LETTER doi:10.1038/nature10665 Predicting mutation outcome from early stochastic variation in genetic interaction partners Alejandro Burga,M.Olivia Casanueva&Ben Lehner.2 Many mutations,including those that cause disease,only have a similar incompletely penetrant defects,and loss ofboth genes results in detrimental effect in a subset of individuals.The reasons for this synthetic lethality in C.elegans67 and Caenorhabditis briggsaes are usually unknown,but may include additional genetic variation If the incomplete penetrance of tbx-9 relates to variation in the and environmental risk factors'.However,phenotypic discordance expression of tbx-8,then increased expression of txb-8 should be able remains even in the absence of genetic variation,for example to compensate for the loss of tbx-9.Expression of TBX-8 from a between monozygotic twins',and incomplete penetrance of muta- transgene indeed reduces the penetrance of a tbx-9(ok2473)null allele tions is frequent in isogenic model organisms in homogeneous (Fig.Ic and Supplementary Fig.2).A further requisite of our model is environments4.Here we proposea modelfor incomplete penetrance the existence of inter-individual variation in the expression of the based on genetic interaction networkss.Using Caenorhabditis genetic interaction partner.We quantified the induction of a fluor- elegans as a model system,we identify two compensation mechan- escent reporter driven by the tbx-8 promoter(ptbx-8::GFP)during isms that vary among individuals and influence mutation outcome. normal development and found that there was substantial variation First,feedback induction of an ancestral gene duplicate differs across among embryos (Fig.2a)with an important extrinsic component individuals,with high expression masking the effects of a mutation. (Supplementary Figs 3a and 4). This supports the hypothesis that redundancy is maintained in We next quantified the induction of the tbx-8 reporter in tbx- genomes to buffer stochastic developmental failure?.Second,during 9(ok2473)mutant animals and found that expression was increased normal embryonic development we find that there is substantial (Fig.2b,1.2-fold upregulation at comma stage,P=1.6x 10).We variation in the induction of molecular chaperones such as Hsp90 confirmed this upregulation by performing single molecule fluor- (DAF-21).Chaperones act as promiscuous buffers of genetic vari- escence in situ hybridization against the endogenous tbx-8 messenger ation5,and embryos with stronger induction of Hsp90 are less likely RNA(mRNA)(Supplementary Fig.5,1.6-fold upregulation in 40-to to be affected by an inherited mutation.Simultaneously quantifying the variation in these two independent responses allows the pheno- Ilsogenic strain typic outcome of a mutation to be more accurately predicted in b Genetic interaction/synthetic lethality Homogeneous individuals.Our model and methodology provide a framework for environment Gene A Gene B Phenotype dissecting the causes of incomplete penetrance.Further,the results WT establish that inter-individual variation in both specific and more general buffering systems combine to determine the outcome WT inherited mutations in each individual. It is well established that isogenic individuals show substantial vari- -50%wT 白十一二Abnormal ation at the molecular level,for example in the expression levels of -50%abnormal phenotype phenotype particular genest.This variation can be an important influence on c 70 d Model for incomplete penetrance signalling and development in wild-type (WT)individuals2.Such 60 molecular variation may also influence the phenotypic consequences 50 GeneA Gene B Phenotype of inherited mutations,but examples of molecular variation have only WT been described that predict phenotypic variation at the level of other 3 20 molecules'or cellular phenotypes'4. 白一十 To identify determinants of mutation outcome,we present a model for incomplete penetrance based on genetic interaction networkss 8ok656 Abnormal (Fig.1b).Our model proposes that in the absence of additional genetic tbx Ex[TBX-9OE variation,it is stochastic variation in the abundance or activity of genetic interaction partners(genes that influence the outcome of a Figure 1 Genetic interactions provide a general model for incomplete mutation when genetically altered)that determines the outcome of a penetrance.a,Inactivation of the gene tbx-9 in C.elegans results in an mutation (Fig.1d).The correlated loss of genetic,environmental and incompletely penetrant defect,with approximately half of embryos hatching stochastic robustness upon gene deletion in yeast is highly consistent with abnormal morphology(small arrow).b,Representation of a negative with the generality of this proposals (synergistic)genetic interaction between two genes A and B.c,Increased To test our model,we first used a null mutation in the T-box tran- expression of TBX-8 reduces the penetrance of a tbx-9(ok2473)null mutation scription factor gene tbx-9(Supplementary Fig.1)that causes an (49%(n=138)compared with 34%(n 139),a 33%decrease in abnormal phenotypes,P=0.011,Fisher's exact test).Likewise,increased TBX-9 incompletely penetrant defect in C.elegans larval morphology owing expression rescues the penetrance of a tbx-8(ok656)null mutation (60% to abnormal development of the epidermis and muscle(Fig la). (n=164)compared with 35%(n=195),a 42%decrease,P=2.8X 10-).d,A Transcription factor tbx-9 is related to another transcription factor, model for incomplete penetrance based on variation in the activity of genetic tbx-8,by an ancestral gene duplication;inactivation of tbx-8 causes interaction partners. EMBL-CRG Systems Biology Unit.Centre for Genomic Regulation(CRG)and Universitat Pompeu Fabra(UPF)Barcelona 03.Spain.2nstituci Catalana de Recercai Estudis Avancats,Centre for Genomic Regulation(CRG)and Universitat Pompeu Fabra (UPF),Barcelona 08003,Spain. 250 NATURE I VOL DECEMBER 2011 2011 Macmillan Publishers Limited.All rights reserved
LETTER doi:10.1038/nature10665 Predicting mutation outcome from early stochastic variation in genetic interaction partners Alejandro Burga1 , M. Olivia Casanueva1 & Ben Lehner1,2 Many mutations, including those that cause disease, only have a detrimental effect in a subset of individuals. The reasons for this are usually unknown, but may include additional genetic variation and environmental risk factors1 . However, phenotypic discordance remains even in the absence of genetic variation, for example between monozygotic twins2 , and incomplete penetrance of mutations is frequent in isogenic model organisms in homogeneous environments3,4. Here we propose a model forincomplete penetrance based on genetic interaction networks5,6. Using Caenorhabditis elegans as a model system, we identify two compensation mechanisms that vary among individuals and influence mutation outcome. First, feedback induction of an ancestral gene duplicate differs across individuals, with high expression masking the effects of a mutation. This supports the hypothesis that redundancy is maintained in genomes to buffer stochastic developmental failure7 . Second, during normal embryonic development we find that there is substantial variation in the induction of molecular chaperones such as Hsp90 (DAF-21). Chaperones act as promiscuous buffers of genetic variation8,9, and embryos with strongerinduction of Hsp90 are less likely to be affected by an inherited mutation. Simultaneously quantifying the variation in these two independent responses allows the phenotypic outcome of a mutation to be more accurately predicted in individuals. Our model and methodology provide a framework for dissecting the causes of incomplete penetrance. Further, the results establish that inter-individual variation in both specific and more general buffering systems combine to determine the outcome inherited mutations in each individual. It is well established that isogenic individuals show substantial variation at the molecular level, for example in the expression levels of particular genes10. This variation can be an important influence on signalling and development in wild-type (WT) individuals11,12. Such molecular variation may also influence the phenotypic consequences of inherited mutations, but examples of molecular variation have only been described that predict phenotypic variation at the level of other molecules13 or cellular phenotypes14. To identify determinants of mutation outcome, we present a model for incomplete penetrance based on genetic interaction networks5,6 (Fig. 1b). Our model proposes that in the absence of additional genetic variation, it is stochastic variation in the abundance or activity of genetic interaction partners (genes that influence the outcome of a mutation when genetically altered) that determines the outcome of a mutation (Fig. 1d). The correlated loss of genetic, environmental and stochastic robustness upon gene deletion in yeast is highly consistent with the generality of this proposal6,15. To test our model, we first used a null mutation in the T-box transcription factor gene tbx-9 (Supplementary Fig. 1) that causes an incompletely penetrant defect in C. elegans larval morphology owing to abnormal development of the epidermis and muscle16,17 (Fig. 1a). Transcription factor tbx-9 is related to another transcription factor, tbx-8, by an ancestral gene duplication; inactivation of tbx-8 causes similar incompletely penetrant defects, and loss of both genes results in synthetic lethality in C. elegans16,17 and Caenorhabditis briggsae18. If the incomplete penetrance of tbx-9 relates to variation in the expression of tbx-8, then increased expression of txb-8 should be able to compensate for the loss of tbx-9. Expression of TBX-8 from a transgene indeed reduces the penetrance of a tbx-9(ok2473) null allele (Fig. 1c and Supplementary Fig. 2). A further requisite of our model is the existence of inter-individual variation in the expression of the genetic interaction partner. We quantified the induction of a fluorescent reporter driven by the tbx-8 promoter (ptbx-8::GFP) during normal development and found that there was substantial variation among embryos (Fig. 2a) with an important extrinsic component (Supplementary Figs 3a and 4). We next quantified the induction of the tbx-8 reporter in tbx- 9(ok2473) mutant animals and found that expression was increased (Fig. 2b, 1.2-fold upregulation at comma stage, P 5 1.6 3 1023 ). We confirmed this upregulation by performing single molecule fluorescence in situ hybridization against the endogenous tbx-8 messenger RNA (mRNA) (Supplementary Fig. 5, 1.6-fold upregulation in 40- to 1 EMBL-CRG Systems Biology Unit, Centre for Genomic Regulation (CRG) and Universitat Pompeu Fabra (UPF), Barcelona 08003, Spain. 2 Institucio´ Catalana de Recerca i Estudis Avançats, Centre for Genomic Regulation (CRG) and Universitat Pompeu Fabra (UPF), Barcelona 08003, Spain. tbx-9(ok2473) tbx-8(ok656) tbx-9(ok2473); Ex[TBX-8]OE tbx-8(ok656); Ex[TBX-9]OE * *** Penetrance (%) c d Genetic interaction/synthetic lethality Gene A Gene B Phenotype Gene A Gene B Phenotype Model for incomplete penetrance tbx-9 - Isogenic strain Developmental noise - Homogeneous environment ~ 50% WT phenotype ~ 50% abnormal phenotype a b WT WT WT Abnormal Abnormal 70 60 50 40 30 20 10 0 Figure 1 | Genetic interactions provide a general model for incomplete penetrance. a, Inactivation of the gene tbx-9 in C. elegans results in an incompletely penetrant defect, with approximately half of embryos hatching with abnormal morphology (small arrow). b, Representation of a negative (synergistic) genetic interaction between two genes A and B. c, Increased expression of TBX-8 reduces the penetrance of a tbx-9(ok2473) null mutation (49% (n 5 138) compared with 34% (n 5 139), a 33% decrease in abnormal phenotypes, P 5 0.011, Fisher’s exact test). Likewise, increased TBX-9 expression rescues the penetrance of a tbx-8(ok656) null mutation (60% (n 5 164) compared with 35% (n 5 195), a 42% decrease, P 5 2.83 1026 ). d, A model for incomplete penetrance based on variation in the activity of genetic interaction partners. 250 | NATURE | VOL 480 | 8 DECEMBER 2011 ©2011 Macmillan Publishers Limited. All rights reserved
