indicated concentrations of decitabine ( 5-aza-2′deoxycytidine , Sigma ) freshly prepared in 50% filter sterilized acetic acid . RNA was extracted using PureLink Micro Kit ( Life Technologies ) . cDNA was prepared using the SuperScript III cDNA Synthesis Kit ( Life Technologies ) as per the manufacturer's instructions . Real-time quantitative reverse transcriptase PCR ( qRTPCR ) was performed in triplicates using 10 ng cDNA using iQ SYBR Green supermix ( Bio-Rad ) and a MyiQ iCycler ( Bio-Rad ) . Please refer to supplementary files Table S1 and Text S1 for primer sequences and sources . Gene expression was normalized to glyceraldehyde-3-phosphate dehydrogenase ( GAPDH [NM_002046] ) or succinate dehydrogenase complex , subunit A ( SDHA [NM_004168] ) . Cells were lysed on ice in 50 mM Tris HCl pH 7 . 5 , 150 mM NaCl , 1% ( v/v ) Triton X-100 , 0 . 5% ( w/v ) deoxycholic acid , 0 . 1% ( w/v ) sodium dodecyl sulphate ( SDS ) and EDTA-free Complete Protease Inhibitor ( Roche ) , vortexed and sonicated followed by centrifugation to pellet cell debris . Protein concentration was determined using the bicinchoninic acid protein assay ( Pierce ) following manufacturer's instructions . Proteins were resolved by SDS-PAGE , transferred to PVDF membrane ( Millipore ) and probed with 1 µg/mL anti-DNMT1 ( R & D systems ) or 9 . 6 ng/mL anti-α-tubulin ( Cell Signaling Technology ) before incubation with anti-IgG HRP . Proteins were visualized by enhanced chemiluminescence using Image Quant TL software and an Image Quant LAS 400 ( GE ) . Native ChIP was performed following micrococcal nuclease digestion of chromatin as described previously [26] . Relative enrichment of histone modifications were assessed using real-time quantitative PCR ( qPCR ) with primers listed in Table S1 . Primers specific to GAPDH and MYOD1 were used as controls for the enrichment of H3K4me3/H3K9ac and H3K27me3 , respectively . Enrichment was normalized to undigested input DNA after subtracting non-specific binding determined using pre-immune IgG . A total of 1×107 cells were lysed on ice in 50 mM Tris HCl pH 7 . 9 , 100 mM KCl , 5 mM MgCl2 , 50% ( v/v ) glycerol , 1 . 5% ( v/v ) β-mercaptoethanol , 0 . 1% ( w/v ) Saponin and Complete Protease Inhibitor with EDTA ( Roche ) followed by equilibration in 50 mM Tris HCl pH 7 . 5 , 0 . 32 mM sucrose , 4 mM MgCl2 , 1 mM CaCl2 , and EDTA-free Complete Protease Inhibitors ( Roche ) . Chromatin was digested using 20 U MNase ( Fermentas ) for 4 min at 37°C to achieve maximal digestion to mononucleosomes . Digestion was stopped by the addition of 20 mM EDTA pH 8 and placed immediately on ice . Cellular debris was pelleted and the supernatant treated with Proteinase K before isolation of DNA by phenol chloroform extraction and ethanol precipitation . Mononucleosomal DNA corresponding to 150 bp was isolated by gel extraction using a QIAquick gel extraction kit ( Qiagen ) . DNA concentration was measured using the Quant-iT PicoGreen dsDNA Assay Kit ( Invitrogen ) . Relative nucleosome levels were measured at nine sites at the MLH1 promoter ( designated Regions I–IX ) using primers listed in Table S1 . Nucleosome levels at each site were normalized to naked genomic DNA . NOMe-Seq was performed as described previously [27] . This involved treatment of intact nuclei with 200 U GpC methyltransferase M . CviPl for 15 min at 37°C followed by termination of the reaction with an equal volume of 20 mM Tris HCl pH 7 . 9 , 600 mM NaCl , 1% ( w/v ) SDS and 10 mM EDTA , and isolation of DNA as described above . DNA was bisulfite converted and amplified using primers listed in Table S1 . M . CviPI enzyme methylates accessible DNA at GpC sites , whereas nucleosome bound DNA is inaccessible and remains refractory to GpC methylation . PCR amplicons were cloned and individual molecules isolated by colony PCR for sequencing , as described above . Regions of M . CviPI inaccessibility of ≥150 bp were identified as nucleosome occupied .
|
Hypomethylating agents are emerging as effective cancer therapies . However , their therapeutic effects are transient and drug resistance inevitably develops . While resistance is associated with resilencing of genes initially demethylated by the drug , the mechanism underlying this resilencing is unknown . We provide evidence that the rapid reassembly of nucleosomes at transcription start sites initiates resilencing and is a prerequisite for promoter remethylation . This finding shows reassembly of nucleosomes at the promoter of critical genes is a potential early marker of resistance to hypomethylating agents . Our findings have implications for the treatment of cancer using epigenetic therapies that target DNA methylation alone , and suggest that overcoming drug resistance will require therapeutic strategies which prevent nucleosome deposition .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"oncology",
"medicine",
"epigenetic",
"therapy",
"gene",
"expression",
"genetics",
"cancer",
"treatment",
"epigenetics",
"biology",
"dna",
"modification",
"molecular",
"cell",
"biology",
"chromatin",
"dna",
"transcription",
"histone",
"modification"
] |
2013
|
Reassembly of Nucleosomes at the MLH1 Promoter Initiates Resilencing Following Decitabine Exposure
|
Reproduction is a defining feature of living systems . To reproduce , aggregates of biological units ( e . g . , multicellular organisms or colonial bacteria ) must fragment into smaller parts . Fragmentation modes in nature range from binary fission in bacteria to collective-level fragmentation and the production of unicellular propagules in multicellular organisms . Despite this apparent ubiquity , the adaptive significance of fragmentation modes has received little attention . Here , we develop a model in which groups arise from the division of single cells that do not separate but stay together until the moment of group fragmentation . We allow for all possible fragmentation patterns and calculate the population growth rate of each associated life cycle . Fragmentation modes that maximise growth rate comprise a restrictive set of patterns that include production of unicellular propagules and division into two similar size groups . Life cycles marked by single-cell bottlenecks maximise population growth rate under a wide range of conditions . This surprising result offers a new evolutionary explanation for the widespread occurrence of this mode of reproduction . All in all , our model provides a framework for exploring the adaptive significance of fragmentation modes and their associated life cycles .
A requirement for evolution—and a defining feature of life—is reproduction [1–4] . Perhaps the simplest mode of reproduction is binary fission in unicellular bacteria , whereby a single cell divides and produces two offspring cells . In more complex organisms , such as colonial bacteria , reproduction involves fragmentation of a group of cells into smaller groups . Bacterial species demonstrate a wide range of fragmentation modes , differing both in the size at which the parental group fragments and the number and sizes of offspring groups [5] . For example , in the bacterium Neisseria , a diplococcus , two daughter cells remain attached forming a two-celled group that separates into two groups of two cells only after a further round of cell division [6] . Staphylococcus aureus , another coccoid bacterium , divides in three planes at right angles to one another to produce grape-like clusters of about 20 cells from which single cells separate to form new clusters [7] . Magnetotactic prokaryotes form spherical clusters of about 20 cells , which divide by splitting into two equally sized clusters [8] . These are just a few examples of a large number of diverse fragmentation modes , but why should there be such a wide range of life cycles ? Do fragmentation modes have adaptive significance or are they simply the unintended consequences of particular cellular processes underpinning cell division ? If adaptive , what selective forces shape their evolution ? Can different life cycles simply provide different opportunities to maximise population growth rate ? A starting point to answer these questions is to consider benefits and costs of group living in cell collectives . Benefits may arise for various reasons . Cells within groups may be better able to withstand environmental stress [9] , escape predation [10 , 11] , or occupy new niches [12 , 13] . Also , via density-dependent gene regulation , cells within groups may gain more of a limiting resource than they would if alone [14 , 15] . On the other hand , cells within groups experience increased competition and must also contend with the build up of potentially toxic waste metabolites [16 , 17] . Thus , it is reasonable to expect an optimal relationship between group size and fragmentation mode that is environment and organism dependent [18–21] . Here we formulate and study a matrix population model [22] that considers all possible modes of group fragmentation . By determining the relationship between life cycle and population growth rate , we show that there is , overall , a narrow class of optimal modes of fragmentation . When the process of fragmentation does not involve costs , optimal fragmentation modes are characterised by a deterministic schedule and binary splitting , whereby groups fragment into exactly two offspring groups . Contrastingly , when a cost is associated with fragmentation , it can be optimal for a group to fragment into multiple propagules . Our results show that the range of life cycles observed in simple microbial populations are likely shaped by selection for intrinsic growth rate advantages inherent to different modes of group fragmentation . While we do not consider complex life cycles , our results may contribute to understanding the emergence of life cycles underpinning the evolution of multicellular life .
We consider a population in which a single type of cell ( or unit or individual ) can form groups ( or complexes or aggregates ) of increasing size by cells staying together after reproduction [18] . We assume that the size of any group is smaller than n , and denote groups of size i by Xi ( see the list of used variables in Table 1 ) . Groups die at rate di and cells within groups divide at rate bi; hence groups grow at rate ibi . The vectors of birth rates b = ( b1 , … , bn−1 ) and of death rates d = ( d1 , … , bn−1 ) make the costs and benefits associated to the size of the groups explicit , thus defining the “fitness landscape” of our model . Groups produce new complexes by fragmenting ( or splitting ) , i . e . , by dividing into smaller groups . We further assume that fragmentation is triggered by growth of individual cells within a given group . Consider a group of size i growing into a group of size i + 1 . Such a group can either stay together or fragment . If it fragments , it can do so in one of several ways . For example , a group of size 4 can give rise to the following five “fragmentation patterns”: 4 ( the group does not split , but stays together ) , 3+1 ( the group splits into two offspring groups: one of size 3 , and one of size 1 ) , 2+2 ( the group splits into two groups of size 2 ) , 2+1+1 ( the group splits into one group of size 2 and two groups of size 1 ) , and 1+1+1+1 ( the group splits into four independent cells ) . Mathematically , such fragmentation patterns correspond to the five partitions of 4 ( a partition of a positive integer i is a way of writing i as a sum of positive integers without regard to order; the summands are called parts [23] ) . We use the notation κ ⊢ ℓ to indicate that κ is a partition of ℓ , for example 2 + 2 ⊢ 4 . The number of partitions of ℓ is given by ζℓ , e . g . , there are ζ4 = 5 partitions of 4 . We consider an exhaustive set of fragmentation modes ( or “fragmentation strategies” ) implementing all possible ways groups of maximum size n can grow and fragment into smaller groups , including both pure and mixed modes ( Fig 1 ) . A pure fragmentation mode is characterised by a single partition κ ⊢ ℓ , i . e . , groups of size i < ℓ grow up to size ℓ and then fragment according to partition κ ⊢ ℓ . The partition κ can then be used to refer to the associated pure strategy . The total number of pure fragmentation strategies is ∑ ℓ = 2 n ( ζ ℓ − 1 ) , which grows quickly with n: There are 128 pure fragmentation modes for n = 10 , but 1 , 295 , 920 for n = 50 . A mixed fragmentation mode is given by a probability distribution over the set of pure fragmentation modes . The relationship between pure and mixed fragmentation modes is hence similar to the one between pure strategies and mixed strategies in evolutionary game theory [24] . One of our main results is that mixed fragmentation modes are always dominated by pure fragmentation modes . Hence , we focus our exposition on pure fragmentation modes , and leave the details of how to specify mixed fragmentation modes to the Supporting Information ( S1 Text , Appendix A ) . Together with the fitness landscape given by the vectors of birth rates b and death rates d , each fragmentation strategy specifies a set of biological reactions . Consider the pure mode κ ⊢ ℓ , whereby groups grow up to size ℓ and then split according to fragmentation pattern κ . A set of reactions X i → d i 0 , i = 1 , … , ℓ - 1 ( 1 ) models the death of groups; an additional set of reactions X i → i b i X i + 1 , i = 1 , … , ℓ - 2 ( 2 ) models the growth of groups ( without splitting ) up to size ℓ − 1 . Finally , one reaction of the type X ℓ - 1 → ( ℓ - 1 ) b ℓ - 1 ∑ i = 1 ℓ - 1 π i ( κ ) X i , ( 3 ) models the growth of the group from size ℓ − 1 to size ℓ and its immediate fragmentation in a way described by fragmentation pattern κ ⊢ ℓ , where parts equal to i appear a number πi ( κ ) of times . For instance , for the pure fragmentation mode 2 + 1 + 1 ⊢ 4 , Eq ( 3 ) becomes X 3 → 3 b 3 X 2 + 2 X 1 , which stipulates that groups of size 3 grow to size 4 at rate 3b3 and split into one group of size 2 and two groups of size 1; here , π1 ( 2 + 1 + 1 ) = 2 , π2 ( 2 + 1 + 1 ) = 1 , π3 ( 2 + 1 + 1 ) = 0 . The sets of reactions ( 1 ) , ( 2 ) and ( 3 ) give rise to the system of differential equations x ˙ 1 = − ( b 1 + d 1 ) x 1 + ( l − 1 ) b l − 1 π 1 ( κ ) x l − 1 , x ˙ i = ( i − 1 ) b i − 1 x i − 1 − ( i b i + d i ) x i + ( l − 1 ) b l − 1 π i ( κ ) x l − 1 , i = 2 , … , l − 1 where xi denotes the abundance of groups of size i . This is a linear system that can be represented in matrix form as x ˙ = A x , ( 4 ) where x = ( x1 , x2 , … , xℓ−1 ) is the vector of abundances of the groups of different size and A = ( − b 1 − d 1 0 ⋯ 0 ( l − 1 ) b l − 1 π 1 ( κ ) b 1 − 2 b 2 − d 2 0 ⋮ ( l − 1 ) b l − 1 π 2 ( κ ) 0 2 b 2 − 3 b 3 − d 3 0 ( l − 1 ) b l − 1 π 3 ( κ ) ⋮ ⋮ ⋱ ⋱ ⋮ 0 0 ⋯ ( l − 2 ) b l − 2 ( l − 1 ) b l − 1 ( π l − 1 ( κ ) − 1 ) − d l − 1 ) is the projection matrix determining the population dynamics . For any fragmentation mode and any fitness landscape , the projection matrix A is “essentially non-negative” ( or quasi-positive ) , i . e . , all the elements outside the main diagonal are non-negative [25] . This implies that A has a real leading eigenvalue λ1 with associated non-negative left and right eigenvectors v and w . In the long term , the solution of Eq ( 4 ) converges to that of an exponentially growing population with a stable distribution , i . e . , lim t → ∞ x ( t ) = e λ 1 t w . The leading eigenvalue λ1 hence gives the total population growth rate in the long term , and its associated right eigenvector w = ( w1 , … , wm−1 ) gives the stable distribution of group sizes so that , in the long term , the fraction xi of complexes of size i in the population is proportional to wi . For a given fitness landscape {b , d} , we can take the leading eigenvalue λ1 ( κ; b , d ) as a measure of fitness of fragmentation mode κ , and consider the competition between two different fragmentation modes , κ1 and κ2 . Indeed , under the assumption of no density limitation , the evolutionary dynamics are described by two uncoupled sets of differential equations of the form ( 4 ) : one set for κ1 and one set for κ2 . In the long term , κ1 is not outcompeted by κ2 if λ1 ( κ1; b , d ) ≥ λ1 ( κ2; b , d ) ; we then say that fragmentation mode κ1 dominates fragmentation mode κ2 . We also say that strategy κi is optimal for given birth rates b and death rates d if it achieves the largest growth rate among all possible fragmentation modes . Fitness landscapes capture the many advantages or disadvantages associated with group living . These advantages may come either in the form of additional resources available to groups depending on their size or as an improved protection from external hazards . For our numerical examples , we consider two classes of fitness landscape , each representing only one of these factors . In the first class , that we call “fecundity landscapes” , group size affects only the birth rates of cells ( while we impose di = 0 for all i ) . In the second class , that we call “survival landscapes” , group size affects only death rates ( and we impose bi = 1 for all i ) . To fix ideas , consider all pure fragmentation modes with a maximum group size n = 3 . These are 1+1 ( “binary fission” , a partition of 2 ) , 2+1 ( “unicellular propagule” , a partition of 3 ) , and 1+1+1 ( “ternary fission” a partition of 3 ) . The three associated projection matrices are given by A 1 + 1 = ( b 1 - d 1 ) , A 2 + 1 = ( - b 1 - d 1 2 b 2 b 1 - d 2 ) , A 1 + 1 + 1 = ( - b 1 - d 1 6 b 2 b 1 - 2 b 2 - d 2 ) . The three growth rates are λ 1 1 + 1= b 1 - d 1 , ( 5a ) λ 1 2 + 1 = − ( b 1 + d 1 + d 2 ) + ( b 1 + d 1 − d 2 ) 2 + 8 b 1 b 2 2 , ( 5b ) λ 1 1 + 1 + 1 = - ( b 1 + 2 b 2 + d 1 + d 2 ) + b 1 2 + 2 b 1 ( 10 b 2 + d 1 - d 2 ) + ( 2 b 2 - d 1 + d 2 ) 2 2 . ( 5c ) In the particular case of a fecundity landscape given by b1 = 1 and b2 = 15/8 ( and d1 = d2 = 0 ) , these growth rates reduce to λ 1 1 + 1 = 1 , λ 1 2 + 1 = 3 / 2 and λ 1 1 + 1 + 1 = 5 / 4 , and we have λ 1 2 + 1 > λ 1 1 + 1 + 1 > λ 1 1 + 1 . We then say that ternary and binary fission are dominated by the unicellular propagule strategy .
Although for simplicity we focus our exposition on pure fragmentation strategies , we also consider mixed fragmentation strategies , i . e . , probabilistic strategies mixing between different pure modes . A natural question to ask is whether a mixed fragmentation mode can achieve a faster growth rate than a pure mode . We find that the answer is no . For any fitness landscape and any maximum group size n , mixed fragmentation modes are dominated by a pure fragmentation mode ( S1 Text , Appendix B ) . Thus , the optimal fragmentation mode for any fitness landscape is pure . As an example , consider fragmentation modes 1+1 and 2+1 , and a mixed fragmentation mode mixing between these two so that with probability q splitting follows fragmentation pattern 2+1 and with probability 1 − q it follows fragmentation pattern 1+1 . For any mixing probability q and any fitness landscape , the growth rate of the mixed fragmentation mode is given by λ 1 q = b 1 ( 1 - 2 q ) - ( d 1 + d 2 ) + ( d 1 + d 2 - ( 1 - 2 q ) b 1 ) 2 + 4 b 1 ( 2 q b 2 + ( 1 - 2 q ) d 2 ) 2 , which can be shown to always lie between the growth rates of the pure fragmentation modes , i . e . , either λ 1 1 + 1 ≤ λ 1 q ≤ λ 1 2 + 1 or λ 1 2 + 1 ≤ λ 1 q ≤ λ 1 1 + 1 holds and the mixed fragmentation mode is dominated ( S1 Text , Appendix C ) . To further illustrate our analytical findings , consider groups of maximum size n = 4 and a fecundity landscape given by b = ( 1 , 2 , 1 . 4 ) . We randomly generated 107 mixed fragmentation modes by drawing the probabilities for growth without splitting from an uniform distribution and letting the probabilities of splitting according to a given fragmentation pattern be proportional to exponential random variables with rate parameter equal to one . We then calculated the growth rate of these mixed strategies together with the growth rate of the seven pure fragmentation modes available for n = 4 , i . e . , 1+1 , 2+1 , 1+1+1 , 3+1 , 2+2 , 2+1+1 , and 1+1+1+1 ( Fig 2A ) . In line with our analysis , a pure fragmentation mode ( namely 2+2 , whereby groups grow up to size 4 and then immediately split into two bicellular groups ) achieves a higher growth rate than the growth rate of any mixed fragmentation mode , and the highest growth rate overall . Having shown that mixed fragmentation modes are dominated , we now ask which pure modes might be optimal . We find that , within the set of pure modes , “binary” fragmentation modes ( whereby groups split into exactly two offspring groups ) dominate “nonbinary” fragmentation modes ( whereby groups split into more than two offspring groups ) . To illustrate this result , consider the simplest case of n = 3 and the three modes 1+1 , 2+1 , and 1+1+1 , out of which 1+1 and 2+1 are binary , and 1+1+1 is nonbinary . Comparing their growth rates ( as given in Eq ( 5 ) , we find that λ 1 1 + 1 ≥ λ 1 1 + 1 + 1 holds if b1 − b2 ≥ d1 − d2 and that λ 1 2 + 1 ≥ λ 1 1 + 1 + 1 holds if b1 − b2 ≤ d1 − d2 . Thus , for any fitness landscape , 1+1+1 is dominated by either 2+1 or by 1+1 . More generally , we can show that for any nonbinary fragmentation mode , one can always find a binary fragmentation mode achieving a greater or equal growth rate under any maximum group size n and fitness landscape ( S1 Text , Appendices D and E ) . Taken together , our analytical results imply that the set of optimal fragmentation modes is countable and , even for large n , relatively small . Consider the proportion of pure fragmentation modes that can be optimal , which is defined by the ratio between the number of binary fragmentation modes and the total number of pure fragmentation modes . While this ratio is relatively high for small n ( e . g . , 2/3 ≈ 0 . 67 for n = 3 or 4/7 ≈ 0 . 57 for n = 4 ) , it decreases sharply with increasing n ( e . g . , 25/128 ≈ 0 . 20 for n = 10 and 625/1295920 ≈ 0 . 00048 for n = 50 ) . Fig 2B shows the growth rate of the seven pure modes for n = 4 for a fecundity landscape given by b = ( 1 , b2 , 1 . 4 ) as a function of b2 . In line with our analysis , only binary fragmentation modes ( 1+1 , 2+1 , 2+2 , and 3+1 ) can be optimal , while nonbinary fragmentation modes ( 1+1+1 , 2+1+1 , and 1+1+1+1 ) are dominated . Which particular binary mode is optimal depends on the particular value of the birth rate of groups of two cells . For small values ( b2 ≲ 0 . 45 ) , the fecundity of such groups is too low , and the optimal fragmentation mode is 1+1 . For intermediary values ( 0 . 45 ≲ b2 ≲ 1 . 11 ) , the reproduction efficiency of groups of three cells mitigates the inefficiency of cell pairs , and the mode 3+1 becomes optimal . For larger values ( 1 . 11 ≲ b2 ≲ 3 . 52 ) , the optimal fragmentation mode is 2+2 , where no single cells are produced . Finally , for very large values ( b2 ≲ 3 . 52 ) , the optimal fragmentation mode is 2+1; this ensures that one offspring group emerges at the most productive bicellular state . More generally , which particular fragmentation mode within the class of binary splitting strategies is optimal depends on all birth rates and death rates characterising the fitness landscape . To further explore this issue , we identified the optimal fragmentation modes for general fecundity and survival landscapes for the simple case of n = 4 ( Fig 3; S1 Text , Appendix F ) . Since we can set b1 = 1 and min ( d ) = 0 without loss of generality ( S1 Text , Appendix D ) , we represent fitness landscapes as points in a two-dimensional parameter space with coordinates b2/b1 and b3/b1 for fecundity landscapes , and coordinates d2 − d1 and d3 − d1 for survival landscapes . The exact boundaries of the parameter regions where a given fragmentation mode is optimal are often nontrivial mathematical expressions . Nevertheless , we identify general patterns dictating which fragmentation mode will be optimal . Consider first the optimality map for fecundity landscapes ( Fig 3A ) . A sufficient condition for the unicellular life cycle 1+1 to be optimal is that the birth rate of single cells is larger than the birth rate of pairs and triplets of cells ( b1 > b2 and b1 > b3 ) . In this case , there is no apparent reason why a fragmentation mode different than 1+1 would be optimal . Perhaps less trivially , 1+1 can also be optimal in cases where single cells are less fertile than groups of three cells , i . e . , even if b1 < b3 holds . This requires the birth rate b2 to be so small that the fecundity benefits accrued when reaching the size of three cells are not enough to compensate for the unavoidable penalty of passing through the less prolific state of two cells . Turning now to fragmentation mode 2+1 , a necessary condition for this mode to be optimal is that pairs of cells have the largest birth rate , i . e . , that b2 > b1 and b2 > b3 holds . Similarly , mode 3+1 can only be optimal if b3 > b1 and b3 > b2 , so that groups of three have the largest birth rate . In these two cases , the optimal fragmentation mode ( either 2+1 or 3+1 ) keeps one of the two offspring groups at the most productive size . Finally , for fragmentation mode 2+2 to be optimal , it is necessary that single cells have the lowest birth rate , i . e . , that b2 > b1 and b3 > b1 holds . In this case , the fragmentation mode ensures that the life cycle of the organism never goes through the least productive unicellular phase . Under survival landscapes , fitness increases as death rates decrease . Taking this qualitative difference into account , the map of optimal fragmentation modes under survival landscapes ( Fig 3B ) follows similar qualitative patterns as the one under fecundity landscapes . So far we have assumed that fragmentation is costless . However , fragmentation processes can be costly to the parental group undergoing division . This is particularly apparent in cases where some cells need to die in order for fragmentation of the group to take place . Examples in simple multicellular forms include Volvox , where somatic cells constituting the outer layer of the group die upon releasing the offspring colonies and are not passed to the next generation [26] , the breaking of filaments in colonial cyanobacteria [27] , and the fragmentation of “snowflake-like” clusters of the yeast Saccharomyces cerevisiae [28] . Fragmentation costs may also be less apparent . For instance , fragmentation may cost resources that would otherwise be available for the growth of cells within a group . To investigate the effect of fragmentation costs on the set of optimal fragmentation modes , we consider two cases: proportional costs and fixed costs . For proportional costs , we assume that π − 1 cells die in the process of a group fragmenting into π parts . This case captures the fragmentation process of filamentous bacteria , where filament breakage entails the death of cells connecting the newly formed fragments [27] . For fixed costs , we assume that exactly one cell is lost upon each fragmentation event . This scenario is loosely inspired by yeast colonies with a tree-like structure , where cells can be connected with many other cells , so the death of a single cell may release more than two offspring colonies [19 , 28] . Mathematically , both cases imply that fragmentation patterns are described by partitions of a number smaller than the size of the parent group ( S1 Text , Appendix G ) . For both kinds of costly fragmentation , we can show that mixed fragmentation modes are still dominated by pure fragmentation modes ( the proof given in S1 Text , Appendix B also holds in this case ) . Moreover , for proportional costs the optimal fragmentation mode is also characterised by binary fragmentation , as it is the case for costless fragmentation ( S1 Text , Appendix H ) . This makes intuitive sense , as the addition of a penalty for splitting into many fragments should further reinforce the optimality of binary splitting ( whereby only one cell per fragmentation event is lost ) . In contrast , we find that under fragmentation with fixed costs the optimal fragmentation mode can involve nonbinary fragmentation , i . e . , division into more than two offspring groups . This result can be readily illustrated for the case of n = 4 where the nonbinary mode 1+1+1 is optimal for a wide range of fitness landscapes ( Fig 4 ) . Another interesting feature of costly fragmentation ( implemented via either proportional or fixed costs ) is that fragmentation modes involving the emergence of large groups can be optimal even if being in a group does not grant any fecundity or survival advantage to cells . If fragmentation is costless , as we assumed before , fitness landscapes for which groups perform worse than unicells ( that is , bi/b1 ≤ 1 for fecundity landscapes or di − d1 ≥ 0 for survival landscapes ) lead to optimal fragmentation modes where splitting occurs at the minimum possible group size i = 2 , so that no multicellular groups emerge in the population ( cf . Fig 3 ) . In contrast , under costly fragmentation some of these fitness landscapes allow for the evolutionary optimality of fragmentation modes according to which groups split at the maximum size n = 4 ( 2+1 under proportional costs , and 1+1+1 under fixed costs ) , and hence for life cycles where multicellular phases are persistent . This seems paradoxical until one realises that by staying together as long as possible groups delay as much as possible the inevitable cell loss associated to a fragmentation event . Thus , even if groups are less fecund or die at a higher rate than independent cells , staying together might be adaptive if splitting apart is too costly . Next , we focus on fitness landscapes for which either the birth rate of cells increases with group size ( fecundity landscapes where larger groups are always more productive ) or the death rate of groups decreases with group size ( survival landscapes where larger groups always live longer ) . In this case , and for a maximum group size n = 4 , the set of optimal modes is given by 2+2 and 3+1 if there are no fragmentation costs ( Fig 3 ) , by 2+1 if fragmentation costs are proportional to the number of fragments ( Fig 3A and 3B ) , and by 2+1 and 1+1+1 if fragmentation involves a fixed cost of one cell ( Fig 4C and 4D ) . To investigate larger maximum group sizes n in a simple but systematic way , we consider fecundity landscapes with birth rates given by bi = 1 + Mgi and survival landscapes with death rates given by di = M ( 1 − gi ) , where gi = [ ( i − 1 ) / ( n − 2 ) ]α [29] models the fecundity or survival benefits associated to group size i and M > 0 is the maximum benefit ( Fig 5 ) . The parameter α is the degree of complementarity between cells; it measures how important the addition of another cell to the group is in producing the maximum possible benefit M [30] . For low degrees of complementarity ( α < 1 ) , the sequence gi is strictly concave and each additional cell contributes less to the per capita benefit of group living [31] and groups of all sizes achieve the same functionality as α tends to zero . If α = 1 , the sequence gi is linear , and each additional cell contributes equally to the fecundity or survival of the group . Finally , for high degrees of complementarity ( α > 1 ) , the sequence gi is strictly convex and each additional cell improves the performance of the group more than the previous cell did . In the limit of large α , the advantages of group living materialise only when complexes achieve the maximum size n − 1 [31] . We numerically calculate the optimal fragmentation modes for n = 20 ( costless fragmentation ) or n = 21 ( costly fragmentation ) and the fitness landscapes described above for parameter values taken from 0 . 01 ≤ α ≤ 100 and 0 . 02 ≤ M ≤ 50 ( Figs 6 and 7 ) . In line with our general analytical results , optimal fragmentation modes are always characterised by binary splitting when fragmentation is costless or when it involves proportional costs , while nonbinary splitting can be optimal only if fragmentation involves a fixed cost . We also find that , for each value of α and M , and for both costless and costly fragmentation , the optimal fragmentation mode is one where fragmentation occurs at the largest possible size . This is expected since the benefit sequence is increasing in group size and thus groups of maximum size perform better , either by achieving the largest birth rate per unit ( fecundity landscapes ) or the lowest death rate ( survival landscapes ) . Which particular fragmentation mode maximizes the growth rate depends nontrivially on whether fragmentation is costless or costly ( and in the latter case also on how such costs are implemented ) , on the kind of group size benefits ( fecundity or survival ) , on the maximum possible benefit M , and on the degree of complementarity α . Despite this apparent complexity , some general patterns can be identified . Let us focus on the case of fecundity landscapes and first fasten attention on the scenario of costless fragmentation ( Fig 6A ) . A salient feature of this case is the prominence of two qualitatively different fragmentation modes: the “equal binary fragmentation” strategy 10+10 ( whereby offspring groups have sizes as similar as possible ) and the “unicellular propagule” strategy 19+1 ( whereby offspring groups have sizes as different as possible ) . A sufficient condition for equal binary fragmentation to be optimal is that increase in size is characterised by diminishing returns . The intuition behind this result is that , if the degree of complementarity is small , then groups ( complexes of size i ≥ 2 ) have similar performance , while independent cells ( i = 1 ) are at a significant disadvantage . Therefore , the optimal strategy is to ensure that both offspring groups are as large as possible , and hence of the same size . However , equal binary fragmentation can be also optimal for synergistic interactions , provided that complementarity is not too high . In contrast , the unicellular propagule strategy is optimal only for relatively high degrees of complementarity . This is because when complementarity is high only the largest group can reap the benefits of group living; in this case , the optimal mode is to have at least one offspring of very large size . Compared to 19+1 and 10+10 , other binary splitting strategies are optimal in smaller regions of the parameter space , and in all cases only for synergetic interactions between cells . Consider now the effects of introducing fragmentation costs proportional to the number of fragments ( Fig 6B ) . Here , the region where the unicellular propagule strategy is optimal shrinks to the corner of the parameter space where benefits of group living and degree of complementarity are maximum , while the region of optimality for equal binary fragmentation expands . This makes intuitive sense . With fragmentation costs , the largest offspring group resulting from fragmenting according to the unicellular propagule strategy is of size 19 , and hence always on the brink of fragmentation ( once it grows to size 21 ) and incurring one cell loss . When group benefits are not high and synergistic enough , the unicellular propagule strategy is dominated by fragmentation modes ( in particular , equal binary fragmentation ) having smaller offspring for which the costs of fragmentation are not so immediate . Finally , if costs of fragmentation are not proportional but fixed ( Fig 6C ) , then two classes of nonbinary splitting become optimal in regions of the parameter space where equal binary fragmentation was optimal under proportional costs: ( i ) “multiple fission” ( 1+1+…+1 ) which is in general favored for small maximum benefit and increasing returns , and ( ii ) “multiple groups” ( modes 2+2+…+2 , 3+3+3+3+3+3+2 , 4+4+3+3+3+3 , 4+4+4+4+4 , 5+5+5+5 , and 7+7+6 ) which are often optimal for diminishing returns . Fig 7 show the results for survival landscapes . The main difference in this case is that the unicellular propagule strategy can be the optimal strategy even when group living is characterised by diminishing returns . In general , fecundity benefits make equal binary fragmentation optimal under more demographic scenarios , while survival benefits make the unicellular propagule strategy optimal under more demographic scenarios .
Reproduction is such a fundamental feature of living systems that the idea that the mode of reproduction may be shaped by natural selection is easily overlooked . Here , we analysed a matrix population model that captures the demographic dynamics of complexes that grow by staying together and reproduce by fragmentation . The costs and benefits associated with group size ultimately determine whether or not a single cell fragments into two separate daughter cells upon cell division , or whether those daughter cells remain in close proximity , with fragmentation occurring only after subsequent rounds of division . We allowed for a vast and complete space of fragmentation strategies , including pure modes ( specifying all possible combinations of size at fragmentation and fragmentation pattern ) and mixed modes ( specifying all probability distributions over the set of pure modes ) , and identified those modes achieving a maximum growth rate for given fecundity and survival size-dependent rates . Our research questions and methodology thus resonates with previous studies in life history theory [32 , 33] . In the language of this field , our fragmentation strategies specify both the size at first reproduction and clutch size , where the latter is subject to a very specific trade-off between the number and size of offspring mathematically given by integer partitions . We found that for any fitness landscape , the optimal life cycle is always a deterministic fragmentation mode involving the regular schedule of group development and fragmentation . This makes intuitive sense given our assumption that the environment is constant . However , this result might not hold if the environment is variable so that the fitness landscape changes over time . In this case different pure fragmentation modes will be optimal at different times , and natural selection might favour life cycles that randomly express a subset of locally optimal fragmentation patterns . Indeed , the evolution of variable phenotypes in response to changing environmental conditions ( also known as bet hedging [34 , 35] ) has been demonstrated in other life history traits , such as germination time in annual plants [36] , and capsulation in bacteria [37] . The extent to which mixed fragmentation modes can evolve via a similar mechanism is beyond the scope of this paper , but it can be addressed in future work by applying existing theory on matrix population models in stochastic environments [22] . We found that when fragmentation is costless , only strategies involving binary splitting ( i . e . , fragmentation into exactly two parts ) are optimal . This result holds for all possible fitness landscapes , and hence any specification of how fecundity or survival benefits might accrue to group living . In particular , the optimal fragmentation mode under monotonic fitness landscapes is generally one of two types: equal binary fragmentation , which involves fission into two equal size groups , and the unicellular propagule strategy , which involves the production of two groups , one comprised of a single cell . Equal fragmentation is favoured when there is a significant advantage associated with formation of even the smallest group , whereas production of a unicellular propagule is favoured when the benefits associated with group size are not evident until groups become large . This makes intuitive sense: when advantages arise when groups are small , it pays for offspring to be in groups ( and not single cells ) . Conversely , when there is little gain until group size is large , it makes sense to maintain one group that reaps this advantage . Interestingly , two bacteria that form groups and are well studied from a clinical perspective , Neisseria gonorrhoeae and Staphylococcus aureus , both show evidence of the above binary splitting fragmentation modes: Neisseria gonorrhoeae divide into groups of two equal sizes [6] , while Staphylococcus aureus divide into one large group plus a unicellular propagule [7] . This leads to questions concerning the nature of the fitness landscape occupied by these bacteria and the basis of any collective level benefit as assumed by our model . Once cell loss upon fragmentation is incorporated as a factor in collective reproduction , a wider range of fragmentation patterns becomes optimal . When fragmentation costs are fixed to a given number of cells , optimal fragmentation modes include those where splitting involves the production of multiple offspring . Among these , a prominent fragmentation strategy is multiple fission , where a group breaks into multiple independent cells . Such a fragmentation mode is reminiscent of palintomy in the volvocine algae [38] . A key difference between our “multiple fission” and palintomy is that the former involves a group of cells growing up to a threshold size at which point fragmentation happens , while the latter involves a single reproductive cell growing to many times its initial size and then undergoing several rounds of division . However , reinterpreting birth rates of cells in groups as growth rates of unicells of different sizes allows us to use our analysis to determine conditions under which such a mode of fragmentation is more adaptive than , say , the more standard strategy of growing to twice the initial size and then dividing in two ( which for arbitrary sizes of offspring groups is equivalent to our “equal binary fragmentation” mode ) . Our results suggest that palintomy is favored over binary fission ( and any other fragmentation mode ) under a wide range of demographic scenarios ( Fig 6C ) . Many multicellular organisms are characterised by a life cycle whereby adults develop from a single cell [39] . Passing through such a unicellular bottleneck is a requirement for sexual reproduction based on syngamy , but life cycles with unicellular stages are also common in asexual reproduction modes such as those used by multicellular algae and ciliates [40] , and colonial bacteria such as S . aureus [7] . If multicellularity evolved because of the benefits associated to group living , why do so many asexual multicellular organisms begin their life cycles as solitary ( and potentially vulnerable ) cells ? Explanatory hypotheses include the purge of deleterious mutations and the reduction of within-organism conflict [39 , 41] . Our results make the case for an alternative ( and perhaps more parsimonious ) explanation: often , a life cycle featuring a unicellular bottleneck is the best way to guarantee that the “parent” group remains as large as possible to reap maximum fecundity and/or survival advantages of group living . Indeed , our theoretical results resonate with previous experimental work demonstrating that single-cell bottlenecks can be adaptive simply because they constitute the life history strategy that maximises reproductive success [42] . Previous theoretical work has explored questions related to the evolution of multicellularity using matrix population models similar to the one proposed in this paper . In a seminal contribution , Roze and Michod [43] explored the evolution of propagule size in the face of deleterious and selfish mutations . In their model , multicellular groups first grow to adult size and then reproduce by splitting into equal size groups , so that fragmentation mode strategies can be indexed by the size of the propagule . In our terminology , this refers to either “multiple fission” or “multiple groups” . An important finding of Roze and Michod [43] is that , even if large groups are advantageous , small propagules can be selected because they are more efficient at eliminating detrimental mutations . We did not study the effects of mutations , but allowed for general fitness landscapes and fragmentation modes , including cases of asymmetric binary division ( e . g . , the unicellular propagule strategy ) neglected by Roze and Michod [43] . Our results indicate that modes of fragmentation involving single cells can lead to growth rate maximisation even when small propagule sizes divide less efficiently or die at a higher rate . In particular , we have shown that if fragmentation is costly , a strategy consisting of a multiple fragmentation mode with a propagule size of one ( i . e . , the small propagule strategy studied by Roze and Michod [43] ) can be adaptive for reasons other than the elimination of deleterious mutations . Closer to our work , Tarnita et al . [18] investigated the evolution of multicellular life cycles via two alternative routes: “staying together” ( ST , whereby offspring cells remain attached to the parent ) and “coming together” ( CT , whereby cells of different origins aggregate in a group ) . In particular , they studied the conditions under which a multicellular strategy that produces groups via ST can outperform a solitary strategy whereby cells always separate after division . The way they modelled group formation and analyzed the resulting population dynamics ( by means of biological reactions and matrix models ) is closely related to our approach . Indeed , their solitary strategy is our binary mode 1+1 , while their ST strategy corresponds to a particular kind of binary mixed fragmentation mode . However , the questions we ask are different . Tarnita et al . [18] were concerned with the conditions under which ( multicellular ) strategies that form groups can invade and replace ( unicellular ) strategies that remain solitary . Contrastingly , we aimed to understand the optimal fragmentation mode out of the vast space of fragmentation strategies comprising all possible deterministic and probabilistic pathways by which complexes can stay together and split apart . A key finding is that , for any fitness landscape and if the environment is constant , mixed fragmentation modes such as some of the ST strategies considered by Tarnita et al . [18] will be outperformed by at least one pure fragmentation mode . More recently , Rashidi et al . [20] developed a conceptual framework to study the competition of life cycles that involved five different life cycles defined by fragmentation patterns of the form 1+1+…+1 and an associated genetic control . Their model , which explicitly considers growing cells of different size , showed that depending on the fitness landscape , each of their five life cycles could prevail . By extending the range of life cycles to encompass all possible fragmentation modes ( albeit with less detailed attributes ) , we have shown that certain life cycles will be suboptimal for any given fitness landscape . In line with many studies in life history theory [32 , 33] , we made the simplifying assumption that the phenotype consists of demographic traits ( in our case , probabilities of fragmenting into given fragmentation patterns ) linked by trade-offs which interact to determine fitness ( growth rate ) . This allowed us to predict the optimal phenotype at equilibrium at the expense of leaving unspecified whether , due to genetic constraints , such an equilibrium will be possible in an actual biological system . The question that inevitably arises is whether , given a presumptive genotype-phenotype mapping , it is possible for evolution to fine tune life cycles with group-level properties ( such as specific fragmentation patterns ) so that optimal fragmentation modes will be obtained as the endpoint of an evolutionary process . While a complete answer requires a more sophisticated analysis , we see no conceptual obstruction preventing seemingly arbitrary fragmentation modes to evolve . Firstly , genotype-phenotype maps of existing organisms can be complex and offer opportunity for adaptation , involving important qualitative behavioral changes [44–46] . Secondly , small genotypic changes can produce major phenotypic changes . For instance , Hammerschmidt et al . [3] observed the emergence of collective-level properties in a previously unicellular organism that was caused by a small number of mutations . Thirdly , even if a current set of genes cannot provide an appropriate template for given phenotypic traits , new genes can emerge de novo [47–51] . Finally , theoretical arguments suggest that genetic constraints can be effectively overcome in phenotypic evolution provided there is a rich variety of new mutant alleles [52] . We thus think that , both in the field and in the laboratory , multicellular organisms will be able to evolve a phenotype close to the optimal fragmentation mode in the ( very ) long run .
|
Mode of reproduction is a defining trait of all organisms , including colonial bacteria and multicellular organisms . To produce offspring , aggregates must fragment by splitting into two or more groups . The particular way that a given group fragments defines the life cycle of the organism . For instance , insect colonies can reproduce by splitting or by producing individuals that found new colonies . Similarly , some colonial bacteria propagate by fission or by releasing single cells , while others split in highly sophisticated ways; in multicellular organisms reproduction typically proceeds via a single-cell bottleneck phase . The space of possibilities for fragmentation is so vast that an exhaustive analysis seems daunting . Focusing on fragmentation modes of a simple kind we parametrise all possible modes of group fragmentation and identify those modes leading to the fastest population growth rate . Two kinds of life cycle dominate: one involving division into two equal size groups , and the other involving production of a unicellular propagule . The prevalence of these life cycles in nature is consistent with our null model and suggests that benefits accruing from population growth rate alone may have shaped the evolution of fragmentation mode .
|
[
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] |
[
"death",
"rates",
"organismal",
"evolution",
"cell",
"division",
"analysis",
"cell",
"cycle",
"and",
"cell",
"division",
"cell",
"processes",
"developmental",
"biology",
"bioassays",
"and",
"physiological",
"analysis",
"population",
"biology",
"fecundity",
"research",
"and",
"analysis",
"methods",
"cell",
"analysis",
"life",
"cycles",
"population",
"metrics",
"birth",
"rates",
"cell",
"biology",
"biology",
"and",
"life",
"sciences",
"evolutionary",
"biology",
"population",
"growth"
] |
2017
|
Fragmentation modes and the evolution of life cycles
|
Systems biology models are used to understand complex biological and physiological systems . Interpretation of these models is an important part of developing this understanding . These models are often fit to experimental data in order to understand how the system has produced various phenomena or behaviour that are seen in the data . In this paper , we have outlined a framework that can be used to perform Bayesian analysis of complex systems biology models . In particular , we have focussed on analysing a systems biology of the brain using both simulated and measured data . By using a combination of sensitivity analysis and approximate Bayesian computation , we have shown that it is possible to obtain distributions of parameters that can better guard against misinterpretation of results , as compared to a maximum likelihood estimate based approach . This is done through analysis of simulated and experimental data . NIRS measurements were simulated using the same simulated systemic input data for the model in a ‘healthy’ and ‘impaired’ state . By analysing both of these datasets , we show that different parameter spaces can be distinguished and compared between different physiological states or conditions . Finally , we analyse experimental data using the new Bayesian framework and the previous maximum likelihood estimate approach , showing that the Bayesian approach provides a more complete understanding of the parameter space .