LETTER RESEARCH d tbx-gpt地x-8:GFP atbx-8:GFP . ype ● 0.6 0 thx-9:pthx-8:GFP p他-8GfP 0 100200300 200 400 Time (min) Time(min) L1 stage phenotype bx-8pt地x-9:GFP 2ph-9.GFP tx-&:ptb-9:GFP 0 0 tbx-8:ptbx-9:GFP 200 400 ptx-9GFP tbx-8: WT Abnormal 200 400 Time (min) ptbx-9..GFP Time (min) L1 stage phenotype fih-1:pfih-2::GFP 34 ■pfh-2:GFp 口h-1:ph2:GFP Bright field Fluorescence 12 3 0.4 3 0.0 0.8 0 fih-1;pfih-2:GFP pfh-2:GFP fh-1 100200300 0.4 0 Time (min) WT Abnorma L1 stage phenotype Figure 2 Early inter-individual variation in the induction of ancestral gene upregulation at comma stage(~375 min,P=3.6x 10).g.Expression of duplicates predicts the outcome of inherited mutations.a,Quantification of tbx-9 reporter in a tbx-8(ok656)mutant background,colour code as in total green fluorescent protein(GFP)expression from a tbx-8 reporter during c.h,Boxplot of g showing tbx-9 expression is higher in tbx-8 embryos that embryonic development in WT(black)and bx-9(ok2473)(green)individuals. develop a WT phenotype(P=0.033).i,Expression ofa pflh-2::GFP reporter in Each individual is a separate line.a.u.,Arbitrary units.b,Boxplot of tbx-8 WT (black)and flh-1(bc374)mutant (green).j,Boxplot of flh-2 reporter reporter expression(a)showing 1.2-fold upregulation in a tbx-9 mutant at expression (i)showing 1.8-fold upregulation in a flh-1 mutant at comma stage comma stage(~290 min,P=1.6X 103,Wilcoxon rank test).c,Expression of (~180 min,P=2.2 X 10).k,Bright-field and fluorescence image of an tbx-8 reporter in a tbx-9(ok2473)background for embryos that hatch with (red) approximate 100-cell flh-1;pflh-2:GFP embryo.Red arrow indicates the local or without (blue,WT)a morphological defect.d,Boxplot of c showing tbx-8 expression of flh-2 reporter quantified for fh-1 phenotypic prediction. expression is higher in tbx-9 embryos that develop a WT phenotype (blue) I,Boxplot showing higher fh-2 reporter expression at approximate 100 cells for compared with those that develop an abnormal(red)phenotype at comma WT(blue)compared with abnormal (red)phenotypes(P=0.014).Boxplots stage(P=6.1X 10-).e,Expression of a ptbx-9::GFP reporter in WT(black) show the median,quartiles,maximum and minimum expression in each and tbx-8(ok656)mutant (green).f,Boxplot of tbx-9reporter showing 4.3-fold data set. 50-cell-stage embryos,P=5.3X 10).Thus,a direct or indirect feed- 4.3-fold upregulation,P=3.6X 10-16)predicted the phenotypic back mechanism exists that upregulates transcription of tbx-8 when its outcome of a tbx-8(ok656)mutation.We found that was indeed the ancestral paralogue is inactivated.This feedback also acts on the tbx-9 case(Fig.2g,h,P=0.033,see Supplementary Table 1). promoter,which is also upregulated in a tbx-9(ok2473)mutant back- The FLYWCH transcription factors flh-1 and flh-2 are an additional ground(Supplementary Fig.6).This is highly consistent with observa- pair of ancestral gene duplicates that redundantly repress embryonic tions that have been made in yeast2 and suggests that compensatory expression of the microRNAs lin-4,mir-48 and mir-241 (ref.23). expression by negative feedback regulation of gene duplicates is Similar to the case for tbx-9 null mutants,flh-1(bc374)mutant embryos probably a conserved phenomenon across species. induced higher levels of a pflh-2::GFP reporter than WT embryos Next,we tested whether variation in the induction of tbx-8 corre- (Fig.2i,j,1.8-fold at comma stage,P<2.2x 101).In contrast,we lated with the outcome of the tbx-9(ok2473)mutation at hatching.We found neither upregulation of a tbx-8 reporter in an flh-1(bc374)back- retrospectively compared the early expression of the reporter in ground,nor upregulation of an fh-2 reporter in a tbx-9(ok2473) embryos that did,and did not,hatch with an abnormal pheno- mutant(Supplementary Fig.8,P=0.74 and P=0.55 respectively). type.Early expression was higher in the second class(Fig.2c,d, flh-1(bc374)embryos that expressed higher flh-2 reporter levels early P=6.1 X 10);whereas 80%of the embryos with reporter expression in development(Fig.2k,see Methods)were,however,more likely to in the highest quartile hatched without a defect,only 40%of those in develop into morphologically WT larvae (Fig.21,P=0.014,see the lowest quartile showed no phenotype(Supplementary Table 1, Supplementary Table 1).In contrast,we found that pflh-2::GFP P=3.3 x 103).In contrast,early expression from a ubiquitously reporter levels did not predict the outcome of a tbx-9(ok2473)null transcribed promoter (plet-858::GFP)did not predict phenotypic mutation (Supplementary Fig.7g,h,P=0.32),nor did induction of outcome in individuals(Supplementary Fig.7a,b,P=0.58),nor did the ptbx-8::GFP reporter predict the outcome of the flh-1(bc374)muta- variation in the expression of a reporter for the transcription factor tion (Supplementary Fig.7e,f,P=0.49). ELT-5,another protein required for epidermal development2 Many ancestral gene duplicates have retained partly redundant func- (Supplementary Fig.7c,d,P=0.97).High expression of the synthetic tions over extensive evolutionary periods2425 One explanation for lethal partner tbx-8,but not high transcription in general,is therefore this could be the selection pressure provided by stochastic develop- partly epistaticto the lossoftbx-9.We also tested the reverse situation: mental errors?('canalization).Our results provide direct empirical that is,whether variation in tbx-9 expression compensation (Fig.2e,f, support for this hypothesis,showing that when one member of a partly 8 DECEMBER 2011 VOL 480 I NATURE 251 2011 Macmillan Publishers Limited.All rights reserved
50-cell-stage embryos, P 5 5.3 3 1024 ). Thus, a direct or indirect feedback mechanism exists that upregulates transcription of tbx-8 when its ancestral paralogue is inactivated. This feedback also acts on the tbx-9 promoter, which is also upregulated in a tbx-9(ok2473) mutant background (Supplementary Fig. 6). This is highly consistent with observations that have been made in yeast19,20 and suggests that compensatory expression by negative feedback regulation of gene duplicates is probably a conserved phenomenon across species. Next, we tested whether variation in the induction of tbx-8 correlated with the outcome of the tbx-9(ok2473) mutation at hatching. We retrospectively compared the early expression of the reporter in embryos that did, and did not, hatch with an abnormal phenotype. Early expression was higher in the second class (Fig. 2c, d, P 5 6.1 3 1023 ); whereas 80% of the embryos with reporter expression in the highest quartile hatched without a defect, only 40% of those in the lowest quartile showed no phenotype (Supplementary Table 1, P 5 3.3 3 1023 ). In contrast, early expression from a ubiquitously transcribed promoter (plet-858::GFP) did not predict phenotypic outcome in individuals (Supplementary Fig. 7a, b, P 5 0.58), nor did variation in the expression of a reporter for the transcription factor ELT-5, another protein required for epidermal development21 (Supplementary Fig. 7c, d, P 5 0.97). High expression of the synthetic lethal partner tbx-8, but not high transcription in general, is therefore partly epistatic22 to the loss of tbx-9.We also tested the reverse situation: that is, whether variation in tbx-9 expression compensation (Fig. 2e, f, 4.3-fold upregulation, P 5 3.6 3 10216) predicted the phenotypic outcome of a tbx-8(ok656) mutation. We found that was indeed the case (Fig. 2g, h, P 5 0.033, see Supplementary Table 1). The FLYWCH transcription factorsflh-1 and flh-2 are an additional pair of ancestral gene duplicates that redundantly repress embryonic expression of the microRNAs lin-4, mir-48 and mir-241 (ref. 23). Similar to the case fortbx-9 null mutants, flh-1(bc374) mutant embryos induced higher levels of a pflh-2::GFP reporter than WT embryos (Fig. 2i, j, 1.8-fold at comma stage, P , 2.2 3 10216). In contrast, we found neither upregulation of a tbx-8 reporter in an flh-1(bc374) background, nor upregulation of an flh-2 reporter in a tbx-9(ok2473) mutant (Supplementary Fig. 8, P 5 0.74 and P 5 0.55 respectively). flh-1(bc374) embryos that expressed higher flh-2 reporter levels early in development (Fig. 2k, see Methods) were, however, more likely to develop into morphologically WT larvae (Fig. 2l, P 5 0.014, see Supplementary Table 1). In contrast, we found that pflh-2::GFP reporter levels did not predict the outcome of a tbx-9(ok2473) null mutation (Supplementary Fig. 7g, h, P 5 0.32), nor did induction of the ptbx-8::GFP reporter predict the outcome of the flh-1(bc374) mutation (Supplementary Fig. 7e, f, P 5 0.49). Many ancestral gene duplicates have retained partly redundant functions over extensive evolutionary periods19,24,25. One explanation for this could be the selection pressure provided by stochastic developmental errors7,26 (‘canalization’27). Our results provide direct empirical support for this hypothesis, showing that when one member of a partly ptbx-9::GFP *** tbx-8; ptbx-9::GFP ptbx-9::GFP fluorescence intensity (a.u.) pflh-2::GFP fluorescence intensity (a.u.) pflh-2::GFP fluorescence intensity at comma stage (a.u.) *** ptbx-9::GFP fluorescence intensity (a.u.) WT Abnormal L1 stage phenotype WT Abnormal L1 stage phenotype tbx-8; ptbx-9::GFP tbx-8; ptbx-9::GFP flh-1; pflh-2::GFP ptbx-9::GFP fluorescence intensity at comma stage (a.u.) ptbx-9::GFP fluorescence intensity at comma stage (a.u.) ptbx-8::GFP fluorescence intensity at comma stage (a.u.) ** ** ptbx-8::GFP fluorescence intensity at comma stage (a.u.) WT Abnormal L1 stage phenotype ptbx-8::GFP tbx-9; ptbx-8::GFP Time (min) Time (min) Time (min) Time (min) Time (min) pflh-2::GFP fluorescence intensity at ~100 cells (a.u.) flh-1; pflh-2::GFP * * a b