Systems biology models are used to understand complex biological and physiological systems comprised of large numbers of individual elements that give rise to emergent behaviours . These complex systems are dependent on both the properties of the whole network and on the individual elements [1] . This inherent complexity within the models can lead to difficulties in determining how best to interpret information obtained through their use . At University College London , the family of BrainSignals models ( and the BRAINCIRC model on which they are based ) are used to understand the brain’s dynamics via a systems biology approach . They bring together a number of mathematical models relating to different aspects of blood circulation , oxygen transport and oxygen metabolism within the brain in order to develop a more complete model that can be used alongside experimental data to simulate physiological phenomena of the brain , such as autoregulation and neural activation . This allows us to understand how our measurements are linked to specific brain physiological and metabolic mechanisms . All of the models were developed to reproduce broadband near-infrared spectroscopy ( NIRS ) measurements of brain tissue concentration changes of haemoglobin ( oxygenation and haemodynamics ) and cytochrome-c-oxidase ( mitochondrial metabolism ) and vary in their complexity and scope . The first model developed was the ‘BRAINCIRC’ model in 2005 [2] , followed by the ‘BrainSignals’ model [3] in 2008 . A number of additional versions were then developed from this , such as the ‘BrainPiglet’ model [4] which was developed to to simulate the physiological and metabolic processes of the piglet brain often used as the neonatal preclinical model . This was extended in BrainPiglet v2 . 0 to incorporate the effects of cell death during injury [5] . In 2015 , Caldwell et al . modified and simplified the BrainSignals model to both reduce model complexity and improve model run time , producing the ‘BrainSignals Revisited’ model [6] . All of these models are run using the Brain/Circulation Model Developer environment ( BCMD ) and are defined in a simple text language . The data collected and analysed with the models primarily consists of broadband NIRS data , providing information about tissue oxygenation , through monitoring of oxy- and deoxy-haemoglobin levels , and cellular metabolism , through the concentration of cytochrome-c-oxidase . This data is then supplemented by systemic information such as blood pressure , arterial oxygen saturation and/or partial pressure of CO2 . One of the main uses of the models is to fit the model simulations to clinical and experimental data and investigate how model parameters are affected . In the case where data is collected from an injured or sick patient , these changes may illuminate what the underlying causes/mechanisms are behind the illness or injury is . The models are currently fit using a maximum likelihood based method , with a single value obtained for each parameter . Sensitivity analysis performed on the models to determine which parameters are most important in influencing each model output for any particular dataset . These parameters are then optimised using the PSwarm method [7] to minimise a given error metric , such as the Euclidean distance , between the modelled and measured signals . Through this each output has a set of optimised parameter values . Parameter values were limited to the same ranges used in the sensitivity analysis [5] . This approach has a number of drawbacks . The models are mechanistic and , if fitted to single value parameter estimates , will produce the same output for the same input . Physiology and biology , however , is unlikely to operate in such a constrained manner . Additionally , this set of best-fit parameters for the model may not be representative of the full parameter space [8] . In an attempt to try to compensate for this potential drawback , Caldwell et al . [5] fit the BrainPiglet model multiple times for two different piglets and found that , whilst parameter values can vary within the same data , separate parameter spaces for each piglet did seem to exist based on the brain physiological status of the piglet following a hypoxic-ischaemic insult . One of the key ways in which these models are used to extract information from data is through the use of parameter estimation and fitting . However , this step remains a difficult mathematical and computational problem , potentially originating in the lack of identifiability [9] . In addition , there has been discussion of ‘universal sloppiness’ within dynamic systems biology models . Gutenkunst et al . [10] proposed that sloppiness , where the parameters of a dynamic model can vary by orders of magnitude without affecting model output , is a universal property of systems biology models . Due to this sloppiness , it may not be possible to make parameter estimations that can be used to make inferences about the system [10 , 11] . Chis et al . have stated however that sloppiness is not equivalent to a lack of identifiability and that a sloppy model can still be identifiable [12] . Apgar et al . note that experimental design can be used to constrain a sloppy parameter space by choosing a set of complementary experiments [13] . The use of a Bayesian methodology , by avoiding point estimates , can allow the full uncertainty of the problem to be captured [8] . In fact , the use of an Approximate Bayesian Computation ( ABC ) approach , discussed below , is particularly well suited to these kinds of problems [14] . There are many examples of Bayesian methods being used to analyse bioinformatics data and systems biology models [15] , including in sequence analysis [16] , gene microarray data [17] and in models of genetic oscillators [18] and DNA network dynamics [19] . There are a number of models that take a systems biology approach towards understanding physiology , particularly oxygen transport and blood flow , including the previously mentioned BrainSignals [2 , 3 , 5] and BrainPiglet [4 , 5] models , the Aubert-Costalat model [20] , and work by Fantini [21–24] and Orlowski and Payne [25 , 26] where Bayesian parameter estimation could also be applied but has yet to be . Bayesian inference utilises Bayes’ rule , p ( θ | y ) = p ( θ , y ) p ( y ) = p ( θ ) p ( y | θ ) p ( y ) , where p ( y ) = ∫θ p ( θ ) p ( y|θ ) dθ is the marginal probability of y and p ( y|θ ) is the likelihood . Typically , p ( y ) is not known and the likelihood will not be known explicitly or may require marginalising over some values of θ . This often leaves the solution analytically intractable . Instead we can try solve for p ( θ|y ) using a Monte Carlo or Markov Chain Monte Carlo ( MCMC ) approach . Where a likelihood function can be defined there are a number of these methods that can be used to infer a posterior distribution , p ( θ|y ) . The simplest is the Gibbs Sampler [27] , which in its most basic form is a special case of the Metropolis-Hastings algorithm [28] . Although the BrainSignals models are deterministic , the model noise is a combination of process noise and experimental error which is expected to depend on the state in a non-trivial manner . This makes formulating an analytical expression for the likelihood difficult . In this case where a likelihood expression is unobtainable a likelihood-free approach using ABC is required instead . There are a number of different methods available with the simplest being the ABC rejection algorithm ( ABC REJ ) approach . This has the additional benefit of allowing us to consider different summary statistics that would not be valid in a likelihood-based approach . It may be that these summary statistics allow us to optimise for specific behaviours that have physiological relevance . The aim of this paper is to introduce the new BayesCMD modelling platform that can be used in systems biology models of physiology such as the BrainSignals models , but that can be replicated beyond these . For this work , we have chosen to use ABC REJ as whilst it is less efficient than the other methods mentioned here , the simplicity with which it can be implemented is a significant factor . The models and modelling environment used are already complex and so this initial work focuses on the use of the simplest method as proof of utility . We will demonstrate the effectiveness of this approach by using it to analyse two simulated datasets chosen to represent healthy and impaired brain states , before then using it on experimental data from a healthy subject undergoing a hypoxia challenge . We will show that the Bayesian approach allows us to extract more information from our data than the previous maximum likelihood approach , with a more complete picture of the parameter space being obtained .
Whilst a brief overview of the history of the BrainSignals models was given in the introduction , in this section we provide more information about the specifics of the different models . Table 1 compares the number of reactions , equations , relations , reactions , variables and parameters in three different models . The BRAINCIRC model from 2005 built on an earlier circulatory model by Ursino and Lodi [29] and combined models for the biophysics of the circulatory system , the brain metabolic biochemistry and the function of vascular smooth muscle . The BrainSignals model which succeeded it simplified the ‘BRAINCIRC’ model and added a submodel of mitochondrial metabolism . As previously mentioned , in order to better simulate the physiological and metabolic processes of the piglet brain , which is often used as the neonatal preclinical model , the ‘BrainPiglet’ model [4] was developed from the BrainSignals model . It involved modifying the default values for 11 of the 107 parameters used and was extended to include simulated measurements for magnetic resonance spectroscopy values that included brain tissue lactate and ATP production , measurements of which are available in piglet studies . Its extension , BrainPiglet v2 , incorporated the effects of cell death during injury in order to to investigate why two piglets showed different recoveries following hypoxia-ischaemia , finding that the differences could be explained by including cell death within the model [5] . The ‘BrainSignals Revisited’ model was produced by making various simplifications to the BrainSignals model by identifying various functions that could be replaced by linear approximators without reducing model applicability . This reduced complexity and decreased the time taken to run a simulation , whilst being able to reproduce the same results and behaviour of the original model . This reduced model of the adult brain was later extended to simulate extracerebral haemodynamics to investigate confounding factors with brain near-infrared spectroscopy measurements , the ‘BSX’ model [30] . The models are driven with input signals , such as the blood pressure and/or oxygen saturation , and simulate brain tissue measurments of oxygenation , blood volume and metabolism , as well as the middle cerebral artery velocity ( Vmca ) and the cerebral metabolic rate of oxygen ( CMRO2 ) . The model can be split into roughly 3 compartments—blood flow , oxygen transport and metabolism—with boundaries chosen to minimise interdependence . Fig 2 outlines this in more detail . All of these models are solved using the BCMD framework and are written in a simple text format that can be translated to executable C code and solved using the RADAU5 solver [31] . The models take a standard differential-algebraic equation representation , of the form: M d y d t = f ( y , θ , t ) ( 1 ) where y is a vector of variables of interest , M is a constant , possibly-singular , mass matrix specifying relations among the differential terms , and f is some vector-valued function , possibly having additional parameters θ . If a row of M is zero , the corresponding equation in f is algebraic rather than differential . In this work we have chosen to use the refactored BrainSignals model [6] , with a minor modification to include the haemoglobin difference ( ΔHbO2 − ΔHHb = ΔHbD ) as a model output alongside the normal outputs of oxyhaemoglobin ( ΔHbO2 ) , deoxyhaemoglobin ( ΔHHb ) , total haemoglobin ( ΔHbO2 + ΔHHb = ΔHbT ) , tissue oxygenation index ( TOI ) , and cytochrome-c-oxidase ( ΔCCO ) . Both ΔHbD and ΔHbT are included in the experimental dataset due to them being good indicators of brain oxygenation changes and brain blood volume changes respectively , with both being easily measured using broadband NIRS . All NIRS outputs , except TOI , are measured as changes relative to an initial value and therefore both data and model outputs are normalised to an initial value of 0 . Three datasets were used to test the new Bayesian model analysis process . Firstly , ‘healthy’ data was simulated using the BrainSignals model with the default parameter settings , as per [2 , 3] . Next , the same inputs were used but with the model modified to represent an ‘impaired’ brain . To do this , a single parameter was changed to reflect a potential pathology or injury , to generate an ‘impaired’ simulated dataset . Finally , we used experimental data from a healthy adult undergoing a hypoxia challenge . When fitting a model as complex as BrainSignals , it is important to reduce the number of parameters that are required to be fit . We expect that not all parameters will have a significant impact on the model output for given set of input data . Instead , we can attempt to reduce the number of considered parameters through sensitivity analysis . We used the Morris method [34 , 35] , which is known to work well with a large number of parameters . The method requires the time series to be reduced to a single number and identifies the parameters that have produce the most variance in this summary value . Previously , we have used the Euclidean distance over the whole time series as our summary value but this has a number of significant drawbacks . If the summary measure is the distance across the whole time series , we’re failing to capture specific changes that we know to be physiologically important . In the case of our hypoxia simulation , for example , we want to select parameters that are important in controlling the overall change from baseline . Taking the Euclidean distance over the time series as a whole however does not prioritise this behaviour . Fig 6a shows three sets of data generated from the same toy model function y i = a x sin ( x ) + b + ϵ , ( 3 ) where a , b are both model parameters and ϵ is random Gaussian noise . Assume that without modification , our model produces data y0 , with the default parameters Θ0: a = 0 , b = 0 , and that the behaviour we want to reproduce is sinusoidal but , for some reason , we don’t know which parameter is most important in producing this specific behaviour . We decide to undertake sensitivity analysis , using a distance measure of some kind as our summary statistic in order to identify the parameter most important in producing sinusoidal behaviour . If when altering a parameter that distance measure increases , then the behaviour summarised by that distance is sensitive to changes in that parameter . In this case , to produce sinusoidal behaviour , we would want parameter a to be identified as important rather than parameter b . To generate our data x was varied from 0 to 2π , producing datasets y1 and y2 for the parameter sets Θ1: a = 1 , b = 0 , where only a is changed from baseline , and Θ2: a = 0 , b = 0 . 707 , where only b is changed from baseline , respectively . y0 and parameter set Θ0 provide our baseline data . This is seen in Fig 6a . It is clear from the figure that the two outputs y1 and y2 show very different behaviour , the behaviour we want to optimise for is seen in y1 . Despite both y1 and y2 being qualitatively very different they are very similar when summarised using only the Euclidean distance , with y1 having a Euclidean distance εeuc , 1 = 10 . 01 and y2 having a Euclidean distance εeuc , 2 = 10 . 03 . This means that we would fail to clearly identify parameter a as being important than parameter b in producing sinusoidal behaviour . Instead we can define a new summary measure , which we will call the “scaled baseline-to-peak” ( SBTP ) distance . We know that we want to find the parameter that determines how sinusoidal our model is . One way to emphasise this behaviour is to find the distance from our baseline to the maximum or minimum ( whichever has the largest absolute value ) of our data , as illustrated in Fig 6b . We then scale this by the range of our ‘default’ signal , y0 , to normalise it and avoid issues comparing data of different magnitudes . This gives us S B T P ( y i ) = max ( { | max ( y i ) − y i ( t = 0 ) | , | min ( y i ) − y i ( t = 0 ) | } ) max ( y 0 ) − min ( y 0 ) ) ( 4 ) We then find the Euclidean distance between the SBTP value for our ‘default’ data , SBTP ( y0 ) , and SBTP ( y1 ) and SBTP ( y2 ) ε S B T P , i = ( S B T P ( y 0 ) − S B T P ( y i ) ) 2 , ( 5 ) where here i ∈ {1 , 2} . If we use εSBTP as our summary measure , we find that y1 has a distance εSBTP , 1 = 240 . 2 and y2 has a distance εSBTP , 2 = 0 . 11 . This would mean that parameter a could be clearly identified as being more important in producing sinusoidal behaviour than parameter b . We scale our baseline-to-peak distance because a number of model outputs significantly vary over different scales . For example , cerebral oxygenation can be measured through TOI which is a percentage and , as seen in Fig 3 can vary over 10-20% . Cytochrome-c-oxidase however , varies over a much smaller range , with a change of less than 1μM being typical . Failing to account for these different scales will lead to parameters that affect larger magnitude outputs being identified as more sensitive than those that affect smaller magnitude outputs , even if the relative change is significant . For example , if changing a parameter θ1 causes the CCO change seen in Fig 3e ) to double to a minimum of -2μM , whilst a change in a parameter θ2 causes TOI to decrease to 55% , without scaling the model seems more sensitive to θ2 because the magnitude of the change is much more , even though the relative change is smaller . If we consider this change proportional to the range of our data however , we account for its relative size . It should also be noted that this choice of metric is specific to the behaviour being optimised for . For example , in the case of a signal that is non-oscillatory , a different summary method would be required based around the behaviour to be replicated within that particular signal . We also acknowledge that there are a variety of different methods for identifying a sinusoidal signal from a linear signal and that our choice of metric here is one of many . We have chosen it as in the case of our hypercapnia data , we expect to see our signal to change from baseline to maxima or minima , depending on the signal , before then returning to baseline . The SBTP distance emphasises this behaviour in a single number whilst also being easily comparable to previous work where the Euclidean distance was used . We used the Morris elementary effect method [34] variant devised by Saltelli et al . [36] . This provides us with two notable statistics: the mean of the absolute values of the changes , μ* , and their standard deviation , σ . The larger the value of μ* , the more influential parameter is on the output , whilst the larger the standard deviation , the more non-linear the influence of the parameter is . The top ten most sensitive parameters , as per μ* were chosen to fit the model . σ was not used to determine which parameters to fit as , whilst knowing the non-linearity of a parameter is useful , in previous work [5 , 6] we have opted to use simply μ* as this gives a good summary of the sensitivity of a single parameter and feel it is pertinent to continue to do so here . The parameter range considered for sensitivity is the default value ±50% . Sensitivities are calculated for each output as well as across all outputs jointly . This joint sensitivity is calculated by summing the SBTP value for each output and then determining variability in this total . After selecting the most important parameters , the model was fit using the rejection algorithm [37] . This is defined , as per [38] , as: Sample a candidate parameter vector θ* from the proposal distribution p ( θ ) . Simulate a dataset yrep from the model described by a conditional probability distribution p ( y|θ* ) . Compare the simulated dataset , yrep , to the experimental dataset , y , using a distance function , d , and tolerance , ϵ . If d ( y , yrep ) ≤ ϵ , accept θ* . The tolerance ϵ ≥ 0 is the desired level of agreement between y and yrep . The output of the ABC algorithm used will be a sample from the distribution p ( θ|d ( y , yrep ) ≤ ϵ ) . If ϵ is sufficiently small , then p ( θ|d ( y , yrep ) ≤ ϵ ) will be a good approximation for the posterior p ( θ|y ) . The choice of d ( ⋅ , ⋅ ) is important , just as with the sensitivity analysis . Previously the Euclidean distance has been used to fit the model but , as in the case of the sensitivity analysis , this fails to account for outputs that vary over different magnitudes . Instead , we have chosen to include a number of other distance metrics including the root-mean-square error ( RMSE ) and the normalised root-mean-square error ( NRMSE ) . These are defined as R M S E= ∑ t = 1 T ( x 1 , t − x 2 , t ) 2 T ( 6 ) N R M S E= R M S E ( x 1 , x 2 ) x 1 , m a x − x 1 , m i n ( 7 ) where x1 and x2 are the two time series being compared , running over t = 1 to t = T , with T being the total number of time points . By dividing the RMSE by the range of the data , the errors for time series that vary over different magnitudes are comparable . Without doing this , parameters that mainly affect outputs that vary over larger magnitudes are preferentially optimised . Normalisation prevents overfitting of one output at the expense of others , providing a more reliable joint posterior distribution after fitting . After an initial exploratory fitting of the different datasets , it was found that setting an absolute tolerance value was not a suitable selection criteria . This was due to massively differing distance values between datasets , with all parameter combinations in the simulated healthy dataset producing NRMSE values smaller than almost all parameter combinations on the impaired dataset . In general , the number of accepted samples that gives an adequate approximation of the posterior distribution is problem dependent; dispersed posterior distributions will ultimately require more samples . Poor estimation of the posterior can in most cases result in a wide posterior predictive distribution which appears to give a poor quality fit because outlier posterior samples cause biases . To address this issue in a pragmatic way , a fixed acceptance rate of 0 . 01% was set . This meant the 0 . 01% parameter combinations with the lowest d ( y , yrep ) were used as the posterior . The posterior was visualised through kernel density estimation on a pairplot using the Seaborn plotting package [39] . The posterior predictive density is then generated by sampling directly from the posterior 25 times and the model simulated for each sample . The results are aggregated and plotted , with the median and 95% credible interval marked on the plot . The model was run in batches of 10 , 000 , 000 and the parameter combinations within the acceptance rate were used as a posterior . This batch size was chosen as a compromise between sufficient sampling of the parameter space and the computational time required to run the batch . The quality of the fit obtained from this posterior determined if the model had been run a sufficient number of times to sample the posterior adequately . If the posterior predictive distribution failed to capture the behaviour seen in the “true” data , then the process was repeated until a more adequate fit was obtained .
We have outlined how this new Bayesian framework for model analysis can be used with models of brain haemodynamics to extract information from physiological data . A more comprehensive picture of the parameter space is obtained , allowing physiological conclusions to be based on a broader picture . This is most clearly seen in the experimental data , where point estimates suggested that the values for a number of parameters had changed significantly during fitting , whilst the Bayesian method showed that the parameters were defined by a broad , roughly uniform distribution . We have also shown , through the use of data simulated from the BrainSignals model in healthy and impaired states , how the Bayesian approach allows us to better distinguish different parameter spaces . Finally , whilst we have focussed on using the BrainSignals model here , any model that can be written in a format compatible with BCMD can use this method to estimate model parameters . A major interest within our research group is to use these models and approaches to understand and investigate further our novel measures of brain tissue physiology and metabolism and how they are linked to brain injury [45 , 46] . In particular , we are interested in neonatal hypoxic ischaemic injury . The Bayesian approach provides a better representation of the parameter space and can inform a better distinction between different brain states , such as between a mild and severe injury . The method will also be adapted to use more efficient methods of parameter estimation , such as ABC SMC , reducing the number of model runs required to obtain a given tolerance .
|
Systems biology models are mathematical representations of biological processes that reproduce the overall behaviour of a biological system . They are comprised by a number of parameters representing biological information . We use them to understand the behaviour of biological systems , such as the brain . We do this by fitting the model’s parameter to observed or simulated data; and by looking at how these values change during the fitting process we investigate the behaviour of our system . We are interested in understanding differences between a healthy and an injured brain . Here we outline a statistical framework that uses a Bayesian approach during the fitting process that can provide us with a distribution of parameters rather than single parameter number . We apply this method when simulating the physiological responses between a healthy and a vascular compromised brain to a drop in oxygenation . We then use experimental data that demonstrates the healthy brain response to an increase in arterial CO2 and fit our brain model predictions to the measurements . In both instances we show that our approach provides more information about the overlap between healthy and unhealthy brain states than a fitting process that provides a single value parameter estimate .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Discussion"
] |
[
"medicine",
"and",
"health",
"sciences",
"oxygen",
"integrative",
"physiology",
"simulation",
"and",
"modeling",
"oxygen",
"metabolism",
"systems",
"science",
"mathematics",
"physiological",
"parameters",
"near-infrared",
"spectroscopy",
"research",
"and",
"analysis",
"methods",
"computer",
"and",
"information",
"sciences",
"spectrum",
"analysis",
"techniques",
"infrared",
"spectroscopy",
"mathematical",
"and",
"statistical",
"techniques",
"bayesian",
"method",
"chemistry",
"systems",
"biology",
"biochemistry",
"physiology",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"metabolism",
"chemical",
"elements"
] |
2019
|
A Bayesian framework for the analysis of systems biology models of the brain
|
The high-risk HPV E6 and E7 proteins cooperate to immortalize primary human cervical cells and the E7 protein can independently transform fibroblasts in vitro , primarily due to its ability to associate with and degrade the retinoblastoma tumor suppressor protein , pRb . The binding of E7 to pRb is mediated by a conserved Leu-X-Cys-X-Glu ( LXCXE ) motif in the conserved region 2 ( CR2 ) of E7 and this domain is both necessary and sufficient for E7/pRb association . In the current study , we report that the E7 protein of the malignancy-associated canine papillomavirus type 2 encodes an E7 protein that has serine substituted for cysteine in the LXCXE motif . In HPV , this substitution in E7 abrogates pRb binding and degradation . However , despite variation at this critical site , the canine papillomavirus E7 protein still bound and degraded pRb . Even complete deletion of the LXSXE domain of canine E7 failed to interfere with binding to pRb in vitro and in vivo . Rather , the dominant binding site for pRb mapped to the C-terminal domain of canine E7 . Finally , while the CR1 and CR2 domains of HPV E7 are sufficient for degradation of pRb , the C-terminal region of canine E7 was also required for pRb degradation . Screening of HPV genome sequences revealed that the LXSXE motif of the canine E7 protein was also present in the gamma HPVs and we demonstrate that the gamma HPV-4 E7 protein also binds pRb in a similar way . It appears , therefore , that the type 2 canine PV and gamma-type HPVs not only share similar properties with respect to tissue specificity and association with immunosuppression , but also the mechanism by which their E7 proteins interact with pRb .
Human papillomaviruses ( HPVs ) mediate the initiation and maintenance of cervical cancer [1] , [2] . Based upon DNA sequence homology , there are more than 150 different HPV genotypes , 40 of which infect anogenital and oral mucosa [3] . In addition to genotyping , HPVs can also be classified as low-risk and high-risk based on the clinical prognosis of their associated lesions . Low-risk HPVs cause benign epithelial hyperplasias while high-risk HPVs cause lesions that can progress to malignancy . Integration of the HPV genome into a host cell chromosome is a frequent event during malignant progression and it may play a significant role in dysregulated expression of the HPV E6 and E7 proteins [4] . The high-risk HPV E6 binds to several cell targets , including p53 , Myc , E6AP , PDZ proteins and other cellular proteins to alter apoptotic/growth regulatory pathways and induce cellular telomerase activity [5] . The E7 protein binds and sequesters pRb and directs its ubiquitin-mediated proteolysis [6] , thereby altering E2F-regulated genes and allowing cells to enter the S phase of the cell cycle . The E7 oncoprotein is approximately 100 amino acids in length and contains two highly conserved regions ( CRs ) , the amino-terminal CR1 and CR2 domains [7] . The E7 CR1 and CR2 domains share strikingly high homology with the CR1 and CR2 regions of adenovirus ( Ad ) E1A and related sequences in simian vacuolating virus 40 ( SV40 ) large tumor antigen ( T Ag ) [4] , [8] . For each of these viruses , the CRs contribute significantly to cell transformation [9] , [10] , [11] , [12] . A conserved Leu-X-Cys-X-Glu ( LXCXE ) motif in the CR2 domain of HPV E7 , as well the ones in Adenovirus E1A and SV40 LT , are necessary and sufficient for association with pRb [13] . The crystal structure of pRb bound to an E7 peptide was resolved , and revealed that LXCXE of HPV E7 binds entirely through the B domain of pRb [14] . For high risk HPV , the LXCXE motif is also required for pRb degradation[15] , [16] . The carboxyl-terminal domain of E7 consists of a metal binding domain composed of two CXXC motifs separated from each other by 29 amino acids [14] . This zinc-binding region is important for E7 dimerization and intracellular stabilization [10] , [17] . The carboxyl-terminal domain also contributes to E7 association with chromatin-modifying enzymes , particularly histone deacetylases and histone acetyl transferases [18] . Although the carboxyl-terminus of high-risk HPV E7 does not appear to have a direct role in the binding and degradation of pRb [15] , [19] , it has been proposed to be important for releasing E2F from pRb [20] , [21] . Papillomavirus can be isolated from a wide range of vertebrates , ranging from birds to manatees [22] , [23] and infection by these viruses is , in general , species-specific . The canine papillomavirus model has been used successfully for vaccine [24] , [25] and therapeutics studies [26] . Recently , our lab isolated and sequenced canine papillomavirus type 2 ( CPV-2 , previously named CfPV2 ) , which we showed to be an epidermotropic virus that occurred frequently in immunosuppressed animals and induced tumors that progressed to aggressive cancers [27] , [28] . The E7 gene of CPV-2 appeared unique in that it lacked the conserved LXCXE motif . In this study we show that this variant E7 protein is still able to bind and degrade pRb and that the primary domain for binding pRb is in its carboxyl-terminus . Interestingly , upon searching the HPV genome database , we observed that the gamma HPVs also contained the variant LXSXE E7 domain and , similar to the canine E7 protein , could still bind and degrade pRb . In addition to this similarity , we also noted that both the type 2 canine PV and gamma HPVs both exhibited a tropism for skin and for immunocompromised hosts .
Wild-type CPV-2 E6 and E7 were generated by PCR using CPV-2 genome as template [27] and subcloned into retrovirus vector pLXSN ( Clontech ) at the sites EcoR I and BamH I . The CPV-2 E7 mutants CPV-2 E7 ΔLXSXE , S26C , and S26G were generated by using the QuikChange Site-Directed Mutagenesis kit ( Stratagene ) , and CR1 , CR2 , CR1CR2 and CT were generated by PCR using pLXSN . CPV-2E7 as template . All the wild type E7 and mutants were cloned into the sites EcoR I and Not I of the pGEX4T-2 ( GE Healthcare ) for GST fusion protein expression . CPV-2 E7 and mutants with a hemagglutinin ( HA ) epitope tag at their amino terminal or carboxyl terminal were generated by PCR . PCR products were then subcloned into the mammalian expression vector pJS55 [29] at the sites EcoR I and BamH I . All plasmids were sequenced to confirm the presence of corresponding mutations . All primer sequences used in subcloning and site-directed mutagenesis please see Supporting Information S1 . All the wild type and mutants of HPV-4 E7 DNA ( GenBank NC_001457 . 1 ) were synthesized ( Celtek Bioscience ) , and cloned into pGEX4T-2 ( GE Healthcare ) for GST fusion protein expression . U2OS cells , Hela cells and SD3443 cells were maintained in Dulbecco's Modified Eagle's Medium ( DMEM ) ( Invitrogen ) supplemented with 10% Fetal Bovine Serum ( FBS ) . Primary human keratinocytes were derived from neonatal foreskins as described previously [30] and were grown in Keratinocyte-SFM medium ( Invitrogen ) . U2OS cells were co-transfected with RcCMV-Rb and pJs55 , pJS55-HA . HPV16 E7 , pJS55-HA . CPV-2 E7ΔLXSXE , pJS55-HA . CPV-2 E7CR1CR2 , pJS55-HA . CPV-2 E7C-RT or pJS55-HA . CPV-2 E7 using Lipofectamine 2000 ( Invitrogen ) as specified by the manufacturer . Hela cells were transfected with pJs55 , pJS55-HA . HPV16 E7 or pJS55-HA . CPV-2 E7 using Lipofectamine 2000 ( Invitrogen ) as specified by the manufacturer . Cell were harvested and lysed by RIPA buffer ( 25 mM Tris•HCl pH 7 . 6 , 150 mM NaCl , 1% NP-40 , 1% sodium deoxycholate , 0 . 1% SDS ) 24 hours post transfection . To prepare retrovirus stocks , SD3443 cells were transfected with E7 retrovirus constructs using Fugen ( Roche applied science , US . ) as specified by the manufacturer . Culture supernatants containing retrovirus were collected 48 h post-transfection . Viral titers of the supernatants were determined using 3T3 cells . The primary HFK cells ( passage 0 ) were infected at a multiplicity of 10 PFU/cell with retrovirus expressing wild type E6 , E7 or E7 mutants . Retrovirus-infected cells were selected in G418 ( 50 ng/ml ) for 2 days . GST and GSTE7 fusion proteins were expressed in BL21pLysS cells ( Invitrogen ) . The cells were induced with 100 µM isopropyl-β-D-thiogalactopyranoside ( IPTG ) 6 hours at 25°C once the optical density at 600 nm reached 0 . 8–1 . 0 . Recombinant CPV-2 E7 and mutants were purified from the supernatant of disrupted cells by glutathione-Sepharose chromatography as previously described [24] . Proteins were extracted from cells and measured concentration as previously described[31] . Proteins were separated on a 4 to 20% Tris-glycine gradient gel ( Novex ) and then were electrophoretically transferred to an Immobilon-P polyvinylidane difluorid ( PVDF ) membrane ( Millipore ) . The primary antibody was used at a dilution of 1∶1 , 000 or 1∶3 , 000 . The secondary antibodies , alkaline phosphatase-conjugated goat anti-mouse IgG and anti-rabbit IgG ( Tropix ) antibodies , were used at a dilution of 1∶2 , 000 . Western blots were visualized by using SuperSignal West Pico Chemiluminescent Substrate ( Thermo Scientific ) . The following commercial antibodies were used: for pRb ( 1∶1000 dilution ) , Rb ( 4H1 ) Mouse mAb ( Cell Signaling technology ) ; for glutathione S-transferase ( 1∶3000 dilution ) , catalog no . 3818-1 ( Clontech ) ; for HA ( 1∶1000 dilution ) , HA . 11 clone 16B12 ( Covance ) . For GST pull-down assays , Jurkat cell or CPEK cell nuclear extract ( 50 µg ) was incubated with 5 µg of GST or GST fusion protein in binding buffer [20 mM Hepes/150 mM KCl/4 mM MgCl2/1 mM EDTA/0 . 02% Nonidet P-40/10% glycerol/0 . 035% 2-mercaptoethanol/1% ( vol/vol ) Sigma protease inhibitor mixture] and rocked for 1 h at 4°C . Glutathione-Sepharose beads ( Amersham Pharmacia Biosciences ) were added to each reaction and rocked for another 1 h at 4°C . The beads were then washed with 1 ml of washing buffer ( 125 mM Tris , 150 mM NaCl , pH 8 . 0 ) four times and boiled with 2× SDS sample buffer , and the proteins were separated by SDS/PAGE . Western blots were used to measure the level of pRb and GSTE7 proteins . The bands of pRb and GSTE7 proteins were quantified by densitometry using Quantity One ( BioRad ) . The relative binding activities were calculated using pRb bound by wild type CPV-2 E7 GST fusion protein as 100% , and normalized with GSTE7 bands . For co-immunoprecipitation assays , N-terminal HA-tagged CPV-2 E7 proteins were immunoprecipitated with a polyclonal anti-pRb antibody ( Santa cruz ) . C-terminal HA-tagged CPV-2 or HPV16 E7 proteins were immunoprecipitated with a polyclonal anti-HA antibody ( Santa Cruz ) . Bead washing buffer was 25 mM Tris•HCl pH 7 . 6 , 150 mM NaCl , 1% NP-40 , 1% sodium deoxycholate , 0 . 1% SDS , 1% Sigma protease inhibitor mixture . The pulled down complexes were resolved on a 4 to 20% gradient gel and then analyzed by Western blotting using either anti-pRb antibody , anti-HA antibody or anti-cullin2 antibody ( Invitrogen ) . Cells in a 3 . 5 cm diameter dish were lysed with 1 ml TRIZOL ( Invitrogen ) . Total RNA was isolated according to manufacturer's protocol . Reverse transcription PCR was performed using ONE STEP RT-PCR KIT ( QIAGEN ) as specified by the manufacturer . All primer sequences and condition please see Supporting Information S1 . Multiple sequence alignments of E7 were prepared using Clustal W . The phyllogenetic analysis was conducted using the Mega version 4 . 0 [32] .