tbx-9; ptbx-8::GFP ptbx-8::GFP fluorescence intensity (a.u.) c d e f g h i j k l ptbx-8::GFP fluorescence intensity (a.u.) tbx-9; ptbx-8::GFP 0 100 200 300 0 1 0.6 1.0 1.4 0 200 400 0 1 0.6 1.0 0 1 0.6 1.0 1.4 1.8 0 200 400 0 1 2 0.4 0.0 0.4 0.8 1.0 1.2 WT phenotype Abnormal phenotype WT phenotype Abnormal phenotype ptbx-8::GFP tbx-9; ptbx-8::GFP ptbx-9::GFP tbx-8; ptbx-9::GFP 0 1 2 0 200 400 4 3 2 1 0 0 100 200 300 pflh-2::GFP flh-1; pflh2::GFP pflh-2::GFP flh-1; pflh-2::GFP Bright field Fluorescence Figure 2 | Early inter-individual variation in the induction of ancestral gene duplicates predicts the outcome of inherited mutations. a, Quantification of total green fluorescent protein (GFP) expression from a tbx-8 reporter during embryonic development in WT (black) and tbx-9(ok2473) (green) individuals. Each individual is a separate line. a.u., Arbitrary units. b, Boxplot of tbx-8 reporter expression (a) showing 1.2-fold upregulation in a tbx-9 mutant at comma stage (,290 min, P 5 1.63 1023 , Wilcoxon rank test). c, Expression of tbx-8 reporter in a tbx-9(ok2473) background for embryos that hatch with (red) or without (blue, WT) a morphological defect. d, Boxplot of c showing tbx-8 expression is higher in tbx-9 embryos that develop a WT phenotype (blue) compared with those that develop an abnormal (red) phenotype at comma stage (P 5 6.1 3 1023 ). e, Expression of a ptbx-9::GFP reporter in WT (black) and tbx-8(ok656) mutant (green). f, Boxplot of tbx-9 reporter showing 4.3-fold upregulation at comma stage (,375 min, P 5 3.63 10216). g, Expression of tbx-9 reporter in a tbx-8(ok656) mutant background, colour code as in c. h, Boxplot of g showing tbx-9 expression is higher in tbx-8 embryos that develop a WT phenotype (P 5 0.033). i, Expression of a pflh-2::GFP reporter in WT (black) and flh-1(bc374) mutant (green). j, Boxplot of flh-2 reporter expression (i) showing 1.8-fold upregulation in a flh-1 mutant at comma stage (,180 min, P 5 2.23 10216). k, Bright-field and fluorescence image of an approximate 100-cell flh-1; pflh-2::GFP embryo. Red arrow indicates the local expression of flh-2 reporter quantified for flh-1 phenotypic prediction. l, Boxplot showing higherflh-2 reporter expression at approximate 100 cells for WT (blue) compared with abnormal (red) phenotypes (P 5 0.014). Boxplots show the median, quartiles, maximum and minimum expression in each data set. LETTER RESEARCH 8 DECEMBER 2011 | VOL 480 | NATURE | 251 ©2011 Macmillan Publishers Limited. All rights reserved
RESEARCH LETTER redundant gene pair is inactivated,expression variation in the other Wetested whether this variability was particular to daf-21 or common gene becomes an important influence on the phenotype of an to other chaperones using a reporter for the gene hsp-4(orthologous to individual. mammalian BiP).Transcription from the hsp-4 promoter also varied The variation in the induction tbx-8 that we can quantify only partly extensively among embryos(Supplementary Fig.13)and was correlated accounts for variation in the outcome of the tbx-9 mutation (Fig.2d). with that from the daf-21 promoter(Fig.3c,Pearson correlation coef- We considered that one additional influence on mutation outcome ficient,r=0.63,P=2.8X 10),similar to the correlation between could be variation in the activity of general buffering systems such as two independent daf-21 reporters (Fig.3b,r=0.77,P=3.4x 10-4). molecular chaperones.In multiple eukaryotes,chaperone inhibition can This and additional correlations between stress response genes enhance the effects of diverse mutations and increased chaperone (Supplementary Fig.14)suggest that,during normal development, activity promiscuously suppresses detrimental mutationsin bacteria20 isogenic embryos differ in a coordinated manner in the transcriptional We constructed a transcriptional reporter for the constitutively induction of multiple chaperones. expressed chaperone daf-21 (Hsp90).Variability in the pdaf- We quantified induction of a pdaf-21::mCherry reporter in a tbx- 21::mCherry reporter was evident during all larval stages and, 9(ok2473)background.The mean induction of this reporter was surprisingly,its induction during embryonic development also varied increased in tbx-9(ok2473)animals(Supplementary Fig.9b,1.8-fold substantially,even in the absence of a mutation or environmental upregulation,P=6.6 10).In addition,embryos that later hatched perturbation(Supplementary Figs 9a and 10).No difference in trans- without a morphological defect expressed higher levels of the reporter gene copy number was detected between animals with high or low daf-21 reporter expression(Supplementary Fig.11),and individuals with high reporter expression have high expression of the DAF-21 2.5 LH HH protein (Fig.3a and Supplementary Fig.12). 2.0 a pdaf-21:mCherry fluorescence intensity 1.5 High Low High Low High Low 1.0 Anti-DAF-21 Anti-mCherry 0.5 Anti-B-actin 0.0 3.0 -0.5 HL r=0.77 r=0.63 立2.5 0.6 0.8 1.0 1.2 14 2.0 ptbx-8::GFP fluorescence 15 0.5 intensity at comma stage (a.u.) 1.0 Bright field ptbx-8::GFP pdaf-21::mCherry WT(%) 0.5 0.5 1.0 1.5 0.40.60.81.0 1.2 pdaf-21::GFP phsp-4::GFP fluorescence intensity (a.u.) fluorescence intensity (a.u.) e tbx-9;pdaf-21::mCherry ■NT phenatype 3 Abnormal phenotype 8 2 9 tbx-9:pdaf-21::mChery 0 100200 300 WT Abnormal Time(min) L1 stage phenotype HH Figure 3Inter-individual variation in chaperone induction predicts the outcome ofa mutation.a,pdaf-21:mCherry:glp-1(e2141)animals were sorted into 'high'and 'low'groups based on their pdaf-21::mCherry expression 24h Figure 4 Simultaneous quantification of inter-individual variation in two after the LI stage at 25C.Western blot analysis confirmed that worms buffering systems accurately predicts the outcome of a mutation. belonging to the high'group had higher levels of endogenous DAF-21 protein. a,Expression levels of tbx-8 and daf-21 reporters in individual tbx-9(ok2473) As a positive control,mCherry protein levels showed the same trend.B-Actin mutant embryos at the comma stage of development are not correlated was used as loading control;results of three replica sortings are shown (r=0.061,P=0.56).Each point represents one individual embryo.Embryos b,c,Correlation in individuals (measured at comma stage)between:pdaf- that hatched into phenotypically WT larvae are blue;embryos that hatched 21::mCherry and pdaf-21::GFP (Pearson correlation coefficient,r=0.77, with a morphological defect are red.The median values for each reporter are P =3.4X 10)(b);pdaf-21::mCherry and phsp-4::GFP (r 0.63, indicated (black lines);embryos with expression above or below the median are P=2.8x 10)(c).d,e,Expression of daf-21 reporter in a tbx-9(ok2473) defined as high (H)and low (L),respectively.b,Percentage of embryos from mutant background.Embryos that hatch into phenotypically WT worms(blue) each expression class hatching with normal morphology.Representative have higher expression than those hatching with a morphological defect(red)at bright-field,green and red fluorescence images are shown.Prediction scores for the comma stage (P=1.9 X 10-3). different time points and measures are given in Supplementary Table 3. 252 NATURE VOL 480 8 DECEMBER 2011 2011 Macmillan Publishers Limited.All rights reserved
redundant gene pair is inactivated, expression variation in the other gene becomes an important influence on the phenotype of an individual. The variation in the induction tbx-8 that we can quantify only partly accounts for variation in the outcome of the tbx-9 mutation (Fig. 2d). We considered that one additional influence on mutation outcome could be variation in the activity of general buffering systems such as molecular chaperones. In multiple eukaryotes, chaperone inhibition can enhance the effects of diverse mutations8,9,28 and increased chaperone activity promiscuously suppresses detrimental mutationsin bacteria29,30. We constructed a transcriptional reporter for the constitutively expressed chaperone daf-21 (Hsp90). Variability in the pdaf- 21::mCherry reporter was evident during all larval stages and, surprisingly, its induction during embryonic development also varied substantially, even in the absence of a mutation or environmental perturbation (Supplementary Figs 9a and 10). No difference in transgene copy number was detected between animals with high or low daf-21 reporter expression (Supplementary Fig. 11), and individuals with high reporter expression have high expression of the DAF-21 protein (Fig. 3a and Supplementary Fig. 12). We testedwhether this variabilitywas particular to daf-21 or common to other chaperones using a reporter for the gene hsp-4 (orthologous to mammalian BiP). Transcription from the hsp-4 promoter also varied extensively among embryos (Supplementary Fig. 13) and was correlated with that from the daf-21 promoter (Fig. 3c, Pearson correlation coefficient, r 5 0.63, P 5 2.8 3 1024 ), similar to the correlation between two independent daf-21 reporters (Fig. 3b, r 5 0.77, P 5 3.4 3 1024 ). This and additional correlations between stress response genes (Supplementary Fig. 14) suggest that, during normal development, isogenic embryos differ in a coordinated manner in the transcriptional induction of multiple chaperones. We quantified induction of a pdaf-21::mCherry reporter in a tbx- 9(ok2473) background. The mean induction of this reporter was increased in tbx-9(ok2473) animals (Supplementary Fig. 9b, 1.8-fold upregulation, P 5 6.6 3 1024 ). In addition, embryos that later hatched without a morphological defect expressed higher levels of the reporter Time (min) *** pdaf-21::mCherry fluorescence intensity at comma stage (a.u.) WT Abnormal L1 stage phenotype pdaf-21::mCherry fluorescence intensity (a.u.) pdaf-21::mCherry fluorescence intensity (a.u.) pdaf-21::mCherry fluorescence intensity (a.u.) pdaf-21::GFP fluorescence intensity (a.u.) phsp-4::GFP fluorescence intensity (a.u.) High Low High Low High Low pdaf-21::mCherry fluorescence intensity Anti-DAF-21 Anti-mCherry Anti-β-actin a b c d e r = 0.77 r = 0.63 tbx-9; pdaf-21::mCherry tbx-9; pdaf-21::mCherry 0 100 200 300 0 1 2 0 1 2 3 3.0 2.5 2.0 1.5 1.0 0.5 0.5 1.0 1.5 1.0 0.5 0.0 0.4 0.6 0.8 1.0 1.2 WT phenotype Abnormal phenotype Figure 3 | Inter-individual variation in chaperone induction predicts the outcome of a mutation. a, pdaf-21::mCherry; glp-1(e2141) animals were sorted into ‘high’ and ‘low’ groups based on their pdaf-21::mCherry expression 24 h after the L1 stage at 25 uC. Western blot analysis confirmed that worms belonging to the ‘high’ group had higher levels of endogenous DAF-21 protein. As a positive control, mCherry protein levels showed the same trend. b-Actin was used as loading control; results of three replica sortings are shown. b, c, Correlation in individuals (measured at comma stage) between: pdaf- 21::mCherry and pdaf-21::GFP (Pearson correlation coefficient, r 5 0.77, P 5 3.43 1028 ) (b); pdaf-21::mCherry and phsp-4::GFP (r 5 0.63, P 5 2.83 1024 ) (c). d, e, Expression of daf-21 reporter in a tbx-9(ok2473) mutant background. Embryos that hatch into phenotypicallyWT worms (blue) have higher expression than those hatching with a morphological defect (red) at the comma stage (P 5 1.93 1023 ). ptbx-8::GFP fluorescence intensity at comma stage (a.u.) pdaf-21::mCherry fluorescence intensity at comma stage (a.u.) a b 0.6 0.8 1.0 1.2 1.4 2.5 2.0 1.5 1.0 0.5 0.0 –0.5 LH HH LL HL Bright field ptbx-8::GFP pdaf-21::mCherry WT (%) 30 68 63 92 LL LH HL HH W Figure 4 | Simultaneous quantification of inter-individual variation in two buffering systems accurately predicts the outcome of a mutation. a, Expression levels of tbx-8 and daf-21 reporters in individual tbx-9(ok2473) mutant embryos at the comma stage of development are not correlated (r 5 0.061, P 5 0.56). Each point represents one individual embryo. Embryos that hatched into phenotypically WT larvae are blue; embryos that hatched with a morphological defect are red. The median values for each reporter are indicated (black lines); embryos with expression above or below the median are defined as high (H) and low (L), respectively. b, Percentage of embryos from each expression class hatching with normal morphology. Representative bright-field, green and red fluorescence images are shown. Prediction scores for different time points and measures are given in Supplementary Table 3. RESEARCH LETTER 252 | NATURE | VOL 480 | 8 DECEMBER 2011 ©2011 Macmillan Publishers Limited. All rights reserved