Several studies have demonstrated that the conserved LXCXE motif in the HPV E7 CR2 domain is necessary and sufficient for binding pRb [4] ) . Studies have also revealed that the substitution of C or E or complete deletion mutation of the LXCXE motif destroys pRb binding . Without binding to the pRb , E7 is unable to degrade pRb [4] . Our laboratory recently isolated CPV-2 from footpad and interdigital papillomas of immunosuppressed dogs . Sequencing revealed that the CPV-2 E7 protein contained the typical two C-X-X-C motifs within its carboxyl terminal half but lacked the conserved pRb binding site ( LXCXE ) which is present in COPV ( CPV-1 ) E7 and most HPVs ( Figure 1A ) . CPV-2 E7 has a serine ( amino acid 26 ) in the position of cysteine in the LXCXE motif . In order to test whether CPV-2 E7 could degrade pRb , canine E7 was transduced into human keratinocytes ( HFKs ) , canine kidney cells ( MDCKs ) or canine keratinocytes ( CPEKs ) by using retrovirus infection . Cell lysates were collected , and pRb levels were measured by western blots . Surprisingly , despite lacking the conserved LXCXE motif , CPV-2 E7 was still able to degrade pRb in HFKs , MDCKs and CPEKs ( Figure 1B , C and D ) . To test whether the lower level of pRb was due to a change at the transcriptional level , RT-PCR was performed to measure the level of pRb mRNA . There was no significant difference between the amount of pRb mRNA in control cells and cells with CPV-2 E7 ( Figure 1E ) . In addition , treatment of the E7 expressing cells with the proteasome inhibitor , MG132 , restored the level of pRb ( Figure 1F ) . These data suggest that the reduction of pRb by CPV-2 E7 occurs at the protein level rather than mRNA level , and that degradation is most likely responsible . The degradation of pRb by HPV16 E7 requires high affinity binding [19] . Since CPV-2 E7 lacks the conserved pRb binding motif , LXCXE , there could be two possibilities for the high affinity binding of pRb by CPV-2 E7 . It could be either that the LXSXE motif has the same binding properties as LXCXE , or that CPV-2 E7 has an alternative dominant binding site . We generated E7 mutants ( Figure 2A ) with mutations within the LXSXE domain to investigate whether the LXSXE motif exhibits similar binding to pRb as LXCXE . The GST E7 wild type and mutant fusion proteins were purified from bacteria ( Figure 2B ) , and tested for their binding to pRb . As demonstrated in Figure 2C , wild type CPV-2 E7 binds well to pRb . Substitution of serine26 to cysteine in CPV-2 E7 significantly increased the interaction between E7 and pRb . This is in agreement with mutagenesis studies showing that mutation of HPV16 E7 cysteine24 to serine substantially decreased the binding activity [13] , [33] . In the studies of HPV16 E7 , the substitution of cysteine in LXCXE for glycine abolished the binding between E7 and pRb [14] , [33] , [34] , [35] . However , for CPV-2 E7 , the pRb binding activity of the glycine mutant protein was still retained ( Figure 2C ) . More importantly , the deletion of the entire LXSXE motif did not significantly affect the ability of CPV-2 E7 to bind pRb . This is very surprising since the deletion of the LXCXE motif in HPV 16 E7 totally abolished pRb binding [14] , [33] , [34] , [35] . Thus , while the CPV-2 LXSXE motif exhibits lower pRb binding than the conserved LXCXE motif , it is clear that there is an alternative pRb binding site in the CPV-2 E7 protein . In order to locate potential alternative pRb binding sites in CPV-2 E7 , several CPV-2 E7 truncation mutants were generated ( Figure 3A ) . CPV-2 E7 protein was divided into CR1 ( amino acids 1 to 15 ) , CR2 ( amino acids 16 to 38 ) and the C terminal domains ( amino acids 39 to 98 ) . The GST E7 wild type and mutant fusion proteins were purified from bacteria ( Figure 3B ) . A binding assay with GSTE7 mutants and pRb was performed . In agreement with earlier studies [4] , [14] , [36] , HPV16 E7CR1CR2 bound pRb much more efficiently than the HPV16 E7 carboxyl-terminus domain ( Figure 3C ) , with 6 times more pRb being bound by the combined CR domains . In contrast , the carboxyl-terminus of CPV-2 E7 has much higher binding than the CR1 and CR2 domains , on the average 7 times higher . Importantly , similar binding affinities were also observed between the canine pRb and CPV-2 E7 ( Figure 3D ) , indicating that these unique binding interactions were not due to differences in the species of Rb used for analysis . Thus , CPV-2 E7 uses its carboxyl-terminus instead of the CR1 and CR2 domains to associate with pRb . To verify that the above in vitro binding studies were relevant to in vivo conditions , U2OS cells transduced by HA-tagged CPV-2 E7 constructs were used to study the in vivo association of CPV-2 E7 protein and pRb . Co-precipitation experiments were performed with the anti-pRb antibody ( Figure 4 ) . In contrast to results with HPV 16 E7 , the CPV-2 E7 mutant deleted of the LXCXE-like motif bound pRb as well as the wild-type E7 did . Furthermore , the carboxy-terminal domain of CPV-2 alone bound to pRb very well . Thus , both in vitro and in vivo studies indicate that the E7 carboxyl-terminus mediates pRb binding . The level of canine E7 CR1CR2 construct was not detectable in the cell since it appears to be unstable . The doublet band noted for both wild-type E7 and the LXCXE deletion E7 proteins is most likely due to alkylation by protease inhibitors during the IP procedure . In previous studies , we also observed two distinct forms of HPV-16 E7 and showed that they were generated in vitro by the alkylating reagents , TPCK and TLCK [37] . These reagents are used as a component of a protease-inhibitor cocktail during IP studies to prevent protein degradation . We identified cysteine 27 near the amino terminus of HPV-16 E7 as the alkylation target . In our current study , we did not observe this modification of HPV-16 E7 , apparently due to interference by the amino-terminal epitope tag . In the published study where we observed alkylated HPV-16 E7 , we had used an E7 protein tagged at its C-terminus . We presume that the CPV-2 E7 protein is being modified in a similar fashion , although different sites might be altered than in HPV-16 E7 . Since the mechanism used by CPV-2 E7 to bind pRb is different from that used by HPV E7 , it was also possible that the two E7 proteins used alternative methods to degrade pRb . To test this possibility , a series of E7 deletion and single amino acid substitution mutants were generated . Keratinocytes were transduced with retrovirus encoding either wild type E7 or E7 mutants . Cell lysates were collected , and the level of pRb was measured with immunoblots . Surprisingly , the LXSXE-deletion mutant ( which retains pRb binding ) lost the ability to degrade pRb ( Figure 5A ) . In addition , when the serine at 26 was changed to cysteine ( which increases pRb binding ) , the degradation of pRb was not enhanced ( Figure 5A ) . Thus , although the primary pRb binding of CPV-2 resides in the carboxyl-terminus , we observed that the amino-terminal LXSXE sequence is necessary for the degradation of pRb . Neither the E7 amino- nor carboxyl-terminal domains could independently degrade pRb ( Figure 5B ) . This is in contrast to studies performed on high risk HPV E7 ( reviewed by Munger [38] ) showing that the sequences important for binding and degradation of pRb localized to CR1 and CR2 . Another difference between HPV16 E7 and CPV-2 E7 is highlighted by previous studies showing that HPV16 E7 associates with the cullin 2 ubiquitin ligase complex and that this association contributes to degradation of pRb [4] , [14] , [36] . This does not appear to be true for CPV-2 E7 protein . Co-immunoprecipitation assays failed to detect any association of Cullin2 with CPV-2 E7 ( Figure 5C ) . Our current results demonstrate that at least one animal papillomavirus uses a different mechanism to bind and degrade pRb . To investigate whether some HPVs might use a similar alternative binding mechanism , we screened a papillomavirus phylogentic tree based upon the E7 protein sequence ( Figure 6A ) . E7 proteins from 33 HPVs and 15 animal papillomaviruses were selected according to their genus [39] and aligned using Clustal W [40] with MEGA version 4 . 0 [32] . Based on the alignment , a phyllogenetic tree was assembled by using the minimum evolution method with MEGA version 4 . 0 [32] . CPV-2 E7 was most closely related to the genus gamma-papillomaviruses ( HPV-4 , 48 , 50 , 60 , 65 , 88 , 95 and 116 ) and only distantly related to the genus lambda-papillomaviruses ( CPV-1 and Felis domesticus papillomavirus ) . Interestingly , the alignment of E7 proteins revealed that all the gamma HPVs lack the LXCXE motif ( Figure 6B ) and , indeed , nearly all these gamma-HPVs contain the same LXSXE sequence found in CPV-2 . One of the gamma-HPVs , HPV60 , contains alanine rather than serine in the LXSXE sequence . To test the binding of a representative gamma-HPV E7 to pRb , HPV-4 E7 was synthesized and cloned into an expression vector . HPV-4 E7 protein was divided into CR1 ( amino acids 1 to 15 ) , CR2 ( amino acids 16 to 38 ) and carboxyl-terminal domains ( amino acids 39 to 100 ) ( Figure 6C ) . Wild type HPV-4 E7 and truncation mutants were expressed as GST fusion proteins ( Figure 6D ) and tested for their abilities to bind pRb . Similar to CPV-2 E7 , the HPV-4 E7 protein contained an LXSXE motif and bound pRb ( Figure 6E ) . More interestingly , as shown in Figure 6E , the carboxyl- terminus of HPV-4 E7 bound pRb more efficiently than the CR1CR2 domains . It appears , therefore , that the gamma-type HPVs and CPV-2 share a mechanism by which their E7 proteins interact with pRb via the carboxyl-terminal domain . Due to the similar ability of the CPV-2 and HPV-4 E7 proteins to bind pRb , we also evaluated whether HPV-4 E7 could degrade pRb . In HFKs , HPV-4 E7 reduced the level of pRb in transduced cells , although somewhat less than observed with CPV-2 or HPV16 E7 ( Figure 7A ) . To test whether the lower level of pRb protein might be due to altered gene transcription , we measured pRb mRNA levels by RT-PCR . There was no significant difference in the amount of pRb mRNA in the control cells compared to cells expressing CPV-2 E7 ( Figure 7B ) , indicating that the pRb protein changes were post-translational . More importantly , treatment of the E7 expressing cells with proteasome inhibitor , MG132 , restored the level of pRb protein ( Figure 7C ) . These data , similar to that for CPV-2 , suggest that the reduction of pRb by HPV-4 E7 is most likely the result of protein degradation .
Small DNA tumor viruses , such as HPV , Adenovirus , and Polyomavirus , produce viral oncoproteins that can interact with pRb and alter its function . Targeting pRb appears important for the ability of these viruses to regulate E2F and cell DNA replication and to complete the virus life cycle . All these oncoproteins , E7 , E1A , and LT , use a conserved CR2 domain and LXCXE motif to bind pRb [41] . However , for CPV-2 E7 , the LXCXE-like motif is not necessary for association with or degradation of pRb . It appears that CfPV has evolved an alternative mechanism to bind pRb by using the E7 carboxyl-terminal domain . Although the HPV E7 carboxyl-terminus has been proposed to have an independent , low affinity pRb binding site [14] , [36] , HPV16 E7 mutants with a deletion of the LXCXE motif in CR2 fail to associate with pRb family members as determined by Western blotting and extensive proteomic analyses of associated cellular protein complexes [4] . In contrast , the carboxyl-terminus of CPV-2 E7 exhibits greater pRb binding than the CR1 and CR2 domains . Even more interesting is the finding that the CPV-2 E7 mutant deleted of the LXSXE domain still can bind to pRb with high efficiency . Furthermore , the carboxyl-terminus of CPV-2 E7 alone can bind pRb in vitro and in vivo . The carboxyl-terminus of HPV E7 has been proposed to be important for releasing E2F from pRb [20] , [21] . It will be interesting to map the association site on pRb , and to determine whether the binding induces the release of E2F from pRb . During preparation of this manuscript , three new canine papillomaviruses were identified [42] . One of the viruses , CPV7 , shares high sequence homology with CPV-2 and its predicted E7 ORF encodes a protein with the LXSXE motif . Overall , however , 5 of the 7 identified canine papillomaviruses contain the LXCXE motif in their E7 protein . CPV-1 E7 , which has LXCXE motif , degrades pRb in cells as anticipated ( data not shown ) . As shown in this study , the LXSXE motif is not limited to canine papillomaviruses; the gamma genus of HPVs also have the same motif . This genus consists of eight HPVs , 7 of which contain the E7 LXSXE motif . Interestingly , the gamma HPVs have several other similarities to CPV-2 . First , they induce cutaneous rather than mucosal lesions . Second , they most likely persist in the population as subclinical infections and induce visually detectable tumors only under conditions of immunosuppression . Third , their tumor cells are characterized by histologically-distinguishable intracytoplasmic inclusion bodies [39] . The canine model may provide a new approach for studying the biology of this unique category of papillomaviruses and their stringent regulation by the host immune response .
|
Human papillomaviruses ( HPVs ) are estimated to cause the most common sexually transmitted infection in the world , and these infections are recognized as the major cause of cervical cancer . One of the papillomavirus oncoproteins , E7 , plays a major role in both the viral life cycle and progression to cancer . In cells E7 associates and inactivates pRb , a tumor suppressor protein . For the vast majority of papillomaviruses , E7 binds to pRb using a small amino acid sequence , LXCXE . However , we have now identified a papillomavirus E7 protein that lacks the LXCXE domain yet still binds and degrades pRb . This E7 protein , derived from a carcinogenic canine virus , uses its C-terminal domain to bind pRb . In addition , we discovered that a family of papillomaviruses , the gamma type HPVs , also lacks the LXCXE domain and binds pRb using a similar mechanism .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] |
[
"virology/viral",
"replication",
"and",
"gene",
"regulation",
"virology/virus",
"evolution",
"and",
"symbiosis",
"virology/animal",
"models",
"of",
"infection",
"virology/emerging",
"viral",
"diseases",
"virology",
"virology/viruses",
"and",
"cancer",
"virology/effects",
"of",
"virus",
"infection",
"on",
"host",
"gene",
"expression"
] |
2010
|
The Canine Papillomavirus and Gamma HPV E7 Proteins Use an Alternative Domain to Bind and Destabilize the Retinoblastoma Protein
|
Malignant carcinomas that recur following therapy are typically de-differentiated and multidrug resistant ( MDR ) . De-differentiated cancer cells acquire MDR by up-regulating reactive oxygen species ( ROS ) –scavenging enzymes and drug efflux pumps , but how these genes are up-regulated in response to de-differentiation is not known . Here , we examine this question by using global transcriptional profiling to identify ROS-induced genes that are already up-regulated in de-differentiated cells , even in the absence of oxidative damage . Using this approach , we found that the Nrf2 transcription factor , which is the master regulator of cellular responses to oxidative stress , is preactivated in de-differentiated cells . In de-differentiated cells , Nrf2 is not activated by oxidation but rather through a noncanonical mechanism involving its phosphorylation by the ER membrane kinase PERK . In contrast , differentiated cells require oxidative damage to activate Nrf2 . Constitutive PERK-Nrf2 signaling protects de-differentiated cells from chemotherapy by reducing ROS levels and increasing drug efflux . These findings are validated in therapy-resistant basal breast cancer cell lines and animal models , where inhibition of the PERK-Nrf2 signaling axis reversed the MDR of de-differentiated cancer cells . Additionally , analysis of patient tumor datasets showed that a PERK pathway signature correlates strongly with chemotherapy resistance , tumor grade , and overall survival . Collectively , these results indicate that de-differentiated cells up-regulate MDR genes via PERK-Nrf2 signaling and suggest that targeting this pathway could sensitize drug-resistant cells to chemotherapy .
Multidrug resistance ( MDR ) is the primary obstacle to treating malignant tumors [1] . Cancer cells develop MDR by overexpressing antioxidant enzymes that neutralize the reactive oxygen species ( ROS ) required for chemotherapy toxicity or by up-regulating drug efflux pumps [2] , [3] . In many cancers , these MDR mechanisms are up-regulated by mutation or amplification of genes encoding antioxidant enzymes or drug efflux pumps . Many other cancers , however , up-regulate these genes through nonmutational mechanisms that remain poorly understood . One nonmutational mechanism by which cancer cells acquire MDR is de-differentiation . De-differentiation is a well-established marker of poor prognosis tumors and can occur when differentiated cells are induced into a more primitive stem-cell–like state [4]–[6] . One mechanism by which both cancerous and noncancerous cells can be de-differentiated is through induction of an epithelial-to-mesenchymal transition ( EMT ) [7]–[14] . De-differentiated cancer cells generated by EMT and cancer stem-like cells are both resistant to a wide range of chemotherapies [15]–[19] . Conversely , cells experimentally induced to differentiate are more sensitive to chemotherapies [20]–[23] . Although de-differentiation is known to up-regulate MDR mechanisms as described above , how this occurs is poorly understood . In this article , we examine this question by employing a global transcriptional profiling approach to identify ROS-induced genes that are preactivated in de-differentiated cells . Many of these genes—which are activated in de-differentiated cells even in the absence of oxidative damage—are regulated by a single signaling pathway . We further show that this pathway is critical for de-differentiated cells to resist chemotherapies .
To study the effects of differentiation state on MDR , we used isogenic pairs of human breast epithelial cells ( HMLE ) that were either differentiated and expressed a control vector , or de-differentiated through induction of an EMT—achieved by expressing the Twist transcription factor [24] , [25] . These de-differentiated HMLE-Twist cells were more resistant to the chemotherapy drugs Paclitaxel ( Tax ) and Doxorubicin ( Dox ) than differentiated HMLE-shGFP cells , consistent with prior reports ( 1 . 5× and 2 . 5× , respectively; Figure 1a ) [26] , [27] . To determine how Twist-induced de-differentiation caused MDR , we assessed whether known mechanisms were up-regulated in these cells . Twist overexpression significantly increased efflux pump activity ( Figure 1b ) and lowered ROS levels—both basal and induced by the oxidizer menadione or Dox ( Figure 1c , d ) [28] . Additionally , HMLE-Twist cells displayed significantly lower amounts of lipid peroxidation compared to HMLE-shGFP cells ( Figure 1e ) . As a measure of overall reducing capacity of the cells , we also show that HMLE-Twist cells had a greater pool of reduced glutathione , which could be maintained even in the presence of menadione ( Figure 1f ) . Finally , Twist overexpression led to a significant increase in expression of enzymes involved in ROS metabolism: superoxide dismutase 1 ( SOD1 ) and catalase ( CAT ) ( Figure 1g ) . We suspected that these MDR mechanisms were up-regulated through a normal regulator of the cellular antioxidant response . To identify putative regulators , we transcriptionally profiled HMLE-shGFP and HMLE-Twist cells treated with vehicle or menadione ( Table S1 ) . In the absence of oxidative stress , 1 , 694 genes were differentially expressed between the two cell types , several of which were ROS and efflux-related genes ( Tables S2 and S3 ) . Treatment with menadione induced the expression of 181 and 170 genes in HMLE-shGFP and HMLE-Twist cells , respectively , with 44 genes being commonly induced in both cell types ( Table S4; hypergeometric test , p value<1 . 0×10−10 ) . Of the 181 genes induced by menadione in HMLE-shGFP cells , 54 were already up-regulated in HMLE-Twist cells in the absence of treatment ( Table S5; hypergeometric test , p value<1 . 0×10−10 , Figure 1h ) . Of these 54 genes , 38 were uniquely induced in HMLE-shGFP but not HMLE-Twist cells treated with menadione . This suggests that some oxidative stress response genes are “preactivated” in de-differentiated HMLE-Twist cells . The most significantly preactivated gene in HMLE-Twist cells was heme oxygenase 1 ( HMOX-1 ) —expressed at 8-fold higher levels in HMLE-Twist cells compared to HMLE-shGFP cells and induced 22-fold in differentiated cells treated with menadione . HMOX-1 is a well-characterized enzyme involved in the metabolism of heme , but is also a major target of master antioxidant regulator Nrf2 [29]–[31] . The Nrf2 transcription factor activates an arsenal of antioxidant genes and ABC transporters , and its up-regulation is associated with acquired MDR [32]–[35] . To test whether Nrf2 might be basally active in HMLE-Twist cells , but not HMLE-shGFP cells , we examined Nrf2 target gene expression . Of 1 , 013 Nrf2 direct-target genes , a significant number—142 genes—were up-regulated in HMLE-Twist cells compared to HMLE-shGFP cells in the absence of oxidative stress ( Table S6; hypergeometric test , p value<1 . 0×10−10 ) [36] . Further , 7 of the 54 oxidative stress response genes “preactivated” in HMLE-Twist cells were Nrf2 direct-target genes , representing a significant enrichment over the number predicted by random chance ( Table S5; hypergeometric test , p value = 4 . 9×10−5 , Figure 1h ) . To confirm Nrf2 activation in HMLE-Twist cells , we assessed its subcellular localization by immunofluorescence . In HMLE-shGFP cells , Nrf2 was sequestered in the cytoplasm and translocated to the nucleus when cells were treated with menadione ( Figure 1i ) . In HMLE-Twist cells , however , Nrf2 was constitutively in the nucleus , and treatment with menadione only modestly increased its nuclear accumulation ( Figure 1i ) . These findings demonstrate that Nrf2 is constitutively active in de-differentiated HMLE-Twist cells—even in the absence of exogenous stress . We next examined why Nrf2 was constitutively active in HMLE-Twist cells , even though basal ROS levels are low . Although ROS activate Nrf2 by oxidation , it can also be activated in the absence of oxidative stress by several kinases [37]–[39] . In particular , Nrf2 is directly phosphorylated and activated by the ER-membrane kinase PERK , which is canonically activated under conditions of ER stress as part of the unfolded protein response ( UPR ) [40]–[42] . In this context , PERK relieves ER stress by slowing protein translation through phosphorylation of eiF2α . We have recently shown that PERK is also activated upon EMT-induced de-differentiation—even in the absence of overt ER stress [43] . Consistent with this , we found that PERK is constitutively phosphorylated in HMLE-Twist cells , but not in HMLE-shGFP cells , and inhibition of PERK with a small-molecule inhibitor blocked its phosphorylation ( Figure 2a ) [44] . To understand if PERK controls constitutive Nrf2 activation in HMLE-Twist cells , we assessed Nrf2 localization following PERK inhibition . We found that inhibition of PERK fully reversed the nuclear localization of Nrf2 in HMLE-Twist cells , but did not prevent oxidative stress-induced nuclear accumulation of Nrf2 in either HMLE-shGFP or HMLE-Twist cells ( Figure 2b ) . As a complementary approach to PERK inhibition and to rule out off-target effects of the small-molecule PERK inhibitor , we also generated cell lines in which PERK expression was stably inhibited by two different shRNAs ( Figure 2c ) . Inhibition of PERK by shRNA significantly decreased Nrf2 nuclear localization in HMLE-Twist cells , mirroring the results obtained with the small-molecule PERK inhibitor ( Figure 2d ) . Collectively , these results demonstrate that Nrf2 nuclear localization is controlled by PERK in de-differentiated HMLE-Twist cells . To confirm that Nrf2 nuclear localization correlated with its activation , we assessed Nrf2 target gene expression following PERK inhibition . We found that PERK inhibition significantly decreased HMOX-1 expression in HMLE-Twist cells , but did not prevent induction of HMOX-1 in response to oxidative stress ( Figure 2e ) . Moreover , using microarray gene expression analyses , we found that PERK inhibition decreased the expression of 58 of the 142 Nrf2-target genes ( 41% ) activated in HMLE-Twist cells ( Table S6 ) . Amongst these PERK-Nrf2-target genes were ABC transporters , enzymes involved in glutathione metabolism and ROS buffering , and several proteins with known roles in drug resistance . These findings confirm that the exit of Nrf2 from the nucleus correlates with down-regulation of its target genes . PERK has previously been shown to bind to , directly phosphorylate , and activate Nrf2 , though the exact phosphorylation sites have not yet been determined [42] . To show that PERK directly regulates Nrf2 in our system , we performed PERK immunoprecipitation followed by western blot with a Nrf2-specific antibody—which confirmed that PERK and Nrf2 directly interact in HMLE-Twist cells ( Figure 2f ) . We also immunoprecipitated Nrf2 in either the presence or absence of the PERK inhibitor , which demonstrated that Nrf2 phosphorylation was markedly reduced by PERK inhibition ( Figure 2g ) . These data , combined with our finding that inhibiting PERK decreases nuclear accumulation of Nrf2 , suggest that PERK directly interacts with Nrf2 to mediate its nuclear translocation and activation . We next tested whether inhibition of PERK would eliminate MDR phenotypes associated with HMLE-Twist cells . PERK inhibition caused a 45% increase in mitochondrial ROS levels in HMLE-Twist cells , but did not affect HMLE-shGFP cells ( Figure 3a ) . PERK inhibition also significantly increased lipid peroxidation in HMLE-Twist cells , but not in HMLE-shGFP cells ( Figure 3b ) . PERK inhibition compromised ROS buffering—cells pretreated with the PERK inhibitor produced 25%–55% more ROS than vehicle-treated cells ( Figure 3c ) . Additionally , PERK inhibition led to a significant decrease in the expression of ROS metabolizing enzymes SOD1 and CAT ( Figure 3d ) . Lastly , inhibition of PERK signaling reduced the percentage of high-effluxing HMLE-Twist cells by 50% and did not affect efflux in HMLE-shGFP cells ( Figure 3e ) . Together these results demonstrate that a simple change in differentiation state confers MDR phenotypes , and these are mediated by constitutive PERK signaling . To understand how this applies in the context of cancer , we expanded our analyses to include several luminal and basal-like breast cancer cell lines , which represent epithelial-like/differentiated and mesenchymal-like/de-differentiated cells , respectively [45] . Previous work has shown that PERK is preferentially activated in basal compared to luminal cell lines [43] . Consistent with our results in the HMLE system , basal breast cancer cells had lower overall ROS than luminal cells , and addition of the PERK inhibitor caused a dramatic increase in ROS levels in basal cells but not luminal cells ( Figure 3f ) . Likewise , inhibition of PERK caused a 25% reduction in the ratio of reduced to oxidized glutathione in only the basal cell lines , indicative of decreased ROS buffering ( Figure 3g ) . This indicates that PERK contributes to the enhanced oxidative stress buffering ability of both noncancerous and cancerous de-differentiated cells . In order to affect chemotherapy resistance , we rationalized that PERK inhibition would need to occur prior to chemotherapy exposure to allow time for reversal of MDR phenotypes ( Figure 4a ) . Pretreatment with the PERK inhibitor greatly sensitized both HMLE-Twist and HMLE-shGFP cells to subsequent treatment with Tax and Dox—the number of surviving cells was reduced significantly in both cell types ( Figure 4b ) . Treatment with a ROS-scavenging agent n-acetyl cysteine ( NAC ) was able to rescue this decreased survival , indicating that PERK pathway activation contributes to chemotherapy resistance in significant part via ROS buffering ( Figure 4c , d ) [46] . We also utilized a small molecule—oltipraz—capable of inducing Nrf2 activation ( Figure 4e , f ) [47] . Activation of Nrf2 significantly rescued PERK-dependent decreases in cell survival . To rule out the possibility that off-target effects of oltipraz were responsible for this effect , we performed the same rescue experiment in cells with stable Nrf2 knockdown achieved by two independent shRNAs . When Nrf2 was inhibited , oltipraz was no longer able to rescue the effects of PERK inhibition , confirming that these effects were mediated by Nrf2 ( Figure 4g–i ) . These results indicate that PERK signaling through Nrf2 is responsible for the acquisition of MDR . These results prompted us to test the effect of PERK inhibition in vivo , utilizing xenografted tumors derived from therapy-resistant basal breast cancer cells . We utilized a treatment plan involving cycles of pretreatment with the PERK inhibitor , followed immediately by treatment with Dox . The combined treatment resulted in significantly smaller tumors compared to single or mock treatments ( Figure 5a ) . To test if PERK inhibition affected ROS buffering in vivo , we harvested tumors from each of the four treatment groups and measured the expression of the ROS-metabolizing enzyme SOD1 . We found that the Dox , PERK inhibitor , and combined treatment groups all had significantly reduced expression of SOD1 compared to control tumors , with the dual-treated tumors having the lowest expression ( Figure 5b ) . Additionally , the combined treatment group had the most necrotic cells compared to the other treatment groups ( Figure 5c ) . We next adjusted the dosage schedule to highlight the synergistic interactions between PERK inhibition and Dox treatment and found that reducing the total dosage and frequency of treatments further emphasized the sensitization effect—dual-treated tumors were 4 times smaller than Dox-treated tumors and >5 times smaller than PERK inhibitor or mock-treated groups ( Figure 5d ) . Although prior research has shown that PERK is critical for tumor growth and angiogenesis [48]–[50] , we found that low-dose inhibition only minimally impacted tumor growth in the absence of chemotherapy . To assess the in vivo effects on ROS buffering , we measured the levels of reduced glutathione ( GSH ) in tumors harvested from each treatment group . Dox , PERK inhibitor , and combined treatment groups all had decreased levels of GSH compared to the control group , with the dual-treated tumors having the lowest amount ( Figure 5e ) . As an important control to demonstrate that the observed in vivo results were not due to off-target effects of the PERK inhibitor , we utilized xenografted tumors derived from luminal breast cancer cells . Although treatment with Dox led to a reduction in tumor size , inhibition of PERK did not provide any additive benefit in the luminal tumors ( Figure 5f ) . This confirms that the effects observed in the basal breast cancer xenografts are not due to off-target effects of the PERK inhibitor , as luminal cells—unlike basal cells—do not constitutively activate PERK and do not significantly respond to PERK inhibition . Together our results suggest that combining Dox treatment with PERK inhibition compromises the ROS-buffering capacity of basal-like breast cancer cells and sensitizes them to chemotherapy-induced cell death . To assess the clinical relevance of our findings , we analyzed primary human breast tumor datasets . Utilizing two independent datasets ( comprised of 413 patient tumors ) , we first tested for correlations between the expression of PERK pathway genes and genes associated with the basal subtype of breast cancer . We found that a PERK gene expression signature correlated positively with a basal breast cancer gene signature , suggesting that the PERK signaling pathway is active in basal breast tumors ( Figure 5g ) [51] . As a negative control , an IRE1 gene expression signature did not show a significant correlation ( Figure 5g ) . Additionally , we found that PERK pathway activity could stratify patient response to therapy—85% of PERK-low tumors displayed complete or partial response to therapy , compared to only 38% of PERK-high tumors ( Figure 5h ) . Finally , PERK pathway expression also correlated to differentiation state and overall survival in invasive high-grade glioma—tumors stratified into a PERK-high group were almost exclusively poorly differentiated grade 4 GBM and had significantly worse overall survival than the PERK-low group ( Figure 5i , j ) . These results highlight the relevance of our work in primary tumors , and suggest that targeting PERK signaling may be beneficial in highly aggressive and malignant tumor types .
These findings identify PERK-Nrf2 signaling as one mechanism by which de-differentiated cells gain MDR . Because they constitutively activate Nrf2 , these de-differentiated cells constitutively express antioxidant enzymes and drug efflux pumps . Remarkably , in this setting , Nrf2 is not activated by oxidation , but rather through a previously reported mechanism involving its phosphorylation by PERK [42] . This finding is of particular interest given Nrf2's known role in promoting chemotherapy survival [34] and its constitutive activation by mutation in a subset of tumors [52]–[54] . Our findings indicate that a change in cellular state , in the absence of mutation or oxidative stress , can also lead to constitutive Nrf2 activation . This enables de-differentiated cells to survive chemotherapy by preventing cellular damage before it occurs . In contrast , differentiated cells activate Nrf2 only after proteins and DNA have been oxidized . Although this defensive response may succeed in neutralizing toxins , the damage to cellular components would have already occurred . Our findings also highlight the importance of stress signaling in cancer . Cancer cells activate stress response pathways to protect themselves from harsh environments encountered during tumor growth and metastasis—for example , hypoxia and nutrient deprivation—and also during the course of chemotherapy . We show that de-differentiated tumor cells preactivate PERK-Nrf2 signaling in the absence of stress and that inhibition of PERK sensitizes these cells to chemotherapy . These observations complement prior studies establishing a role for the UPR and its downstream targets in chemosensitization [55]–[67] . Collectively , our findings provide mechanistic insights into how cellular de-differentiation promotes MDR and suggest that inhibiting PERK-Nrf2 signaling may reverse the MDR of cancer cells that are otherwise drug resistant .
This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health . The protocol was approved by the Animal Care and Use Committee of the Massachusetts Institute of Technology ( Protocol No . 0611-071-14 ) . All surgery was performed under isoflurane anesthesia , and every effort was made to minimize suffering . HMLE-shGFP and HMLE-Twist cell lines were a kind gift of Dr . Robert Weinberg and cultured as described previously [8] . Basal breast cancer ( MDA . MB . 231 , Hs578t ) and luminal breast cancer ( MCF7 , T47D , MDA . MB . 361 ) cell lines were purchased from ATCC and cultured in DMEM+10% FBS . The SUM159 basal breast cancer cell line was purchased from Asterand and cultured in Ham's F-12+5% FBS , Insulin and Hydrocortisone . Chemical oxidizer menadione , Nrf2 activator oltipraz , and ER-stress inducer thapsigargin were purchased from Sigma-Aldrich . Cumene hydroperoxide ( CH ) was purchased from Life Technologies . The PERK inhibitor ( PERKi ) was described previously and purchased from EMD Millipore [44] . Lentiviral short hairpin RNA ( shRNA ) constructs were generated as described previously [68] . Lentiviral integration was selected with 1 µg/ml puromycin or 10 µg/ml blasticidin for 7 d , and knockdown efficiency was measured by quantitative RT-PCR . ROS production was measured by fluorescent imaging or flow cytometry analysis of MitoSOX or CellROX probes ( Life Technologies ) according to manufacturer instructions . Lipid peroxidation was assessed using the Click-iT Lipid Peroxidation Imaging Kit ( Life Technologies ) according to the manufacturer instructions . Total , reduced , and oxidized glutathione were determined using the GSH/GSSG-Glo™ Assay ( Promega ) according to manufacturer instructions . MDR1 efflux ability was measured by flow cytometry quantification of DiOC2 ( 3 ) -dye efflux ( EMD Millipore ) . Efflux assays were conducted according to manufacturer instructions . Briefly , cells were loaded with DiOC2 ( 3 ) -dye for 10 min and either kept on ice or placed at 37°C for 1 . 5 h to allow efflux of the dye . Control and efflux samples were then immediately analyzed by flow cytometry . Cells were lysed with cold RIPA buffer plus complete protease inhibitor cocktail ( Roche Applied Science ) . The signal was detected using the SuperSignal ECL system ( Thermo Scientific ) . The following antibodies were used for immunoblotting: SOD1 and Nrf2 ( Santa Cruz Biotechnology ) and CAT , HMOX-1 , pan-phospho , and β-actin ( Cell Signaling Technologies ) . HMLE-Twist cells grown in the presence or absence of 1 µM PERK inhibitor for 48 h were lysed in nondenaturing lysis buffer ( 20 mM Tris pH 7 . 5 , 150 mM NaCl , 2 mM EDTA , 1% NP-40 , supplemented with cocktails for phosphatase and protease inhibition ) . Equal protein amounts were used for immunoprecipitation using PERK or Nrf2 antibody as per the vendors' instructions . Samples were analyzed by immunoblotting using antibodies to Nrf2 and phospho-Ser/Thr–containing proteins . Anti-Nrf2 ( C-2 ) and anti–phospho-PERK ( pPERK ) antibodies was purchased from Santa Cruz Biotechnology . Cells were fixed on glass chamber slides in 4% PFA for 5 min , blocked with 5% BSA in PBS , and incubated with primary antibody at a 1∶50 dilution for 2 h . Slides were then washed with PBS and incubated with an Alexa Fluor 488 anti-rabbit secondary antibody . The nuclei were then stained with DAPI prior to analysis . Immediately after harvest , tumors were fixed in 4% PFA for 24 h and paraffin-embedded . For staining , slides were deparaffinized in xylene and then rehydrated with ethanol and double distilled water . Hydrogen peroxide was used to block nonspecific sites , and Diva Decloaker ( BioCare Medical ) solution and microwaving were used for antigen retrieval . Sections were incubated for 2 h at room temperature with a SOD1 antibody ( Santa Cruz Biotechnology ) . Expression was detected using HRP anti-rabbit secondary antibody ( BioCare Medical ) and betazoid DAB ( BioCare Medical ) . The slides were counterstained with hematoxylin . HMLE-shGFP and HMLE-Twist cells were treated with 1 µM of PERKi or DMSO for 48 h and then treated with 40 µM menadione or DMSO for 2 h immediately before RNA extraction . Total RNA were extracted using Qiagen RNeasy kit , and integrity and quality verified prior to analysis . Gene expression analyses were conducted using Affymetrix GeneChip Human Genome U133 Plus 2 . 0 Arrays according to standard Affymetrix protocols , with normalization as described previously [69] . Alteration of gene expression by PERKi and/or menadione was calculated by comparing the expression of each gene across treatment groups for each cell type . The gene-expression data have been deposited in the NCBI Gene Expression Omnibus public database ( GEO; GSE59780 ) . Cells were seeded and treated according to the schedule described in Figure 4a . Briefly , cells were seeded on day 0 , pretreated with PERKi ( 1 µM ) or DMSO for 48 h on days 1–3 , and rescued with NAC ( 3 mM ) , oltipraz ( 25 µM ) , or DMSO for 24 h on day 3 . Cells were then treated with Dox ( 30 nM ) , Tax ( 2 nM ) , or DMSO on day 4 . Cells were followed for an additional 5 d , with complete media and drug replacement on day 7 . Cell survival was assessed on day 9 by manual cell count and normalized as described for each experiment . Female NOD/SCID mice were purchased from Jackson Labs . The Animal Care and Use Committee of the Massachusetts Institute of Technology approved all animal procedures . For tumor regression studies , 1×106 cells were injected bilaterally into the mammary fat pad of 6–8-wk-old female NOD/SCID mice . After reaching 60–80 mm3 , tumors were treated with PERKi ( 7 . 5 mg/kg/tumor ) or DMSO by intratumoral injection on days 1 , 2 , 4 , 5 , 8 , 9 , 11 , and 12 , and Dox ( 2 . 5 mg/kg ) or PBS by intraperitoneal ( IP ) injection on days 2 , 5 , 9 , and 12 unless otherwise specified . Tumor volume over time and average tumor weight at sacrifice were measured and presented as the average ± standard error of mean for 10 tumors per treatment group . For correlation analyses , the PERK gene expression signature was defined as the top 500 genes down-regulated in de-differentiated cells treated with 1 µM PERKi for 48 h . The PERK signature scores were calculated for each patient sample from human breast cancer ( GSE3143 , GSE41998 ) and glioma ( GSE4412 ) datasets by summing the log-transformed normalized expression values for each probe in the signature set . High- and low-PERK signature groups were defined as the top or bottom 15% of samples within each group . Spearman's rho was used to measure correlation , and a p value was determined by Monte Carlo sampling as described previously [43] . All data are presented as mean ± standard error of mean unless otherwise specified . Student t test ( two-tailed ) was used to calculate p values , and p<0 . 05 was considered significant .
|
The development of multidrug resistance is the primary obstacle to treating cancers . High-grade tumors that are less differentiated typically respond poorly to therapy and carry a much worse prognosis than well-differentiated low-grade tumors . Therapy-resistant cancer cells often overexpress antioxidants or efflux proteins that pump drugs out of the cell , but how the differentiation state of cancer cells influences these resistance mechanisms is not well understood . Here we used genome-scale approaches and found that the PERK kinase and its downstream target , Nrf2—a master transcriptional regulator of the cellular antioxidant response—are key mediators of therapy resistance in poorly differentiated breast cancer cells . We show that Nrf2 is activated when cancer cells de-differentiate and that this activation requires PERK . We further show that blocking PERK-Nrf2 signaling with a small-molecule inhibitor sensitizes drug-resistant cancer cells to chemotherapy . Our results identify a novel role for PERK-Nrf2 signaling in multidrug resistance and suggest that targeting this pathway could improve the responsiveness of otherwise resistant tumors to chemotherapy .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cellular",
"stress",
"responses",
"cell",
"biology",
"gene",
"expression",
"genetics",
"cellular",
"types",
"biology",
"and",
"life",
"sciences",
"cell",
"processes",
"molecular",
"cell",
"biology"
] |
2014
|
De-Differentiation Confers Multidrug Resistance Via Noncanonical PERK-Nrf2 Signaling
|
The Arabidopsis COP1/SPA E3 ubiquitin ligase is a key negative regulator that represses light signaling in darkness by targeting transcription factors involved in the light response for degradation . The COP1/SPA complex consists of COP1 and members of the four-member SPA protein family ( SPA1-SPA4 ) . Genetic analysis indicated that COP1/SPA2 function is particularly strongly repressed by light when compared to complexes carrying the other three SPAs , thereby promoting a light response after exposure of plants to extremely low light . Here , we show that the SPA2 protein is degraded within 5–15 min after exposure of dark-grown seedlings to a pulse of light . Phytochrome photoreceptors are required for the rapid degradation of SPA2 in red , far-red and also in blue light , whereas cryptochromes are not involved in the rapid , blue light-induced reduction in SPA2 protein levels . These results uncover a photoreceptor-specific mechanism of light-induced inhibition of COP1/SPA2 function . Phytochrome A ( phyA ) is required for the severe blue light responsiveness of spa triple mutants expressing only SPA2 , thus confirming the important role of phyA in downregulating SPA2 function in blue light . In blue light , SPA2 forms a complex with cryptochrome 1 ( cry1 ) , but not with cryptochrome 2 ( cry2 ) in vivo , indicating that the lack of a rapid blue light response of the SPA2 protein is only in part caused by a failure to interact with cryptochromes . Since SPA1 interacts with both cry1 and cry2 , these results provide first molecular evidence that the light-regulation of different SPA proteins diverged during evolution . SPA2 degradation in the light requires COP1 and the COP1-interacting coiled-coil domain of SPA2 , supporting that SPA2 is ubiquitinated by COP1 . We propose that light perceived by phytochromes causes a switch in the ubiquitination activity of COP1/SPA2 from ubiquitinating downstream substrates to ubiquitinating SPA2 , which subsequently causes a repression of COP1/SPA2 function .
As sessile organisms plants continuously monitor the ambient light conditions and adjust their growth and development with the aim to optimize growth and—ultimately—seed production in a competitive environment . Plants sense the intensity , color , direction and periodicity of light . Responses to these light parameters include seedling deetiolation ( inhibition of hypocotyl elongation , opening of cotyledons and apical hook , greening ) , phototropism , shade avoidance , the accumulation of anthocyanins and the induction of flowering in particular day lengths [1] . To sense the light , plants have evolved several classes of photoreceptors [1 , 2] . The phytochrome photoreceptors sense red light ( R ) and far-red light ( FR ) and exist in two R/FR photointerconvertible conformations . Among the five phytochromes in Arabidopsis ( phyA-phyE ) , the relatively light-stable phyB is the primary phytochrome controlling FR-reversible responses to R . These responses are also named low fluence responses ( LFR ) . phyA is rapidly degraded in R and primarily mediates high-irradiance responses ( HIR ) to continuous FR ( FRc ) and very low fluence responses ( VLFR ) [3 , 4] . Blue light ( B ) is sensed by cryptochromes , phototropins and the ZEITLUPE family , but also by phyA . Cryptochromes are encoded by two genes in Arabidopsis , CRY1 and CRY2 . Both mediate seedling deetiolation in B , while primarily cry2 is responsible for B-induced flowering in long days [5 , 6] . For both , phytochromes and cryptochromes , mutant photoreceptor variants have been identified that are constitutively active and thus signal also in darkness [7–10] . Recently , UVR8 was identified as the long-sought UV-B receptor [11 , 12] . In Arabidopsis , the phytochrome and cryptochrome photoreceptors act to inhibit a key repressor of light signaling that prevents light responses in darkness . This repressor , the CONSTITUTIVELY PHOTOMORPHOGENIC1/SUPPRESSOR OF PHYA-105 ( COP1/SPA ) complex , functions as an E3 ubiquitin ligase which ubiquitinates positively-acting light signaling intermediates , mainly transcription factors , thereby targeting them for proteolytic degradation in the 26S proteasome . In the light , photoreceptors directly interact with the COP1/SPA complex , leading to its inactivation which subsequently allows the target transcription factors to accumulate and to initiate vast reprogramming of gene expression [13 , 14] . The degradation of the light-labile photoreceptors phyA and cry2 is also in part dependent on COP1 and/or SPA genes [15–18] . The Arabidopsis COP1/SPA complex is likely a tetramer consisting of two COP1 and two SPA subunits [19] . COP1 is a single-copy gene in higher plants , while SPA proteins are encoded by a small gene family of four genes in Arabidopsis ( SPA1-SPA4 ) and 2 genes in rice [13 , 20] . Mutations in either COP1 or all four SPA genes lead to constitutive photomorphogenesis in Arabidopsis , with seedlings showing the features of light-grown seedlings in complete darkness [21 , 22] . While cop1 null mutants arrest growth at the seedling stage , spa null mutants are viable . cop1 spa quintuple null mutants can complete embryogenesis , indicating that the COP1/SPA complex is not necessary for embryogenesis [23] . Apart from controlling seedling growth , the COP1/SPA complex also plays an important role during other light-induced responses , such as anthocyanin biosynthesis , elongation responses during shade avoidance , leaf expansion and the suppression of flowering under non-inductive short-day conditions . These responses are mediated through a number of COP1/SPA substrates including CO , HFR1 , PAP1 , PAP2 and BBX family proteins [24–32] . Moreover , COP1/SPA is a positive regulator in UV-B mediated photomorphogenesis [11 , 12] . The four SPA genes have overlapping but also distinct functions in controlling the various light responses during plant development [22 , 24–26 , 33] . The COP1/SPA complex acts as part of a CULLIN4 ( CUL4 ) -based E3 ubiquitin ligase . CUL4-associated E3 ligases consist of CUL4 , RBX1 , DDB1 as well as a variable WD repeat protein which recognizes the substrate and binds DDB1 [34 , 35] . The WD repeat proteins COP1 and SPA are substrate adaptors in CUL4-DDB1COP1/SPA E3 ligase ( s ) [36] . Both COP1 and SPAs contain a central coiled-coil domain responsible for the formation of the COP1/SPA complex via homo- and heterodimerization [19 , 37 , 38] . In their C-termini , both COP1 and SPAs carry a WD-repeat domain which mediates interaction with substrates as well as with DDB1 [36 , 39] . The N-termini of COP1 and SPA are distinct , with COP1 harboring a RING finger domain and SPA proteins carrying a kinase-like domain [40 , 41] . Light is the key factor controlling COP1/SPA activity . Genetic studies showed that the SPA2 protein is particularly strongly inactivated by light when compared to the other three SPAs , making SPA2 a particularly interesting SPA when analyzing light-mediated inhibition of COP1/SPA activity [22 , 42] . How light inactivates the COP1/SPA complex is not fully understood . Evidence indicates that phytochrome and cryptochrome photoreceptors converge on COP1/SPA to promote light signaling in R , FR and B . Such light-induced inactivation of COP1/SPA occurs via multiple mechanisms . First , after light exposure , COP1 translocates from the nucleus into the cytoplasm [43 , 44] . Second , the B-dependent interaction of cry1 with SPA1 reduces the COP1/SPA1 interaction [45–47] . Similarly , an interaction of light-activated phytochromes A and B with members of the SPA family reduces the interaction between COP1 and SPA proteins [48 , 49] . For cry2 , B acts to promote the interaction of cry2 with COP1 [50] . A third mechanism which reduces COP1/SPA activity in FRc-grown plants involves the degradation of SPA1 and SPA2 in the proteasome [42] . Here , we have analyzed the molecular mechanism of SPA2-degradation in different light qualities and uncover a photoreceptor-specific mechanism of light-induced COP1/SPA repression via COP1 .
To investigate the dynamics and wave-length dependency of light-induced SPA2 degradation , we determined SPA2 protein levels in dark-grown seedlings that were briefly exposed to R , FR or B . These seedlings expressed HA-tagged SPA2 under the control of the 5′ and 3′ regulatory sequences of SPA2 ( SPA2::SPA2-HA ) [42] . The SPA2 promoter expresses at the same level in dark-grown and light-exposed seedlings [42 , 51] . Therefore , light-induced differences in SPA2-HA protein levels in these lines are due to changes in protein stability , as shown previously [42] . Exposure of dark-grown seedlings to a short , 200-second pulse of R ( Rp ) was sufficient to strongly reduce SPA2-HA protein levels within 5 min after subsequent transfer to darkness ( Fig 1A ) . Ten minutes after the Rp , there was barely any SPA2-HA protein detectable . Similarly , when dark-grown seedlings were irradiated with a pulse of FR ( FRp ) or B ( Bp ) , SPA2-HA protein abundance decreased to a very low level . The response time to FRp and Bp was also very rapid , but slightly longer when compared to Rp . To determine whether the native SPA2 protein behaves like the SPA2-HA protein , we analyzed SPA2 protein levels in wild-type seedlings using an α-SPA2 antibody . Because SPA2 levels are very low , we enriched the protein preparations through nuclear extracts to detect the constitutively nuclear-localized SPA2 protein [22 , 42] . Fig 1B shows that a pulse of FR , R or B rapidly and strongly reduced SPA2 protein abundance . Again , Rp was more effective in reducing SPA2 levels than FRp and Bp . We subsequently asked what fluences are necessary for the reduction of SPA2 protein levels . Fluences of 0 . 002 μmol m-2 of R , i . e . a 200-s-pulse of R with a fluence rate of 10−5 μmol m-2 s-1 , was sufficient to reduce SPA2 protein levels to almost undetectable levels ( Fig 1C ) , indicating that degradation in R is extremely sensitive to light and likely involves a VLFR . FRp and Bp were again less effective than Rp ( Fig 1C ) . We subsequently asked whether R , FR and B also cause a decrease in SPA1 levels . Here , the induction of SPA1 gene expression by light [40] precluded a specific analysis of protein stability using α-SPA1 antibodies . Therefore , we used transgenic lines expressing SPA1-HA under the control of the constitutive SPA2 promoter . These lines showed a strong reduction in SPA1-HA abundance in FR , R and B ( Fig 2 ) . We asked which photoreceptor ( s ) are responsible for degradation of SPA2 in different light qualities and quantities . To this end we investigated SPA2 levels in various photoreceptor mutants . Degradation of SPA2 in response to FRc was fully abolished in a phyA mutant , in both Col and RLD accessions ( Fig 3A ) . Similarly , a pulse of FR had no effect on SPA2 protein levels in a phyA mutant ( Fig 3B ) . Hence , phyA is responsible for SPA2 degradation in FR . After a pulse of R with low fluence rates , SPA2 protein levels were not reduced in a phyA mutant nor in a phyA phyB double mutant . A phyB mutant , in contrast , showed a reduction in SPA2 levels after Rp and thus exhibited a similar response as the wild type ( Fig 3C ) . The phyA-requirement for a response to low Rp confirms that this treatment initiates a phyA-perceived VLFR [3] . After a pulse with higher fluence rates of R , only the phyA phyB double mutant lacked a reduction in SPA2 protein abundance when compared to dark-grown seedlings ( Fig 3D ) . Hence , phyA and phyB mediate the degradation of SPA2 after high Rp . This suggests that both VLFR and LFR responses trigger SPA2 degradation in red light . R/FR reversibility is a hall-mark of an LFR [4] . Indeed , SPA2 degradation after Rp was reversible by a pulse of FR in a phyA mutant background which would lack the VLFR ( Fig 3E ) . Deetiolation in blue light is mediated by the cryptochromes cry1 and cry2 as well as by phyA . We therefore investigated B-induced degradation of SPA2 in cry1 cry2 and in phyA mutants . After a pulse of B , the decrease in SPA2 levels was abolished in phyA mutant seedlings but was normal in the cry1 cry2 mutant ( Fig 4A ) . These results indicate that a pulse of B only triggered phyA-mediated SPA2 degradation . Also after irradiation with continuous B of very high fluence rates ( 50 μmol m-2 s-1 ) for 30 min high SPA2 levels were retained in phyA mutant seedlings . In cry1 cry2 mutant seedlings , SPA2 levels were again strongly reduced similar to wild-type seedlings ( Fig 4B ) . Only after prolonged irradiation with B of high fluence rates for 24 h , SPA2 levels decreased in a phyA-deficient mutant ( Fig 4C ) . These results show that the rapid B-induced reduction in SPA2 levels is exclusively mediated by phyA . Only after very long irradiation with B of high fluence rates other photoreceptor ( s ) become active in reducing SPA2 levels . In an attempt to uncover a possible role of cryptochromes in the response to long-term B irradiation , we analyzed SPA2 protein levels in a cry1 cry2 phyA-201 triple mutant background ( Ler accession ) . SPA2 protein levels still decreased in this mutant after prolonged exposure to blue light ( S1A Fig ) . However , SPA2 levels in phyA-201 also decreased after FRc ( S1B Fig ) . Hence , degradation of SPA2 in the cry1 cry2 phyA-201 triple mutant may either be due to residual phyA activity or , alternatively , the regulation of SPA2 stability may be different in the Ler accession than in the Col and RLD accessions . Constitutively active photoreceptor variants have been described that initiate light signaling even in darkness . We therefore investigated whether these photoreceptor variants also cause a constitutive reduction in SPA2 protein abundance , i . e . also in dark-grown seedlings . To this end , we analyzed SPA2 protein levels in transgenic lines expressing the constitutively active phytochrome mutants phyBY276H and phyAY242H [7] . As reported previously , phyBY276H-expressing seedlings showed very strong constitutive photomorphogenesis , both in the PHYB wild-type and the phyB-5 mutant background [7] ( Fig 5A ) . These phyBY276H lines showed very low SPA2 protein levels in dark-grown seedlings ( Fig 5B ) , suggesting that the SPA2 protein is destabilized in darkness by the constitutively active phyB photoreceptor . phyAY242H-expressing seedlings exhibit weaker constitutive photomorphogenesis than phyBY276H-expressing seedlings [7] . These seedlings had a shorter hypocotyl and a partially opened hook , especially in the phyA mutant background , when compared to the wild type ( Fig 5A ) . This phenotype was somewhat weaker than reported previously which is likely due to the younger age of our seedlings and the absence of sucrose in the culture medium when compared to [7] . SPA2 abundance in dark-grown seedlings was strongly reduced in lines expressing phyAY242H in a phyA background , while it was similar to wild type in lines expressing phyAY242H in a PHYA wild-type background ( phyAY242H/Ler ) ( Fig 5B ) . Hence , SPA2 levels were constitutively reduced in the presence of phyAY242H , but this effect is outcompeted by the presence of wild-type phyA . In summary , these mutations in both phyA and phyB cause constitutive degradation of SPA2 in darkness . Expression of the N-terminal 406 amino acids of phyA fused to an artificial dimerization domain has also been shown to cause constitutive photomorphogenesis in darkness [10] ( Fig 5C ) . However , this phyA variant did not alter SPA2 protein levels in darkness ( Fig 5D ) , indicating that this constitutively active phyA variant was not capable of inducing SPA2 degradation in darkness . Fusion of the cry1 C-terminal extension ( CCT1 ) to an artificial dimerization domain ( GUS ) leads to a constitutively active cry1 photoreceptor . Similarly , a cry1G380R variant is constitutively active . Hence , seedlings expressing CCT1 or cry1G380R exhibit strong constitutive photomorphogenesis in darkness [8 , 9] . SPA2 protein levels were unaltered in dark-grown GUS-CCT1- and cry1G380R-expressing seedlings when compared to the wild type , despite the constitutive photomorphogenesis displayed by these seedlings ( Fig 5E and 5F ) . Hence , none of the constitutively active cry1 variants affected SPA2 protein levels in darkness . This is in agreement with the primary roles of phytochromes in SPA2 degradation . Since rapid degradation of SPA2 in B was exclusively dependent on phyA , we predicted that phyA is of particular importance in inactivating SPA2 function in B . To test this hypothesis , we generated a phyA-deficient spa1 spa3 spa4 phyA mutant which only expresses functional SPA2 among the four SPA proteins . Hence , we can observe the effect of light on SPA2 activity in the absence of any other SPAs , and in the presence or absence of phyA . We had shown previously that spa1 spa3 spa4 mutant seedlings etiolate normally in darkness but are very hypersensitive to R , FR and B when compared to the wild type , thus resembling a spa quadruple mutant already at extremely low fluence rates of light [22 , 42] ( Fig 6A–6C ) . Hence , SPA2 is sufficient for full repression of photomorphogenesis in darkness but is extremely effectively inactivated by light . In B , spa1 spa3 spa4 phyA mutant seedlings displayed much longer hypocotyls than spa1 spa3 spa4 mutant seedlings , indicating that the lack of phyA dramatically reduced the responsiveness of the spa1 spa3 spa4 mutant to B . The hypocotyl length of the spa1 spa3 spa4 phyA mutant in B was very similar to that of the phyA single mutant . Hence , in the absence of phyA , the mutations in SPA1 , SPA3 and SPA4 had no detectable effect ( Fig 6A ) . In Rc , the phyA mutation abolished the hypersensitivity of the spa1 spa3 spa4 mutant to lower fluence rates of Rc but not to higher fluence rates of Rc ( Fig 6B ) . This is consistent with our finding that SPA2 degradation in lower fluence rates of R requires phyA , while in higher fluence rates of R phyB in addition to phyA mediates SPA2 degradation . As expected , the responsiveness of spa1 spa3 spa4 mutant seedlings to FRc was fully dependent on phyA ( Fig 6C ) . Taken together , these results show that the hypersensitivity of the spa1 spa3 spa4 mutant to B fully depends on phyA . This agrees with our observation that rapid SPA2 degradation in B was exclusively dependent on phyA . Since the spa1 spa3 spa4 phyA mutant retained responsiveness to B , as indicated by the inhibition of hypocotyl elongation in B of higher fluence rates , additional phyA-independent mechanisms of SPA2 inactivation by B exist . These are likely mediated by the cryptochromes . The lack of cryptochrome activity in B-induced SPA2 degradation might be caused by a failure of cryptochromes to rapidly interact with SPA2 . Indeed , FRET/FLIM studies in transfected tobacco leaves failed to show an interaction between cry2 and SPA2 . Similarly , recombinantly produced cry2 and SPA2 did not interact in in vitro pull-down assays [15] . On the contrary , SPA2 was shown to weakly interact with cry2 in B in the yeast two-hybrid system [50] . For cry1 , no significant interaction with SPA2 was observed in the yeast two-hybrid assay [46] . To reinvestigate this question in planta , we conducted co-immunoprecipitation experiments using transgenic Arabidopsis seedlings expressing SPA2-HA and , as a positive control , SPA1-HA ( Fig 7 ) . To obtain similar protein levels of SPA1-HA and SPA2-HA in B , SPA1-HA was expressed under the control of the weaker SPA2 promoter ( SPA2::SPA1-HA ) and SPA2-HA from the stronger SPA1 promoter ( SPA1::SPA2-HA ) . Moreover , seedlings were treated with proteasome inhibitor to reduce SPA degradation in B . Fig 7A shows that upon B-exposure both SPA1-HA and SPA2-HA co-immunoprecipitated higher-mobility cry1 isoforms which are formed in B . Hence , B induced the formation of a SPA2/cry1 complex , as it was previously reported for a SPA1/cry1 complex [46 , 47] . In addition , a lower-mobility cry1 which likely represents the non-phosphorylated isoform of cry1 showed weak constitutive interactions with SPA1-HA and SPA2-HA in B and darkness . The association of higher-mobility cry1 with SPA2 was very rapid . It occurred within 5 min of B-exposure ( S2 Fig ) . cry2 , in contrast , was not co-immunoprecipitated by SPA2-HA , neither in darkness nor in B . The positive control SPA1-HA showed the expected B-dependent association with cry2 ( Fig 7B ) . In summary , in B , SPA2 associates with cry1 but not with cry2 in planta . To identify the E3 ubiquitin ligase that mediates SPA2 degradation in the light , we asked whether the COP1/SPA E3 ligase itself may be responsible for ubiquitination of SPA2 . We therefore investigated SPA2 protein levels in the hypomorphic cop1-4 mutant and in the cop1-5 null mutant , using light conditions that cause full degradation of SPA2 . In a cop1-4 hypomorphic background , considerable SPA2 protein levels were retained in seedlings irradiated with FRc ( Fig 8A and 8B ) . Hence , the FRc-induced reduction in SPA2 protein abundance was strongly attenuated , but not abolished , by the partial-loss-of-function cop1-4 mutation . We subsequently analyzed SPA2 protein levels in the cop1-5 null mutant . Because cop1 null mutants arrest growth at the very early seedling stage and , moreover , mostly fail to break the seed coat during germination , we could not obtain enough tissue for nuclear-enriched protein preparations which are necessary to detect the native SPA2 protein with α-SPA2 antibodies . We therefore crossed the SPA2::SPA2-HA transgene into a cop1-5 mutant background and detected the SPA2-HA protein using α-HA antibodies . This transgene-encoded SPA2-HA fully mimics function and behavior of the native SPA2 protein [42] ( this study ) . As shown in Fig 8C and asreported above , SPA2-HA protein levels in the progenitor SPA2::SPA2-HA line decreased to almost undetectable levels upon irradiation with Rc . In a homozygous cop1-5 mutant background , in contrast , SPA2-HA levels were not reduced in Rc when compared to darkness . As an additional control , we also determined SPA2-HA protein levels in COP1 wild-type siblings that segregated in a progeny derived from the cross of cop1-5 with the SPA2::SPA2-HA line . In these siblings , SPA2-HA protein levels decreased upon Rc irradiation as in the progenitor SPA2::SPA2-HA line . Hence , the Rc-induced reduction in SPA2-HA protein abundance was fully dependent on COP1 . Because the cop1 null mutations severely affect seedling growth and cause growth arrest , we wished to exclude the possibility that premature lethality is an indirect reason for the lack of SPA2 degradation in cop1-5 . To do so , we made use of previous findings showing that degradation of phyA-Pfr is in part COP1-independent [16 , 18] and thus should occur in cop1-5 . Indeed , phyA levels strongly decreased upon Rc irradiation ( Fig 8C ) . Hence , the cop1-5 tissue used was clearly still capable of light perception and light response . The analysis of phyA abundance indicated that phyA levels were considerably lower in cop1-5 than in the wild type in both dark-grown and light-exposed tissues . The reasons for this are unknown . It may relate to the low efficiency of protein extraction using cop1-5 when compared to using the wild type . Hence , normalization to HSC70 levels may be unreliable . Consistent with this idea , SPA2-HA levels were also unexpectedly lower in cop1-5 than in the wild type . Since the four SPA proteins heterodimerize in the tetrameric COP1/SPA complex [19] , we asked whether the presence of SPA1 , SPA3 and SPA4 affects SPA2 protein levels in the light . The light-induced reduction in SPA2 protein levels was also dramatic in the spa1 spa3 spa4 mutant , but slightly higher SPA2 protein levels consistently remained in FRc in spa1 spa3 spa4 when compared to the wild type ( Fig 8D ) . Hence , the COP1/SPA2 complex which forms in the spa1 spa3 spa4 mutant is sufficient to allow SPA2 degradation in the light . Whether SPA2 is required for its own degradation cannot be determined from the presented experiments . The finding that the other three SPAs slightly increase SPA2 degradation in FRc hints at the possibility that SPA2 is involved in its own degradation . Our finding that COP1 is required for SPA2 degradation in the light suggests that SPA2 is directly ubiquitinated by the COP1 or COP1/SPA2 ubiquitin ligase . If so , it is expected that interaction of SPA2 with COP1 is necessary for SPA2 degradation to occur . To test this hypothesis , we expressed a SPA2 deletion derivative that lacks the COP1-interacting coiled-coil domain under the control of the native SPA2 promoter ( SPA2::ΔCC SPA2-HA; Fig 9A ) . Indeed , the ΔCC SPA2-HA protein failed to co-immunoprecipitate COP1 in extracts of transgenic plants , confirming that ΔCC SPA2-HA does not incorporate into a COP1/SPA complex ( Fig 9C ) . Consistent with this finding , the ΔCC SPA2-HA transgene did not complement the spa1 spa2 spa3 mutant phenotype , whereas the full-length SPA2-HA transgene did ( S3 Fig ) . ΔCC SPA2-HA protein abundance did not change in response to light ( Fig 9B ) . The levels of full-length SPA2-HA , in contrast , decreased to undetectable levels in FRc . This difference in the behavior of the SPA2-HA and ΔCC SPA2-HA proteins is not due to any differences in SPA2-HA and ΔCC SPA2-HA transcript levels because transcript levels were not regulated by light , as expected for a gene expressed from the SPA2 promoter ( S4 Fig ) . These results show that the COP1-interacting coiled-coil domain of SPA2 is necessary for SPA2 degradation in the light .
The four SPA proteins are components of the COP1/SPA E3 ubiquitin ligase and have redundant but also distinct functions in regulating plant growth and development in response to the light environment . The phenotypic analysis of spa mutants showed that SPA2 , among the four SPA proteins , exhibits the greatest difference in activity between dark- and light-grown seedlings and is therefore a particularly interesting SPA protein when investigating light-induced inactivation of COP1/SPA activity [22 , 42] . Here , we have analyzed the molecular mechanism of SPA2 degradation in different light qualities and have uncovered a photoreceptor-specific mechanism of light-induced COP1/SPA repression via COP1 . Our results demonstrate that the SPA2 protein is degraded very rapidly , i . e . within 5–15 min after dark-grown seedlings were exposed to a brief pulse of R , FR or B . Since COP1 function depends on SPA proteins , this rapid , light-induced degradation of SPA2 provides a very effective mechanism to inactivate COP1/SPA2 activity in light-grown plants . We and others have shown previously that COP1 levels do not significantly change in response to R , FR or B [41 , 42] . Hence , light does not affect the stability of the whole COP1/SPA2 complex but only that of SPA2 . This shows that the presence of SPA2 in the COP1/SPA2 E3 ubiquitin ligase provides a means for light-induced inactivation of the COP1/SPA2 complex . Though both phytochrome and cryptochrome photoreceptors inactivate COP1/SPA function in the respective light qualities [14] , we found that the rapid degradation of SPA2 specifically required phytochromes not only in R and FR , but also in B . Thus , this mechanism of rapid COP1/SPA2 inactivation is specific to phytochrome action . In summary , our analysis shows that a photoreceptor-specific mechanism of COP1/SPA2 inactivation developed during evolution . Evidence indicates that multiple mechanisms have evolved that inactivate COP1/SPA function in the light . Another mechanism of inactivation was found to be common to phytochromes and cry1 since phyA , phyB and cry1 induce a dissociation of COP1 from SPA1 in R or B , respectively [46–49] . A third mechanism , the light-induced exclusion of COP1 from the nucleus also occurs in R , FR and B and is primarily mediated by phyA , phyB and cry1 in FR , R and B , respectively [52] . On the other hand , B-control of COP1 nuclear abundance was found to also require biosynthesis of the phytochrome chromophore [53] , suggesting an essential role of phytochromes also in B . In total , evidence indicates that photoreceptor-specific mechanisms and common mechanisms induced by both phy and cry photoreceptors co-act to allow an appropriate response to a changing light environment . Our results show that rapid SPA2 degradation in R involves a phyA-dependent VLFR and a phyB-dependent LFR which is also reversible by FR . In FR , SPA2 degradation was fully dependent on phyA . This demonstrates that the responsiveness of the SPA2 protein to R and FR directly correlates with our current knowledge on phyA and phyB activities in R and FR [3 , 4] and thus appears to be an immediate output of light-induced phytochrome action . Previous findings showing that SPA2 directly interacts with phyA and phyB [49] are in good agreement with this conclusion . phyA is also a well-known B-photoreceptor that together with cry1 and cry2 is responsible for seedling deetiolation in B [4] . The particular biological significance of phyA in B-induced repression of SPA2 function is supported by our finding that the extreme hypersensitivity to B in spa1 spa3 spa4 triple mutants which only have functional SPA2 was indeed fully dependent on phyA . We therefore suggest that light inactivates COP1/SPA2 function in B primarily through rapid , phyA-induced degradation of SPA2 . Residual SPA2 protein that escapes degradation may be inactivated by additional mechanisms , such as cry1-mediated dissociation from COP1 , as it has been described for SPA1 [46 , 47] , and phyA-mediated dissociation from COP1 [49] . The latter , however , has not been analyzed in B so far . Since the SPA1 protein is also degraded in R , FR and B , albeit with lower efficiency than SPA2 , a SPA1-containing COP1 complex may also be inactivated through phytochrome-mediated degradation of SPA1 , i . e . via the same or a very similar mechanism as the light-induced degradation of SPA2 . Interestingly , the mutant phenotypes of spa single mutant seedlings defective in SPA1 , SPA3 or SPA4 are also fully dependent on phyA , even in R . These single mutants etiolate normally in darkness , but exhibit hypersensitivity in the light in a PHYA wild-type background only [33 , 54 , 55] . The mechanistic reason for this observation has so far remained unknown but could be explained by a phyA-mediated de-stabilization of these SPA proteins in light-grown seedlings . Hence , a stabilization of SPA1 , SPA3 and SPA4 in a phyA mutant background might lead to the complete rescue of the spa single mutant phenotypes . The failure of other B receptors than phyA , such as cryptochromes , to cause rapid degradation of SPA2 in B is not due to a general lack of SPA2-cry interactions in vivo . However , our results demonstrate that SPA2 only associates with cry1 and not with cry2 in B-treated seedlings . Hence , the lack of a cry2-SPA2 interaction is likely in part responsible for the observed stability of SPA2 in B-treated phyA mutant seedlings . On the other hand , our results also show that SPA2 rapidly interacts with cry1 in B without causing rapid SPA2 degradation . Based on this finding we conclude that the failure of cry1 to cause rapid degradation of SPA2 is not due to a lack of a SPA2-cry1 interaction , especially since the SPA2-cry1 interaction is observed rapidly in vivo , i . e . within 5 min of B irradiation . Thus , cry1 interacting with SPA2 in B does not induce rapid degradation of SPA2; cry1 action thereby strongly differs from phytochrome actions on the SPA2 protein . In contrast to SPA2 which only interacted with cry1 in our in vivo co-immunoprecipitation experiments , SPA1 interacted with both cryptochromes , as shown previously [46 , 47 , 50] . Hence , SPA1 and SPA2 clearly differ in their interaction capacity with cry2 . cry2 was shown to interact with the N-terminal domain of SPA1 [50] . Though we do not know the cry2-interacting domain in SPA2 , it is possible that the relatively high sequence divergence between the N-terminal domains of SPA1 and SPA2 might be the cause for their differential interaction capacities with cry2 . cry1 , in contrast , interacts with the WD-repeat domains of SPA1 and SPA2 [47] , and this domain is highly conserved between SPA1 and SPA2 [55] . The mechanism of SPA2 degradation may essentially reflect ubiquitination by the COP1 ( or COP1/SPA2 ) E3 ubiquitin ligase or the action of another E3 ligase . Recently , the COP1-interacting E3 ubiquitin ligase COP1 SUPPRESSOR1 ( CSU1 ) was reported to de-stabilize COP1 and SPA1 in darkness , but not in the light . SPA2 , SPA3 and SPA4 protein levels were not altered in csu1 mutants , neither in dark-grown nor in light-grown seedlings [56] . It is therefore unlikely that CSU1 is involved in the light-dependent degradation of SPA2 . Indeed , light-induced SPA2 degradation was absent in a cop1-5 null mutant . Hence , ubiquitination of SPA2 by COP1 or the COP1/SPA2 ubiquitin ligase is the likely mechanism . This is supported by our finding that a ΔCC SPA2 deletion derivative which does not interact with COP1 in vivo is not degraded in the light . We therefore propose that light influences the E3 ligase activity of COP1/SPA2 in two ways: it inhibits COP1/SPA2 E3 ligase activity towards its substrate transcription factors , while it enhances COP1 ( or COP1/SPA2 ) ( auto ) -ubiquitination activity towards SPA2 and , possibly , SPA1 as well ( Fig 10 ) . However , we cannot fully exclude the possibility that SPA2 is ubiquitinated by an indirect COP1-dependent mechanism . For example , COP1 might be a scaffolding protein required for SPA2 degradation or control the activity of another E3 ubiquitin ligase . Whether SPA3 and SPA4 protein stability is controlled by light remains to be determined . In humans , DNA damage increases COP1 autodegradation by ATM-mediated phosphorylation of COP1 , followed by stabilization of the COP1 substrate p53 as a cell cycle check point [57] . Though the phosphorylated residue in human COP1 is not conserved neither in Arabidopsis COP1 nor in the SPA proteins , this finding shows that autodegradation of components of this E3 ligase is a regulatory mechanism used in both humans and plants .
Wild-type Arabidopsis thaliana accessions Col-0 , RLD and Ler were used in this study . Photoreceptor mutants phyA-211 ( Col-0 ) [58] , phyA-101 ( RLD ) [59] , phyB-1 ( introgressed into RLD ) [60 , 61] , phyA-101 phyB-1 ( RLD ) , phyA-201 ( Ler ) [58] , cry1 cry2 ( Ler ) and cry1 cry2 phyA-201 ( Ler ) [62] were described previously . The transgenic lines with constitutively active photoreceptors expressed the phytochromes AY242H and BY276H [7] , PHYA406-YFP-DD/NLS [10] , CRY1G380R [8] or GUS-CCT1 [9] . The transgenic lines SPA2::SPA1-HA 28 , SPA2::SPA1-HA 70 , SPA1::SPA2-HA 64 and SPA2::SPA2-HA 32 were described previously [42] . The mutants spa1-7 spa2-1 spa3-1 and spa1-7 spa3-1 spa4-1 [51] were used whenever no allele information is provided . spa1-100 spa3-1 spa4-3 [23] , cop1-4 [63] and cop1-5 [64] were described . The spa1-7 spa3-1 spa4-1 phyA-211 quadruple mutant was generated by crossing the spa1-7 spa3-1 spa4-1 triple mutant with the phyA-211 single mutant and was confirmed in the F2 and F3 progenies by the phyA phenotype and a genotypic analysis using molecular markers that can distinguish between mutant and wild-type spa alleles . To obtain SPA2::SPA2-HA cop1-5 ( -/- ) seed , the transgenic line SPA2::SPA2-HA 32 was crossed with cop1-5 ( +/- ) . Transgenic homozygous cop1-5 seeds were selected in a segregating F4 population based on their black seed phenotype which was scored using a stereo microscope . Seeds with normal seed color served as a control that is homo- or heterozygous for the wild-type COP1 allele . LED light sources and seedling growth conditions were as described previously [22 , 54] . Growth conditions for the SPA2::SPA2-HA cop1-5 ( -/- ) experiment were as follows: after stratification of imbibed seeds for 3 days at 4°C , seeds were irradiated with white light for 3 h to break the dormancy and were subsequently kept in darkness for another 21 h . Seeds were then transferred from darkness to Rc ( 40 μmol m–2 s–1 ) for 6 h . SPA2::ΔCC SPA2-HA lines express a deletion derivative lacking the amino acids 580–702 in the SPA2 protein . To generate the construct , two PCR fragments were amplified from the full-length SPA2 ORF lacking the stop codon using the primer pairs SC_SPA2deltaCC_ApaI_F1 and SC_SPA2deltaCC_R1 or SC_SPA2deltaCC_F2 and SPA2deltaN-NotI-R . Both PCR products were purified , combined and subsequently used as templates for amplifying the ΔCC SPA2 sequence using the primers SC_SPA2deltaCC_ApaI_F1 and SPA2 delta N NotI R , thereby also introducing a 5’ ApaI restriction site and a 3’ NotI restriction site . After PCR-amplification of ΔCC SPA2 , the resulting fragment was introduced into the pJET1 . 2 vector ( Thermo Scientific ) . After sequencing of the insert to confirm the correct sequence , the deletion construct was digested with ApaI and NotI and ligated into the ApaI and NotI sites of the pBS vector carrying the SPA2 5’ and 3’ regulatory sequences as described previously [42] , resulting in the SPA2::ΔCC SPA2 construct in pBS . The 3xHA tag with stop codon was subsequently cloned into the NotI site and the complete insert was cloned into the pJHA212 binary vector [65] as described in [42] to generate SPA2::ΔCC SPA2-HA . This construct was transformed into spa1-7 spa2-1 spa3-1 mutant plants by floral dipping . T2 plants were used for analysis . In order to detect the native SPA2 protein using an α-SPA2 antibody [42] , nuclear proteins were enriched from seedlings as described previously [66] . Approximately 200 mg of seedlings or , for cop1-5 related experiments , approximately 20 μl volume-equivalents of imbibed seeds were homogenized to a fine powder using liquid nitrogen . Lysis buffer [50 mM Tris pH 7 . 5 , 150 mM NaCl , 1 mM EDTA , 10% glycerol , 0 . 1% Triton X-100 , 5 mM DTT , 1% protease inhibitor cocktail ( Sigma-Aldrich ) , 10 μM MG132] was added to the ground tissue at a ratio of 150 μl per 100 mg tissue . The mixtures were thawed on ice and centrifuged at 20 . 000 g at 4°C for 12 min . 5x Laemmli buffer was added to the supernatant to a final concentration of 1x before heating at 96°C for 5 min . Protein concentrations were determined by Bradford assay ( Bio-Rad ) . For separating nuclear-enriched protein extracts by SDS-PAGE , equal volumes of nuclear-enriched extracts were loaded . To separate total protein extracts , equal amounts of protein were resolved by SDS-PAGE . Protein samples were subsequently blotted onto PVDF membranes . After blotting , membranes were blocked with Rotiblock ( Roth ) reagent and incubated with the respective primary antibody followed by a horseradish peroxidase ( HRP ) -conjugated secondary antibody . HRP activity was detected using the SuperSignal West Femto Maximum Sensitivity kit ( Thermo Scientific ) and visualized by a LAS-4000 Mini bioimager ( GE Healthcare Life Sciences ) . Signal intensities were quantified using Multi-Gauge software ( GE Healthcare Life Sciences ) . Commercial antibodies used were HRP-conjugated α-HA ( Roche ) , α-Histone H3 ( Abcam ) , α-HSC70 ( Stressgen ) , α-α-Tubulin ( Sigma-Aldrich ) , α-rabbit IgG-HRP ( Sigma-Aldrich ) and α-mouse IgG-HRP ( Sigma-Aldrich ) . α-SPA2 and α-COP1 antibodies were described previously in [42] . α-cry1 [67] and α-cry2 [68] antibodies were used to detect cry1 and cry2 , respectively . Co-immunoprecipitation experiments were performed using μMACS Anti-HA Starting Kits ( Miltenyi Biotec ) according to the manufacturer’s protocol with minor modification . Total proteins were extracted as described above . Protein lysates were incubated with 10 μl μMACS Anti-HA MicroBeads . After incubation on ice for 30 min , the mixture was applied onto prepared μ Columns which were placed in the magnetic field of μMACS Separator attached to a MACS MultiStand . The columns were washed four times with lysis buffer and once with Wash Buffer 2 provided by the kit . Elution was performed at 95°C with Elution Buffer according to the manufacturer’s manual . For cry1 and cry2 pull-down experiments , seedlings were pre-infiltrated with 100 μM MG132 and 10 μM clasto-Lactacystin β-lactone twice , 15 min each , before light treatment . Furthermore , five times more protein extract was used for the SPA2-HA immunoprecipitation than for the SPA1-HA immunoprecipitation . Seedlings were flattened on the surface of solid MS plates and photographed with a Nikon D5000 digital camera . Images were analyzed by ImageJ 1 . 43u ( Wayne Rasband , National Institutes of Health ) to obtain hypocotyl lengths . Total RNA isolation , DNase I treatment , first-strand cDNA synthesis and qRT-PCR were performed as described in [42] . Primers used to amplify HA-tag and UBQ10 were previously described [42] . Two biological replicates were included . Relative transcript levels were calculated using the ΔΔCt method with UBQ10 as a normalization transcript . COP1 ( At2g32950 ) , SPA1 ( At2g46340 ) , SPA2 ( At4g11110 ) , SPA3 ( At3g15354 ) , SPA4 ( At1g53090 ) , cry1 ( AT4G08920 ) , cry2 ( AT1G04400 ) , phyA ( AT1G09570 ) , phyB ( AT2G18790 ) .
|
Plants have evolved photoreceptors that initiate a signaling cascade to adjust growth and development to the ambient light environment . The CUL4-dependent COP1/SPA E3 ubiquitin ligase is a key negative regulator of light signaling whose function is repressed by light . Recent research has identified mechanisms that are common to both phytochrome and cryptochrome photoreceptors . Here , we have identified a mechanism of light-induced COP1/SPA repression that is specific to phytochrome photoreceptors . We show that the SPA2 protein is very rapidly degraded in red , far-red and blue light in a phytochrome-dependent fashion . We further show that SPA2 degradation in the light depends on COP1 and on the interaction of SPA2 with COP1 . Hence , our results suggest a light-induced degradation of SPA2 , but not of COP1 , by the COP1/SPA2 ubiquitin ligase . The human ortholog of COP1 , which functions without the plant-specific SPA proteins , is known to be regulated by autodegradation following DNA damage . Hence , autodegradation of components of this E3 ligase is a regulatory mechanism used in both humans and plants .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[] |
2015
|
Photoreceptor Specificity in the Light-Induced and COP1-Mediated Rapid Degradation of the Repressor of Photomorphogenesis SPA2 in Arabidopsis
|
Adaptation in response to selection on polygenic phenotypes may occur via subtle allele frequencies shifts at many loci . Current population genomic techniques are not well posed to identify such signals . In the past decade , detailed knowledge about the specific loci underlying polygenic traits has begun to emerge from genome-wide association studies ( GWAS ) . Here we combine this knowledge from GWAS with robust population genetic modeling to identify traits that may have been influenced by local adaptation . We exploit the fact that GWAS provide an estimate of the additive effect size of many loci to estimate the mean additive genetic value for a given phenotype across many populations as simple weighted sums of allele frequencies . We use a general model of neutral genetic value drift for an arbitrary number of populations with an arbitrary relatedness structure . Based on this model , we develop methods for detecting unusually strong correlations between genetic values and specific environmental variables , as well as a generalization of comparisons to test for over-dispersion of genetic values among populations . Finally we lay out a framework to identify the individual populations or groups of populations that contribute to the signal of overdispersion . These tests have considerably greater power than their single locus equivalents due to the fact that they look for positive covariance between like effect alleles , and also significantly outperform methods that do not account for population structure . We apply our tests to the Human Genome Diversity Panel ( HGDP ) dataset using GWAS data for height , skin pigmentation , type 2 diabetes , body mass index , and two inflammatory bowel disease datasets . This analysis uncovers a number of putative signals of local adaptation , and we discuss the biological interpretation and caveats of these results .