ETTER RESEARCH early on in development(Fig.3d,e,P=1.9 X 103,see Supplementary 10.Elowitz,M.B.Levine,A.J.,Siggia,E.D.&Swain,P.S.Stochastic gene expression in Table 1).The daf-21 reporter also predicted tbx-8(ok656)mutation a single cell.Science 297,1183-1186(2002). 11.Wernet,M.F.et al.Stochastic spineless expression creates the retinal mosaic for outcome (Supplementary Figs 15 and 16).Higher chaperone expres- colour vision.Nature 440,174-180(2006). sion during early embryonic development therefore predicts a reduced 12.Chang H.H.,Hemberg,M.,Barahona,M.,Ingber,D.E.Huang,S.Transcriptome- effect of the inherited mutation. wide noise controls lineage choice in mammalian progenitor cells.Nature 453, 544-547(2008). We next constructed a strain allowing us to quantify simultaneously 13.Raj,A.,Rifkin,S.A,Andersen,E.van Oudenaarden,A Variability in gene inter-individual variation in the expression of both buffering systems. expression underlies incomplete penetrance.Nature 463,913-918(2010). Expression from the tbx-8 and daf-21 reporters does not correlate 14. Eldar,A.et al.Partial penetrance facilitates developmental evolution in bacteria. across individuals (Fig.4a,r=0.061,P=0.56),showing that the Nature460.510-514(2009). 15.Lehner,B.Genes confer similar robustness to environmental,stochastic,and two buffering systems vary independently.We retrospectively divided genetic perturbations in yeast.PLoS ONE 5,e9035(2010). tbx-9(ok2473)mutation embryos into four approximately equally 16. Andachi,Y.Caenorhabditis elegans T-box genes tbx-9 and tbx-8 are required for populated groups,depending upon whether their expression was formation of hypodermis and body-wall muscle in embryogenesis.Genes Cells 9, 331-344(2004) above or below the median for each of the reporters,and examined 17.Pocock,R.Ahringer,J.Mitsch,M.Maxwell.S.Woollard.AA regulatory network the proportion of individuals hatching as WT larvae in each group of T-box genes and the even-skipped homologue vab-7 controls patterning and (Fig.4b).Whereas 30%of embryos with below-median expression of morphogenesis in C.elegans.Development 131,2373-2385(2004). both reporters hatched without an abnormal phenotype('LL'embryos, 18.Baugh,L R.et al.Synthetic lethal analysis of Caenorhabditis elegans posterior embryonic patterning genes identifies conserved genetic interactions.Genome Fig.4a,b,P=1.7 X 10),68%and 63%of embryos with above- 8iol6,R45(2005) median expression of a single reporter (daf-21 or tbx-8,respectively) 19.Kafri,R.,Bar-Even,A.Pilpel,Y.Transcription control reprogramming in genetic showed no morphological defect(Fig.4b).Strikingly,92%of embryos backup circuits.Nature Genet 37,295-299(2005). 20.DeLuna,A,Springer,M.Kirschner,M.W.Kishony,R.Need-based up-regulation with above median expression of both reporters hatched as pheno- of protein levels in response to deletion of their duplicate genes.PLoS Biol.8, typically WT larvae (HH embryos,Fig.4a,b).That is,non-genetic e1000347(2010). variation in two buffering systems can be almost completely epistatic2 21 Koh,K.Rothman,J.H.ELT-5 and ELT-6 are required continuously to regulate epidermal seam cell differentiation and cell fusion in C.elegans.Development 128. to the tbx-9 mutation.Receiver operating characteristic curve analysis 2867-2880(2001) confirmed the independent contribution of both reporters to pheno- 22.Bateson.W.Mendel's Principles of Heredity(Cambridge Univ.Press,1909). typic predictions (Supplementary Fig.17). 23.Ow,M.C.et al.The FLYWCH transcription factors FLH-1,FLH-2,and FLH-3 repress In summary,we have shown here that incomplete penetrance is not embryonic expression of microRNA genes in C.elegans.Genes Dev.22, just a direct consequence of failure caused by mutationsbut that 2520-2534(2008) 24.Vavouri,T.,Semple,J.I.Lehner,B.Widespread conservation of genetic organisms have plastic compensatory responses in both specific yduring a billion years of eukaryotic evolution.Trends Genet.24, and more promiscuous genetic interaction partners that vary among individuals.It is the combination of this variation that determines the 25.Tischler,J.Lehner,B.Chen,N.Fraser,A.G.Combinatorial RNA interference in Caenorhabditis elegans reveals that redundancy between gene duplicates can be outcome of each mutation in each individual. maintained for more than 80 million years of evolution.Genome Biol7,R69 Inherited mutations also frequently have variable consequences in (2006) other species,including in human disease2.Based on our findings in 26.Lehner,B.Conflict between noise and plasticity in yeast PLoSGenet 6,e1001185 2010). C.elegans,we propose that incomplete genetic compensation may 27.Waddington,C.H.Canalization of de opment and the inheritance of acquired play a role in human disease,influencing the outcome of inherited characters.Nature 150,563-565(1942). polymorphisms in each individual. 28.Bobula,J.et al.Why molecular chaperones buffer mutatio age:a case study with a yeast Hsp40/70 system.Genetics 174,937-944(2006). METHODS SUMMARY 29.Van Dyk,T.K G.A LaRossa RA.Demonstration by genetic su pp ression of interaction of GroE products with many proteins.Nature 342,451-453(1989). The expression from individual developing embryos was quantified using 30.Tokuriki,N.Tawfik,D.S.Chaperonin overexpression promotes genetic variation customized imaging and analysis protocols.Embryos were released from adult and enzyme evolution.Nature 459,668-673(2009). worms by dissection and sorted according to their developmental stage. Supplementary Information is linked to the online version of the paper at Phenotypes were scored at hatching.Reporter constructs were generated by micro www.nature.com/nature. injection or bombardment. Acknowledgements This work was funded by grants from the European Research Full Methods and any associated references are available in the online version of Council,Institucio Catalana de Recerca i Estudis Avancats,Ministerio de Ciencia e the paper at www.nature.com/nature. Innovacion Plan Nacional BFU2008-00365,Agencia de Gestio d'juts Universitarisi de Recerca,ERASysBio+,the European Molecular Biology Organization Young Received 15 February:accepted 21 October 2011. Personal Investigac 1. Badano,J.L.Katsanis,N.Beyond Mendel:an evolving view of human genetic Beatriu de Pinos Fellowship to M.O.C.We thank I.Hope,V.Ambros and S.Kim for disease transmission.Nature Rev.Genet 3,779-789(2002). providing strains.Additional strains were obtained from the Caenorhabditis Genetics 2. Baranzini,S.E.et al Genome,epigenome and RNA sequences of monozygotic Center,which is funded by the National Institutes of Health National Center for twins discordant for multiple sclerosis.Nature 464,1351-1356(2010). Research Resources.We thank T.Zimmermann and R.Garcia from the CRG Advanced 3. Horvitz,H.R.Sulston,J.E.Isolation and genetic characterization of cell-lineage Light Microscopy Unit for advice and assistance,J.Miwa and Y.Yamaguchi for 608F mutants of the nematode Caenorhabditis elegans.Genetics 96,435-454 (1980) antibody,J.Semple for providing complementary DNA clones,A Marchetti and 4 Gartner,K.A third component causing random variability beside environment and R.Garcia-Verdugo for technical assistance,J.Tischler and C.Kiel for advice on genotype.A reason for the limited success of a 30 year long effort to standardize single-molecule fluorescence in situ hybridization and western blotting.respectively, laboratory animals?Lab.Anim.24,71-77(1990). and L.Serrano,M.Isalan and J.Semple for comments on the manuscript. 5.Lehner.B..Crombie.C.Tischler,J.Fortunato,A.Fraser,A.G.Systematic Author Contributions AB.performed all experiments,developed the method and mapping of genetic interactions in Caenorhabditis elegans identifies common analysed the data;M.O.C.demonstrated that increased chaperone activity can modifiers of diverse signaling pathways.Nature Genet 38,896-903(2006). suppress mutation outcome in C.elegans;AB.and B.L designed experiments 6. Costanzo,M.etal.The genetic landscape of a cell.Science 327,425-431 (2010) conceived the model and wrote the manuscript 7 Nowak,M.A Boerlijst,M.C.,Cooke,J.Smith,J.M.Evolution of genetic redundancy.Nature 388,167-171(1997). Author Information Reprints and permissions information is available at 8. Rutherford,S.L.Lindquist,S.Hsp90 as a capacitor for morphological evolution. www.nature.com/reprints.The authors declare no competing financial interests. Nature396.336-342(1998). Readers are welcome to comment on the online version of this article at 9 Queitsch,C.Sangster,T.A.Lindquist,S.Hsp90 as a capacitor of phenotypic www.nature.com/nature.Correspondence and requests for materials should be variation.Nature 417,618-624(2002). addressed to B.L (ben.lehner@crg.eu). 8 DECEMBER 2011I VOL 480 I NATURE I253 2011 Macmillan Publishers Limited.All rights reserved