Population and quantitative genetics were in large part seeded by Fisher's insight [1] that the inheritance and evolution of quantitative characters could be explained by small contributions from many independent Mendelian loci [2] . While still theoretically aligned [3] , these two fields have often been divergent in empirical practice . Evolutionary quantitative geneticists have historically focused either on mapping the genetic basis of relatively simple traits [4] , or in the absence of any such knowledge , on understanding the evolutionary dynamics of phenotypes in response to selection over relatively short time-scales [5] . Population geneticists , on the other hand , have usually focused on understanding the subtle signals left in genetic data by selection over longer time scales [6]–[8] , usually at the expense of a clear relationship between these patterns of genetic diversity and evolution at the phenotypic level . Recent advances in population genetics have also allowed for the genome-wide identification of individual recent selective events either by identifying unusually large allele frequency differences among populations and environments or by detecting the effects of these events on linked diversity [9] . Such approaches are nonetheless limited because they rely on identifying individual loci that look unusual , and thus are only capable of identifying selection on traits where an individual allele has a large and/or sustained effect on fitness . When selection acts on a phenotype that is underwritten by a large number of loci , the response at any given locus is expected to be modest , and the signal instead manifests as a coordinated shift in allele frequency across many loci , with the phenotype increasing alleles all on average shifting in the same direction [10]–[14] . Because this signal is so weak at the level of the individual locus , it may be impossible to identify against the genome-wide background without a very specific annotation of which sites are the target of selection on a given trait [15] , [16] . The advent of well-powered genome wide association studies with large sample sizes [17] has allowed for just this sort of annotation , enabling the mapping of many small effect alleles associated with phenotypic variation down to the scale of linkage disequilibrium in the population . The development and application of these methods in human populations has identified thousands of loci associated with a wide array of traits , largely confirming the polygenic view of phenotypic variation [18] . Although the field of human medical genetics has been the largest and most rapid to puruse such approaches , evolutionary geneticists studying non-human model organisms have also carried out GWAS for a wide array of fitness-associated traits , and the development of further resources is ongoing [19]–[21] . In human populations , the cumulative contribution of these loci to the additive variance so far only explain a fraction of the narrow sense heritability for a given trait ( usually less than 15% ) , a phenomenon known as the missing heritability problem [22] , [23] . Nonetheless , these GWAS hits represent a rich source of information about the loci underlying phenotypic variation . Many investigators have begun to test whether the loci uncovered by these studies tend to be enriched for signals of selection , in the hopes of learning more about how adaptation has shaped phenotypic diversity and disease risk [24]–[27] . The tests applied are generally still predicated on the idea of identifying individual loci that look unusual , such that a positive signal of selection is only observed if some subset of the GWAS loci have experienced strong enough selection to make them individually distinguishable from the genomic background . As noted above , it is unlikely that such a signature will exist , or at least be easy to detect , if adaptation is truly polygenic , and thus many selective events will not be identified by this approach . Here we develop and implement a general method based on simple quantitative and population genetic principals , using allele frequency data at GWAS loci to test for a signal of selection on the phenotypes they underwrite while accounting for the hierarchical structure among populations induced by shared history and genetic drift . Our work is most closely related to the recent work of Turchin et al [28] , Fraser [29] and Corona et al [30] , who look for co-ordinated shifts in allele frequencies of GWAS alleles for particular traits . Our approach constitutes an improvement over the methods implemented in these studies as it provides a high powered and theoretically grounded approach to investigate selection in an arbitrary number of populations with an arbitrary relatedness structure . Using the set of GWAS effect size estimates and genome wide allele frequency data , we estimate the mean genetic value [31] , [32] for the trait of interest in a diverse array of human populations . These genetic values may often be poor predictors of the actual phenotypes for reasons we address below and in the Discussion . We therefore make no strong claims about their ability to predict present day observed phenotypes . We instead focus on population genetic modeling of the joint distribution of genetic values , which provides a robust way of investigating how selection may have impacted the underlying loci . We develop a framework to describe how genetic values covary across populations based on a flexible model of genetic drift and population history . In Figure 1 we show a schematic diagram of our approach to aid the reader . Using this null model , we implement simple test statistics based on transformations of the genetic values that remove this covariance among populations . We judge the significance of the departure from neutrality by comparing to a null distribution of test statistics constructed from well matched sets of control SNPs . Specifically , we test for local adaptation by asking whether the transformed genetic values show excessive correlations with environmental or geographic variables . We also develop and implement a less powerful but more general test , which asks whether the genetic values are over-dispersed among populations compared to our null model of drift . We show that this overdispersion test , which is closely related to [33] , [34] and a series of approaches from the population genetics literature [35]–[39] , gains considerable power to detect selection over single locus tests by looking for unexpected covariance among loci in the deviation they take from neutral expectations . Lastly , we develop an extension of our model that allows us to identify individual populations or groups of populations whose genetic values deviate from their neutral expectations given the values observed for related populations , and thus have likely been impacted by selection . While we develop these methods in the context of GWAS data , we also relate them to recent methodological developments in the quantitative genetics of measured phenotypes ( as opposed to allele frequencies ) [40] , [41] , highlighting the useful connection between these approaches . An implementation of the methods described here in the form of a collection of R scripts is available at https://github . com/jjberg2/PolygenicAdaptationCode .
Consider a trait of interest where loci ( e . g . biallelic SNPs ) have been identified from a genome-wide association study . We arbitrarily label the phenotype increasing allele and the alternate allele at each locus . These loci have additive effect size estimates , where is the average increase in an individual's phenotype from replacing an allele with an allele at locus . We have allele frequency data for populations at our SNPs , and denote by the observed sample frequency of allele at the locus in the population . From these , we estimate the mean genetic value in the population as ( 1 ) and we take to be the vector containing the mean genetic values for all populations . We are chiefly interested in developing a framework for testing the hypothesis that the joint distribution of is driven by neutral processes alone , with rejection of this hypothesis implying a role for selection . We first describe a general model for the expected joint distribution of estimated genetic values ( ) across populations under neutrality , accounting for genetic drift and shared population history . A simple approximation to a model of genetic drift is that the current frequency of an allele in a population is normally distributed around some ancestral frequency ( ) . Under a Wright-Fisher model of genetic drift , the variance of this distribution is approximately , where is a property of the population shared by all loci , reflecting the compounded effect of many generations of binomially sampling [42] . Note also that for small values , is approximately equal to the inbreeding coefficient of the present day population relative to the defined ancestral population , and thus has an interpretation as the correlation between two randomly chosen alleles relative to the ancestral population [42] . We can expand this framework to describe the joint distribution of allele frequencies across an arbitrary number of populations for an arbitrary demographic history by assuming that the vector of allele frequencies in populations follows a multivariate normal distribution ( 2 ) where is an by positive definite matrix describing the correlation structure of allele frequencies across populations relative to the mean/ancestral frequency . Note again that for small values it is also approximately the matrix of inbreeding coefficients ( on the diagonal ) and kinship coefficients ( on the off-diagonals ) describing relatedness among populations [38] , [43] . This flexible model was introduced , to our knowledge , by [44] ( see [45] for a review ) , and has subsequently been used as a computationally tractable model for population history inference [42] , [46] , and as a null model for signals of selection [38] , [39] , [47] , [48] . So long as the multivariate normal assumption of drift holds reasonably well , this framework can summarize arbitrary population histories , including tree-like structures with substantial gene flow between populations [46] , or even those which lack any coherent tree-like component , such as isolation by distance models [49] , [50] . Recall that our estimated genetic values are merely a sum of sample allele frequencies weighted by effect size . If the underlying allele frequencies are well explained by the multivariate normal model described above , then the distribution of is a weighted sum of multivariate normals , such that this distribution is itself multivariate normal ( 3 ) where and are respectively the expected genetic value and additive genetic variance of the ancestral ( global ) population . The covariance matrix describing the distribution of therefore differs from that describing the distribution of frequencies at individual loci only by a scaling factor that can be interpreted as two times the contribution of the associated loci to the additive genetic variance present in a hypothetical population with allele frequencies equal to the grand mean of the sampled populations . The assumption that the drift of allele frequencies around their shared mean is normally distributed ( 2 ) may be problematic if there is substantial drift . However , even if that is the case , the estimated genetic values may still be assumed to follow a multivariate normal distribution by appealing to the central limit theorem , as each estimated genetic value is a sum over many loci . We show in the Results that this assumption often holds in practice . It is useful here to note that the relationship between the model for drift at the individual locus level , and at the genetic value level , gives an insight into where most of the information and statistical power for our methods will come from . Each locus adds a contribution to the vector of deviations of the genetic values from the global mean . If the allele frequencies are unaffected by selection then the frequency deviation of an allele at locus in population will be uncorrelated in magnitude or sign with both the effect at locus and the allele frequency deviation taken by other unlinked loci . Thus the expected departure of the genetic value of a population from the mean is zero , and the noise around this should be well described by our multivariate normal model . The tests described below will give positive results when these observations are violated . The effect of selection is to induce a non-independence of the allele frequency deviation ( ) across loci , determined by the sign and magnitude of the effect sizes [10]–[14] and as we demonstrate below , all of our methods rely principally on identifying this non-independence . This observations has important considerations for the false positive profile of our methods . Specifically , false positives will arise only if the GWAS ascertainment procedure induces a correlation between the estimated effect size of an allele ( ) and the deviation that this allele takes across populations . This should not be the case if the GWAS is performed in a single population which is well mixed compared to the populations considered in the test . False positives can occur when a GWAS is performed in a structured population and fails to account for the fact that the phenotype of interest is correlated with ancestry in this population . We address this case in greater depth in the Discussion . These observation also allows us to exclude certain sources of statistical error as a cause of false positives . For example , simple error in the estimation of , or failing to include all loci affecting a trait cannot cause false positives , because this error has no systematic effect on across loci . Similarly , if the trait of interest truly is neutral , variation in the true effects of an allele across populations or over time or space ( which can arise from epistatic interactions among loci , or from gene by environment interactions ) will not drive false positives , again because no systematic trends in population deviations will arise . This sort of heterogeneity can , however , reduce statistical power , as well as make straightforward interpretation of positive results difficult , points which we address further below . As described above , we obtain the vector by summing allele frequencies across loci while weighting by effect size . We do not get to observe the ancestral genetic value of the sample , so we assume that this is simply equal to the mean genetic value across populations . This assumption costs us a degree of freedom , and so we must work with a vector , which is the vector of estimated genetic values for the first populations , centered at the mean of the ( see Methods for details ) . Note that this procedure will be the norm for the rest of this paper , and thus we will always work with vectors of length that are obtained by subtracting the mean of the vector and dropping the last component . The information about the dropped population is retained in the mean of the length vectors , and thus the choice of which population to drop is arbitrary and does not affect the inference . To estimate the null covariance structure of the populations we sample a large number K random unlinked SNPs . In our procedure , the SNPs are sampled so as to match certain properties of the GWAS SNPs ( the specific matching procedure is described in more depth below and in the Methods section ) . Setting to be the mean sample allele frequency across populations at the SNP , we standardize the sample allele frequency in population as . We then calculate the sample covariance matrix ( ) of these standardized frequencies , accounting for the rank of the matrix ( see Methods ) . We estimate the scaling factor of this matrix as ( 4 ) We now have an estimated genetic value for each population , and a simple null model describing their expected covariance due to shared population history . Under this multivariate normal framework , we can transform the vector of mean centered genetic values ( ) so as to remove this covariance . First , we note that the Cholesky decomposition of the matrix is ( 5 ) where is a lower triangular matrix , and is its transpose . Informally , this can be thought of as taking the square root of , and so can loosely be thought of as analogous to the standard deviation matrix . Using this matrix we can transform our estimated genetic values as: ( 6 ) If then , where is the identity matrix . Therefore , under the assumptions of our model , these standardized genetic values should be independent and identically distributed random variates [39] . It is worth spending a moment to consider what this transformation has done to the allele frequencies at the loci underlying the estimated genetic values . As our original genetic values are written as a weighted sum of allele frequencies , our transformed genetic values can be written as a weighted sum of transformed allele frequencies ( which have passed through the same transform ) . We can write ( 7 ) and so we can define the vector of transformed allele frequencies at locus to be ( 8 ) This set of transformed frequencies exist within a set of transformed populations , which by definition have zero covariance with one another under the null , and are related by a star-like population tree with branches of equal length . As such , we can proceed with simple , straightforward and familiar statistical approaches to test for the impact of spatially varying selection on the estimated genetic values . Below we describe three simple methods for identifying the signature of polygenic adaptation , which arise naturally from this observation . We first test if the genetic values are unusually correlated with an environmental variable across populations compared to our null model . A significant correlation is consistent with the hypothesis that the populations are locally adapted , via the phenotype , to local conditions that are correlated with the environmental variable . However , the link from correlation to causation must be supported by alternate forms of evidence , and in the lack of such evidence , a positive result from our environmental correlation tests may be consistent with many explanations . Assume we have a vector , containing measurements of a specific environmental variable of interest in each of the populations . We mean-center this vector and put it through a transform identical to that which we applied to the estimated genetic values in ( 7 ) . This gives us a vector , which is in the same frame of reference as the transformed genetic values . There are many possible models to describe the relationship between a trait of interest and a particular environmental variable that may act as a selective agent . We first consider a simple linear model , where we model the distribution of transformed genetic values ( ) as a linear effect of the transformed environmental variables ( ) ( 9 ) where under our null is a set of normal , independent and identically distributed random variates ( i . e . residuals ) , and can simply be estimated as . We can also calculate the associated squared Pearson correlation coefficient ( ) as a measure of the fraction of variance explained by our variable of choice , as well as the non-parametric Spearman's rank correlation , which is robust to outliers that can mislead the linear model . We note that we could equivalently pose this linear model as a mixed effects model , with a random effect covariance matrix . However , as we know both and , we would not have to estimate any of the random effect parameters , reducing it to a fixed effect model as in ( 9 ) [51] . In the Methods ( section “The Linear Model at the Individual Locus Level” ) we show that the linear environmental model applied to our transformed genetic values has a natural interpretation in terms of the underlying individual loci . Therefore , exploring the environmental correlations of estimated genetic values nicely summarizes information in a sensible way at the underlying loci identified by the GWAS . In order to assess the significance of these measures , we implement an empirical null hypothesis testing framework , using , , and as test statistics . We sample many sets of SNPs randomly from the genome , again applying a matching procedure discussed below and in the Methods . With each set of SNPs we construct a vector , which represents a single draw from the genome-wide null distribution for a trait with the given ascertainment profile . We then perform an identical set of transformations and analyses on each , thus obtaining an empirical genome-wide null distribution for all test statistics . As an alternative to testing the hypothesis of an effect by a specific environmental variable , one might simply test whether the estimated genetic values exhibit more variance among populations than expected due to drift . Here we develop a simple test of this hypothesis . As is composed of independent , identically distributed standard normal random variables , a natural choice of test statistic is given by ( 10 ) This statistic represents a standardized measure of the among population variance in estimated genetic values that is not explained by drift and shared history . It is also worth noting that by comparing the rightmost form in ( 10 ) to the multivariate normal likelihood function , we find that is proportional to the negative log likelihood of the estimated genetic values under the neutral null model , and is thus the natural measurement of the model's ability to explain their distribution . Multivariate normal theory predicts that this statistic should follow a distribution with degrees of freedom under the null hypothesis . Nonetheless , we use a similar approach to that described for the linear model , generating the empirical null distribution by resampling SNPs genome-wide . As discussed below , we find that in practice the empirical null distribution tends to be very closely matched by the theoretically predicted distribution . Values of this statistic that are in the upper tail correspond to an excess of variance among populations . This excess of variance is consistent with the differential action of natural selection on the phenotype among populations ( e . g . due to local adaptation ) . Values in the lower tail correspond a paucity of variance , and thus potentially to widespread stabilizing selection , with many populations selected for the same optimum . In this paper we report upper tail p-values from the empirical null distribution of both for our power simulations and empirical results . A two tailed test would be appropriate in cases where stabilizing selection is also of interest , however such signals are likely to be difficult to spot with GWAS data because the we are missing the large effect , low frequency alleles most likely to reveal a signal of stabilizing selection . Having detected a putative signal of selection for a given trait , one may wish to identify individual regions and populations which contribute to the signal . Here we rely on our multivariate normal model of relatedness among populations , along with well understood methods for generating conditional multivariate normal distributions , in order to investigate specific hypotheses about individual populations or groups of populations . Using standard results from multivariate normal theory , we can generate the expected joint conditional distribution of genetic values for an arbitrary set of populations given the observed genetic values in some other set of populations . These conditional distributions allow for a convenient way to ask whether the estimated genetic values observed in certain populations or groups of populations differ significantly from the values we would expect them to take under the neutral model given the values observed in related populations . Specifically , we exclude a population or set of populations , and then calculate the expected mean and variance of genetic values in these excluded populations given the values observed in the remaining populations , and the covariance matrix relating them . Using this conditional mean and variance , we calculate a Z-score to describe how well fit the estimated genetic values of the excluded populations are by our model of drift , conditional on the values in the remaining populations . In simple terms , the observation of an extreme Z-score for a particular population or group of populations may be seen as evidence that that group has experienced directional selection on the trait of interest ( or a correlated one ) that was not experienced by the related populations on which we condition the analyses . The approach cannot uniquely determine the target of selection , however . For example , conditioning on populations that have themselves been influenced by directional selection may lead to large Z-scores for the population being tested , even if that population has been evolving neutrally . We refer the reader to the Methods section for a mathematical explication of these approaches . We conducted power simulations and an empirical application of our methods based on the Human Genome Diversity Panel ( HGDP ) population genomic dataset [58] , and a number of GWAS SNP sets . To ensure that we made the fullest possible use of the information in the HGDP data , we took advantage of a genome wide allele frequency dataset of 3 million SNPs imputed from the Phase II HapMap into the 52 populations of the HGDP . These SNPs were imputed as part of the HGDP phasing procedure in [59]; see our Methods section for a recap of the details . We applied our method to test for signals of selection in six human GWAS datasets identifying SNPs associated with height , skin pigmentation , body mass index ( BMI ) , type 2 diabetes ( T2D ) , Crohn's Disease ( CD ) and Ulcerative Colitis ( UC ) . To assess the power of our methods in comparison to other possible approaches , we conducted a series of power simulations . There are two possible approaches to simulate the effect of selection on large scale allele frequency data of the type for which our methods are designed . The first is to simulate under some approximate model of the evolutionary history ( e . g . full forward simulation under the Wright-Fisher model with selection ) . The second is to perturb real data in such a way that approximates the effect of selection . We choose to pursue the latter , both because it is more computationally tractable , and because it allows us to compare the power of our different approaches for populations with evolutionary histories of the same complexity as the real data we analyze . Each of our simulations will thus consist of sampling 1000 sets of SNPs matched to the height dataset ( in much the same way we sample SNPs to construct the null distributions of our test statistics ) , and then adding slight shifts in frequency in various ways to mimic the effect of selection . Below we first describe the set of alternative statistics to which we compare our methods . We then describe the manner in which we add perturbations to mimic selection , and lastly describe a number of variations on this theme which we pursued in order to better demonstrate how the power of our statistics changes as we vary parameters of the trait of interest , evolutionary process , or the ascertainment . We estimated genetic values for each of six traits from the subset of GWAS SNPs that were present in the HGDP dataset , as described above . We discuss the analysis of each dataset in detail below , and address general points first . For each dataset , we constructed the covariance matrix from a sample of approximately appropriately matched SNPs , and the null distributions of our test statistics from a sample of sets of null genetic values , which were also constructed according to a similar matching procedure ( as described in the Methods ) . In an effort to be descriptive and unbiased in our decisions about which environmental variables to test , we tested each trait for an effect of the major climate variables considered by Hancock et al ( 2008 ) [63] in their analysis of adaptation to climate at the level of individual SNPs . We followed their general procedure by running principal components ( PC ) analysis for both seasons on a matrix containing six major climate variables , as well as latitude and longitude ( following Hancock et al's rationale that these two geographic variables may capture certain elements of the long term climatic environment experienced by human populations ) . The percent of the variance explained by these PCs and their weighting ( eigenvectors ) of the different environmental variables are given in Table 1 . We view these analyses largely as a descriptive data exploration enterprise across a relatively small number of phenotypes and distinct environmental variables , and do not impose a multiple testing penalty against our significance measures . A multiple testing penalization or false discovery rate approach may be needed when testing a large number phenotypes and/or environmental variables . We also applied our test to identify traits whose underlying loci showed consistent patterns of unusual differentiation across populations , with results reported in Table 2 . In Figure 3 we show for each GWAS set the observed value of and its empirical null distribution calculated using SNPs matched to the GWAS loci as described above . We also plot the expected null distribution of the statistic ( ) . The expected null distribution closely matches the empirical distribution in all cases , suggesting that our multivariate normal framework provides a good null model for the data ( although we will use the empirical null distribution to obtain measures of statistical significance ) . For each GWAS SNP set we also separate our statistic into its -like and LD-like terms , as described in ( 14 ) . In Figure 4 we plot the null distributions of these two components for the height dataset as histograms , with the observed value marked by red arrows ( Figures S6–S10 give these plots for the other five traits we examined ) . In accordance with the expectation from our power simulations , the signal of selection on height is driven entirely by covariance among loci in their deviations from neutrality , and not by the deviations themselves being unusually large . Lastly , we pursue a number or regionally restricted analyses . For each trait and for each of the seven geographic/genetic clusters described by Rosenberg et al ( 2002 ) [62] , we compute a region specific statistic to get a sense for the extent to which global signals we detect can be explained by variation among populations with these regions , and to highlight particular populations and traits which may merit further examination as more association data becomes available . The results are reported in Table 3 . We also apply our conditional Z-score approach at two levels of population structure: first at the level of Rosenberg's geographic/genetic clusters , testing each cluster in turn for how differentiated it is from the rest of the world , and second at the level of individual populations . The regional level Z-scores are useful for identifying signals of selection acting over broad regional scale or on deeper evolutionary timescales , while the population specific Z-scores are useful for identifying very recent selection that has only impacted a single population . We generally employ these regional statistics as a heuristic tool to localize signatures of selection uncovered in global analyses , or in cases where there is no globally interesting signal , to highlight populations or regions which may merit further examination as more association data becomes available . The result of these analyses are depicted in Figure 5 , as well as Tables S3–S14 .
Among the most significant potential pitfalls of our analysis ( and the most likely cause of a false positive ) is the fact that the loci used to test for the effect of selection on a given phenotype have been obtained through a GWAS ascertainment procedure , which can introduce false signals of selection if potential confounds are not properly controlled . We condition on simple features of the ascertainment process via our allele matching procedure , but deeper issues may arise from artifactual associations that result from the effects of population structure in the GWAS ascertainment panel . Given the importance of addressing this issue to the broader GWAS community , a range of well developed methods exist for doing GWAS in structured populations , and we refer the reader to the existing literature for a full discussion [80]–[86] . Here , we focus on two related issues . First , the propensity of population structure in the GWAS ascertainment panel to generate false positives in our selection analysis , and second , the difficulties introduced by the sophisticated statistical approaches employed to deal with this issue when GWAS are done in strongly structured populations . The problem of population structure arises generally when there is a correlation in the ascertainment panel between phenotype and ancestry such that SNPs that are ancestry informative will appear to be associated with the trait , even when no causal relationship exists [81] . This phenomenon can occur regardless of whether the correlation between ancestry and phenotype is caused by genetic or environmental effects . To make matters worse , multiple false positive associations will tend to line up with same axis of population structure . If the populations being tested with our methods lie at least partially along the same axis of structure present in the GWAS ascertainment panel , then the ascertainment process will serve to generate the very signal of positive covariance among like effect alleles that our methods rely on to detect the signal of selection . The primary takeaway from this observation is that the more diverse the array of individuals sampled for a given GWAS are with respect to ancestry , the greater the possibility that failing to control for population structure will generate false associations ( or bias effect sizes ) and hence false positives for our method . What bearing do these complications have on our empirical results ? The GWAS datasets we used can be divided into those conducted within populations of European descent and the skin pigmentation dataset ( which used an admixed population ) . We will first discuss our analysis of the former . The European GWAS loci we used were found in relatively homogeneous populations , in studies with rigorous standards for replication and control for population structure . Therefore , we are reasonably confident that these loci are true positives . Couple this with the fact that they were ascertained in populations that are fairly homogenous relative to the global scale of our analyses , and it is unlikely that population structure in the ascertainment panels is driving our positive signals . One might worry that we could still generate false signals by including European populations in our analysis , however many of the signals we see are driven by patterns outside of Europe ( where the influence of structure within Europe should be much lessened ) . For height , where we do see a strong signal from within Europe , we use a set of loci that have been independently verified using a family based design that is immune to the effects of population structure [28] . We further note that for a number of GWAS datasets , including some of those analyzed here , studies of non-European populations have replicated many of the loci identified in European populations [87]–[93] , and for many diseases , the failure of some SNPs to replicate , as well as discrepancies in effect size estimate , are likely due to simple considerations of statistical power and differences in patterns of LD across populations [94] , [95] . This suggests that , at least for GWAS done in relatively homogenous human populations , structure is unlikely to be a major confounding factor . The issue of population structure may be more profound for our style of approach when GWAS are conducted using individuals from more strongly structured populations . In some cases it is desirable to conduct GWAS in such populations as locally adaptive alleles will be present at intermediate frequencies in these broader samples , whereas they may be nearly fixed in more homogeneous samples . A range of methods have been developed to adjust for population structure in these setting [96]–[98] . While generally effective in their goal , these methods present their own issues for our selection analysis . Consider the extreme case , such as that of Atwell et al ( 2010 ) [19] , who carried out a GWAS in Arabidopsis thaliana for 107 phenotypes across an array of 183 inbred lines of diverse geographical and ecological origin . Atwell and colleagues used the genome-wide mixed model program EMMA [83] , [96] , [97] to control for the complex structure present in their ascertainment panel . This practice helps ensure that many of the identified associations are likely to be real , but also means that the loci found are likely to have unusual frequencies patterns across the species range . This follows from the fact that the loci identified as associated with the trait must stand out as being correlated with the trait in a way not predicted by the individual kinship matrix ( as used by EMMA and other mixed model approaches ) . Our approach is predicated on the fact that we can use genome-wide patterns of kinship to adjust for population structure , but this correction is exactly the null model that loci significantly associated with phenotypes by mixed models have overcome . For this reason , both the theoretical distribution of the statistic , as well as the empirical null distributions we construct from resampling , may be inappropriate . The Cape Verde skin pigmentation data we used may qualify as this second type of study . The Cape Verde population is an admixed population of African/European descent , and has substantial inter-individual variation in admixture proportion . Due to its admixed nature , the population segregates alleles which would not be at intermediate frequency in either parental population , making it an ideal mapping population . Despite the considerable population structure , the fact that recombination continues to mix genotypes in this population means that much of the LD due to the African/European population structure has been broken up ( and the remaining LD is well predicted by an individual's genome-wide admixture coefficient ) . Population structure seems to have been well controlled for in this study , and a number of the loci have been replicated in independent admixed populations . While we think it unlikely that the four loci we use are false associations , they could in principle suffer from the structured ascertainment issues described above , so it is unclear that the null distributions we use are strictly appropriate . That said , provided that Beleza and colleagues have appropriately controlled for population structure , under neutrality there would be no reason to expect that the correlation among the loci should be strongly positive with respect to the sign of their effect on the phenotype , and thus the pattern observed is at least consistent with a history of selection , especially in light of the multiple alternative lines of evidence for adaptation on the basis of skin pigmentation [68]–[70] , [99]–[101] . Further work is needed to determine how best to modify the tests proposed herein to deal with GWAS performed in structured populations . Our understanding of the genetic basis of variation in complex traits remains very incomplete , and as such the results of these analyses must be interpreted with caution . That said , because our methods are based simply on the rejection of a robust , neutral null model , an incomplete knowledge of the genetic basis of a given trait should only lead to a loss of statistical power , and not to a high false positive rate . For all traits analyzed here except for skin pigmentation , the within population variance for genetic value is considerably larger than the variance between populations . This suggests that much of what we find is relatively subtle adaptation even on the level of the phenotype , and emphasizes the fact that for most genetic and phenotypic variation in humans , the majority of the variance is within populations rather than between populations ( see Figures S14–S19 ) . In many cases , the influence of the environment likely plays a stronger role in the differences between populations for true phenotypes than the subtle differences we find here ( as demonstrated by the rapid change in T2D incidence with changing diet , e . g . [102] ) . That said , an understanding of how adaptation has shaped the genetic basis of a wide variety of phenotypes is clearly of interest , even if environmental differences dominate as the cause of present day population differences , as it informs our understanding of the biology and evolutionary history of these traits . The larger conceptual issues relate to the interpretation of our positive findings , which we detail below . A number of these issues are inherent to the conceptual interpretation of evidence for local adaptation [103] .
Written in matrix notation , the procedure of mean centering the estimated genetic values and dropping one population from the analysis can be expressed as ( 16 ) where is an by matrix with on the main diagonal , and elsewhere . In order to calculate the corresponding expected neutral covariance structure about this mean , we use the following procedure . Let be an by matrix , where each column is a vector of allele frequencies across the populations at a particular SNP , randomly sampled from the genome according to the matching procedure described below . Let and be the mean allele frequency in columns and of respectively , and let be a matrix such that . With these data , we can estimate as ( 17 ) This transformation performs the operation of centering the matrix at the mean value , and rooting the analysis with one population by dropping it from the covariance matrix ( the same one we dropped from the vector of estimated genetic values ) , resulting in a covariance matrix describing the relationship of the remaining populations . This procedure thus escapes the singularity introduced by centering the matrix at the observed mean of the sample . As we do not get to observe the population allele frequencies , the entries of are the sample frequencies at the randomly chosen loci , and thus the covariance matrix also includes the effect of finite sample size . Because the noise introduced by the sampling of individuals is uncorrelated across populations ( in contrast to that introduced by drift and shared history ) , the primary effect is to inflate the diagonal entries of the matrix by a factor of , where is the number of chromosomes sampled in population ( see the supplementary material of [46] for discussion ) . This means that our population structure adjusted statistics also approximately control for differences in sample size . As described in the Results , we can use our multivariate normal model of relatedness to obtain the expected distribution of genetic values for an arbitrary set of populations , conditional on the observed values in some other arbitrary set . We first partition our populations into two groups , those for which we want to obtain the expected distribution of genetic values ( group 1 ) , and those on which we condition in order to obtain this distribution ( group 2 ) . We then re–estimate the covariance matrix such that it is centered on the mean of group 2 . This step is necessary because the amount of divergence between the populations in group 1 and the mean of group 2 will always be greater than the amount of divergence from the global mean , even under the neutral model , and our covariance matrix needs to reflect this fact in order to make accurate predictions . We can obtain this re-parameterized matrix as follows . If is the total number of populations in the sample , then let be the number of populations in group one , and let be the number of populations in group 2 . We then define a new matrix such that the columns corresponding the populations in group one have 1 on the diagonal , and 0 elsewhere , while the columns corresponding to group two have on the diagonal , and elsewhere . We can then re–estimate a covariance matrix that is centered at the mean of the populations in group 2 . Recalling our matrices and from ( 17 ) , this matrix is calculated as ( 19 ) where we write to indicate that it is a covariance matrix that has been re-centered on the mean of group two . Once we have calculated this re–centered covariance matrix , we can use well known results from multivariate normal theory to obtain the expected joint distribution of the genetic values for group one , conditional on the values observed in group two . We partition our vector of genetic values and the re–centered covariance matrix such that ( 20 ) and ( 21 ) where and are vectors of genetic values in group 1 and 2 respectively , and , and are the marginal covariance matrices of populations within group 1 , within group 2 , and across the two groups , respectively . Letting ( i . e . the sum of the elements of ) , we wish to obtain the distribution ( 22 ) where and give the expected means and covariance structure of the populations in group 1 , conditional on the values observed in group 2 . These can be calculated as ( 23 ) and ( 24 ) where the one vectors in line ( 23 ) are of length and respectively . This distribution is itself multivariate normal , and as such this framework is extremely flexible , as it allows us to obtain the expected joint distribution for arbitrary sets of populations ( e . g . geographic regions or continents ) , or for each individual population . Further , ( 25 ) and ( 26 ) where denotes the elements of . In words , the conditional expectation of the mean estimated genetic value across group 1 is equal to the mean of the conditional expectations , and its variance is equal to the mean value of the elements of the conditional covariance matrix . As such we can easily calculate a Z score ( and corresponding p value ) for group one as a whole as ( 27 ) This Z score is a normal random variable with mean zero , variance one under the null hypothesis , and thus measures the divergence of the genetic values between the two populations relative to the null expectation under drift . Note that the observation of a significant Z score in a given population or region cannot necessarily be taken as evidence that selection has acted in that population or region , as selection in some of the populations on which we condition ( especially the closely related ones ) could be responsible for such a signal . As such , caution is warranted when interpreting the output of these sort of analyses , and is best done in the context of more explicit information about the demographic history , geography , and ecology of the populations . As with our excess variance test , explored in the main text , it is natural to ask how our environmental correlation tests can be written in terms of allele frequencies at individual loci . As noted in ( 8 ) , we can obtain for each underlying locus a set of transformed allele frequencies , which have passed through the same transformation as the estimated genetic values . We assume that each locus has a regression coefficient ( 28 ) where is shared across all loci so that ( 29 ) where the are independent and identically distributed residuals . We can find the maximum likelihood estimate by treating as the linear predictor , and taking the regression of the combined vector , across all populations and loci , on the combined vector . As such ( 30 ) we can decompose this into a sum across loci such that ( 31 ) As noted in ( 8 ) , our transformed genetic values can be written as ( 32 ) and so the estimated slope ( ) of our regression ( ) is ( 33 ) Comparing these equations , the mean regression coefficient at the individual loci ( 31 ) and the regression coefficient of the estimated genetic values ( 33 ) are proportional to each other via a constant that is given by one over two times the sum of the effect sizes squared ( i . e . ) . Our test based on estimating the regression of genetic values on the environmental variable is thus mathematically equivalent to an approach in which we assume that the regression coefficients of individual loci on the environmental variable are proportional to one another via a constant that is a function of the effect sizes . Such a relationship can also be demonstrated for the correlation coefficient ( ) calculated at the genetic value level and at the individual locus level ( this is not necessarily true for the rank correlation ) , however the algebra is more complicated , and thus we do not show it here . This is in contrast to the enrichment statistic we compute for the power simulations , in which we assume that the correlations of individual loci with the environmental variable are independent of one another , and then perform a test for whether more loci individually show strong correlations with the environmental variable than we would expect by chance . We used imputed allele frequency data in the HGDP , where the imputation was performed as part of the phasing procedure of [59] , as per the recommendations of [124] . We briefly recap their procedure here: Phasing and imputation were done using fastPHASE [125] , with the settings that allow variation in the switch rate between subpopulations . The populations were grouped into subpopulations corresponding to the clusters identified in [62] . Haplotypes from the HapMap YRI and CEU populations were included as known , as they were phased in trios and are highly accurate . HapMap JPT and CHB genotypes were also included to help with the phasing . Various components of our procedure involve sampling random sets of SNPs from across the genome . While we control for biases in our test statistics introduced by population structure through our matrix , we are also concerned that subtle ascertainment effects of the GWAS process could lead to biased test statistics , even under neutral conditions . We control for this possibility by sampling null SNPs so as to match the joint distribution of certain properties of the ascertained GWAS SNPs . Specifically , we were concerned that the minor allele frequency ( MAF ) in the ascertainment population , the imputation status of the allele in the HGDP datasets , and the background selection environment experienced at a given locus , as measured by B value [61] , might influence the distribution of allele frequencies across populations in ways that we could not predict . We partitioned SNPs into a three way contingency table , with 25 bins for MAF ( i . e . a bin size of 0 . 02 ) , 2 bins for imputation ( either imputed or not ) , and 10 bins for B value ( B values range from 0 to 1 , and thus our bin size was 0 . 1 ) . For each set of null genetic values , we sampled one null SNP from the same cell in the contingency table as each of the GWAS SNPs , and assigned this null SNP the effect size associated with the GWAS SNP it was sampled to match . While we do not assign effect sizes to sampled SNPs used to estimate the covariance matrix ( instead simply scaling by a weighted sum of squared effect sizes , which is mathematically equivalent under our assumption that all SNPs have the same covariance matrix ) , we follow the same sampling procedure to ensure that describes the expected covariance structure of the GWAS SNPs . For the skin pigmentation GWAS [67] we do not have a good proxy present in the HGDP population , as the Cape Verdeans are an admixed population . Cape Verdeans are admixed with African ancestry , and European ancestry in the sample obtained by [67] ( Beleza , pers . comm . , April 8 , 2013 ) . As such , we estimated genome wide allele frequencies in Cape Verde by taking a weighted mean of the frequencies in the French and Yoruban populations of the HGDP , such that . We then used these estimated frequencies to assign SNPs to frequency bins . [67] also used an admixture mapping strategy to map the genetic basis of skin pigmentation . However , if they had only mapped these loci in an admixture mapping setting we would have to condition our null model on having strong enough allele frequency differentiation between Africans and Europeans at the functional loci for admixture mapping to have power [126] . The fact that [67] mapped these loci in a GWAS framework allows us to simply reproduce the strategy , and we ignore the results of the admixture mapping study ( although we note that the loci and effect sizes estimated were similar ) . This highlights the need for a reasonably well defined ascertainment population for our approach , a point which we comment further on in the Discussion .
|
The process of adaptation is of fundamental importance in evolutionary biology . Within the last few decades , genotyping technologies and new statistical methods have given evolutionary biologists the ability to identify individual regions of the genome that are likely to have been important in this process . When adaptation occurs in traits that are underwritten by many genes , however , the genetic signals left behind are more diffuse , and no individual region of the genome is likely to show strong signatures of selection . Identifying this signature therefore requires a detailed annotation of sites associated with a particular phenotype . Here we develop and implement a suite of statistical methods to integrate this sort of annotation from genome wide association studies with allele frequency data from many populations , providing a powerful way to identify the signal of adaptation in polygenic traits . We apply our methods to test for the impact of selection on human height , skin pigmentation , body mass index , type 2 diabetes risk , and inflammatory bowel disease risk . We find relatively strong signals for height and skin pigmentation , moderate signals for inflammatory bowel disease , and comparatively little evidence for body mass index and type 2 diabetes risk .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"genome-wide",
"association",
"studies",
"genetic",
"polymorphism",
"natural",
"selection",
"genome",
"analysis",
"genetics",
"biology",
"and",
"life",
"sciences",
"evolutionary",
"adaptation",
"population",
"genetics",
"computational",
"biology",
"evolutionary",
"biology",
"evolutionary",
"processes",
"genetic",
"drift",
"human",
"genetics"
] |
2014
|
A Population Genetic Signal of Polygenic Adaptation
|
Mosquitoes host communities of microbes in their digestive tract that consist primarily of bacteria . We previously reported that Aedes aegypti larvae colonized by a native community of bacteria and gnotobiotic larvae colonized by only Escherichia coli develop very similarly into adults , whereas axenic larvae never molt and die as first instars . In this study , we extended these findings by first comparing the growth and abundance of bacteria in conventional , gnotobiotic , and axenic larvae during the first instar . Results showed that conventional and gnotobiotic larvae exhibited no differences in growth , timing of molting , or number of bacteria in their digestive tract . Axenic larvae in contrast grew minimally and never achieved the critical size associated with molting by conventional and gnotobiotic larvae . In the second part of the study we compared patterns of gene expression in conventional , gnotobiotic and axenic larvae by conducting an RNAseq analysis of gut and nongut tissues ( carcass ) at 22 h post-hatching . Approximately 12% of Ae . aegypti transcripts were differentially expressed in axenic versus conventional or gnotobiotic larvae . However , this profile consisted primarily of transcripts in seven categories that included the down-regulation of select peptidases in the gut and up-regulation of several genes in the gut and carcass with roles in amino acid transport , hormonal signaling , and metabolism . Overall , our results indicate that axenic larvae exhibit alterations in gene expression consistent with defects in acquisition and assimilation of nutrients required for growth .
Like most animals , mosquitoes host communities of microbes in their digestive tract that consist primarily of bacteria [1–3] . Both field and laboratory studies indicate that most of these bacteria are aerobes or facultative anaerobes [3–12] . Analysis of 16S rRNA gene amplicons of select species indicates that larvae primarily contain a subset of the bacteria in their aquatic environment , while some but not all of these bacteria are present in adults [4 , 7–9 , 13] . In contrast , controlled experiments show that larvae contain no gut bacteria if they hatch from surface sterilized eggs and are maintained in a sterile environment [7] . Taken together , these findings indicate that mosquito larvae acquire most if not all of their microbiota from their environment and that they transstadially transmit some members of the bacterial community to adults . Aedes aegypti is a key vector of several human pathogens including filarial nematodes and the viruses that cause yellow fever , Dengue fever , Zika fever and Chikungunya [14 , 15] . Ae . aegypti is also an important model for many fundamental studies on mosquito development , immunity and behavior [16–18] . Larvae reared under conventional ( non-sterile ) conditions and fed a nutritionally complete diet molt through four instars before pupating and emerging as adults [19] . Studies dating back to the 1920s noted that Ae . aegypti and other species of mosquito larvae contain bacteria in their gut [20–23] , but conclusions regarding the role of these bacteria in development vary . Some report that bacteria are a source of nutrients or provide other factors that are required for development [23 , 24] while others report that larvae develop on both undefined and defined diets in the absence of bacteria [20 , 25 , 26] . A key challenge in interpreting these variable findings is that researchers during this period lacked the molecular tools needed to characterize the gut microbiota in mosquitoes or determine whether larvae reported to lack bacteria actually were ‘germ free’ . As a result , it is also difficult to evaluate the accuracy of the findings reported . Using high-throughput sequencing approaches , we previously determined that a laboratory population of Ae . aegypti ( UGAL strain ) contains ~100 bacterial operational taxonomic units ( OTUs ) during the larval stage with lower bacterial diversity in adults [7] . Our experiments also indicated that axenic larvae , conclusively shown to have no bacteria , die as first instars when fed a standardized diet and maintained under sterile conditions [7 , 27] . Axenic larvae also die as first instars if standard diet is supplemented with dead bacteria or is preconditioned by co-culture with living bacteria before feeding . However , axenic larvae develop into adults if colonized by bacteria from water containing conventionally reared larvae [7] . Gnotobiotic Ae . aegypti larvae colonized individually by several members of the bacterial community in conventionally reared larvae or the non-community member Escherichia coli also develop normally with adults showing no morphological defects or reductions in fitness as measured by development time , size and fecundity [7 , 27] . Lastly , offspring from field collected Ae . aegypti and several other mosquito species host communities of bacteria that differ from laboratory cultures but exhibit the same dependency on living bacteria for development as UGAL strain Ae . aegypti [28] . Altogether , we conclude from these results that several mosquito species fail to develop if reared under axenic conditions but larvae develop normally into adults if living bacteria are present in the digestive tract . Our results further indicate that development does not depend on a particular OTU or community of bacteria in the larval digestive tract . These findings are important because they implicate gut bacteria as a key factor in the development of larvae into adults , which is the life stage that transmits vector borne pathogens to humans . Understanding the interactions between larval stage mosquitoes and gut bacteria is also important because many of the OTUs in larvae are transstadially transmitted to adults where they can affect vector competence to transmit Plasmodium and arboviruses ( summarized by [2 , 29] ) . In this study , we further assessed Ae . aegypti development by comparing the growth and abundance of bacteria in conventional larvae , gnotobiotic larvae colonized by only E . coli and axenic larvae during the first instar . Based on these data , we then performed a transcriptome analysis of larvae in each treatment as a first step to understanding how bacteria in the gut affect gene expression in first instars . Our results indicated that conventional and gnotobiotic first instars grow similarly , whereas axenic larvae do not attain the critical size associated with molting of conventional and gnotobiotic larvae to the second instar . Our transcriptome analysis further indicated that a number of genes with functions in nutrient acquisition , metabolism , and stress were differentially expressed in axenic larvae when compared to the conventional and gnotobiotic treatments .
Animal care and use are described in Animal Use Protocol A2014 12-013-R1 ( renewal 1/28/2016 ) , which was approved by The University of Georgia Institutional Animal Care and Use Committee ( IACUC ) . The UGA IACUC oversees and provides veterinary care for all campus animal care facilities and is licensed by the US Department of Agriculture ( USDA ) and maintains an animal welfare Assurance , in compliance with Public Health Service policy , through the NIH Office of Laboratory Animal Welfare , and registration with the USDA APHIS Animal Care , in compliance with the USDA Animal Welfare Act and Regulations , 9 CFR . IACUC personnel attend to all rodent husbandry under strict guidelines to insure careful and consistent handling . The University of Georgia’s animal use policies and operating procedures facilitate compliance with applicable federal regulations , guidance , and state laws governing animal use in research and teaching including the: 1 ) The Animal Welfare Act , 2 ) Public Health Service ( PHS ) Policy on the Humane Care and Use of Laboratory Animals , 3 ) United States Government Principles for the Utilization and Care of Vertebrate Animals Used in Testing , Research and Training , 4 ) Guide for the Care and Use of Laboratory Animals , 5 ) Guide for the Care and Use of Agricultural Animals in Research and Teaching , 6 ) American Veterinary Medical Association Guidelines for the Euthanasia of Animals , and 7 ) Applicable Georgia laws . UGAL Ae . aegypti were maintained as previously described by feeding larvae a standardized , nutritionally complete diet ( 1:1:1 rat chow: lactalbumin: torula yeast ) and blood-feeding adult females on an anesthetized rat [30] . Anesthetization of rats ( Sprague-Dawley strain ) obtained from Charles Rivers Laboratories for mosquito blood feeding was performed and monitored by trained personnel as in Animal Use Protocol A2014 12-013-R1 . All larvae used in the study hatched from eggs that were surface sterilized using previously developed methods [7] . In brief , eggs laid 5–7 days previously were submerged in a sterile petri dish containing 70% ethanol in water for 5 min followed by transfer to a second petri dish containing a solution of 3% bleach and 0 . 1% ROCCAL-D ( Pfizer ) in sterile water for 3 min , followed by a second wash in 70% ethanol for 5 min . Surface sterilized eggs were then transferred to a new sterile petri dish and washed 3 times with 10 ml of sterile water followed by transfer to a sterile 10 cm2 culture flask containing 15 ml sterile water and allowed to hatch for 1 hour . Axenic larvae that hatched from eggs were transferred to culture flasks that contained 10 mg of our standard rearing diet that had been sterilized by gamma-irradiation [7] . Conventional larvae were produced by adding 1 ml of water from the general lab culture to a culture flask containing axenic larvae . Gnotobiotic larvae colonized by only E . coli were produced by adding 108 CFUs from an overnight culture of the K12 strain ( National BioResource Project: E . coli/B . subtilis , National Institute of Genetics , Shizuoka , Japan ) to culture flasks containing axenic larvae . When fed a nutritionally complete diet under controlled temperature and photoperiod , Ae . aegypti larvae molt at predictable intervals with each instar being distinguished by the width of the head capsule [19] . To distinguish key traits within the first instar we monitored the growth of conventional , gnotobiotic and axenic larvae by placing newly hatched individuals in 24 well culture plates containing sterilized diet and water . Cohorts of larvae were then observed every 2 h for behavioral and morphological characters associated with feeding , apolysis , and ecdysis . Larval length was measured from the anterior border of the head to the posterior border of the last abdominal segment , which precedes the siphon tube . We also measured the width of the head capsule and prothorax from the dorsal side at their widest point . All measures were made using a Leica stereomicroscope fitted with an ocular micrometer . Critical size , which is defined as the point within an instar when a larva achieved sufficient size to molt , was confirmed by transferring larvae from wells containing diet at specific times post-hatching to wells containing only sterile water . The number of larvae that molted to the second instar was then determined . We estimated the number of bacteria in conventional , gnotobiotic and axenic first instars by two methods: colony count analysis of culturable bacteria and quantitative real time PCR ( qPCR ) . Colony count data were generated as previously described [7] by collecting and surface sterilizing larvae at 18 h post-hatching followed by homogenization in LB broth and culturing on LB plates at 27° for 72 h . The number of bacterial colonies was then counted . For qPCR assays , an absolute standard curve was generated by PCR amplification using the universal bacterial 16S primers HDA1 ( ACTCCTACGGGAGGCAGCAGT ) and HDA2 ( GTATTACCGCGGCTGCTGGCA ) [31] and bacterial DNA from K12 E . coli as template followed by TOPO-TA cloning of the product as previously described [32] . After propagation in E . coli , plasmid was purified using the GeneJet Miniprep kit ( Thermo Scientific ) . A standard curve was then generated by serial dilution of the plasmid ( 107–102 copies ) and qPCR analysis . Bacterial DNA was then isolated from individual conventional , gnotobiotic and axenic larvae as previously described [7] followed by qPCR using the same primers and fitting the data to the standard curve to estimate bacterial abundance via amplicon copy number [32] . Digestive tracts were dissected for immunofluorescence microscopy from conventional , gnotobiotic and axenic larvae at 18 h post-hatching in phosphate buffer saline ( PBS , pH 7 . 4 ) . Samples were fixed in 4% paraformaldehyde in PBS for 20 min at room temperature . After rinsing three times in PBS , guts were dehydrated in ethanol , permeabilized for 20 min in PBS plus 0 . 2% Triton X-100 ( PBT ) for 20 min , and then rewashed three times in PBT . After blocking for 1 h in PBS containing 5% goat serum ( Sigma ) and 0 . 1% Tween 20 ( vol/vol ) ( PBS-GS-T ) , samples were incubated overnight at 4°C with a mouse anti-peptidoglycan primary antibody ( GTX39437 GeneTex ) diluted 1:200 in PBS-GS-T . After washing three times for 10 min in PBS-GS-T , samples were incubated at room temperature for 2 h with an Alexa Fluor 488 goat anti-mouse secondary antibody ( Thermo Fisher ) diluted 1: 2000 in PBS-GS-T . After three washes in PBS , samples were incubated overnight at 4°C with a Cy3-labeled chitin binding protein [33] diluted 1:5 , followed by rinsing in PBS , and mounting on slides in 50% glycerol diluted in PBS containing 1 μg/ml HOECHST 33342 ( Sigma ) . Samples were then examined using a Zeiss LSM 710 inverted confocal microscope with acquired images processed using Adobe Photoshop CS4 . Gnotobiotic larvae colonized by K12 strain E . coli that constitutively expressed green fluorescent protein ( GFP ) were also processed and examined as described above . Flasks of larvae containing conventional , gnotobiotic or axenic larvae were prepared and then used to produce RNA samples for sequencing libraries . This was done by dissecting 50 larvae per biological replicate at 22 h post-hatching in sterile PBS . Larval heads were removed and the digestive tract from each larva was collected to produce a gut sample , while the remainder of each larva formed a non-gut ( carcass ) sample , which consisted primarily of fat body , cuticular epithelium , the nervous system , and trachea . Each gut and carcass from a given larva was transferred to an RNase-free 1 . 5 ml tube . Total RNA was then extracted from each sample using TRIZol ( Life Technologies ) according to the manufacturer’s instructions followed by two DNAase treatments using the Turbo-DNAfree kit ( Life Technologies ) . RNA integrity was assessed on a BioAnalyzer ( Agilent ) using a Eukaryotic Total mRNA Nano chip . Stranded , paired-end libraries ( 75 bp ) were constructed at the University of Georgia Genomics Core Facility for each of 18 samples: three replicates per treatment ( axenic , conventional , and gnotobiotic ) for each tissue ( gut and carcass ) . Each library was barcoded and equal amounts of the libraries were pooled and sequenced on an Illumina NextSeq mid-output flowcell . Resulting FASTQ sequences were de-multiplexed and quality filtered using the FASTX-toolkit ( http://hannonlab . cshl . edu/fastx_toolkit/ ) . Reads with Phred-equivalent scores of < 30 ( corresponding to a per-base error rate of 0 . 1% ) for any base were omitted from further analysis . Reads were then re-paired and mapped to the Ae . aegypti genome ( [34]; assembly AaegL3 , geneset AaegL3 . 3 ) using TopHat2 [35] . Read counts and differential expression were determined using the Cufflinks package [36] . This generated fragments per kilobase of transcript per million reads mapped ( FPKM ) values for Ae . aegypti gene expression . This analysis also identified novel transcripts not present in the L3 . 3 annotation of the Ae . aegypti genome [36] . Un-annotated transcripts were further analyzed using TransDecoder , which is part of the Trinity package [37] that identifies potential protein-coding genes . Gene Ontology ( GO ) terms were obtained from VectorBase annotations . Larval growth and bacterial colony count assays were analyzed by either one-way analysis of variance ( ANOVA ) followed by post-hoc Tukey-Kramer Honest Significant Difference ( HSD ) tests or Fisher’s Exact Test using R ( http://www . r-project . org/ ) . Pairwise analyses between treatments and tissues of transcript abundance data were performed in Cufflinks and significance cutoffs were made at a false discovery corrected p ≤ 0 . 05 [35] .
All first instars hatched with an average head-capsule diameter of 281 . 7 ± 9 . 8 ( SE ) μm . Conventional and gnotobiotic larvae began feeding within 1 h of hatching ( 0 h ) which continued for ~16 h post-hatching as evidenced by the presence of food in the gut and a corresponding increase in body size as measured by length ( Fig 1A ) . We also noted that the width of the prothorax was less than the width of the head capsule at hatching but by 16 h was greater than the width of the head capsule ( Fig 1A ) . These morphological features at 16 h post-hatching were associated with individuals becoming somewhat more sedentary and also not increasing further in length until after molting to the second instar ( Fig 1A ) . Ecdysis to the second instar occurred on average at 23 . 5 ± 1 . 2 h for conventional and 23 . 4 ± 0 . 9 h for gnotobiotic larvae ( t = 0 . 3; P > 0 . 1 ) . Collectively , we interpreted these data as suggesting that conventional and gnotobiotic larvae achieved critical size and initiated apolysis at a similar time in the first instar ( ~16 h ) , which resulted in larvae from both treatments also molting to the second instar at near identical times . Experimental support for these conclusions derived from transferring conventional and gnotobiotic larvae at different times post-hatching to wells without food and assessing whether or not they could molt to the second instar . Results showed that no larvae in either treatment molted if transferred to wells without food prior to 16 h , whereas ~50% molted if transferred at 18 h , and >85% molting if transferred at 20 h ( Fig 1B ) . Prior results showed that axenic larvae consume food [7] . However , measurements made in the current study showed that axenic first instars exhibited less growth as measured by length and the ratio of thorax: head capsule width , which remained <1 ( Fig 1A ) . In turn , no axenic larvae ever molted , which resulted in all individuals ultimately dying as first instars . Previous studies indicated that conventionally reared Ae . aegypti larvae contain gram negative aerobes or facultative anaerobes that are obtained from the water where they feed [7 , 28] . Several of these OTUs as well as E . coli used to colonize gnotobiotic larvae can also be cultured on Luria Broth ( LB ) plates at 27° [7] . We therefore used a colony count assay as a first step to estimating the number of bacteria in individual larvae at 18 h post-hatching . Results indicated that the mean number of bacteria culturable on LB plates was higher in conventional ( 5374 . 9 ± 550 ( SE ) ) than gnotobiotic larvae ( 2632 . 6 ± 414 . 4 ) but this difference was not significant due to inter-individual variation ( Fig 2A ) . As expected , no culturable bacteria were present in axenic larvae ( Fig 2A ) . Since some bacteria in conventional larvae are potentially not culturable on LB plates , we also estimated bacterial abundance using culture-independent qPCR and universal primers that amplify a conserved region of the bacterial 16S rRNA gene . 16S gene copy number did not significantly differ between conventional ( 19 , 852 ± 3 , 841 16S copies ) and gnotobiotic ( 15 , 418 ± 3 , 841 16S copies ) larvae , and no 16S amplicons were generated from axenic larvae ( Fig 2B ) . However , mean values generated by qPCR were also 3 . 7x higher for conventional and 5 . 86x higher for gnotobiotic larvae than colony count estimates . This likely reflected that many bacteria encode multiple 16S operons [38 , 39] and individual cells can be polyploid [40] . qPCR can also capture DNA from both living and dead bacteria . The impact of copy number is well illustrated by K12 E . coli , which is fully culturable on LB plates but contain 7 16S rRNA operons [38] . Dividing the mean 16S copy number for gnotobiotic larvae by 7 yielded a value of 2203 , which was very similar to the estimate generated by colony count . We did not know 16S copy numbers for each of the OTUs in conventional larvae but the same reasoning suggested qPCR estimates were consistent with colony count data . It also suggested that the higher values generated by qPCR versus colony counts more likely reflects 16S copy number than an abundance of bacteria that were not culturable under the conditions we used . We examined the distribution of bacteria in the digestive tract of conventional and gnotobiotic larvae using an anti-peptidoglycan antibody , a Cy3-labeled chitin binding protein that labeled the peritrophic matrix , and Hoechst 33342 that labeled gut cell nuclei ( Fig 3 ) . In the case of gnotobiotic larvae , distribution was also visualized using E . coli that constitutively expressed GFP ( S1 Fig ) . Results showed the presence of bacteria in the foregut , midgut and hindgut of conventional and gnotobiotic larvae ( Fig 3 , S1 Fig ) . All bacteria in the midgut also resided within the endoperitrophic space formed by the peritrophic matrix ( Fig 3 , S1 Fig ) . Anti-peptidoglycan and GFP signal intensity were similar between conventional and gnotobiotic larvae , which was consistent with our colony count and qPCR data that did not detect any differences in bacteria abundance ( Fig 3 , S1 Fig ) . Higher magnification images also clearly indicated that anti-peptidoglycan bound to particles in the endoperitrophic space that morphologically appeared to be rod-shaped bacteria ( Fig 3 ) . In contrast , anti-peptidoglycan did not detect any bacteria that were in contact with midgut cells ( Fig 3 ) . As expected , anti-peptidoglycan did not bind to any particles in the guts of axenic larvae but binding of Cy3-labeled chitin binding protein clearly showed that the midgut of axenic larvae was lined with a peritrophic matrix ( Fig 3 ) . We used Illumina sequencing to transcriptionally profile conventional , gnotobiotic and axenic first instars at 22 h post-hatching which was a time point that preceded molting of conventional and gnotobiotic first instars , whereas axenic larvae remained below critical size ( see Fig 1 ) . We also profiled the gut and carcass in each of these treatments separately . Three biological replicates per treatment and two tissue sources ( gut and carcass ) resulted in a total of 18 samples for which sequencing libraries were produced and analyzed . An average of 45 . 2 million reads were generated per sample ( range: 166–10 . 7 ) , which was reduced to an average of 6 . 3 million paired reads ( range 9 . 9–4 . 4 ) after quality filtering ( S1 Table ) . This resulted in a total of 15 . 8 to 22 . 9 million quality filtered reads per treatment ( S1 Table ) of which 67 . 8% on average mapped to the current assembly of the Ae . aegypti genome ( AaegL3 ) using Tophat ( S2 Table ) . Of the 18 , 293 transcripts that are annotated in the Ae . aegypti reference genome , 13 , 551 had an FPKM ≥1 in one or more of our samples . A total of 1 , 353 transcripts were identified that did not map to the L3 annotation of the Ae . aegypti genome ( Fig 4A ) . Using TransDecoder , 164 of these had predicted open reading frames that were > 100 amino acids ( AA ) , which we searched against the NCBI nr database . BLAST results detected a hit to an annotated insect gene with a bit score > 100 for 125 of these transcripts , which we interpreted as evidence they likely derive from protein coding genes that are absent from the current annotation of the Ae . aegypti genome ( S2 Table ) . However , only 3 of these likely protein-coding transcripts were differentially expressed among treatments ( Fig 4A ) . One of these was a conserved hypothetical protein that was more abundant in the gut and carcass of axenic versus conventional and gnotobiotic larvae . The second was a putative structural component of cuticle that was also more abundant in the carcass of axenic larvae . The third was a transcript significantly upregulated in the gut of axenic larvae that was most similar to the Culex quinquefaciatus gene schnurri: a regulatory factor in the decapentaplegic pathway implicated as a negative regulator of intestinal stem cell proliferation in the midgut of D . melanogaster [41] . The remaining 1 , 228 unannotated transcripts were presumptive non-coding RNAs of which 253 were classified using PLEK [42] as long , non-coding RNAs ( Fig 4A ) . To examine the number of genes that were differentially expressed between treatments , we first limited our consideration to loci with an FPKM of 10 or higher in one condition . Among the three treatments , this resulted in the number of significantly differentially expressed genes ranging from 1 , 328 between conventional and axenic carcasses to 228 between axenic and gnotobiotic carcasses ( Fig 4B ) . We noted that more genes were significantly up-regulated ( 995 ) than down-regulated ( 84 ) in the carcasses of axenic larvae when compared to conventional larvae ( Fig 4B ) . This was also the case when comparing the carcasses of axenic and gnotobiotic larvae ( Fig 4B ) . In contrast , the number of up-regulated versus down-regulated genes was less distinctly different between the carcasses of conventional and gnotobiotic larvae or the guts of axenic , conventional , and gnotobiotic larvae ( Fig 4B ) . Transcripts with an FPKM that was > 10 in axenic but < 1 in gnotobiotic or conventional larvae were classified as preferentially and highly up-regulated under axenic rearing conditions . Only 21 loci met these criteria with 6 being detected in the gut , 15 in the carcass , and none in both tissues . Moreover , only 3 of these loci mapped to annotated genes while 2 generated significant BLAST hits to known insect proteins . These included one acyl-CoA transferase expressed in the gut ( AAEL006672 ) a second acyl-CoA transferase expressed in the carcass ( AAEL000466 ) , and a heat-shock 70 ( HSP70 ) gene ( AAEL017978 ) also expressed in the carcass . The two unannotated transcripts with significant BLAST hits were a predicted diacylglycerol kinase and an asparagine synthetase that were both expressed in the gut . The other 17 loci were unannotated with no significant BLAST hits , which suggested they were non-coding RNAs . We further assessed large-scale differences between treatments and tissues by conducting a principle components analysis ( PCA ) that included all genes with an FPKM value ≥ 1 that were differentially expressed ( log2 fold change ≥ 2 ) in at least one of the comparisons shown in Fig 4B ( see also S3–S5 Tables ) . The first component , explaining 44 . 8% of the variation in our data , separated the samples by tissue type , which not surprisingly showed within each treatment that the differentially expressed genes identified in gut and carcass samples largely did not overlap ( Fig 4C ) . The second component , which explained 28 . 9% of the variation in the data , separated the samples by treatment ( Fig 4C ) . This indicated that the gut and carcass samples from axenic larvae most differed from conventional larvae . However , the pool of differentially expressed genes in conventional and gnotobiotic larvae also did not overlap even though larvae in both treatments grew and molted to the second instar near identically . By extracting global classification of gene ontology ( GO ) terms from VectorBase , we determined that most differentially expressed genes ( log2 fold change ≥ 2 ) in Fig 4B belonged to 7 functional categories: cell cycle , chitin/cuticle formation , metabolism , oxidoreductases , peptidases , signaling , and transport . Up-regulated genes in the guts and carcasses of axenic larvae were most enriched in the categories of metabolism , transport , and oxidoreductases . Most up-regulated genes in the category of oxidoreductases were cytochrome p450 enzymes ( CYPs ) rather than genes associated with the formation or neutralization of reactive oxygen species ( S3–S5 Tables ) . Down-regulated genes in the guts of axenic larvae were most enriched for peptidases , while in the carcass they were most enriched for the category of chitin/cuticle ( Fig 4D ) . Altogether , these results indicated the absence of bacteria in axenic larvae as well as the type of bacteria in conventional versus gnotobiotic larvae affected gene expression in Ae . aegypti first instars . They also indicated gene expression was affected in both gut and non-gut tissues . We next focused on genes in a subset of the categories shown in Fig 4D to gain additional insights into factors that potentially contribute to the disabled growth of axenic larvae . The Ae . aegypti genome contains hundreds of peptidases but this category was of interest because of the known role peptidases play in digestion and the finding that several peptidase genes were significantly down-regulated in axenic larvae . The functional literature on digestive peptidases in Ae . aegypti is restricted to adult females where the principal enzymes identified in bloodmeal digestion are select trypsin-like serine peptidases [43–47] . However , additional trypsins or trypsin-like genes expressed in larvae have also been identified through PCR-based , expressed sequence tag ( EST ) , or transcriptome data sets prepared from whole body samples [48–51] . The first important feature our data set revealed was that most peptidases previously identified in bloodmeal digestion were not expressed in the guts of conventional , gnotobiotic or axenic first instars ( Fig 5A ) . Instead , several other peptidase genes exhibited FPKM values ≥50 in the gut of each treatment , while all of the peptidases with significantly lower FPKM values in axenic versus conventional and gnotobiotic larvae were serine or leukotriene-C4-hydrolases ( Fig 5A ) . Comparing these results with another RNAseq data set [16] indicated these down-regulated peptidase genes are not expressed in the guts or carcasses of adults either before or after consumption of a blood meal . In addition , none of these genes with the exception of AAEL007926 had previously been reported to be differentially expressed in larvae [49] . The second category of interest from the perspective of digestion and nutrient acquisition was transmembrane transporters . Due potentially to lower expression of certain peptidases , several heavy subunit and proton-coupled amino acid ( AA ) transporter genes plus one glucose transporter had significantly higher mean FPKM values in the guts of axenic versus conventional or gnotobiotic larvae ( Fig 5B ) . In contrast , transcript abundance of one sugar transporter was much higher in the guts of gnotobiotic than conventional or axenic larvae ( Fig 5B ) . Several AA transporter genes as well as select neurotransmitter and sterol transporter genes were also significantly up-regulated in the carcasses of axenic larvae relative to conventional and/or gnotobiotic larvae ( Fig 5B ) . Neurotransmitter transporters are involved in the degradation of neurotransmitters in the nervous system , and sterol transporters aid uptake and incorporation of sterols into cell and organelle membranes . While many genes with metabolic or signaling functions were differentially expressed between treatments , the proportion of these genes that were significantly up- or down-regulated exhibited no obvious patterns when examined by GO category distribution alone ( Fig 4D ) . However , certain patterns did emerge when we focused on genes within these categories with essential roles in growth and molting . The first of these gene groups that we examined focused on ecdysteroids , which regulate molting and affect larval growth [52] , juvenile hormone ( JH ) , which influences ecdysteroid function and also affects growth [53] , and select other peptide hormones with roles in ecdysone and JH biosynthesis or other aspects of molting [54] . Cholesterol either stored or from the diet is converted into ecdysteroids through early steps catalyzed by shroud , a short-chain dehydrogenase/reductase , and neverland , a Rieske oxygenase , and later by the Halloween CYPs ( shadow , spook , disembodied , phantom , and shade ) [55] . Only shroud exhibited higher transcript abundances in the carcasses of conventional and gnotobiotic larvae when compared to axenic larvae ( Fig 6A ) . In contrast , shade , which catalyzes the conversion of ecdysone to 20-hydroxyecdysone in target tissues , was significantly more abundant in the carcasses of axenic larvae as were several downstream components of the ecdysone signaling pathway such as the ecdysteroid receptor ( ecr ) , its partner ultraspiracle , and the downstream factor e75 ( Fig 6A ) . Other peptide hormones and associated receptor genes with roles in regulating ecdysone biosynthesis such as prothracicotropic hormone ( ptth ) , or molting such as bursicon and eclosion hormone , were not differentially expressed ( Fig 6A ) . No significant differences were detected in mean FPKM values of allatotropin , allatostatins , or their receptors , which positively and negatively regulate JH biosynthesis in Ae . aegypti [56–58] ( Fig 6A ) . Genes for key JH biosynthetic and metabolic enzymes including putative 3-hydroxy-3-methylglutaryl CoA reductase ( hmgr ) , farnesoic acid O-methyltransferase ( famet ) , and multiple predicted JH esterases also exhibited few differences among treatments ( Fig 6A ) . In contrast , Ae . aegypti encodes multiple members of the takeout gene family , several of which are annotated as JH binding proteins ( JHBPs ) in VectorBase ( jhbp-to ) and were among the most strongly upregulated genes in the carcasses and guts of axenic larvae when compared to conventional or gnotobiotic larvae ( Fig 6A ) . However , takeout genes overall share similarity with odorant binding proteins ( OBPs ) , lipocalins and a putative JHBP ( JP29 ) in Manduca sexta . Thus Takeout proteins are more broadly classified as putative hydrophobic ligand binding proteins [59] . The actual ligands for takeout gene family members are unknown in any insect , but studies in Drosophila implicate takeout in feeding and longevity , while also showing that starvation strongly upregulates takeout expression [60] . In addition to ecdysteroids and JH , growth and metabolism in insects involves the insulin signaling pathway , which converges with amino acid sensing and the target of rapamycin ( TOR ) pathway . FPKM values for several genes in the insulin and TOR pathways were significantly higher in the guts and carcasses of axenic versus conventional or gnotobiotic larvae ( Fig 6B ) . Particularly striking were the increases in mean FPKM values for the insulin receptor ( mir ) , foxo , and the FOXO target 4e-bp , which are up-regulated in several vertebrates and invertebrates including Ae . aegypti in response to starvation or reduced nutrient availability [61–64] . No differences in expression of mir and foxo were detected when conventional and gnotobiotic larvae were compared to one another . However , select other insulin and TOR pathway genes exhibited higher mean FPKM values in gnotobiotic than conventional larvae , although fold differences were usually smaller than in comparisons between axenic and conventional or gnotobiotic larvae ( Fig 6B ) . Altered expression of genes in the insulin and TOR pathways in association with starvation is often coupled with up-regulated expression of genes in energy-producing metabolic pathways such as glycolysis , fatty acid metabolism , and fatty acid oxidation [61] . Mean FPKM values for several genes in each of these processes were significantly up-regulated in the guts and carcasses of axenic larvae when compared to conventional larvae ( Fig 6C–6E ) . A lesser number of these genes were also significantly up-regulated in axenic larvae when compared to gnotobiotic larvae ( Fig 6C–6E ) . Insects including mosquitoes encode a diversity of cuticular proteins ( CPs ) that interact with chitin to form cuticle and/or the peritrophic matrix of the midgut [65] . A total of ten CP families are currently recognized on the basis of different motifs . These include two families distinguished by Rebers and Riddiford ( RR ) consensus sequences ( CPR1 , 2 ) [66] , two others that are classified as Cuticular Proteins Analogous to Peritrophins ( CPAP1 , 3 ) , four CP families of low complexity ( CPLCA , G , W , C ) , and two families designated as CPF and CPT ( = Tweedle ) ( Fig 7 ) . Using the CP accessions curated by Ioannidou et al . [65] , we determined that each had at least one member that was differentially expressed between treatments , which suggested gut bacteria broadly affect CP gene expression ( Fig 7 ) . Transcript abundance of many CP genes was significantly higher in the carcasses of axenic versus conventional and gnotobiotic larvae . However , several of the same CP genes were also differentially expressed between conventional and gnotobiotic larvae ( Fig 7 ) . Prior work establishes that bacteria in the gut induce basal level expression of genes in both the Toll and Imd pathways in adult mosquitoes [67–69] while only basal expression of the Imd pathway is induced in the digestive tract of adult Drosophila [70 , 71] . We thus anticipated that several immune genes would likely be differentially expressed in the guts of axenic , conventional and gnotobiotic first instars . However , immune genes were not among the categories that were significantly enriched in any of our treatments ( Fig 4D , S3–S5 Tables ) . Among the few immune genes that were differentially expressed ( log2 fold change ≥ 2 ) were pgrp-le , which activates the Imd pathway [72 , 73] , and was significantly down-regulated in the guts of axenic versus conventional and gnotobiotic larvae . However , no other components of the Imd pathway were differentially expressed among treatments in either the gut or carcass ( S3–S5 Tables ) . Three späetzle genes ( spz2 , 4 and 6 ) which encode predicted ligands for the Toll receptor , were also down-regulated in the carcasses of axenic versus conventional larvae , but almost no other genes in or regulated by the Toll pathway , including effector proteins , were differentially expressed among treatments .
Our previous results indicated that several species of mosquitoes including Ae . aegypti fail to develop when fed a nutritionally complete diet and cultured under axenic conditions [7 , 28] . This outcome notably contrasts with studies of Drosophila and mice , which show defects in maturation of the digestive tract and immune system but do not require gut microbes for development since axenic cultures of both can be maintained over multiple generations if fed a nutritionally complete diet [71 , 74–77] . Only under conditions of low nutrient availability do axenic Drosophila larvae exhibit delays in development , which can be rescued in gnotobiotic larvae that are singly colonized by particular members of the gut community [78 , 79] . Development of axenic Ae . aegypti can also be rescued in gnotobiotic larvae that are singly colonized by different species of bacteria . Unlike Drosophila , however , several different species of bacteria identified as community members as well as some non-community members such as E . coli rescue development of Ae . aegypti larvae , which develop at the same rate as conventionally reared larvae [7 , 28] . Adult Ae . aegypti produced from gnotobiotic larvae singly colonized by E . coli also show no morphological defects or reductions in fitness when compared to adults produced from conventional larvae [27] . Altogether , these findings suggest an essential role for living microbes in development of Ae . aegypti . Axenic larvae will not develop when provided diet along with dead bacteria or diet that has been pre-conditioned by living bacteria [7] . Along with our current findings , these data argue against bacteria being an essential food source or providing a particular nutrient essential to larval development . In contrast , the absence of living bacteria in the gut could adversely affect physiological processes in larvae with roles in nutrient acquisition or assimilation . Thus , the primary goal of this study was to assess whether axenic larvae exhibit alterations consistent with this possibility or alternatively exhibit defects that point to other factors that could potentially underlie their inability to develop . We first assessed whether conventional and gnotobiotic larvae exhibit any fine scale differences in growth during the first instar , and also whether axenic larvae exhibit specific traits that help explain why they do not molt . Our results identified no differences in growth or timing of molting between conventional and gnotobiotic first instars . The statistically similar number and distribution of bacteria in conventional and gnotobiotic larvae suggests the digestive tract of both contains sufficient space to host a finite number of bacterial cells that E . coli occupied when alone but which multiple species occupied in conventional larvae . The observation that all bacteria in conventional and gnotobiotic larvae reside inside the endoperitrophic space further suggests their essential role in growth does not involve direct contact with midgut cells . In contrast , our results indicate that axenic larvae grow a small amount but never reach the critical size associated with apolysis and other events that precede molting by conventional and gnotobiotic larvae . Studies of several insects indicate that individual species often increase in size by approximately the same factor through the penultimate instar [80 , 81] . Within each instar , larvae also initiate a molt upon reaching a particular critical size , which is often associated with allometries such as the ratio between head capsule width and weight . In the first through penultimate instar , reaching critical size stimulates ecdysteroid hormone release , which induces the epidermis to produce a new cuticle while digesting most of the old endocuticle ( apolysis ) . This is followed by ecdysis , which refers to shedding of the old exo- and epicuticle and the beginning of the next instar . In the final instar related events result in pupation . The aquatic habit and small size of Ae . aegypti first instars precluded using the ratio between head capsule width and weight to estimate when larvae achieved critical size . However , we determined that the ratio of prothorax width to head capsule width exceeds 1 when conventional and gnotobiotic Ae . aegypti first instars achieve critical size . This measure also supported the conclusion that axenic larvae do not achieve critical size . Our transcriptome analysis at 22 h post-hatching indicated that approximately 12% of the annotated genes in the Ae . aegypti genome are differentially expressed in axenic larvae when compared to conventional or gnotobiotic larvae . However , this profile consisted primarily of genes in seven categories that included the down-regulation of select peptidases in the gut and up-regulation of several genes in the gut and carcass with roles in amino acid transport , signaling through the ecdysteroid , insulin and TOR pathways , and fatty acid oxidation . Reduced expression of select peptidases suggests the absence of bacteria may adversely affect digestion , while the increased transcription of amino acid transporters , genes associated with insulin and TOR signaling , and fatty acid oxidation suggests a response to acquire additional nutrients and use lipid reserves from embryogenesis for nourishment . Similar patterns have been observed in mammals , Drosophila and mosquitoes in response to starvation stress [62 , 82–84] . Insulin and TOR signaling have also been implicated in affecting JH synthesis , ecdysteroid synthesis , and ecdysteroid signaling in several insects including Ae . aegypti [62 , 85–89] . That Ae . aegypti encodes multiple takeout orthologs , which are up-regulated in axenic larvae , is also intriguing given evidence showing that takeout expression is strongly upregulated in Drosophila larvae subjected to starvation but not other stress factors [60] . As previously noted , takeout gene products exhibit features of OBPs , JP29 , a predicted JH binding protein , and lipocalins that transport a diversity of hydrophobic molecules including retinoids , steroids , lipids and pheromones [60 , 90] . The actual ligands Takeout proteins bind , however , are unknown . A number of CP genes are differentially expressed in axenic larvae relative to conventional and gnotobiotic larvae as are several CYPs assigned to the category of oxidoreductases . The significance of these differences in regard to growth and molting are uncertain although other studies have noted the differential expression of both CPs and CYPs in response to stress factors including starvation , heat , cold , and ionizing radiation [82 , 83] . Insects also continuously deposit endocuticle during the intermolt period [52] , which could explain why CP transcripts are detected in both axenic larvae , which never molt , and conventional or gnotobiotic larvae that were post-critical size and in the process of molting when tissue samples were collected . In contrast , we are uncertain why so few differences were detected among our treatments in regard to expression of immune genes . At minimum our results suggest differences between first instars and prior studies conducted in adult mosquitoes [67–69] . Why such differences exist , however , will require future study . While our primary goal was to identify differentially expressed genes in axenic larvae , our results also identified several differences between conventional and gnotobiotic larvae . This indicates that colonization of larvae by E . coli alone does not fully recapitulate gene expression patterns in conventional larvae , and that the community of bacteria in the gut affects gene activity in larvae . On the other hand the differences in gene expression detected between conventional and gnotobiotic Ae . aegypti larvae are insufficient to substantially alter growth given the similarities in when larvae molted to the second instar and recently completed results showing that conventional and gnotobiotic larvae develop into adults that exhibit no differences in size or fecundity [7 , 27] . In summary , this study indicates that living bacteria in first instar Ae . aegypti affect growth and alter the expression of several genes with roles in nutrient acquisition , nutrient assimilation and stress . Since we examined only a single time point in the first instar , our transcriptome data do not identify when axenic larvae first exhibit changes in gene expression relative to conventional or gnotobiotic larvae . However , given that axenic first instars grow minimally beyond their size at hatching suggests the absence of living bacteria in the digestive tract adversely affects nutrient acquisition and/or assimilation almost immediately after hatching . We also recognize that our study did not include a treatment where conventional and gnotobiotic larvae were deprived of food to ascertain whether similar patterns are exhibited when compared to axenic larvae . We did not do this because at the onset of the investigation we did not know the key patterns our transcriptome data would identify . However , the results reported here position us to study select genes in this manner , while also providing information that will be used in functional studies of axenic larvae . In terms of disease control , the current study advances prior results by suggesting that the absence of gut bacteria disables growth at least in part by altering the metabolism of mosquito larvae and nutrient uptake . If correct , these findings further suggest that disruption of the microbial factors larvae require could potentially be used to reduce vector abundance and disease transmission [91] .
Transcriptome data have been deposited in the Short Read Archive under accession PRJNA340082 .
|
Several mosquito species including Aedes aegypti transmit pathogens as adults that cause disease in humans and other vertebrates . It has also long been known that mosquitoes host bacteria in their digestive tract , which are primarily acquired during the larval stage and transstadially transmitted to adults . Our recent results indicate that axenic larvae , which lack bacteria , do not develop beyond the first instar , whereas larvae with living bacteria develop into adults . To better understand the effects of bacteria on mosquito development , we compared growth , molting and gene expression in larval Ae . aegypti that contained several species of bacteria , only one species of bacterium ( Escherichia coli ) , or no bacteria . Results showed that larvae containing several species or only E . coli grew and molted very similarly while larvae with no bacteria grew minimally and never molted . A number of Ae . aegypti genes with roles in regulating growth were differentially expressed in larvae without bacteria when compared to larvae with bacteria . Overall , our results indicate that mosquito larvae without bacteria do not grow or molt because of defects in assimilating nutrients .
|
[
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion",
"Data",
"Deposition"
] |
[
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"enzymes",
"enzymology",
"diet",
"animals",
"physiological",
"processes",
"developmental",
"biology",
"nutrition",
"insect",
"vectors",
"bacteria",
"digestive",
"system",
"molting",
"aedes",
"aegypti",
"proteins",
"epidemiology",
"gene",
"expression",
"life",
"cycles",
"disease",
"vectors",
"insects",
"arthropoda",
"gastrointestinal",
"tract",
"biochemistry",
"mosquitoes",
"anatomy",
"physiology",
"genetics",
"biology",
"and",
"life",
"sciences",
"proteases",
"larvae",
"organisms"
] |
2017
|
Transcriptome Sequencing Reveals Large-Scale Changes in Axenic Aedes aegypti Larvae
|
Echolocating bats rely on active sound emission ( echolocation ) for mapping novel environments and navigating through them . Many theoretical frameworks have been suggested to explain how they do so , but few attempts have been made to build an actual robot that mimics their abilities . Here , we present the ‘Robat’—a fully autonomous bat-like terrestrial robot that relies on echolocation to move through a novel environment while mapping it solely based on sound . Using the echoes reflected from the environment , the Robat delineates the borders of objects it encounters , and classifies them using an artificial neural-network , thus creating a rich map of its environment . Unlike most previous attempts to apply sonar in robotics , we focus on a biological bat-like approach , which relies on a single emitter and two ears , and we apply a biological plausible signal processing approach to extract information about objects’ position and identity .