early on in development (Fig. 3d, e, P 5 1.9 3 1023 , see Supplementary Table 1). The daf-21 reporter also predicted tbx-8(ok656) mutation outcome (Supplementary Figs 15 and 16). Higher chaperone expression during early embryonic development therefore predicts a reduced effect of the inherited mutation. We next constructed a strain allowing us to quantify simultaneously inter-individual variation in the expression of both buffering systems. Expression from the tbx-8 and daf-21 reporters does not correlate across individuals (Fig. 4a, r 5 0.061, P 5 0.56), showing that the two buffering systems vary independently. We retrospectively divided tbx-9(ok2473) mutation embryos into four approximately equally populated groups, depending upon whether their expression was above or below the median for each of the reporters, and examined the proportion of individuals hatching as WT larvae in each group (Fig. 4b). Whereas 30% of embryos with below-median expression of both reporters hatched without an abnormal phenotype (‘LL’ embryos, Fig. 4a, b, P 5 1.7 3 1025 ), 68% and 63% of embryos with abovemedian expression of a single reporter (daf-21 or tbx-8, respectively) showed no morphological defect (Fig. 4b). Strikingly, 92% of embryos with above median expression of both reporters hatched as phenotypically WT larvae (‘HH’ embryos, Fig. 4a, b). That is, non-genetic variation in two buffering systems can be almost completely epistatic22 to the tbx-9 mutation. Receiver operating characteristic curve analysis confirmed the independent contribution of both reporters to phenotypic predictions (Supplementary Fig. 17). In summary, we have shown here that incomplete penetrance is not just a direct consequence of failure caused by mutations13,14, but that organisms have plastic compensatory responses19 in both specific and more promiscuous genetic interaction partners that vary among individuals. It is the combination of this variation that determines the outcome of each mutation in each individual. Inherited mutations also frequently have variable consequences in other species, including in human disease2 . Based on our findings in C. elegans, we propose that incomplete genetic compensation may play a role in human disease, influencing the outcome of inherited polymorphisms in each individual. METHODS SUMMARY The expression from individual developing embryos was quantified using customized imaging and analysis protocols. Embryos were released from adult worms by dissection and sorted according to their developmental stage. Phenotypes were scored at hatching. Reporter constructs were generated by microinjection or bombardment. Full Methods and any associated references are available in the online version of the paper at www.nature.com/nature. Received 15 February; accepted 21 October 2011. 1. Badano, J. L. & Katsanis, N. Beyond Mendel: an evolving view of human genetic disease transmission. Nature Rev. Genet. 3, 779–789 (2002). 2. Baranzini, S. E. et al. Genome, epigenome and RNA sequences of monozygotic twins discordant for multiple sclerosis. Nature 464, 1351–1356 (2010). 3. Horvitz, H. R. & Sulston, J. E. Isolation and genetic characterization of cell-lineage mutants of the nematode Caenorhabditis elegans. Genetics 96, 435–454 (1980). 4. Gartner, K. A third component causing random variability beside environment and genotype. A reason for the limited success of a 30 year long effort to standardize laboratory animals? Lab. Anim. 24, 71–77 (1990). 5. Lehner, B., Crombie, C., Tischler, J., Fortunato, A. & Fraser, A. G. Systematic mapping of genetic interactions in Caenorhabditis elegans identifies common modifiers of diverse signaling pathways. Nature Genet. 38, 896–903 (2006). 6. Costanzo, M. et al. The genetic landscape of a cell. Science 327, 425–431 (2010). 7. Nowak, M. A., Boerlijst, M. C., Cooke, J. & Smith, J. M. Evolution of genetic redundancy. Nature 388, 167–171 (1997). 8. Rutherford, S. L. & Lindquist, S. Hsp90 as a capacitor for morphological evolution. Nature 396, 336–342 (1998). 9. Queitsch, C., Sangster, T. A. & Lindquist, S. Hsp90 as a capacitor of phenotypic variation. Nature 417, 618–624 (2002). 10. Elowitz, M. B., Levine, A. J., Siggia, E. D. & Swain, P. S. Stochastic gene expression in a single cell. Science 297, 1183–1186 (2002). 11. Wernet, M. F. et al. Stochastic spineless expression creates the retinal mosaic for colour vision. Nature 440, 174–180 (2006). 12. Chang, H. H., Hemberg, M., Barahona, M., Ingber, D. E. & Huang, S. Transcriptomewide noise controls lineage choice in mammalian progenitor cells. Nature 453, 544–547 (2008). 13. Raj, A., Rifkin, S. A., Andersen, E. & van Oudenaarden, A. Variability in gene expression underlies incomplete penetrance. Nature 463, 913–918 (2010). 14. Eldar, A. et al. Partial penetrance facilitates developmental evolution in bacteria. Nature 460, 510–514 (2009). 15. Lehner, B. Genes confer similar robustness to environmental, stochastic, and genetic perturbations in yeast. PLoS ONE 5, e9035 (2010). 16. Andachi, Y. Caenorhabditis elegans T-box genes tbx-9 and tbx-8 are required for formation of hypodermis and body-wall muscle in embryogenesis. Genes Cells 9, 331–344 (2004). 17. Pocock, R., Ahringer, J., Mitsch, M., Maxwell, S. & Woollard, A. A regulatory network of T-box genes and the even-skipped homologue vab-7 controls patterning and morphogenesis in C. elegans. Development 131, 2373–2385 (2004). 18. Baugh, L. R. et al. Synthetic lethal analysis of Caenorhabditis elegans posterior embryonic patterning genes identifies conserved genetic interactions. Genome Biol. 6, R45 (2005). 19. Kafri, R., Bar-Even, A. & Pilpel, Y. Transcription control reprogramming in genetic backup circuits. Nature Genet. 37, 295–299 (2005). 20. DeLuna, A., Springer, M., Kirschner, M. W. & Kishony, R. Need-based up-regulation of protein levels in response to deletion of their duplicate genes. PLoS Biol. 8, e1000347 (2010). 21. Koh, K. & Rothman, J. H. ELT-5 and ELT-6 are required continuously to regulate epidermal seam cell differentiation and cell fusion in C. elegans. Development 128, 2867–2880 (2001). 22. Bateson, W. Mendel’s Principles of Heredity (Cambridge Univ. Press, 1909). 23. Ow, M. C. et al. The FLYWCH transcription factors FLH-1, FLH-2, and FLH-3 repress embryonic expression of microRNA genes in C. elegans. Genes Dev. 22, 2520–2534 (2008). 24. Vavouri, T., Semple, J. I. & Lehner, B. Widespread conservation of genetic redundancy during a billion years of eukaryotic evolution. Trends Genet. 24, 485–488 (2008). 25. Tischler, J., Lehner, B., Chen, N. & Fraser, A. G. Combinatorial RNA interference in Caenorhabditis elegans reveals that redundancy between gene duplicates can be maintained for more than 80 million years of evolution. Genome Biol. 7, R69 (2006). 26. Lehner, B. Conflict between noise and plasticity in yeast. PLoS Genet. 6, e1001185 (2010). 27. Waddington, C. H. Canalization of development and the inheritance of acquired characters. Nature 150, 563–565 (1942). 28. Bobula, J. et al. Why molecular chaperones buffer mutational damage: a case study with a yeast Hsp40/70 system. Genetics 174, 937–944 (2006). 29. Van Dyk, T. K. G. A. LaRossa RA. Demonstration by genetic suppression of interaction of GroE products with many proteins. Nature 342, 451–453 (1989). 30. Tokuriki, N. & Tawfik, D. S. Chaperonin overexpression promotes genetic variation and enzyme evolution. Nature 459, 668–673 (2009). Supplementary Information is linked to the online version of the paper at www.nature.com/nature. Acknowledgements This work was funded by grants from the European Research Council, Institucio´ Catalana de Recerca i Estudis Avançats, Ministerio de Ciencia e Innovacio´n Plan Nacional BFU2008-00365, Age`ncia de Gestio´ d’juts Universitaris i de Recerca, ERASysBio1, the European Molecular Biology Organization Young Investigator Programme, the EMBL-CRG Systems Biology Program, by a Formacio´n de Personal Investigador–Ministerio de Ciencia e Innovacio´n fellowship to A.B. and by a Beatriu de Pino´s Fellowship to M.O.C. We thank I. Hope, V. Ambros and S. Kim for providing strains. Additional strains were obtained from the Caenorhabditis Genetics Center, which is funded by the National Institutes of Health National Center for Research Resources. We thank T. Zimmermann and R. Garcı´a from the CRG Advanced Light Microscopy Unit for advice and assistance, J. Miwa and Y. Yamaguchi for 608F antibody, J. Semple for providing complementary DNA clones, A. Marchetti and R. Garcı´a-Verdugo for technical assistance, J. Tischler and C. Kiel for advice on single-molecule fluorescence in situ hybridization and western blotting, respectively, and L. Serrano, M. Isalan and J. Semple for comments on the manuscript. Author Contributions A.B. performed all experiments, developed the method and analysed the data; M.O.C. demonstrated that increased chaperone activity can suppress mutation outcome in C. elegans; A.B. and B.L. designed experiments, conceived the model and wrote the manuscript. Author Information Reprints and permissions information is available at www.nature.com/reprints. The authors declare no competing financial interests. Readers are welcome to comment on the online version of this article at www.nature.com/nature. Correspondence and requests for materials should be addressed to B.L. (ben.lehner@crg.eu). LETTER RESEARCH 8 DECEMBER 2011 | VOL 480 | NATURE | 253 ©2011 Macmillan Publishers Limited. All rights reserved
RESEARCH LETTER METHODS different genotypes)for each time-lapse.For visualization,five time-point Expression profiling:image acquisition.Synchronous populations of tbx- smoothing was used for 10 min acquisition frequency recordings,and three 9(ok2473)animals were obtained by treating worms with sodium hypochlorite time-point smoothing used when the frequency was 25 min and then allowing embryos to develop at 25C.For a typical time-lapse,around Data analysis,statistics and figures were generated using R(http://www.r-project. 