The growing use of autonomous robots emphasizes the need for new sensory approaches to facilitate tasks such as obstacle avoidance , object recognition and path planning . One of the most challenging tasks , faced by many robots , is the problem of generating a map of an unknown environment , while simultaneously navigating through this environment for the first time [1] . This problem , is routinely solved by echolocating bats that perceive their surroundings acoustically ( other animals also solve this task on a daily basis using a range of sensory modalities ) [2] . By emitting sound signals and analyzing the returning echoes , bats can orient through a new environment and probably also map it [3] [4] . Inspired by this ability , we present the ‘Robat’—a fully autonomous terrestrial robot that solely relies on bat-like SONAR to orient through a novel environment and map it . Using a biologically plausible system with two receivers ( ears ) and a single emitter ( mouth ) which produced frequency modulated ( FM ) chirps at a typical bat rate , the Robat managed to move through a large out-doors novel environment and map it in real-time . There have been many attempts to use airborne sonar for mapping the environment and moving through it using non-biological approaches; for example by using an array of multiple narrow-band speakers [5 , 6] [7] and or multiple microphones [8] . These studies proved , that by using multiple emitters , or by carefully scanning the environment with a sonar beam , as if it were a laser , one can map the environment acoustically , but these approaches are very far from the biological solution [9] . A bat emits relatively few sonar emissions towards an object , and it must rely on two receivers only ( its ears ) in order to extract spatial information from its very wide bio-sonar beam which can reach 60 degrees ( 6 dB double side drop in amplitude [10] [11] [12] ) . Unlike the narrow-band signals typically used in robotic applications , the bat’s wide-band signals provide ample spatial information allowing it to localize multiple reflectors within a single beam . This is the approach we aimed to test and mimic in this study . Numerous studies have shown that echoes generated by emitting bat-like sonar signals contain spatial information that can be exploited for localization and identification of objects [13] [14] [15] [16] [17] [18] . Several previous attempts have been made to model and mimic bats’ spatial abilities of localization and mapping [19] [20] . One of the most comprehensive attempts to use a biological approach to map the environment was ‘BatSLAM’ [21] , which relied on mammalian brain-like computation for simultaneous localization and mapping of a novel environment using biomimetic sonar . Using a biological representation of the data ( the cochleogram ) the BatSLAM algorithm generated topological maps in which the nodes represent unique places in the environment and the edges represent the robot’s displacements between them . The approach of recognizing a location based on its unique acoustic signature was further broadened by Vanderelst et al . [6] who classified a wide range of natural scenes based on their acoustic statistics , once again , without extraction of their spatial characteristics . Vanderelst et al . limited the information extracted from the echoes to the acoustic resolution available to a bat , and they were still successful in achieving useful scene recognition . Our work differs from these former studies in two important respects: ( 1 ) Our Robat moved through the environment autonomously while the previous robots were driven by the user . ( 2 ) We mapped the 2D structure of the environment , while they mapped the position of the robot in the environment . Namely , in our approach the outline of the objects that were encountered by the Robat were delineated so that paths ( free of obstacles ) were revealed for future use . In these previous studies , objects in the environment were mapped as locations with a unique acoustic representation so that when encountered again , the agent could localize itself on the acoustic-map , but no spatial information about objects’ size or orientation was extracted . When moving autonomously , such information is essential for movement planning . In addition to mapping , our Robat had to autonomously move through the environment while avoiding obstacles . Some previous attempts were made to model orientation and obstacle avoidance using a biological echolocation-based approach . For example , Vanderelst et al . [9] , suggested a simple sensorimotor approach for obstacle avoidance based on turning away from the louder of the two echoes received by the ears . They showed that a simulated agent can move through a novel environment without any mapping of the positions or borders of the objects within it . This approach might be beneficial when an animal wants to move fast through the environment without an intention of returning to specific locations within it , but if the animal needs to find its way back to some point in this environment ( e . g . , to its roost ) , or to plan its movement to a specific location , some mapping must be performed . For example , the robust low-level sensorimotor heuristic presented in [9] could be combined with higher level mapping algorithms ( e . g . , [22] ) . To our best knowledge , our Robat is the first fully autonomous bat-like biologically plausible robot that moves through a novel environment while mapping it solely based on echo information—delineating the borders of objects and the free paths between them and recognizing their type .
The Robat moved through the environment emitting echolocation signals every 0 . 5m thus mimicking a bat flying at 5m/s while emitting a signal every 100ms which is within the range of flight-speeds and echolocation-rates used by many foraging bats [23] [24 , 25] [25] [26] . Every 0 . 5m , the Robat emitted three bat-like wide-band frequency-modulated sound signals while pointing its sensors ( emitter and receivers ) in three different headings: -60 , 0 , 60 degrees relative to the direction of movement ( Fig 1A ) . This procedure aimed to overcome the narrow acoustic beam of the Robat and to better mimic a bat beam which is typically much wider than that of our speaker ( see Methods ) [27] [28] [10] [29] [11] [12] . Following echo acquisition , acoustic peaks of interest ( representing objects ) were identified in the echoes ( Fig 1B ) . Equivalent peaks—i . e . , peaks returning from the same object—received by the two ears were matched and the reflecting objects were localized . The time-delay between the emission and the arrival of the echoes was used to determine the distance of an object and the difference between the time of arrival of the echo to the two ears was used to determine its azimuth ( i . e . , Mapping was performed in 2D , Fig 1C , Turquoise points depict objects’ location , see Methods for full details ) . Importantly , the Robat was able to localized multiple objects whose echoes were received within a single beam ( S1 Fig ) . This ability has not been reported in previous studies and bats are likely able to do so . After every 5 steps ( i . e . , 2 . 5m ) the Robat applied an inflation and interpolation algorithm that incorporated the newly mapped objects into the map that has been created so far ( based on the previous echoes , Fig 1C , yellow shaded area , see Methods ) . At each time step , following echo acquisition and object localization , the Robat planned its next movement according to the iterative map that has been created so far and according to the objects detected in the most recent acquisition . Movement planning was based on the bug algorithm [30] which can be simply described as turning 90 degrees to the right , whenever an obstacle is encountered ahead , and then turning left to maneuver around the obstacle . The movement and mapping algorithms were tested in two outdoor environments: ( 1 ) The pteridophyte greenhouse ( 5m x 12m ) and ( 2 ) The palm greenhouse ( 40m x 5m ) both situated in the Tel Aviv University Botanical Garden . The Robat successfully moved through both new environments without hitting objects and while mapping their locations and contour line ( see Robat’s trajectory depicted in black in Fig 2A ) . When an obstacle was placed in the Robat’s way , it moved around it ( Fig 2B ) . To quantify the mapping performance , we compared the contour of the objects as it was estimated by the Robat to the real contour ( which we estimated from drone images in the Palm greenhouse and measured manually in the Pteridophyte greenhouse , see Methods ) . In the palm greenhouse , the mean distance between the two contours was 0 . 42 ± 0 . 74 ( mean + STD ) [m] meaning that along the 35m trail that the Robat passed and mapped in the Palm greenhouse , the estimated borders of the objects on both sides of the trail , were off by 42cm on average , relative to their real position . This might seem inaccurate when considering bats’ ability to estimate range with an accuracy of less than 1cm in a highly controlled experiment , [31] [32] but it should be emphasized that the Robat only detected and localized parts of the objects while their borders were delineated based on our inflation an interpolation algorithm ( Methods ) . Moreover , note that many of the objects in our environment were plants with multiple branches so that the exact borders of the objects were inherently difficult to define ( even in the drone images ) . Similar performance ( 0 . 44 ± 0 . 25 ( mean + STD ) [m] ) was observed in the second environment ( the Pteridophyte greenhouse , S2 Fig ) . When moving through the environment , a real bat can probably use echoes in order to classify objects into categories ( e . g . , rocks , trees , bushes ) and even to identify specific objects ( e . g . , a specific beech tree in its favorite foraging site ) . Such recognition would greatly assist the bat to navigate , for example , by recognizing specific landmarks at important turning points along its flight route and it could also assist its foraging , for example , by recognizing specific vegetation that is rich in fruit or insects [33] [17] . So far we demonstrated that the Robat can translate a novel natural environment into a binary map of open spaces and obstacles . In order to improve the mapping , we added a classification step to the algorithm , which was performed using a neural-network that was trained to distinguish between two object categories—plants and non-plants . To this end , a set of acoustic features were extracted from the echoes and used as input for the network ( Methods ) . The Robat was able to classify objects as plants or non plants significantly above chance level ( Fig 2C and Table 1 ) with a balanced accuracy of 68% ( chance was 50% , P = 0 . 01 , based on a permutation test with 100 permutations , the balanced accuracy is the number of correct classifications in each class , divided by the number of examples in each class , averaged over all classes . This measurement mitigates biases which could rise from unbalanced class sizes , see Methods ) . The classification performance is also shown in Fig 2a where colored points depict plants ( green ) and non-plants ( gray ) . Finally , we tested the functionality of this classification ability by purposefully driving the Robat into a dead end where it faced obstacles in all directions ahead ( i . e . , right , left and straight ahead , S7 Fig ) . The Robat had to determine which of the three obstacles was a plant , through which it could drive , and it did so successfully at ~70% of the cases ( in accordance with its ~70% accurate classification rate , see movie: https://www . youtube . com/watch ? v=LzGGuzvYSH8-second 49 and onward ) .
In this study , we managed to build an autonomous robot that moves through a novel environment and maps it acoustically using bat-like Bio-sonar . We achieved high mapping accuracy , despite our simple approach , proving the great potential of using active wide-band sound emissions to map the environment . We created a ( 2D ) topographic map which would allow us to plan future movements through the environment ( and not a topological map ) . The statistical approach presented in [9] is therefore complementary to ours , allowing classifying specific locations based on their echoes . For example , when navigating back to a specific location using the map created by the Robat , their approach could be used to validate the arrival at the desired location and also to help adjust the map to improve its accuracy . The Robat was much slower than a real bat , stopping for ca . 30 seconds every 0 . 5m to acquire echoes . This slowness was however , merely a result of the mechanical limitations of our system and mainly the gimbal that was slow . Using a speaker with a wider beam ( that eliminates the need to turn at each location ) would allow the Robat to acquire echoes on the move , while moving as fast as a bat . Importantly , despite our stopping for echo recording , the acoustic information we acquired did not differ from that received by a bat , except for the fact that a bat’s echoes would also be slightly Doppler-shifted ( but this would probably not affect any of our results ) . In some respects , our processing was not fully bat-like . We used a sampling rate of 250kHz , which is higher than the theoretical time precision of the auditory system [34] . Bats and other small mammals have been shown to estimate azimuth with an accuracy of <10degrees ( the exact accuracy depends on the azimuth , ( e . g . [35 , 36] ) ) . This accuracy accounts for an inter-aural time difference of <10μs which is in accordance with our sampling rate ( sampling at 250kHz is equivalent to an error of ~5μs when estimating time differences between two ears ) . Therefore , even if our computation was different from that of a bat ( which does not cross-correlate two highly sampled time signals ) the overall accuracy allowed by our approach was not better than that of a bat . Moreover , due to the inflation and interpolation method that we used in order to delineate the borders of the objects , the effective accuracy of our mapping was much lower than that allowed by this high sampling rate , and probably much lower than that available to bats [31 , 32] . Therefore , we hypothesize that using an auditory preprocessing model like that used in Batslam for example [21] would probably not change our results dramatically . Another advantage that we had over real bats was the relatively large distance between the two ears which were spaced 7cm apart—ca . two times more than in a large bat . This probably allowed more accurate azimuth estimations , but once again , we hypothesize that because of the use of inflation , this did not improve our performance dramatically . Importantly , we managed to extract information about multiple objects within a single sonar beam . On average , in each echo that contained reflections ( some echoes did not ) we detected 4 . 1 objects positioned in a range of azimuths between -50—50 degrees . Another important difference between the Robat and an actual bat is the lack of an external ear in the Robat . The angle-dependent frequency response of the external ear allows bats ( and other animals ) to gain information about the location of a sound source in three dimensions . Because we relied on temporal information for object localization , we used a first approximation of an ear . Adding a structure mimicking the external ear could have further improved our localization performance and it would be essential in order to expand our mapping to 3D . In order to better mimic the bat’s beam , we used three beams ( directed 60 degrees apart ) , but this made our task easier than a bat’s because we could analyze the echoes returning from each direction separately . We therefore also tested an approach in which we sum the three echoes collected ( with different headings ) at each acquisition point , thus mimicking a wider beam . Even with this degraded data , we were able to map the environment with a decent accuracy of 1 . 14 ± 0 . 70 [m] ( mean + STD , S6b Fig ) , an accuracy that would allow future planning of trajectories while avoiding obstacles on the way . In some respects our approach was probably much more simplistic than a bat . For example , the obstacle avoidance algorithm was very simple and a better approach would probably use control-theory to steer the Robat around obstacles [37] . In terms of mission priority , we used serial processing where the Robat first processes new incoming sensory information; it then performs the urgent low-level task of obstacle avoidance and path planning , and only every several acquisitions , it performs the high-level process of map integration . There is much evidence that the mammalian brain also performs sensory tasks sequentially ( e . g . , [38] ) but it would be interesting to test some procedures for parallel processing in the future . In addition to mapping the positions of objects in the environment , a complete map should also include information about the objects such as their type or identity . To show that such information is available in the echoes , we developed a classifier that can categorize objects based on their echo . We hypothesize that the medium classification performance that we achieved ( 68% ) was a result of our choice of categories . We trained the classifier to distinguish between plant and non-plant objects but these are not always two well distinct groups . For example , the echo of an artificial object such as a fence will have vegetation-like acoustic features and indeed most of the classifier’s mistakes were recognition of non-plants as plants . Bats might thus divide their world of objects differently , perhaps to diffusive vs . glint-reflecting objects . Altogether , we show how a rather simple signal processing approach allows to autonomously move and map a new environment based on acoustic information . Our work thus proves the great potential of using acoustic echoes to map and navigate , a potential that is translated into action by echolocating bats on a daily basis .
The Robat was based on the ‘Komodo’ robotic platform ( Robotican , Israel ) . The Bio-sonar sensor was mounted on a DJI Ronin gimbal which allowed turning the sensing unit relatively to the base of the robot in a stable manner . The sensing unit included an ultrasonic speaker acting as the bat’s mouth ( VIFA XT25SC90-04 ) and 2 ultrasonic microphones acting as the bat’s ears spaced 7cm apart ( Avisoft-Bioacoustics CM16/CMPA40-5V Condenser ) . The speaker and the microphones were connected to A/D and D/A converters which were based on the USB-1608GX-2AO NI DAQ board , sampling at 250KS/s at each ear . The emitted signal was a 10ms FM chirp sweeping between 100-20kHz . It was amplified using a Sony Amplifier ( XM-GS4 ) . An uEye RGB camera , was used for image collection for validation purposes only . Three 2 . 4GHz/5 . 8GHz antennae were mounted at the rear of the Robat for wireless communication between the Robat and a stationary station . This allowed viewing the map created by the Robat in real time , but importantly , all calculations and decisions were performed on the Robat itself . While moving , the Robat stopped every 0 . 5m ( based on its odometry measurements ) and the sonar system ( emitter and receivers ) was rotated to three different headings [0 , 60 , -60 degrees] relative to the direction of movement , a sound signal ( see above ) was emitted , and echoes were recorded . Each recording was 0 . 035 sec long , equivalent to a range of ca . 6 meters ( farther objects were thus ignored at each emission ) . The signal-to-echo delay time and the time of arrival differences of the echoes to the two ears ( i . e . , the Interaural Time Difference ) were used together in order to map the environment . To this end , the received signals were cross-correlated with the theoretical emitted signal . The cross-correlated signal was normalized relative to the maximum value of the recording , and a peak detection function was used to find peaks of interest ( python peakutil with a minimal peak distance of 0 . 002 sec , and a min amplitude of 0 . 3 . ) . To match peaks arriving at the right and left ears , for each peak detected in one ear , an equivalent peak was searched for in the other ear within a window of +/- 0 . 001 sec . If a peak was found , the Pearson correlation was used to determine if the two echoes were reflections of the same object . For this purpose , a segment of 0 . 01 seconds around each peak was cut and the correlation between the two time signals ( one from each ear ) was computed . Only correlations higher than 0 . 9 were accepted . This threshold was conservative thus potentially resulting in missing of objects , but it reduced the localization of artifact non-existent objects . Because the Robat emitted very 0 . 5 m—there was much overlap between echoes of consecutive emissions . We were therefore likely to detect an object several times , so a conservative approach was chosen . In addition to its position , each object on the map was defined by three parameters: “C |T |P” , where C is the Pearson correlation coefficient between the left and right ears for the specific point , T is the object’s type based on its acoustic classification—either artificial or a plant , and P is the classification probability ( see more below about the classification process ) . Results in the in-doors controlled environment showed that using two ears , the mean error in distance estimation was 1 . 3 ± 2 . 1 [cm] ( mean + STD , S3 Fig ) and the mean azimuth estimation error was 1 . 2 ± 0 . 7 [degrees] ( mean+STD , S3 Fig ) . Importantly , these are the results for a single reflector , so accuracy in the real environment where many reflections are received at each point will be lower . Every 5 Robat-steps , newly localized objects were integrated into the map that was created so far . This was done using an Iterative-Object-Inflation algorithm , which inflated points into squares and connected them . To this end , the entire area around the Robat was divided into a grid with 2000x2000 pixels ( 5x5cm2 each ) . Each detected object was placed in the corresponding pixel on the map and was inflated to an area of 20x20 pixels around its center ( i . e . , 1x1 m2 , S4 Fig ) . This procedure creates a binary map with 1’s depicting objects and 0’s depicting a free path . Pixels along the trajectory that the Robat previously moved through always received the value 0 depicting an open path ( even if they were within the 20x20 window of a detected object ) . We chose a very simple obstacle avoidance approach also known as the ‘bug algorithm’ [39] . During the exploration process , the Robat moved forward in steps of 0 . 5m between consecutive acquisition points . When detecting an obstacle less than 1 . 2m in front of it , the Robat turned 90 degrees towards the right , and performed a 1m step towards the right ( after checking that there is no obstacle ahead ) . After performing a 1m step to the right , the Robat turned 90 degrees to the left and acquired an echo . If no obstacle was detected ( meaning that the obstacle has been passed ) the Robat continued straight ( i . e . , in its previous direction before turning right ) . If the way was still blocked ( i . e . , the obstacle was not passed ) , the Robat turned again to the right and kept moving towards the right ( 90 degrees relative to its original direction ) . In order to better mimic the bat , that has a beam much wider than Robat’s beam , we examine an approach of summing the echoes returning from the three different headings ( mentioned above ) into one superposition echo , and then running the same ( detection , localization and mapping ) algorithms as described above . In order to examine the acoustic map generated by the Robat , inspired by [40] , we collected aerial images using a drone ( DJI Phantom 4 , DJI ) , to construct a complete ground truth map of the area . This procedure was only performed for the large palm greenhouse ( 40x5 m2 ) . The contour of the objects on both sides of the trail in the greenhouse was extracted and compared to the contour of the inflated map that was acoustically reconstructed by the Robat ( both contours were marked manually ) . Each of the two contours was fit by a 55-coefficient order polynomial function which was then sampled at 500 points to get a high resolution description of the contour . The two contours ( real and Robat-estimated ) were compared by calculating the root-mean-square distance between them ( the average over these 500 points , S5 Fig ) . Acoustic based object classification was performed using a neural-network that was trained on a binary task—classifying whether and object was a plant or not . Only objects that were located closer than 3[m] from the sensing unit were classified . 0 . 035 s long echoes were used from both the right and left ear . These recordings were passed through three band pass filters , without the transmitted echo , ( 20-40kHz , 40-60kHz and 60-100kHz ) . Each echo was represented by 6 signals—3 filters x two ears . Next , a set of 21 acoustic features ( Table 2 ) were extracted from each band-passed recording following T . Giannakopoulos [41] . Each echo was divided into seven windows equally spaced with an overlap of 40ms and the 21 features were extracted for each window generating a total of 147 dimensions per signal ( 21 features x 23 windows ) . The classifier was thus fed with 6 signals ( 483 dimensions each ) and the decision of the majority of the six classifiers was used . The data was fed into a neural network with the following architecture: We used Python’s TensorFlow to construct and train a three-layer neural-network ( using the Keras directory ) . The training sets included 788 plant examples and 628 non-plant examples collected on several sites on campus . We used the camera that was on the Robat to label the echoes . Finally , to assess the statistical significance of our classification , we ran 100 permutations in which we assigned the training data randomly into the two classes ( plants and non-plants ) , trained a classifier for each permutation and tested it on the same test-data . We also tested several additional classification methods before choosing the neural-network . We tested a KNN ( K nearest neighbors ) classifier with five different distance measurements: Mahalanobis , Euclidean , Correlation , Minkowski and Canberra . We also tested two additional approaches for dimensionality reduction ( before using the KNN ) including PCA and LDA . In addition , we also tested a linear SVM classifier . For all classifiers , we used the same input features ( see above ) . The results were similar for most classifiers , but the neural network performed slightly better than the other ( S8 Fig ) .
|
Many animals are able of mapping a new environment even while moving through it for the first time . Bats can do this by emitting sound and extracting information from the echoes reflected from objects in their surroundings . In this study , we mimicked this ability by developing a robot that emits sound like a bat and analyzes the returning echoes to generate a map of space . Our Robat had an ultrasonic speaker mimicking the bat’s mouth and two ultrasonic microphones mimicking its ears . It moved autonomously through novel out-doors environments and mapped them using sound only . It was able to negotiate obstacles and move around them , to avoid dead-ends and even to recognize if the object in front of it is a plant or not . We show the great potential of using sound for future robotic applications .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] |
[
"acoustics",
"medicine",
"and",
"health",
"sciences",
"engineering",
"and",
"technology",
"ears",
"vertebrates",
"robots",
"animals",
"mammals",
"echoes",
"remote",
"sensing",
"robotics",
"bioacoustics",
"head",
"physics",
"mechanical",
"engineering",
"eukaryota",
"anatomy",
"bats",
"biology",
"and",
"life",
"sciences",
"physical",
"sciences",
"amniotes",
"organisms",
"sonar",
"acoustic",
"signals"
] |
2018
|
A fully autonomous terrestrial bat-like acoustic robot
|
Macrophage cells that are stimulated by two different ligands that bind to G-protein-coupled receptors ( GPCRs ) usually respond as if the stimulus effects are additive , but for a minority of ligand combinations the response is synergistic . The G-protein-coupled receptor system integrates signaling cues from the environment to actuate cell morphology , gene expression , ion homeostasis , and other physiological states . We analyze the effects of the two signaling molecules complement factors 5a ( C5a ) and uridine diphosphate ( UDP ) on the intracellular second messenger calcium to elucidate the principles that govern the processing of multiple signals by GPCRs . We have developed a formal hypothesis , in the form of a kinetic model , for the mechanism of action of this GPCR signal transduction system using data obtained from RAW264 . 7 macrophage cells . Bayesian statistical methods are employed to represent uncertainty in both data and model parameters and formally tie the model to experimental data . When the model is also used as a tool in the design of experiments , it predicts a synergistic region in the calcium peak height dose response that results when cells are simultaneously stimulated by C5a and UDP . An analysis of the model reveals a potential mechanism for crosstalk between the Gαi-coupled C5a receptor and the Gαq-coupled UDP receptor signaling systems that results in synergistic calcium release .
The G-protein-coupled signal transduction system integrates a wide range of intercellular signals and actuates downstream pathways . G-protein-coupled receptors ( GPCRs ) are composed of seven α-helices that span the plasma membrane , an extracellular domain that is activated by an agonist and an intracellular domain that binds a guanine nucleotide heterotrimer made up of different α , β , and γ subunit isoforms . This receptor system accounts for 40–50% of modern medicinal drug targets but only 10% of the known receptors are targeted by drugs [1] . Though the system is physiologically and pharmacologically important , the mechanism by which the system integrates multiple signals is not well understood [2] . We address the G-protein-mediated route to calcium release in RAW264 . 7 cells . When activated by a specific ligand , the G protein heterotrimer dissociates to free Gα-GTP and Gβγ . Specific Gα and Gβγ isoforms are able to bind specific isoforms of phospholipase C β ( PLCβ ) and catalyze the synthesis of inositol ( 1 , 4 , 5 ) -triphosphate ( IP3 ) and diacylglycerol ( DAG ) from phosphatidylinositol ( 4 , 5 ) -bisphosphate ( PIP2 ) [3] , [4] . In addition to its catalytic activity , PLCβ acts as a GTPase for Gα-GTP [5] . IP3 binds to specific receptor-channels on the membrane of the ER to release Ca2+ into the cytosol [6] . DAG and Ca2+ bind to and activate protein kinase C ( PKC ) which may phosphorylate and inactivate specific PLCβ isoforms [7] . G protein receptor kinase ( GRK ) is activated once it is phosphorylated by PKC [8] and is localized to the plasma membrane by Gβγ [9] . Though phosphorylation has not been shown to be necessary for GRK activation , we have assumed so in our model because phosphorylation by PKC may release the inhibition of GRK2 by being bound to calmodulin [8] . Activated GRK can then phosphorylate specific GPCRs which leads to receptor inactivation—perhaps directly or by arrestin activity [8] . In this complex signal transduction network , Gα and Gβγ subunits have different patterns of specificity for PLCβ isoforms and calcium is an important cofactor in several important feedback loops [10] . The two extracellular signaling ligands we consider here are C5a and UDP . The small peptide C5a is a potent anaphylotoxin and a strong chemoattractant for many immune system components [11] . The calcium response due to stimulation by C5a is predominantly coupled through Gαi-linked heterotrimers . Macrophage cells and their precursors , monocytes , express several receptors that are specific to extracellular nucleotides and it has been shown that the P2Y6 receptor , which is sensitive to UDP , regulates the production and secretion of the chemokine interleukin 8 ( IL-8 ) in monocytes [12] . The UDP response is mediated by Gαq-linked heterotrimers , but other receptors in the P2Y family may respond to UDP and couple the signal through other G protein isoforms [13] . Four recent models have sought to explore various aspects of the G protein coupled signal transduction system in detail . Lukas et al . compare measured calcium response over a range of bradykinin doses to their model predictions [14] . Mishra and Bhalla built a model to investigate the role of IP4 as a signal coincidence detector in the GPCR pathway [15] . The model by Lemon et al . predicts the calcium response to UTP stimulation and is the closest in focus to our model [16] . A recent model of calcium dynamics in RAW cells has been proposed that is quite similar to this model , but does not deal with crosstalk between receptors or formal statistical uncertainty in model predictions [17] , [18] . Several hypotheses for the mechanism of crosstalk and synergy among GPCR-mediated pathways have been proposed . Crosstalk among GPCR-mediated pathways is important both physiologically and pharmaceutically . Quitterer et al . propose that crosstalk is mediated by Gβγ exchange between Gαi-coupled and Gαq-coupled receptors [19] . Zhu et al . speculated that PLC is under either conditional or dual regulation of Gβγ and Gα [20] . Though these hypothetical mechanisms for crosstalk among G protein coupled receptor systems are conceptually plausible we have not found these or any other of the many competing hypothetical mechanisms tested in the context of a quantitative mathematical model [2] . In this paper Bayesian statistical inference is used to provide a rigorous connection between the mathematical model derived from mass-action kinetics , prior information from in-vitro biochemical studies and heterogeneous experimental data . The prior distribution over the parameters represents our uncertainty before observing a set of experimental data . A broad , high variance , prior distribution means we are quite uncertain and a concentrated , low variance , prior means we are more certain about the parameter a priori . The objective of our inference is the posterior distribution over the parameters because it is an informed estimate of both the value of the parameter and the uncertainty in the parameter value . The posterior distribution over the parameters is then used as a tool for experiment design to estimate the model-based posterior distribution over observable quantities such as the cytosolic calcium concentration and to drive the design of new experiments . This statistical approach is possible in a model of this size because of the abundance and quality of the data collected for this study .
Complement factor 5a activates the C5a receptor which is a Gαi-coupled receptor [28] . The released Gβγ dimer activates PLCβ2 and PLCβ3 which are lumped and called PLCβ3 in our model because: ( i ) the activity of Gβγ-activated PLCβ3 has been shown to be greater than Gβγ-activated PLCβ2 in in-vitro studies and ( ii ) Gαq activates both PLCβ2 and PLCβ3 so the structural connections from Gβγ and Gαq to PLCβ2 and PLCβ3 in the model are identical [4] , [29] . PLCβ1 is activated by Gβγ and Gαq , but RAW264 . 7 macrophage cells do not express this isoform , so we have not included it in the model . PLCβ3 then catalyzes the hydrolysis of phosphatidylinositol ( 4 , 5 ) -bisphosphate ( PIP2 ) into inositol 1 , 4 , 5-trisphosphate ( IP3 ) and diacylglycerol ( DAG ) . UDP stimulates the P2Y6 receptor and the associated Gαq-GTP activates both PLCβ3 [30] and PLCβ4 [31] . The GTPase rate of Gαq is increased 1000-fold when bound to PLCβ [5] . Due to this rapid hydrolysis rate , we have assumed , in our model , that PLCβ3 or PLCβ4 bound Gαq-GTP may only hydrolyze one molecule of PIP2 before releasing Gαq-GDP . Additionally , the Gβγ released by the P2Y6 receptor also activates PLCβ3 [30] , but does not activate PLCβ4 [32] . Our model assumes that PLCβ3 does not simultaneously bind Gβγ and Gαq . Indeed , a biochemical study of PLCβ2 activity in reconstituted membrane fractions strongly argues that Gαq and Gβγ do not simultaneously bind this effector [33] . While this was specifically demonstrated for PLCβ2 , we implicitly assume the same holds for PLCβ3 because we lump the two in our model . This is a mechanistic assumption of our model and an interesting issue for future testing with directed experiments . Though important for response specificity , the dynamical control of calcium release is not limited to the forward pathway in this system . Calcium participates in feedback processes that both enhance and inhibit its own release at multiple points in the pathway . There are four main nodes of calcium-dependent feedback control in our model: PLCβ , IP3 receptor , protein kinase C ( PKC ) and G protein receptor kinase ( GRK ) . Calcium enhances its own release by binding to the EF-hand domain on PLCβ and is required for PLCβ to hydrolyze PIP2 into IP3 and DAG [34] . Because the dissociation constant for PLCβ-Ca2+ in our model is larger than the basal concentration of cytosolic calcium , as more Ca2+ is released from the ER , more PLCβ-Ca2+ becomes available to bind Gαq or Gβγ . This positive feedback mechanism accelerates the release of Ca2+ . In our model , Ca2+ and IP3 cooperatively open the channel between the ER and the cytosol . It is believed that Ca2+ initially stimulates the IP3 receptor with maximal stimulatory effect at 100–300 nM [6] . At higher concentrations , Ca2+ has an inhibitory effect . We use the IP3 receptor model structure in the Keizer and DeYoung model for this component [25] . Protein kinase C ( PKC ) has been shown to phosphorylate PLCβ3 which inhibits PLCβ3 activation due to Gαq and Gβγ [35] , [36] . PKC is activated when bound to DAG and Ca2+ [7] , [37] . Because the preferred order of binding is not entirely known , PKC , DAG and Ca2+ form a thermodynamic cycle of reversible reaction with only the PKC-DAG-Ca2+ form active . In our model , the dissociation constant of PKC and Ca2+ is much greater than the basal Ca2+ concentration , and upon binding DAG , the PKC-DAG complex has a higher affinity for Ca2+ making the order of binding preferentially PKC to DAG then PKC-DAG to Ca2+ . It is not known whether PLCβ4 is also regulated by PKC . We have assumed , in our model , the same mechanism of PKC regulation of PLCβ3 and PLCβ4 . The final key calcium-dependent feedback loop in our model is mediated by G protein receptor kinase ( GRK ) . GRK2 phosphorylates and inactivates ligand-bound C5a receptors when activated by PKC and Gβγ . In sequence , PKC phosphorylates GRK2 which causes translocation to the plasma membrane [8] . When properly localized , GRK2 may bind Gβγ and then phosphorylate the C5a-C5a receptor complex to inactivate it [38] . This simplified representation of the receptor desensitization mechanism does not include arrestin activity , multiple receptor phosphorylation sites and other fine grain or slower biochemical interactions that may be present in-vivo . Having specified the structure of our model , we direct our attention to the parameters . We estimate 20 of the 84 parameters in our model using a dataset composed of 96 Fura-2 time series measurements as described in the Materials and Methods section . Each experiment consists of 3–4 samples from different wells in a 96 well plate . There are 15 experiments spanning 9 doses of C5a and 14 experiments spanning 11 doses of UDP on wild-type cells in the dataset ( see Figure S3 ) . The dataset also contains calcium measurements on 5 different shRNAi knockdown cell lines constructed by lentiviral infection ( see Figure S4 ) . The time interval between samples is approximately 3–4 seconds and each time series is approximately 100–300 seconds of post-stimulation data . Table 1 shows a summary of the knockdown data used for statistical parameter estimation for this model in addition to the wild-type experiments . We find that our model is generally quantitatively consistent with the experimental data within measurement uncertainty . Where the model is less consistent with the data – specifically for the GRK knockdown experiment – we find the deviation has a reasonable biological explanation . The summary of the dataset and the fit of the model to each single ligand experiment are available in the Supporting Information . We briefly discuss some issues relating to goodness of fit and the Bayesian parameter estimation here . While most optimization procedures produce a point estimate of the parameters that maximize the goodness of fit of the model to the observed data , the Bayesian procedure we have employed here estimates the entire posterior distribution of the parameters given the data . This information is valuable for qualitatively and quantitatively evaluating the precision of the parameters estimates . Figure 2 shows , as a qualitative evaluation , that while the a-priori forward and reverse binding rates for the receptors ( C5aR and P2YR ) are uncorrelated they are correlated in the posterior distribution . The calcium measurements have informed and constrained the posterior estimates of the dissociation constants to be approximately 5 nM and 250 nM for the C5aR and P2YR respectively . We have quantitatively computed marginal highest posterior density ( HPD ) confidence intervals for each of the twenty parameters we have estimated from the data . Those estimates are shown in Table S3 . Those parameters with large HPD intervals are not well informed by the measurements and are candidates for directed biochemical experiments . The calcium response to C5a adapts and returns to the basal level , but the UDP response has a sustained elevated calcium level that slowly decays . Figure 3 shows two representative experiments of the response of the wild-type cell to stimulation with C5a and UDP . We expect that the fit to this data will be good because 20 key model parameters were fit using an experimental dataset that included these experiments – the fit is indeed accurate . The point estimate curve is constructed from the maximum a-posteriori parameters from an MCMC chain . The prediction intervals are estimated by Monte Carlo sampling from the posterior parameter distribution and the measurement error distribution conditional on the parameters . The prediction confidence intervals generally cover the observed data . Lentiviral infection is used to introduce small hairpin RNAs to interfere with the translation of the key signaling proteins GRK2 , Gαi2 , Gαq , PLCβ3 and PLCβ4 [39] . There are three main sources of uncertainty in the knockdown experiment model predictions: parametric uncertainty , measurement uncertainty and knockdown efficiency uncertainty . We have dealt with the first two sources in the previous section on wild-type experiments . Here we address prediction variability due to knockdown efficiency uncertainty by using nominal parameter values . Figure 4 shows simulations and experimental data for three representative knockdown experiments . The upper-left panel of Figure 4 shows a GRK knockdown line stimulated with 250 nM C5a . Because GRK2 desensitizes the C5a receptor , we expect that by eliminating the feedback mechanism , the calcium peak will be higher and more sustained . The experimental data as well as the model indeed show that effect . Quantitatively , the model prediction shows a greater effect than the experimental data . A likely reason is that the model only considers one isoform of GRK while there are four isoforms expressed in the RAW264 . 7 cell line ( GRK1 , 2 , 4 , 6 ) . If more than one isoform can desensitize the C5a receptor , the effective knockdown in desensitization function will be less than as measured by western blot analysis on GRK2 . While GRK does not desensitize the P2Y receptor in our model , it is a buffer for Gβγ released from Gαq . Reducing the amount of GRK will shift the equilibrium towards more Gβγ bound to PLCβ3 and thus more calcium release even though GRK does not directly feed back on the P2Y6 receptor . The top-right panel in Figure 4 shows that , based on the model , the peak intracellular calcium concentration is expected to be very slightly higher in the GRK2 knockdown line when stimulated by 25 µM UDP . A comparison of the experimental peak heights of the wild-type and GRK knockdown cell line data by t-test cannot reject the null hypothesis that the peak heights are equal ( p = 0 . 9963 ) . The effect of the GRK knockdown is expected to be so slight that the effect size is overwhelmed by the measurement error in the data . The effect of the uncertainty in the GRK2 knockdown fraction impacts the range of the confidence intervals of the predicted C5a response much more than the confidence intervals of the predicted UDP response which is consistent with GRK2 being a more significant component of the C5a response . Our model structure has PLCβ3 stimulated by either Gβγ or Gαq . Because the C5a response signals only through PLCβ3 the effect of the knockdown is expected to be more pronounced for the C5a response than for the UDP response . The bottom-left panel of Figure 4 confirms that the model prediction is consistent with the representative experiment . The UDP response activates PLCβ3 through Gβγ , but also activates PLCβ3 and PLCβ4 with Gαq . Therefore , we expect that the calcium response should be more robust to perturbations in just one of the PLCβ isoforms . The UDP response in the PLCβ3 knockdown line ( bottom right panel of Figure 4 ) shows that our model predicts the knockdown effect to be small relative to the total magnitude of the response in part due to the redundancy in the use of PLCβ isoforms in the UDP response . Because this dataset was used for parameter estimation , the fit of model to the data may overstate the accuracy of the model . Nonetheless , the good fit does suggest that the model warrants being tested in truly predictive experiments; we describe such experiments in the following section . We examine our model response to a simultaneous stimulation by C5a and UDP because it has been shown experimentally that macrophage cells respond synergistically to such conditions [40] . To quantify the amount of synergy or non-additivity that is present in the calcium response , a synergy ratio is computed for each ligand dose pair . The numerator of the ratio is the peak offset from baseline of the intracellular calcium concentration . The denominator of the ratio is the sum of the peak offsets when the cell or model is stimulated with only one ligand . A synergy is present when the ratio is greater than one implying the peak height is greater than expected from an additive combination of ligand effects . While this is certainly not the only possible measure of synergy it is widely adopted and has been used in previous studies on calcium synergy [40] . The left panel of Figure 5 shows the results of model simulations at nominal parameters for a grid of doses of C5a and UDP . In the dose response surface , there is a ridge of synergistic calcium release for a moderate dose of UDP . We tested the model prediction with the experiment design measuring the synergy ratio at the points denoted as black open circles in the left panel of Figure 5 . A χ2 goodness-of fit test comparing the model expected synergy ratio to the observed synergy ratio fails to reject the null hypothesis that the data were generated by the model mechanism ( p-value≈1 . 0 ) . The root-mean-squared error ( RMSE ) deviation between the predicted and actual experimental data is 0 . 492 . By way of comparison , the RMSE between the data and the null model of no synergy is 1 . 044 . We therefore conclude that the model predictions are consistent with the experimental observations . It should be noted that measurements of synergy in RAW cells are noisy and the ridge occurs at low doses of UDP . Notwithstanding , the phenomenon has been reported [40] and has been observed by us in this cell line . The right panel of Figure 5 shows the same synergy dose response surface but for a GRK knockdown cell line . The synergy ridge observed in the wild-type cell simulation is changed in the GRK knockdown simulation indicating the C5a receptor desensitization mechanism mediated by GRK is important for the synergistic release of calcium . In the next section we pursue this conclusion in more detail , developing a conceptual explanation of the mechanism of crosstalk and synergy within our model .
G-protein-coupled receptors form a complex network of interacting proteins that generally exhibits the properties of a system in which each receptor signal is buffered from the others . For a minority of ligand combinations , however , crosstalk between pairs of receptors is apparent . Due to the complexity and importance of the system many hypothetical mechanisms have been proposed to explain the crosstalk [2] . In particular , simultaneous Gβγ and Gαq binding to PLCβ [20] and Gβγ exchange between Gαi and Gαq-coupled receptors have been proposed as potential mechanisms [19] . While our model does not eliminate these potential mechanisms , we do show that the mechanism represented in our model is consistent with a full range of experimental data including a variety of doses of C5a and UDP , C5a and UDP stimulation of five different knockdown cell-lines and double-ligand dose response experiments . To our knowledge , this is the first multireceptor GPCR model and the first to address the complex phenomenon of crosstalk between GPCR receptor pathways that has been statistically estimated and validated with experimental data . This important phenomenon plays a role in processes as diverse as chemotaxis and perhaps drug interactions . In our model , the primary mechanism of synergy is due to the cooperative opening of the IP3 receptor . The robustness of the synergy is due to the feedback of GRK on the C5a receptor and the specificity of the synergy is due to the interaction patterns between specific Gα isoforms and PLCβ isoforms . The simultaneous binding model [20] accounts for the specificity of synergy , but not the robustness pattern of the synergy . We observe in the model that if the Gαq-PLCβ3-Ca2+ and Gαq-PLCβ4-Ca2+ binding reactions are inhibited , the system still exhibits synergy . We conclude from this observation that the crosstalk mechanism is mediated by Gβγ . If the binding reaction of Gβγ to phosphorylated GRK2 is removed , the synergy is eliminated . Furthermore , if the GRK2-mediated phosphorylation of complexed C5a receptors is removed , the double ligand response is additive . We deduce then that the synergy mechanism involves GRK2 phosphorylation of complexed C5a receptors . However , GRK2 phosphorylation does not entirely explain the synergy mechanism . In our model , the calcium released from the IP3 receptor is a function of the number of receptor molecules complexed to IP3 raised to the fourth power [41] . Therefore , for a small range of IP3 concentration , the amount of Ca2+ released is more than additive ( see Figure S8 ) . We conclude from our analysis of the model that the synergy ridge in Figure 5 arises because the GRK2 mediated mechanism holds the IP3 concentration in this non-additive region for most concentrations of C5a . The UDP response does not have the GRK2 mediated feedback and thus only shows a synergistic response for a small range of UDP concentration . If the GRK2 desensitization is removed from the model , the synergy ridge is removed and synergy is only present at low doses of C5a and UDP ( see Figure 5 ) . The Bayesian method we have used for this model has several advantages for the estimation of model parameters in complex mechanistic system models . We have used an informative prior to exclude negative rate constants from the permitted parameter space . We have also used the prior distribution to center our a priori expectations of the rate constant at values obtained from in-vitro and other biochemical experiments . The Bayesian update rule allowed us to estimate parameters with our best current dataset and then update those estimates as new data became available from the calcium assay . In this way , we were able to iteratively refine and recalibrate our model with the most recent data available during data collection period for this project . The posterior distribution provides not only an estimate of the rate constants , but the entire distribution , from which we can calculate highest posterior confidence intervals and posterior correlations between parameters . For example , the posterior correlation between the binding and unbinding rates for the UDP-P2Y receptor complex were highly correlated , but those two constants were uncorrelated with the corresponding rates for the C5a-C5a receptor complex reaction even though we imposed no correlations a priori . Finally , the algorithmic methods for collecting ensembles of samples from the posterior distribution have improved considerably in recent years in terms of speed and robustness We have shown that the signal transduction system as it is represented by our model does not require simultaneous binding of Gαq and Gβγ to PLCβ3 to cause a synergistic Ca2+ response due to simultaneous stimulation by C5a and UDP . We have shown that our representative model is consistent with this experimental dataset in RAW264 . 7 macrophage cells , but we have not excluded all other potential mechanisms that may be absent or regulated differently in this cell line compared to other macrophage cell lines . Indeed there are a few examples of statistical discrepancies between the model and experiments in our dataset ( Table S4 ) . These differences are substrate for further experimentation and modeling . The purpose of our model is to provide a quantitative tool to aid in reasoning about such complex interacting systems so that meaningful experiments can be designed to explore and understand the biological mechanism .