30-50 gravid worms were picked and then washed three times with M9 buffer in a org).Receiver operating characteristic curves for dual reporter predictions were watch glass.Worms were then transferred to a 50 mm glass-bottomed culture dish constructed using the ROCR package Expression levels of tbx-8 and daf-21 at (MatTek Corporation).Adults were dissected and released embryos were manu- comma stage were normalized by subtracting the mean and dividing by the standard ally sorted.In a period of time not exceeding 15 min,4-to 12-cell-stage embryos deviation of each population.The sum of the normalized values for each embryo was were collected and carefully placed with an eyelash-pick into the centre ofthe dish; used to calculate the receiver operating characteristic curve of the joint contribution. the rest of embryos and debris were removed using a pipette.A cover slip was Phenotypes.The phenotype of each embryo was scored as WT or abnormal by placed on top of the worms and sealed with paraffin.Fluorescence and bright-field visual inspection of the hatching morphology.Any abnormal morphology was images in multiple positions were acquired with a X10 Plan-NEOFLUAR 0.3 NA defined as non-WT.To study the prediction of mutation expressivity,tbx- objective on a Zeiss Cell Observer HS system consisting of an inverted microscope 9(ok2473)animals that hatched with an abnormal phenotype were further classified (AxioObserver.Z1)equipped with an automated stage,a C9100-13 IMAG-EM into two classes.'Mild'abnormality indicated individuals with a characteristic Dual Mode EM-CCD camera(Hamamatsu)and a Sutter DG-4 fast switching posterior depression but that conserve the general body structure.'Severe'embryos xenon light source (Sutter Instrument Company).Temperature control was were those with severe malformations along most of the posterior axis or the entire achieved using a custom microscope incubator box and temperature control body (Supplementary Fig.19). device developed by the European Molecular Biology Laboratory (http://www. Nematode strains and growth conditions.Worms were grown at 20Con NGM embl-em.de)coupled to a Unichiller cooling unit (Huber Kaltemaschinenbau) plates using Escherichia coli Op50 as a food source unless otherwise noted.The for temperature control below room temperature. strains used in and constructed for this study are listed in Supplementary Table 2. At the start of imaging,approximately 25-30 min had elapsed since the dissec- Wild type was Bristol N2(ref.34).Reporter strains ptbx-8::GFP(UL2520)and pelt- tion of worms.During the first 20 min of the time-lapse no fluorescence illumina- 5:GFP (SD1434)were provided by I.Hope and S.Kim,respectively.The strain tion was applied,to avoid phototoxicity as early embryos are very sensitive to VT1343(flh-1(bc374))was a gift from V.Ambros.Some deletion mutations used illumination.Subsequently bright-field and fluorescence images were obtained in this work were provided by the C.elegans Gene Knockout Consortium. every 10 min for 8h,followed by just bright-field images collected every 2 min We optimized worm growth and embryo imaging conditions for each mutant, for 4 more hours or until most of the worms had hatched.The increase in bright- to obtain a penetrance with a reasonable representation of embryos hatching with field frequency facilitated the phenotypic score ofhatching worms.Exposure times both WT and morphologically abnormal phenotypes.The tbx-9(ok2473)mutants and frequency were adjusted to maximize the signal-to-noise ratio while avoiding were maintained at 25Cand embryos were allowed to develop at 27.5C to give a phototoxicity (evaluated as approximately 100%hatching of WI (N2)embryos). penetrance of approximately50%;tbx-8(ok656)were grown at 16Cand embryos The imaging conditions for the daf-21,let-858,elt-5 and flh-2reporters in a tbx-9 allowed to develop at 20C;flh-1(bc374)mutants were grown at 20C and mutant background were exactly the same as for the tbx-8 reporter.In the other embryos allowed to develop at 26C.For flh-2(bc375)mutants,a growth condition genetic backgrounds exposure times and acquisition characteristics were adjusted was not found where the penetrance was high enough to be usefully studied. according to the characteristics of the reporter construct.The tbx-9 reporter was Reporter gene constructs.We generated transcriptional reporters for the genes expressed earlier than the tbx-8 reporter and at lower levels,and the exposure time tbx-9,fh-2 and daf-21.Promoter regions of these genes were amplified from N2 was increased 2.5X to compensate for this,while reducing the frequency of genomic DNA by PCR using the following primer pairs:FW-ptbx-9, acquisition to once every 25 min.The fh-2 reporter was also expressed at lower 5'-TTGGGTTCAGATAACAATTTGG-3';RV-ptbx-9,5'-ATTTTTTGTCTGA levels,but the induction was later in development.For this reporter the exposure AAACGTTAAAATA-3;FW-pfh-2,5'-GCGCTTCTCGTGGGCTCT-3':RV. time was increased(2.5X compared with that for the tbx-8 reporter),fluorescence pflh-2,5'-ATACAGGCGGTCTGAAAA-3';FW-pdaf-21,5-CGAAACGGTCG acquisition was initiated I h later than bright field,but images were still obtained AATTTCATAA-3';RV-pdaf-21,5'-ATGGTTCTGGAAAAATATCAATTA-3' once every 10 min. The lengths of the amplified regions were 1.5kb,1.5kb and 1.9kb,respectively. When comparing the expression of a particular reporter in two different genetic The promoters were cloned into the MultiSite Gateway Entry vector pENTR backgrounds,embryo isolation was performed in 10 min,one strain immediately 5'-TOPO following the manufacturer's instructions (Invitrogen)and then trans- after the other(total time of 20 min),and image acquisition was performed simul- ferred by three-fragment MultiSite Gateway Pro LR reaction(Invitrogen)into the taneously using the conditions previously established for each strain. CFJ150-pDESTttTi5605[R4-R3]destination vector.Middle Entry vectors used Expression profiling:quantification of gene expression.Total fluorescence were pENTRwGFP or pENTRmCherry;3'Entry vector was pCM5.37 containing from each embryo was quantified using Image](http://rsb.info.nih.gov/ij/) the unc-54 3'untranslated region (UTR).Final constructs were confirmed by PCR First,we performed a global background correction of all images by subtracting amplification and sequencing. an image taken in a region of no fluorescence for each time point.For the pelt- Strain construction.Strains carrying an extrachromosomal array expressing ptbx- 5:mCherry reporter data set,an additional flat field correction step was necessary. 8:TBX-8::unc-54 3'UTR or ptbx-9::TBX-9::unc-54 3'UTR were generated by Background-subtracted images were divided by the homogeneous fluorescence microinjection as previously describedss using pmyo-2::mCherry as a co-injection reference to correct for non-homogeneous illumination,and the resulting images marker.The transcriptional reporter ptbx-9:GFP was generated by microinjection used for quantification.The homogeneous fluorescent reference was created by using pmyo-2::mCherry as a co-injection marker,followed by integration using the imaging a drop of 0.05mgml Nile Red solution in M9 buffer.All embryos were altraviolet irradiation method.The transcriptional reporters pdaf-21:mCherry, visually identified and tagged so that no embryo was counted twice.A region of pdaf-21:GFP,phsp-4::mCherry,ptbx-8::mCherry and pflh-2::GFP were generated by interest(ROI)was manually drawn around each embryo and an equally sized ROI bombardment in an unc-119(ed3)background37.All transgenic strains were back- drawn in the background vicinity;finally,the total fluorescence from each embryo crossed at least four times.No transgenic strains displayed abnormal phenotypes. was calculated as the integrated density difference between the embryo containing Quantitative PCR.Worms carrying the pdaf-21::mCherry transgene were syn- ROIand the localbackground ROI.For the particular case of the fh-2reporter,at the chronized by hypochlorite treatment,plated as LI larvae and allowed to develop 100-cell stage most of the reporter signal came from a very restricted area of the for 24h at 25C.Larvae were then sorted by hand using a MVX10 Macro Zoom embryo.Because embryo background fluorescence was masking this specific source fluorescence microscope (Olympus)into 'high'and low'groups based on their of expression,we quantified it by first drawing an equally sized circle-shaped ROI in mCherry fluorescence levels.Each group had approximately 20 larvae and sorting all embryos around the signal source and measuring the integrated density.This was performed in triplicate.Worms were washed in M9 and transferred to 1.5 ml quantified expression was used for fh-1 phenotypic prediction. Eppendorf tubes containing 200 ul of lysis Buffer(50 mM KCl,10 mM Tris-HCl All embryos were staged using three developmental landmarks:(1)number of (pH 8),2.5 nM MgCl,0.45%Nonidet P-40,0.45%Tween20)supplemented with cells at beginning of time-lapse,(2)elapsed time until comma stage and (3)elapsed 20 ug proteinase K and 20 ug RNase A.Tubes were frozen in liquid nitrogen, time until twofold-stage.Only those embryos starting the time-lapse at the followed by incubation at 65C for 1h and 95C for 30 min.Genomic DNA approximately 6-to 15-cell stage and reaching the next development landmarks was extracted by adding 500 ul Phenol:Chloroform:Isoamyl Alcohol 25:24:1 with a delay ofless than or equal to 30 min were considered for further analysis.In (Sigma)and precipitated using 600 ul of ice cold ethanol and 50ul of 3 M a typical time-lapse,around 10-15 embryos fulfilled those requirements.Data for NaAc.Quantitative PCR was performed on a LightCycler 480 machine with each reporter-mutant pair prediction included embryos from at least four inde- SYBR green detection(Roche).The transgene was amplified with primers target- pendent experiments.To correct for day-to-day technical variation(Supplemen- ing the mCherry coding region:FW-mCherry,5'-CTACGACGCTGAGGTC tary Fig.18),the expression of each embryo was normalized using the mean AAGA-3';RV-mCherrry,5'-CGATGGTGTAGTCCTCGTTG-3'.Transgene expression of the population (or populations when comparing the expression of levels were normalized by the mean level of two other loci in the genome using 2011 Macmillan Publishers Limited.All rights reserved
METHODS Expression profiling: image acquisition. Synchronous populations of tbx- 9(ok2473) animals were obtained by treating worms with sodium hypochlorite and then allowing embryos to develop at 25 uC. For a typical time-lapse, around 30–50 gravid worms were picked and then washed three times with M9 buffer in a watch glass. Worms were then transferred to a 50 mm glass-bottomed culture dish (MatTek Corporation). Adults were dissected and released embryos were manually sorted. In a period of time not exceeding 15 min, 4- to 12-cell-stage embryos were collected and carefully placed with an eyelash-pick into the centre of the dish; the rest of embryos and debris were removed using a pipette. A cover slip was placed on top of the worms and sealed with paraffin. Fluorescence and bright-field images in multiple positions were acquired with a 310 Plan-NEOFLUAR 0.3 NA objective on a Zeiss Cell Observer HS system consisting of an inverted microscope (AxioObserver.Z1) equipped with an automated stage, a C9100-13 IMAG-EM Dual Mode EM-CCD camera (Hamamatsu) and a Sutter DG-4 fast switching xenon light source (Sutter Instrument Company). Temperature control was achieved using a custom microscope incubator box and temperature control device developed by the European Molecular Biology Laboratory (http://www. embl-em.de) coupled to a Unichiller cooling unit (Huber Ka¨ltemaschinenbau) for temperature control below room temperature. At the start of imaging, approximately 25–30 min had elapsed since the dissection of worms. During the first 20 min of the time-lapse no fluorescence illumination was applied, to avoid phototoxicity as early embryos are very sensitive to illumination. Subsequently bright-field and fluorescence images were obtained every 10 min for 8 h, followed by just bright-field images collected every 2 min for 4 more hours or until most of the worms had hatched. The increase in brightfield frequency facilitated the phenotypic score of hatching worms. Exposure times and frequency were adjusted to maximize the signal-to-noise ratio while avoiding phototoxicity (evaluated as approximately 100% hatching of WT (N2) embryos). The imaging conditions for the daf-21, let-858,elt-5 and flh-2 reporters in a tbx-9 mutant