Intracellular free calcium in cultured adherent RAW264 . 7 cells was measured in a 96-well plate format using the Ca2+-sensitive fluorescent dye Fura-2 [43] , [44] . A Molecular Devices FLEXstation scanning fluorometer was used to measure fluorescence using a bottom read of a 96-well plate . Each well was sampled approximately every 4 seconds . The measurement protocol is described in AfCS experimental protocol ID #PP00000211 ( available from http://www . signaling-gateway . org ) . The parameters in ligand concentration model were estimated using FITC solution in the FLEXstation scanning fluorometer as described in Molecular Devices Maxline Application Note #45 and in Protocol S1 ( see also Figure S5 ) . Twenty of the 84 parameters were chosen to be estimated from data based on relevance to the experimental hypothesis . Only those parameters that related to the knockdown experiments in the dataset were estimated and are denoted with a star in Table S2 . We used data to estimate only the two forward rate constants in the enzymatic mass-action equations because the forward and reverse rate constants for a given reaction will be highly correlated in the posterior distribution making estimation by Markov chain methods computationally expensive . An analysis of the sensitivity of the model to each parameter is shown in Figure S9 . For each estimated parameter we constructed an independent Gaussian prior on a log scale with a mean chosen based on relevant literature and a standard deviation of 0 . 25 . We found that this prior variance was sufficiently permissive to allow exploration of the space while still constraining the rates to be physically reasonable . The prior distribution over the parameters allows the incorporation of both soft and hard constraints in the parameter estimates . Parameter sets with zero measure are not permitted in the posterior distribution and parameter sets with small measure must be assigned a large likelihood in order to have a large posterior probability . The likelihood is a function of the parameters ( θ ) and links the prior distribution with the posterior distribution under Bayes rulewhere y denotes the observed data . In our model , the likelihood function is a Gaussian distribution according to the non-linear regression equation y = f ( θ ) +ε , ε∼N ( 0 , σ2 ) , where f ( θ ) is the deterministic model prediction . The posterior distribution is of interest because it informs us as to the most probable setting of the parameters as well as the uncertainty in the values . The Metropolis-Hastings algorithm [45] was used to estimate the posterior density of the parameters Pr ( θ|y ) . Three independent chains were simulated from different initial parameter values ( see Figure S1 ) . To assess convergence of the posterior distribution estimate , we used the Gelman-Rubin potential scale reduction factor ( PSRF ) [46] . The multivariate PSRF is 2 . 44 and 95% of the individual PSRFs were less than 1 . 5 . A PSRF value of one indicates that the distribution has converged and values near one are close to converged . Posterior prediction confidence intervals were constructed using the percentiles from the predictive distribution approximated with 2000 Monte Carlo samples from Pr ( ynew|θi ) at each of 100 simple random samples from Pr ( θ|y ) obtained fromwhere Pr ( ynew|θi ) ∼N ( f ( θ ) , s2 ) and s2 is the pooled variance estimate , which is computed as an average of the variances of all the time points in each of the 29 wild-type experiments . These average variances were weighted by the number of technical replicates in each experiment and then averaged to yield the estimate s2 . A small factor of 1 nM2 was added to each variance estimate to bound variance estimates away from zero .
|
The G protein signal transduction system transmits a wide variety of extracellular signals including light , odors , and hormones , to intracellular effectors in diverse cell types in eukaryotes . G-protein-coupled receptors are involved in many diseases including inflammation , cardiac dysfunction , and diabetes , and are the targets of 40–50% of modern drugs . Despite the physiological and pharmacological importance of this signal transduction system it is not known how the system buffers and integrates information at a biochemical level . The multiple receptors expressed by every cell pass their signals through a common set of downstream effectors distinguished by multiple isoforms with slightly different specificities and activities . The coupling among these pathways causes interactions among the signals sent by the different classes of receptors . We have developed a mechanistic model of the G protein signal transduction system from the receptor to the central intracellular second-messenger calcium . We have used statistical methods to integrate a diverse set of experimental data into our model and quantify confidence in our model predictions . We used this model , trained on single receptor data , to predict the signal processing of two G-protein-coupled-receptor signals . Validation experiments support our hypothesized mechanism for dual receptor signal processing and the predictions of the model .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] |
[
"cell",
"biology/cell",
"signaling",
"biophysics/theory",
"and",
"simulation",
"mathematics/statistics",
"biophysics/cell",
"signaling",
"and",
"trafficking",
"structures",
"computational",
"biology/systems",
"biology"
] |
2008
|
A Dual Receptor Crosstalk Model of G-Protein-Coupled Signal Transduction
|
Attenuated viral vaccines can be generated by targeting essential pathogenicity factors . We report here the rational design of an attenuated recombinant coronavirus vaccine based on a deletion in the coding sequence of the non-structural protein 1 ( nsp1 ) . In cell culture , nsp1 of mouse hepatitis virus ( MHV ) , like its SARS-coronavirus homolog , strongly reduced cellular gene expression . The effect of nsp1 on MHV replication in vitro and in vivo was analyzed using a recombinant MHV encoding a deletion in the nsp1-coding sequence . The recombinant MHV nsp1 mutant grew normally in tissue culture , but was severely attenuated in vivo . Replication and spread of the nsp1 mutant virus was restored almost to wild-type levels in type I interferon ( IFN ) receptor-deficient mice , indicating that nsp1 interferes efficiently with the type I IFN system . Importantly , replication of nsp1 mutant virus in professional antigen-presenting cells such as conventional dendritic cells and macrophages , and induction of type I IFN in plasmacytoid dendritic cells , was not impaired . Furthermore , even low doses of nsp1 mutant MHV elicited potent cytotoxic T cell responses and protected mice against homologous and heterologous virus challenge . Taken together , the presented attenuation strategy provides a paradigm for the development of highly efficient coronavirus vaccines .
Coronaviruses are vertebrate pathogens mainly associated with respiratory and enteric diseases [1] . They can cause severe diseases in livestock animals and lead thereby to high economic losses . In humans , coronavirus infections manifest usually as mild respiratory tract disease ( common cold ) that may cause more severe symptoms in elderly or immune-compromised individuals [2 , 3] . In 2002–2003 , the appearance of severe acute respiratory syndrome ( SARS ) , caused by a formerly unknown coronavirus ( SARS-CoV ) , exemplified the potential of coronaviruses to seriously affect human health [4–7] . The frequent detection of SARS-like coronaviruses in horseshoe bats ( Rhinolophus sp . ) and the broad range of mammalian hosts that are susceptible to SARS-CoV infection may facilitate a potential reintroduction into the human population [8] . Therefore , the development of efficacious coronavirus vaccines is of high medical and veterinary importance . Effective vaccines controlling virus spread and disease are available for a number of infections , such as smallpox , poliomyelitis , measles , mumps , rubella , influenza , hepatitis A , and hepatitis B [9 , 10] . Some of these vaccines consist of virus subunits or inactivated virus preparations that mainly induce the production of pathogen-specific antibodies . In contrast , live attenuated vaccines consist of replication-competent viruses that induce broad cellular and humoral immune responses without causing disease [10] . The most prominent live attenuated vaccines are vaccinia virus [11] , poliovirus [12] , and yellow fever virus ( YF-17D ) [13] . Despite their documented efficacy , it is still not fully understood why and how successful vaccines work [10 , 14] . However , recent concepts in immunology provide a link between innate and adaptive immune responses and suggest that the quality , quantity , and longevity of adaptive immune responses is determined very early after infection or vaccination [14] . Of major importance are professional antigen-presenting cells ( pAPCs ) such as dendritic cells ( DCs ) and macrophages , which play a major role in ( i ) sensing pathogen-associated molecular patterns , ( ii ) inducing innate immune responses , and ( iii ) shaping the upcoming adaptive immune response . Efficient live attenuated vaccines should therefore not only lack significant pathogenicity , but should also deliver antigens to pAPCs and activate the innate immune system . Notably , the majority of currently available attenuated vaccines have been derived empirically . Given the recent proceedings in the areas of virus reverse genetics and virus–host interactions , the time should be ripe for more rational approaches in vaccine development . An attractive strategy is to target virally encoded pathogenicity factors , such as interferon ( IFN ) antagonists [15] , to attenuate virulence while retaining immunogenicity . This concept has been proposed for the generation of live attenuated influenza virus vaccines encoding altered NS1 proteins [16 , 17] . Our rudimentary knowledge on coronavirus-encoded pathogenicity factors is reflected by the fact that only a few putative coronaviral pathogenicity factors have been identified and that functional analyses are still limited to the description of in vitro effects [18–20] . For a number of reasons , the non-structural protein 1 ( nsp1 ) is of particular interest in this context . First , coronaviruses are positive-stranded RNA viruses , and the replicase-encoded nsps are expressed from the viral genomic RNA immediately after virus entry by translation of two large polyproteins . nsp1 is encoded at the 5′ end of the replicase gene and is therefore the first mature viral protein expressed in the host cell cytoplasm [21] . Second , a recent in vitro study suggests that SARS-CoV nsp1 may be associated with host cell mRNA degradation and may counteract innate immune responses [18] . Finally , nsp1 is encoded by all mammalian coronaviruses known to date ( coronavirus groups 1 , 2a , and 2b ) [22] , and recent structural data on SARS-CoV ( group 2b ) nsp1 suggest functional similarities to mouse hepatitis virus ( MHV; group 2a ) nsp1 [23] . Using a reverse genetics approach , we show here that nsp1 is a major pathogenicity factor . Recombinant MHV mutants encoding a deletion in nsp1 replicated as efficiently as wild-type virus in cell culture , but displayed an unprecedented degree of attenuation in mice . Interference with the type I IFN system appears to be the dominant mode of action of murine coronavirus nsp1 . Vaccination with the nsp1 mutant virus elicited efficient memory cytotoxic T cell responses and protected against homologous and heterologous virus infections . Our study will pave the way for the generation of novel coronavirus vaccines based on modified coronavirus replicase genes .
We assessed several replicase-encoded nsps of MHV ( strain A59 ) , SARS-CoV , and human coronavirus 229E ( HCoV-229E ) for their ability to interfere with host cell gene expression . Using transient gene expression studies , we found that MHV-A59 , SARS-CoV , and HCoV-229E nsp1 significantly reduced luciferase reporter gene expression under the control of IFN-β , IFN-stimulated response element ( ISRE ) , and SV40 promoters ( Figure 1 ) . This is consistent with a recent report suggesting that SARS-CoV nsp1 induces general host cell mRNA degradation [18] . Nevertheless , it should be noted that the nsp1-mediated reduction in reporter protein expression appeared more robust for ISRE and SV40 than for IFN-β promoter-driven expression . Our data also support the hypothesis that MHV encodes a SARS-CoV nsp1 homolog that displays the same function [23] . Although comparative sequence analyses suggest that nsp1 of group 2a/2b coronaviruses ( e . g . , MHV and SARS-CoV , respectively ) and the nsp1 of group 1 coronaviruses ( e . g . , HCoV-229E ) may belong to different protein families [22 , 23] , we also observed reduced reporter gene expression in HCoV-229E nsp1-transfected cells ( Figure 1 ) . Whether functional similarities may exist between nsp1 molecules encoded by coronaviruses of different phylogenetic lineages remains to be established in future studies . Importantly , our data also revealed that reporter gene expression from all tested promoters was not affected when C-terminally truncated MHV nsp1 molecules were tested ( Figure 1 ) . To assess the role of nsp1 in the context of virus replication , we constructed a recombinant MHV encoding a truncated nsp1 protein using our reverse genetic system [24] . Based on the results shown in Figure 1 , we decided to delete MHV nucleotides ( nts ) 829–927 ( 99 nts ) . In the resulting mutant virus , MHV-nsp1Δ99 , the replicase gene start codon , the translational reading frame , and the residues required for proteolytic release of nsp1 from the replicase polyprotein were maintained ( Figure 2A ) . As reported for a set of similar MHV mutants by Brockway et al . [25] , viral growth and peak titers of MHV-nsp1Δ99 in murine 17Clone1 cells were indistinguishable from that of wild-type virus ( Figure 2B ) . To assess the stability of the recombinant MHV-nsp1Δ99 , we analyzed the nsp1-coding region by RT-PCR sequencing after seven passages in tissue culture and no nucleotide changes were detected ( unpublished data ) . Infection of conventional DCs ( cDCs ) is an early and crucial event for the generation of protective antiviral immunity [26] . MHV productively infects cDCs and activates plasmacytoid DCs ( pDCs ) to generate a first wave of protective type I IFN [27] . To assess whether the mutant MHV-nsp1Δ99 has retained the ability to infect pAPCs , peritoneal macrophages ( Figure 2C ) , bone marrow–derived CD11c+ cDCs ( Figure 2D ) , and splenic , FACS-sorted CD11c+ cDCs ( Figure 2E ) were exposed to MHV-nsp1Δ99 and wild-type control virus . Similar to replication kinetics in cell lines ( Figure 2B ) , MHV-nsp1Δ99 showed no significant growth defect in primary pAPCs ( Figure 2C–2E ) , indicating that the deletion of nsp1 did not alter the pronounced tropism of MHV for cDCs and macrophages . MHV-A59 is a hepatotropic and neurotropic virus that can cause acute hepatitis and encephalitis . Following intraperitoneal infection , virus replication is first detectable in spleen and liver , followed by virus spread to other organs , including the central nervous system . Hepatitis is the first clinical sign of disease , accompanied by elevated liver enzyme values in serum . Associated with the appearance of cytotoxic T cell responses approximately at day 5 post infection ( p . i . ) , virus titers usually decline and are no longer detectable after day 7 p . i . Infections with a high dose ( ≥ 5 × 106 pfu , intraperitoneal ) may , however , occasionally result in fatality . To evaluate the importance of nsp1 for virus replication and viral pathogenicity in vivo , C57BL/6 mice were infected intraperitoneally with different doses of wild-type MHV or MHV-nsp1Δ99 . Both viruses replicated in the spleen , whereby MHV-nsp1Δ99 titers were consistently lower than wild-type virus titers ( Figure 3A ) . Furthermore , MHV-nsp1Δ99 was rapidly cleared and not detectable after day 2 p . i . ( Figure 3A ) . Wild-type , but not mutant virus , was detectable in the liver at low and intermediate dose ( 50 pfu and 5 , 000 pfu , respectively ) ( Figure 3B ) . When high virus doses ( 5 × 106 pfu ) were applied , MHV-nsp1Δ99 eventually reached the liver at day 2 p . i . , but was not detectable at later time points ( Figure 3B ) . MHV-nsp1Δ99 was not detectable in other non-hematopoietic organs , such as lung and central nervous system ( unpublished data ) . Mice infected with wild-type virus showed acute liver disease with elevated liver enzyme values in serum . Furthermore , after high dose infection with wild-type virus ( 5 × 106 pfu ) , a significant weight loss that peaked at approximately 10%–15% at day 4 was observed ( Figure 3E ) . In contrast , mice infected with the nsp1 mutant virus remained healthy after low , intermediate , or high dose infections . Even at the highest dose applied ( 5 × 106 pfu ) , MHV-nsp1Δ99–infected mice did not lose weight ( Figure 3E ) , and no elevated liver enzyme values were detected in the serum ( Figure 3C ) . This observation correlated well with the absence of hepatocyte necrosis and parenchymal inflammation following MHV-nsp1Δ99 infection ( Figure 3D ) . To further assess the attenuation of the MHV nsp1 mutant , mice were infected intracranially with 200 pfu and 20 , 000 pfu of MHV-nsp1Δ99 or MHV-A59 . All mice infected with 200 pfu survived for at least 30 d ( unpublished data ) . Mice infected with 20 , 000 pfu of MHV-A59 succumbed to the infection , whereas mice infected with 20 , 000 pfu of MHV-nsp1Δ99 survived and showed no signs of clinical disease ( Figure 3F ) . Collectively , these data demonstrate that MHV-nsp1Δ99 is strongly attenuated in vivo , but has retained the ability to replicate in secondary lymphoid organs , such as the spleen . We have previously shown that pDCs are the major source of IFN-α in the early stages of MHV infection and that type I IFN responses in CD11c+ cDCs are only weakly triggered by MHV [27] . To test whether nsp1 has an influence on the induction of IFN-α , we infected both cDCs and pDCs with MHV-nsp1Δ99 or wild-type MHV . Both viruses elicited rapid and high IFN-α production in Flt3-L–differentiated bone marrow–derived pDCs ( Figure 4A ) and FACS-sorted primary pDCs ( Figure 4B ) . Furthermore , both wild-type and mutant MHV elicited only a late and weak IFN-α production in cDCs ( Figure 4A and 4B ) . These results suggest that nsp1 does not affect the induction of type I IFN . To assess a potential impact of nsp1 on type I IFN signaling and antiviral effector mechanisms in target cells that efficiently support MHV replication , cDCs and macrophages were pretreated with different dosages of IFN-α and infected with MHV-nsp1Δ99 or wild-type MHV . In cDCs , IFN-α treatment had a comparable effect on the replication of both MHV-nsp1Δ99 and the wild-type control virus ( Figure 4C ) . However , replication of MHV-nsp1Δ99 was , in a dose-dependent manner , more vulnerable to IFN-α treatment in macrophages ( Figure 4D ) , suggesting that nsp1 might counteract IFN signaling and/or the antiviral activities of IFN-induced effector proteins . Indeed , in vivo experiments in type I IFN receptor-deficient ( IFNAR−/− ) mice [28] strongly support this interpretation . Infection of IFNAR−/− mice with wild-type MHV led to high titers in all tested organs ( Figure 5A–5D ) , indicating that signals transmitted via the IFNAR are crucial for preventing uncontrolled spread of the virus [27] . Surprisingly , the severe attenuation of MHV-nsp1Δ99 in wild-type 129Sv mice was not present in IFNAR−/− mice ( Figure 5A–5D ) . Replication of MHV-nsp1Δ99 in IFNAR−/− mice was largely restored and virus titers reached about 104–105 pfu/g tissue in several organs after only 36 h ( figures 5A–5D ) . These data strongly suggest that nsp1 has a pivotal role in counteracting type I IFN host responses and provide an explanation for the rapid clearance of MHV-nsp1Δ99 in wild-type mice . Interestingly , liver damage , measured as alanine 2-oxoglutarate-aminotransferase ( ALT ) levels in serum , was not yet detectable in MHV-nsp1Δ99–infected IFNAR−/− mice at 36 h p . i . ( Figure 5E ) . At 72 h p . i . , MHV-nsp1Δ99 reached titers and ALT levels in IFNAR−/− mice comparable to those observed in MHV-A59–infected IFNAR−/− mice at 36 h p . i . , demonstrating that MHV-nsp1Δ99 replication in IFNAR−/− is , although with slower kinetics , restored . The phenotypic analysis of MHV-nsp1Δ99 revealed a number of features that are advantageous for live attenuated vaccines . MHV-nsp1Δ99 grows to high titers in cell culture , infects pAPCs , replicates almost exclusively in secondary lymphoid organs , and is strongly attenuated in vivo . To assess the potential of MHV-nsp1Δ99 as an attenuated live vaccine , we replaced accessory gene 4 of MHV-nsp1Δ99 and wild-type MHV-A59 by a gene encoding a fusion protein of the immunodominant cytotoxic T lymphocyte ( CTL ) epitope ( KAVYNFATC ) of the lymphocytic choriomeningitis virus ( LCMV ) and the enhanced green fluorescent protein ( GP33-GFP ) [29] . The resulting recombinant viruses , MHV-nsp1Δ99-GP33-GFP and MHV-GP33-GFP , were used to infect C57BL/6 mice with different doses ( 50 and 5 , 000 pfu , intraperitoneal ) , and CD8+ T cell responses were assessed using flow cytometry–based detection of intracellular IFN-γ following antigen-specific short-term in vitro restimulation . As shown in Figure 6A and 6B , infection with as few as 50 pfu of MHV-nsp1Δ99-GP33-GFP elicited strong CD8+ T cell responses against both the H2-Db–restricted GP33 and the H2-Kb–restricted MHV S598 epitope . To assess the level of protection against homologous MHV-A59 challenge , groups of C57BL/6 mice were immunized ( 5 , 000 pfu ) with MHV-nsp1Δ99-GP33-GFP , MHV-GP33-GFP , or treated with PBS . Sixteen days p . i . , mice were challenged with wild-type MHV ( 5 , 000 pfu ) and viral titers were determined 5 d post challenge infection . Viral titers were below the limit of detection in MHV-nsp1Δ99-GP33-GFP– and MHV-GP33-GFP–immunized mice ( Figure 6C ) . Together with the absence of elevated liver enzyme values in immunized mice ( Figure 6D ) , these data indicate that vaccination with the attenuated MHV nsp1 mutant provides complete protection against homologous virus challenge . The reverse genetic system facilitates incorporation of antigens derived from other infectious organisms . In order to determine whether the attenuated nsp1 mutant virus could confer protection against heterologous virus infection , MHV-nsp1Δ99-GP33-GFP–immunized C57BL/6 mice were challenged after 4 wk with LCMV ( 200 pfu , intravenous ) . LCMV titers in spleens were significantly reduced both in mice vaccinated with MHV-GP33-GFP and the attenuated MHV-nsp1Δ99-GP33-GFP virus ( Figure 6E ) . Remarkably , only 50 pfu of nsp1 mutant virus expressing the GP33 epitope were sufficient to achieve a reduction of LCMV titers by more than 4 orders of magnitude , indicating that nsp1 mutant viruses are well-suited to serve as attenuated recombinant virus vectors against heterologous viral infections .
The rational design of live attenuated viral vaccines is greatly facilitated by the identification and targeting of pathogenicity factors . This study demonstrates an unprecedented level of attenuation of a murine coronavirus through a 99-nt deletion in nsp1 . The nsp1 mutant virus was rapidly cleared in mice and did not induce clinical signs of disease in immunocompetent mice . These findings in the murine coronavirus model demonstrate that nsp1 is a major pathogenicity factor . In a stepwise approach , we made use of these observations to provide a blueprint for the construction and evaluation of live attenuated coronavirus vaccines encoding a truncated nsp1 . The presented results indicate that nsp1 plays a crucial role in the MHV life cycle by interfering with host innate immune responses . In accordance with the recent report by Kamitani et al . [18] , we observed reduced reporter gene expression in transient nsp1 expression studies . The suggestion that SARS-CoV nsp1 may play a role in SARS-CoV pathogenesis by promoting host cell mRNA degradation [18] has now received support through the analysis of a coronavirus nsp1 mutant in a murine model . The MHV nsp1 mutant phenotype led us to conclude that nsp1 mainly affects IFN signaling pathways or other downstream events . The influence on IFN-α induction appears to be limited . These conclusions are based on several observations . First , the analysis of IFN-α production by pDCs and cDCs revealed no significant differences between wild-type and nsp1 mutant virus infections . Second , treatment of macrophages with IFN-α revealed a more efficient reduction of MHV-nsp1Δ99 replication compared with that of wild-type MHV . Finally , and most strikingly , IFNAR−/− mice were highly permissive for the mutant virus , and organ titers almost reached those of wild-type MHV-infected IFNAR−/− mice . Nevertheless , it should be noted that the nsp1 mutant virus replication was still slightly delayed in IFNAR−/− mice . Therefore , further studies are required to define molecular target ( s ) and the precise function ( s ) of coronavirus nsp1 . Likewise , further studies are required to assess the impact of other coronaviral gene products on coronavirus pathogenicity . Recent reports indicate that coronaviruses most likely express a number of proteins , such as MHV and SARS-CoV nucleocapsid proteins , and SARS-CoV ORF3b and ORF6 proteins , that may interact with innate immune responses [19 , 20] . Also , the coronavirus replicase gene may harbor additional functions that play a role in virus–host interactions . It has been shown that the MHV and SARS-CoV nsp2 proteins [30] , and the highly conserved ADP-ribose-1′′-monophosphatase activity [31] encoded in nsp3 , are both dispensable for virus replication in tissue culture , and that a single point mutation in the MHV nsp14 confers a strong attenuation of MHV in mice [32] . Clearly , the murine model , with MHV as a natural mouse pathogen , will be highly advantageous in the examination of this issue , because it allows the use of well-characterized inbred and transgenic mice in combination with well-established immunological techniques required to assess the full range of coronavirus–host interactions . The most remarkable finding of this study is the level of attenuation of the nsp1 mutant virus and its restricted replication in secondary lymphoid organs . It may well be that other coronaviral nsp1 molecules exert similar functions as the MHV nsp1 . The coronavirus nsp1 has been suggested as a group-specific marker to differentiate group 1 coronaviruses from group 2a/2b coronaviruses [22] . Our transient nsp1 expression data indeed support the notion that SARS-CoV and MHV may encode evolutionarily conserved nsp1 homologs [22 , 23] . Nevertheless , further in vivo studies are required to determine whether the group 2b SARS-CoV nsp1 is indeed a functional equivalent to the structurally highly conserved group 2a nsp1 molecules encoded by MHV , bovine coronavirus , porcine hemagglutinating encephalomyocarditis virus , HCoV-OC43 , and HCoV-HKU1 . Likewise , it will be important to clarify in vivo , whether , despite the apparent lack of any sequence homology [22 , 23] , the nsp1 of group 1 coronaviruses ( e . g . , HCoV-229E ) may represent a functional correlate to the nsp1 of group 2a/2b coronaviruses . Recent progress in the establishment of suitable mouse models for SARS-CoV [33–35] and HCoV-229E [36] will enable researchers to address these questions in future studies . The chosen attenuation strategy has resulted in the generation of a recombinant virus that fulfills important criteria of a live virus vaccine candidate: ( i ) growth to high titers in cell culture , which facilitates vaccine production , and ( ii ) generation of immunological memory that mediates efficient protection against viral challenge . One important aspect of protection against viral infections is the induction of specific cytotoxic T cells by pAPCs in secondary lymphoid organs [14] . A number of coronaviruses , such as MHV , HCoV-229E , feline infectious peritonitis virus , and SARS-CoV , have been shown to infect pAPCs and to replicate in the secondary lymphoid organs [27 , 37–42] . Because of their pronounced tropism for pAPCs and the induction of strong CTL responses , we propose that coronaviruses represent promising vectors for the expression of heterologous antigens . The identification of nsp1 as a major pathogenicity factor will significantly increase the safety of coronavirus-based vectors [43] . For example , the deletion of accessory genes ( i . e . , not replicase or structural genes ) has been described for some coronaviruses to confer attenuation in the natural host [44–46] , and the deletion of the structural envelope protein E has resulted in the development of replication-competent , but propagation-deficient , coronavirus vectors [47 , 48] . Now , with an accompanying deletion in the nsp1-coding sequence , such vectors would be considered “recombination proof” , because the deletions are located at opposite genomic regions ( i . e . , within the replicase gene at the 5′ end and within the structural gene region at the 3′ end of the coronavirus genome ) , which make the reconstruction of virulent viruses by recombination unlikely . We therefore suggest that accessory gene , E gene , and partial nsp1 gene deletions will result in particular safe vectors with the potential to express multiple heterologous antigens [40 , 49] . Taken together , our results describe a novel type of coronavirus vaccines based on impaired function of a replicase gene product . We expect that our approach is applicable to most , if not all , mammalian coronaviruses and that it will enable the development of long-desired live attenuated vaccines for important coronavirus-induced diseases in humans and animals .
C57BL/6 mice were obtained from Charles River Laboratories ( http://www . criver . com/ ) . 129Sv and type I IFN receptor-deficient mice ( IFNAR−/− ) [28] were obtained from the Institut für Labortierkunde ( University of Zürich ) and bred in our facilities . All mice were maintained in individually ventilated cages and were used between 6 and 9 wk of age . All animal experiments were performed in accordance with the Swiss Federal legislation on animal protection . MC57 , BHK-21 , L929 , 293 , and CV-1 cells were purchased from the European Collection of Cell Cultures ( http://www . ecacc . org . uk/ ) . D980R cells were a kind gift from G . L . Smith , Imperial College , London , United Kingdom . 17Clone1 cells were a kind gift from S . G . Sawicki , Medical University of Ohio , Toledo , Ohio , United States . BHK-MHV-N cells , expressing the MHV-A59 nucleocapsid protein under the control of the TET/ON system ( Clontech , http://www . clontech . com/ ) , have been described previously [24] . All cells were maintained in minimal essential medium supplemented with fetal bovine serum ( 5%–10% ) and antibiotics . Murine cDCs and pDCs were obtained from spleens of C57BL/6 , 129Sv , or IFNAR−/− mice following digestion with collagenase type II ( Gibco-BRL , http://www . invitrogen . com/ ) for 20 min at 37 °C . Cells were resuspended in PBS supplemented with 2% FCS and 2 mM EDTA and overlaid on 20% Optiprep density gradient medium ( Sigma-Aldrich , http://www . sigmaaldrich . com/ ) . After centrifugation at 700g for 15 min , low density cells were depleted of CD3- and CD19-positive cells using DYNAL magnetic beads according to the instructions of the manufacturer ( Invitrogen , http://www . invitrogen . com/ ) . The DC-enriched preparations were stained with α-PDCA-1 , α-CD11b , and α-CD11c , and the distinct pDC and cDC populations were sorted using a FACS ARIA ( BD Biosciences , http://www . bdbiosciences . com/ ) sorter . Purity of both cell preparations was always >98% . Murine bone marrow–derived cDCs or pDCs were generated by 6 to 7 d of culture with either granulocyte-monocyte colony stimulating factor ( GM-CSF ) -containing supernatant from the cell line X63-GM-CSF ( kindly provided by Antonius Rolink , University of Basel , Basel , Switzerland ) or Flt3-L ( R&D Systems , http://www . rndsystems . com/ ) at 20 ng/ml , respectively . Bone marrow–derived cDCs were further purified using Optiprep density gradient centrifugation . Bone marrow–derived pDCs were purified using the mouse pDC isolation kit ( Miltenyi Biotec , http://www . miltenyibiotec . com/ ) adapted for the isolation of in vitro–derived pDCs by adding CD11b-biotin to the negative selection cocktail . Antibodies used in this study were purchased from BioLegend ( http://www . biolegend . com/ ) : CD11c-PE , B220-APC , CD11b-FITC; or from Miltenyi Biotec: mPDCA-1-FITC and CD11c-APC . Thioglycolate-elicited macrophages were collected from the peritoneal cavity of mice and cultured at 4 × 105 cells per well in DMEM ( with 10% FCS , L-glutamine , and penicillin/streptomycin ) for 2 h at 37 °C . Non-adherent cells were removed by washing with cold PBS . LCMV-WE strain , originally obtained from F . Lehmann-Grube ( Hamburg , Germany ) , was propagated on L929 cells . MHV A59 was generated from a molecularly cloned cDNA [24] based on the Albany strain of MHV A59 . Coronaviruses and recombinant vaccinia viruses were propagated , titrated , and purified as described [24 , 50 , 51] . Mutant vaccinia viruses are based on the recombinant vaccina virus vMHV-inf-1 ( containing the full-length MHV-A59 cDNA ) and were generated using our reverse genetic system as described previously [24] . Briefly , the gene to be mutated was replaced by the Escherichia coli guanine-phosphoribosyltransferase ( GPT ) gene through vaccinia virus–mediated homologous recombination , and GPT-positive clones were selected by three rounds of plaque purification on CV-1 cells in the presence of xanthine , hypoxanthine , and mycophenolic acid ( GPT-positive selection ) [50] . In a second round , the GPT gene was replaced by the mutated gene , and GPT-negative clones , containing the mutated gene , were selected by three rounds of plaque purification on D980R cells in the presence of 6-thioguanine ( GPT-negative seletion ) [50] . To construct the recombinant vaccinia virus encoding the MHV-nsp1Δ99 cDNA , the 5′ end of the MHV-A59 cDNA in vMHV-inf-1 was replaced by GPT using the plasmid pRec4 . This plasmid is based on pGPT-1 [50] , and the GPT gene is flanked to its left by 500 bp of vaccinia DNA and to its right by an internal ribosomal entry sequence ( IRES ) followed by MHV-A59 nts 952-1315 . The GPT-negative selection was carried out using the plasmid pRec15 . This plasmid contains 250 bp of vaccinia DNA followed by the bacteriophage T7 RNA polymerase promoter , one G nucleotide , and MHV nts 1–828 linked to MHV nts 928-1315 . To replace the MHV-A59 accessory gene 4 in vMHV-inf-1 and vMHV-nsp1Δ99 by a gene encoding a fusion protein of EGFP and the LCMV-derived CTL epitope KAVYNFATC ( GP33-GFP ) [29] , the plasmid pRec8 was used for recombination with vaccinia viruses vMHV-inf-1 and vMHV-nsp1Δ99 . This plasmid contains the GPT gene flanked to its left by MHV nts 27500–27967 and to its right by MHV nts 28265–28700 . GPT-negative selection was carried out using the plasmid pRec9 . This plasmid contains MHV bp 27500–27967 , the GP33-GFP gene , and MHV nts 28265–28700 . Further cloning details , plasmid maps , and sequences are available from the authors upon request . Recombinant coronaviruses were rescued from cloned cDNA using purified vaccinia virus DNA as template for the in vitro transcription of recombinant MHV genomes as described [51] . The firefly luciferase ( FF-Luc ) plasmid for monitoring IFN-β promoter activation ( p125-Luc ) was kindly provided by Takashi Fujita , Tokyo Metropolitan Institute of Medical Science , Japan [52] . The FF-Luc reporter construct for monitoring ISRE activation ( p ( 9–27 ) 4tkΔ ( −39 ) lucter ) [53] was kindly provided by Stephen Goodbourn , St . George's Hospital Medical School , London , UK . The control plasmid pRL-SV40 ( Promega , http://www . promega . com/ ) encodes the renilla luciferase ( REN-Luc ) gene under control of the constitutive SV40 promoter . The negative control expression plasmid contained the reading frame of the N-terminus of the human MxA protein . To construct the coronavirus nsp1 expression plasmids , the MHV nts 1–951 ( pMHV-nsp1 ) , MHV nts 1–902 ( pMHV-Δ49 ) , MHV nts 1–851 ( pMHV-Δ100 ) , HCoV-229E nts 293–625 ( pHCoV-229E-nsp1 ) , and SARS-CoV nts 265–804 ( pSARS-CoV-nsp1 ) were amplified by standard PCR techniques and cloned downstream of a CMV promoter between the SmaI and XhoI sites of the eukaryotic high-level expression plasmid pI . 18 ( kindly provided by Jim Robertson , National Institute for Biological Standards and Control , Hertfordshire , UK ) . Subconfluent cell monolayers of 293 cells seeded in 12-well plates were transfected with 250 ng p125-Luc reporter plasmid , 50 ng pRL-SV40 , and 1 μg of expression plasmid in 200 μl of OPTIMEM ( Gibco-BRL ) containing 3 . 9 μl of Fugene HD ( Roche , http://www . roche . com ) . At 8 h post transfection , cells were induced with either 0 . 2 μg of viral ssRNA containing 5′ triphosphates [54] ( p125-Luc ) , 2 . 5 μg of poly ( I:C ) ( Sigma ) , or 500 U/ml IFN-α ( p ( 9–27 ) 4tkΔ ( −39 ) lucter ) , or left uninduced . After an incubation period of 16 h , cells were harvested and lysed in 100 μl of Reporter Lysis Buffer ( Promega ) . An aliquot of 10 μl lysate was used to measure luciferase activity as decribed by the manufacturer ( Promega ) . Mice were injected intraperitoneally or intracranially with indicated pfu of MHV A59 or intravenously with the indicated pfu of LCMV and sacrificed at the indicated time points . Organs were stored at −70 °C until further analysis . Blood was incubated at RT to coagulate , centrifuged , and serum was used for ALT measurements using a Hitachi 747 autoanalyzer ( http://www . hitachi . com/ ) . Peritoneal exudates cells ( PECs ) were isolated from the peritoneal cavity by flushing with 4 ml of ice-cold PBS . MHV titers were determined by standard plaque assay using L929 cells . LCMV titers in the spleens were determined 4 d after intravenous challenge in an LCMV infectious focus assay as previously described [55] . Organs were fixed in 4% formalin and embedded in paraffin . Sections were stained with hematoxylin and eosin . Images were acquired using a Leica DMRA microscope ( Leica , http://www . leica . com/ ) with a 25×/0 . 65 NA objective ( total magnification , ×162 ) . Images were processed using Adobe Photoshop ( Adobe Systems , http://www . adobe . com ) . Mouse IFN-α concentration in cell culture supernatants was measured by ELISA ( PBL Biomedical Laboratories , http://www . interferonsource . com/ ) according to the manufacturer's instructions . IFN-α treatment of cells prior to MHV infection was performed using universal type I IFN ( IFN-αA/D , Sigma ) . Specific ex vivo production of IFN-γ was determined by intracellular cytokine staining . Organs were removed at the indicated time points following infection with recombinant MHV . For intracellular cytokine staining , single cell suspensions of 1 × 106 splenocytes were incubated for 5 h at 37 °C in 96-well round-bottom plates in 200 μl of culture medium containing 25 U/ml IL-2 and 5 μg/ml Brefeldin A ( Sigma ) . Cells were stimulated with phorbolmyristateacetate ( PMA , 50 ng/ml ) and ionomycin ( 500 ng/ml ) ( both purchased from Sigma ) as positive control or left untreated as a negative control . For analysis of peptide-specific responses , cells were stimulated with 10−6 M GP33 peptide or 10−4 M MHV S598 peptide . The percentage of CD8+ T cells producing IFN-γ was determined using a FACSCalibur flow cytometer ( BD Biosciences ) . Both S598 ( RCQIFANI ) and GP33 ( KAVYNFATC ) peptides were purchased from Neosystem ( http://www . neomps . com/ ) . All statistical analyses were performed with Prism 4 . 0 ( GraphPad Software , http://www . graphpad . com/ ) . Data were analyzed with the paired Student's t-test assuming that the values followed a Gaussian distribution . A p-value of < 0 . 05 was considered significant .
The GenBank ( http://www . ncbi . nlm . nih . gov/Genbank/ ) accession numbers for the viruses and sequences discussed in this paper are HCoV-229E ( AF304460 ) , MHV-A59 ( AY700211 ) , and SARS-CoV Frankfurt-1 ( AY291315 ) .
|
Prevention of viral diseases by vaccination aims for controlled induction of protective immune responses against viral pathogens . Live viral vaccines consist of attenuated , replication-competent viruses that are believed to be superior in the induction of broad immune responses , including cell-mediated immunity . The recent proceedings in the area of virus reverse genetics allows for the rational design of recombinant vaccines by targeting , i . e . , inactivating , viral pathogenicity factors . For coronaviruses , a major pathogenicity factor has now been identified . The effect of coronavirus non-structural protein 1 on pathogenicity has been analyzed in a murine model of coronavirus infection . By deleting a part of this protein , a recombinant virus has been generated that is greatly attenuated in vivo , while retaining immunogenicity . In particular , the mutant virus retained the ability to replicate in professional antigen-presenting cells and fulfilled an important requirement of a promising vaccine candidate: the induction of a protective long-lasting , antigen-specific cellular immune response . This study has implications for the rational design of live attenuated coronavirus vaccines aimed at preventing coronavirus-induced diseases of veterinary and medical importance , including the potentially lethal severe acute respiratory syndrome .
|
[
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] |
[
"viruses",
"infectious",
"diseases",
"virology",
"immunology",
"mus",
"(mouse)",
"homo",
"(human)",
"mammals"
] |
2007
|
Coronavirus Non-Structural Protein 1 Is a Major Pathogenicity Factor: Implications for the Rational Design of Coronavirus Vaccines
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.