background were exactly the same as for the tbx-8 reporter. In the other genetic backgrounds exposure times and acquisition characteristics were adjusted according to the characteristics of the reporter construct. The tbx-9 reporter was expressed earlier than the tbx-8 reporter and at lower levels, and the exposure time was increased 2.53 to compensate for this, while reducing the frequency of acquisition to once every 25 min. The flh-2 reporter was also expressed at lower levels, but the induction was later in development. For this reporter the exposure time was increased (2.53 compared with that for the tbx-8 reporter), fluorescence acquisition was initiated 1 h later than bright field, but images were still obtained once every 10 min. When comparing the expression of a particular reporter in two different genetic backgrounds, embryo isolation was performed in 10 min, one strain immediately after the other (total time of 20 min), and image acquisition was performed simultaneously using the conditions previously established for each strain. Expression profiling: quantification of gene expression. Total fluorescence from each embryo was quantified using ImageJ31 (http://rsb.info.nih.gov/ij/). First, we performed a global background correction of all images by subtracting an image taken in a region of no fluorescence for each time point. For the pelt- 5::mCherry reporter data set, an additional flat field correction step was necessary. Background-subtracted images were divided by the homogeneous fluorescence reference to correct for non-homogeneous illumination, and the resulting images used for quantification32. The homogeneous fluorescent reference was created by imaging a drop of 0.05 mg ml21 Nile Red solution in M9 buffer. All embryos were visually identified and tagged so that no embryo was counted twice. A region of interest (ROI) was manually drawn around each embryo and an equally sized ROI drawn in the background vicinity; finally, the total fluorescence from each embryo was calculated as the integrated density difference between the embryo containing ROI and the local background ROI. For the particular case of the flh-2 reporter, at the 100-cell stage most of the reporter signal came from a very restricted area of the embryo. Because embryo background fluorescence was masking this specific source of expression, we quantified it by first drawing an equally sized circle-shaped ROI in all embryos around the signal source and measuring the integrated density. This quantified expression was used for flh-1 phenotypic prediction. All embryos were staged using three developmental landmarks: (1) number of cells at beginning of time-lapse, (2) elapsed time until comma stage and (3) elapsed time until twofold-stage. Only those embryos starting the time-lapse at the approximately 6- to 15-cell stage and reaching the next development landmarks with a delay of less than or equal to 30 min were considered for further analysis. In a typical time-lapse, around 10–15 embryos fulfilled those requirements. Data for each reporter–mutant pair prediction included embryos from at least four independent experiments. To correct for day-to-day technical variation (Supplementary Fig. 18), the expression of each embryo was normalized using the mean expression of the population (or populations when comparing the expression of different genotypes) for each time-lapse. For visualization, five time-point smoothing was used for 10 min acquisition frequency recordings, and three time-point smoothing used when the frequency was 25 min. Data analysis, statistics and figures were generated using R (http://www.r-project. org). Receiver operating characteristic curves for dual reporter predictions were constructed using the ROCR package33. Expression levels of tbx-8 and daf-21 at comma stage were normalized by subtracting the mean and dividing by the standard deviation of each population. The sum of the normalized values for each embryo was used to calculate the receiver operating characteristic curve of the joint contribution. Phenotypes. The phenotype of each embryo was scored as WT or abnormal by visual inspection of the hatching morphology. Any abnormal morphology was defined as non-WT. To study the prediction of mutation expressivity, tbx- 9(ok2473) animalsthat hatched with an abnormal phenotype were further classified into two classes. ‘Mild’ abnormality indicated individuals with a characteristic posterior depression but that conserve the general body structure. ‘Severe’ embryos were those with severe malformations along most of the posterior axis or the entire body (Supplementary Fig. 19). Nematode strains and growth conditions. Worms were grown at 20 uC on NGM plates using Escherichia coli OP50 as a food source unless otherwise noted. The strains used in and constructed for this study are listed in Supplementary Table 2. Wild type was Bristol N2 (ref. 34). Reporter strains ptbx-8::GFP (UL2520) and pelt- 5::GFP (SD1434) were provided by I. Hope and S. Kim, respectively. The strain VT1343 (flh-1(bc374)) was a gift from V. Ambros. Some deletion mutations used in this work were provided by the C. elegans Gene Knockout Consortium. We optimized worm growth and embryo imaging conditions for each mutant, to obtain a penetrance with a reasonable representation of embryos hatching with both WT and morphologically abnormal phenotypes. The tbx-9(ok2473) mutants were maintained at 25 uC and embryos were allowed to develop at 27.5 uC to give a penetrance of approximately 50%; tbx-8(ok656) were grown at 16 uC and embryos allowed to develop at 20 uC; flh-1(bc374) mutants were grown at 20 uC and embryos allowed to develop at 26 uC. Forflh-2(bc375) mutants, a growth condition was not found where the penetrance was high enough to be usefully studied. Reporter gene constructs. We generated transcriptional reporters for the genes tbx-9, flh-2 and daf-21. Promoter regions of these genes were amplified from N2 genomic DNA by PCR using the following primer pairs: FW-ptbx-9, 59-TTGGGTTCAGATAACAATTTGG-39; RV-ptbx-9, 59-ATTTTTTGTCTGA AAACGTTAAAATA-39; FW-pflh-2, 59-GCGCTTCTCGTGGGCTCT-39; RVpflh-2, 59-ATACAGGCGGTCTGAAAA-39; FW-pdaf-21, 59-CGAAACGGTCG AATTTCATAA-39; RV-pdaf-21, 59-ATGGTTCTGGAAAAATATCAATTA-39. The lengths of the amplified regions were 1.5 kb, 1.5 kb and 1.9 kb, respectively. The promoters were cloned into the MultiSite Gateway Entry vector pENTR 59-TOPO following the manufacturer’s instructions (Invitrogen) and then transferred by three-fragment MultiSite Gateway Pro LR reaction (Invitrogen) into the pCFJ150-pDESTttTi5605[R4-R3] destination vector. Middle Entry vectors used were pENTRwGFP or pENTRmCherry; 39 Entry vector was pCM5.37 containing the unc-54 39 untranslated region (UTR). Final constructs were confirmed by PCR amplification and sequencing. Strain construction. Strains carrying an extrachromosomal array expressing ptbx- 8::TBX-8::unc-54 39 UTR or ptbx-9::TBX-9::unc-54 39 UTR were generated by microinjection as previously described35 using pmyo-2::mCherry as a co-injection marker. The transcriptional reporter ptbx-9::GFP was generated by microinjection using pmyo-2::mCherry as a co-injection marker, followed by integration using the ultraviolet irradiation method36. The transcriptional reporters pdaf-21::mCherry, pdaf-21::GFP, phsp-4::mCherry, ptbx-8::mCherry and pflh-2::GFPwere generated by bombardment in anunc-119(ed3) background37,38. All transgenic strains were backcrossed at least four times. No transgenic strains displayed abnormal phenotypes. Quantitative PCR. Worms carrying the pdaf-21::mCherry transgene were synchronized by hypochlorite treatment, plated as L1 larvae and allowed to develop for 24 h at 25 uC. Larvae were then sorted by hand using a MVX10 Macro Zoom fluorescence microscope (Olympus) into ‘high’ and ‘low’ groups based on their mCherry fluorescence levels. Each group had approximately 20 larvae and sorting was performed in triplicate. Worms were washed in M9 and transferred to 1.5 ml Eppendorf tubes containing 200 ml of lysis Buffer (50 mM KCl, 10 mM Tris-HCl (pH 8), 2.5 nM MgCl2, 0.45% Nonidet P-40, 0.45% Tween20) supplemented with 20 mg proteinase K and 20 mg RNase A. Tubes were frozen in liquid nitrogen, followed by incubation at 65 uC for 1 h and 95 uC for 30 min. Genomic DNA was extracted by adding 500 ml Phenol:Chloroform:Isoamyl Alcohol 25:24:1 (Sigma) and precipitated using 600 ml of ice cold ethanol and 50 ml of 3 M NaAc. Quantitative PCR was performed on a LightCycler 480 machine with SYBR green detection (Roche). The transgene was amplified with primers targeting the mCherry coding region: FW-mCherry, 59-CTACGACGCTGAGGTC AAGA-39; RV-mCherrry, 59-CGATGGTGTAGTCCTCGTTG-39. Transgene levels were normalized by the mean level of two other loci in the genome using RESEARCH LETTER ©2011 Macmillan Publishers Limited. All rights reserved
LETTER RESEARCH the following primers:FW-daf-21ORF,5'-TCCAATGACTGGGAAGATCA-3'; Embryos were staged by counting the total number of nuclei stained with RV-daf-21ORF,5'-CGAACGTAGAGCTTGATGGA-3';FW-csq-1,5'-AACTG 4',6-diamidino-2-phenylindole(DAPI). AGGTTCTGACCGAGAAG-3';RW-csq-1,5'-TACTGGTCAAGCTCTGAGT Rescue experiments.Worms mutant for tbx-9 were crossed with a line carrying CGTC-3 an extrachromosomal array overexpressing TBX-8 protein and a pmyo. Western blotting.To test whether expression of the pdaf-21::mCherry predicted 2:mCherry reporter as a marker.Adult gravid worms growing at 25C were differences in the levels of DAF-21 protein expressed from the endogenous gene, manually sorted under an MVX10 Macro Zoom fluorescence microscope pdaf-21::mCherry:glp-1(e2141)animals were synchronized as LI larvae by hypo- (Olympus)into two groups,one carrying the fluorescent marker,the other lacking chlorite treatment and starvation at 20C overnight.Worms were transferred to it.This latter group served as an internal negative control.Extrachromosomal plates and allowed to develop for 24h at 25C (glp-1(e2141)mutation restrictive array transmission was around approximately 50%,meaning that not all embryos temperature;this mutation prevents the development of the germ line;the pdaf- in the overexpression group actually carried the transgene,so rescue scores are 21:mCherry reporter does not express in germline cells).Larvae were then sorted probably an underestimation.Four-to 12-cell stage embryos were collected in using an MVX10 Macro Zoom fluorescence microscope (Olympus)into high 50 mm glass-bottomed dishes following the same protocol as for time-lapses,and and low'groups based on their mCherry fluorescence.Each group had approxi- incubated at 27.5C.Phenotypes of worms were scored 15h later.In the case of mately 30 larvae and sorting was performed in triplicate.Worms were washed in M9 buffer,frozen in liquid nitrogen,dissolved in 30 ul of SDS loading buffer, tbx-8 mutation rescue,worms were grown at 16C and incubated at 20C. Control experiments were performed crossing the tbx-9(ok2473)and tbx- sonicated for 5 min at 4C and then boiled at 95C for 5 min.Samples were loaded into a 12.5%SDS polyacrylamide gel and transferred to a nitrocellulose 8(ok656)mutations to a strain carrying the extrachromosomal array membrane.Primary antibodies were incubated overnight at 4C.The following wwEx26[pmir-42-44:GFP+unc-119(+)].This transcriptional reporter is antibodies and dilutions were used:rabbit polyclonal anti-B-actin (ab8227) expressed at high levels during early and late development,and its transmission (Abcam)1:1,000;rabbit polyclonal anti-DsRed (ClonTech)1:1,000;mouse is approximately 95%.Rescue experiments were performed at least in triplicate for each pair.Finally,repeats were pooled and significance was tested using Fisher's monoclonal anti-DAF-21 608F hybridoma supernatant(a gift from J.Miwa and Y.Yamaguchi)1:40;anti-rabbit IgG peroxidase conjugate (Jackson exact test. ImmunoResearch)1:40,000 and anti-mouse IgG peroxidase conjugate (Sigma- 31.Abramoff,M.D.Magelhaes,P.J.Ram,S.J.Image processing with ImageJ. Aldrich)1:8,000.Quantification of DAF-21 levels in N2 WT and daf-21(+/-) 8 iophoton.lnt11,36-42(2004). was performed in a similar fashion;30 L4 larvae of each genotype were picked in 32.Wolf,D.E.,Samarasekera.C.Swedlow,J.R.Quantitative analysis of digital triplicate.We detected peroxidase activity with SuperSignal West Femto Substrate microscope images.Methods Cell Biol.81,365-396(2007). (Thermo Scientific)and blot imaging was performed using a Fujifilm LAS-3000 33.Sing.T.,Sander,O.,Beerenwinkel,N.Lengauer,T.ROCR:visualizing classifier performance in R.Bioinformatics 21,3940-3941(2005). luminescent image analyser with exposures times in the lineal range of detection. 34.Brenner,S.The genetics of Caenorhabditis elegans.Genetics 77,71-94(1974). Finally,band intensities were quantified using ImageJ. 35.Kelly,W.G.,Xu,S.,Montgomery,M.K.Fire,A.Distinct requirements for somatic Single molecule fluorescence in situ hybridization.We followed the protocol of and germline expression of a generally expressed Caemorhabditis elegans gene. Raj et al40 with the following modifications.Probes against tbx-8 mRNA were Genetics146,227-238(199万. generated with the online tool available at www.singlemoleculefish.com and syn- 36.Mitani,S.Genetic regulation of mec-3 gene expression implicated in the thesized by Biosearch Technologies.Probes were coupled to CAL Fluor Red 590 specification of the mechanosensory neuron cell types in Caenorhabditis elegans. fluorophore.The formamide concentration in hybridization and wash buffer was Dev.Growth Differ.37,551-557 (2003). 37.Praitis,V.Casey,E..Collar,D.Austin,J.Creation of low-copy integrated 20%and the total concentration of pooled probes was 25 nM.Hybridization was performed overnight at 30C.We imaged embryos using an oil immersion X100 transgenic lines in Caenorhabditis elegans.Genetics 157,1217-1226(2001) 38.Wilm,T.Demel,P.Koop,H.U.,Schnabel,H.Schnabel,R.Ballistic transformation objective on a Leica DMI4000 inverted microscope equipped with an Evolve 512 of Caenorhabditis elegans.Gene 229,31-35(1999). EMCCD camera (Photometrics)and a Lumen 200 metal arc lamp (Prior 39.Yamaguchi,Y.Murakami,K.,Furusawa,M.Miwa,J.Germline-specific antigens Scientific).We collected stacks of typically 25 images spaced by 0.3 um for each identified by monoclonal antibodies in the nematode Caenorhabditis elegans.Dev. embryo.Stacks were processed applying a Laplacian of Gaussian filter with the Growth Differ.25.121-131(1983). LoG3D plugin"for Imagel.Then dimensionality was reduced by applying a 40.Raj.A,van den Bogaard,P.Rifkin,S.A,van Oudenaarden,A.Tyagi,S.Imaging Z-maximum intensity projection(the final image contains the maximum value individual mRNA molecules using multiple singly labeled probes.Nature Methods 5,877-879(2008. over all images in the stack at the particular pixel location),and a single threshold 41.Sage,D.Neumann,F.R.,Hediger,F.,Gasser,S.M.Unser,M.Automatic tracking was applied to each embryo by visual inspection as recommended.The total of individual fluorescence particles:application to the study of chromosome number of particles was counted with the 'Analyze particles'Imagel tool dynamics.IEEE Trans.Image Process.14,1372-1383(2005). 2011 Macmillan Publishers Limited.All rights reserved
the following primers: FW-daf-21ORF, 59-TCCAATGACTGGGAAGATCA-39; RV-daf-21ORF, 59-CGAACGTAGAGCTTGATGGA-39; FW-csq-1, 59-AACTG AGGTTCTGACCGAGAAG-39; RW-csq-1, 59-TACTGGTCAAGCTCTGAGT CGTC-39. Western blotting. To test whether expression of the pdaf-21::mCherry predicted differences in the levels of DAF-21 protein expressed from the endogenous gene, pdaf-21::mCherry; glp-1(e2141) animals were synchronized as L1 larvae by hypochlorite treatment and starvation at 20 uC overnight. Worms were transferred to plates and allowed to develop for 24 h at 25 uC (glp-1(e2141) mutation restrictive temperature; this mutation prevents the development of the germ line; the pdaf- 21::mCherry reporter does not express in germline cells). Larvae were then sorted using an MVX10 Macro Zoom fluorescence microscope (Olympus) into ‘high’ and ‘low’ groups based on their mCherry fluorescence. Each group had approximately 30 larvae and sorting was performed in triplicate. Worms were washed in M9 buffer, frozen in liquid nitrogen, dissolved in 30 ml of SDS loading buffer, sonicated for 5 min at 4 uC and then boiled at 95 uC for 5 min. Samples were loaded into a 12.5% SDS polyacrylamide gel and transferred to a nitrocellulose membrane. Primary antibodies were incubated overnight at 4 uC. The following antibodies and dilutions were used: rabbit polyclonal anti-b-actin (ab8227) (Abcam) 1:1,000; rabbit polyclonal anti-DsRed (ClonTech) 1:1,000; mouse monoclonal anti-DAF-21 608F hybridoma supernatant39 (a gift from J. Miwa and Y. Yamaguchi) 1:40; anti-rabbit IgG peroxidase conjugate (Jackson ImmunoResearch) 1:40,000 and anti-mouse IgG peroxidase conjugate (SigmaAldrich) 1:8,000. Quantification of DAF-21 levels in N2 WT and daf-21(1/2) was performed in a similar fashion; 30 L4 larvae of each genotype were picked in triplicate. We detected peroxidase activity with SuperSignal West Femto Substrate (Thermo Scientific) and blot imaging was performed using a Fujifilm LAS-3000 luminescent image analyser with exposures times in the lineal range of detection. Finally, band intensities were quantified using ImageJ. Single molecule fluorescence in situ hybridization. We followed the protocol of Raj et al.40 with the following modifications. Probes against tbx-8 mRNA were generated with the online tool available at www.singlemoleculefish.com and synthesized by Biosearch Technologies. Probes were coupled to CAL Fluor Red 590 fluorophore. The formamide concentration in hybridization and wash buffer was 20% and the total concentration of pooled probes was 25 nM. Hybridization was performed overnight at 30 uC. We imaged embryos using an oil immersion 3100 objective on a Leica DMI4000 inverted microscope equipped with an Evolve 512 EMCCD camera (Photometrics) and a Lumen 200 metal arc lamp (Prior Scientific). We collected stacks of typically 25 images spaced by 0.3 mm for each embryo. Stacks were processed applying a Laplacian of Gaussian filter with the LoG3D plugin41 for ImageJ. Then dimensionality was reduced by applying a Z-maximum intensity projection (the final image contains the maximum value over all images in the stack at the particular pixel location), and a single threshold was applied to each embryo by visual inspection as recommended. The total number of particles was counted with the ‘Analyze particles’ ImageJ tool. Embryos were staged by counting the total number of nuclei stained with 49,6-diamidino-2-phenylindole (DAPI). Rescue experiments. Worms mutant for tbx-9 were crossed with a line carrying an extrachromosomal array overexpressing TBX-8 protein and a pmyo- 2::mCherry reporter as a marker. Adult gravid worms growing at 25 uC were manually sorted under an MVX10 Macro Zoom fluorescence microscope (Olympus) into two groups, one carrying the fluorescent marker, the other lacking it. This latter group served as an internal negative control. Extrachromosomal array transmission was around approximately 50%, meaning that not all embryos in the overexpression group actually carried the transgene, so rescue scores are probably an underestimation. Four- to 12-cell stage embryos were collected in 50 mm glass-bottomed dishes following the same protocol as for time-lapses, and incubated at 27.5 uC. Phenotypes of worms were scored 15 h later. In the case of tbx-8 mutation rescue, worms were grown at 16 uC and incubated at 20 uC. Control experiments were performed crossing the tbx-9(ok2473) and tbx- 8(ok656) mutations to a strain carrying the extrachromosomal array wwEx26[pmir-42-44::GFP1unc-119(1)]. This transcriptional reporter is expressed at high levels during early and late development, and its transmission is approximately 95%. Rescue experiments were performed at least in triplicate for each pair. Finally, repeats were pooled and significance was tested using Fisher’s exact test. 31. Abramoff, M. D., Magelhaes, P. J. & Ram, S. J. Image processing with ImageJ. Biophoton. Int. 11, 36–42 (2004). 32. Wolf, D. E., Samarasekera, C. & Swedlow, J. R. Quantitative analysis of digital microscope images. Methods Cell Biol. 81, 365–396 (2007). 33. Sing, T., Sander, O., Beerenwinkel, N. & Lengauer, T. ROCR: visualizing classifier performance in R. Bioinformatics 21, 3940–3941 (2005). 34. Brenner, S. The genetics of Caenorhabditis elegans. Genetics 77, 71–94 (1974). 35. Kelly, W. G., Xu, S., Montgomery, M. K. & Fire, A. Distinct requirements for somatic and germline expression of a generally expressed Caernorhabditis elegans gene. Genetics 146, 227–238 (1997). 36. Mitani, S. Genetic regulation of mec-3 gene expression implicated in the specification of the mechanosensory neuron cell types in Caenorhabditis elegans. Dev. Growth Differ. 37, 551–557 (2003). 37. Praitis, V., Casey, E., Collar, D. & Austin, J. Creation of low-copy integrated transgenic lines in Caenorhabditis elegans. Genetics 157, 1217–1226 (2001). 38. Wilm, T., Demel, P., Koop, H. U., Schnabel, H. & Schnabel, R. Ballistic transformation of Caenorhabditis elegans. Gene 229, 31–35 (1999). 39. Yamaguchi, Y., Murakami, K., Furusawa, M. & Miwa, J. Germline-specific antigens identified by monoclonal antibodies in the nematode Caenorhabditis elegans. Dev. Growth Differ. 25, 121–131 (1983). 40. Raj, A., van den Bogaard, P., Rifkin, S. A., van Oudenaarden, A. & Tyagi, S. Imaging individual mRNA molecules using multiple singly labeled probes. Nature Methods 5, 877–879 (2008). 41. Sage, D., Neumann, F. R., Hediger, F., Gasser, S. M. & Unser, M. Automatic tracking of individual fluorescence particles: application to the study of chromosome dynamics. IEEE Trans. Image Process. 14, 1372–1383 (2005). LETTER RESEARCH ©2011 Macmillan Publishers Limited. All rights reserved