PDP-1 Links the TGF-β and IIS Pathways to Regulate Longevity, Development, and Metabolism

The insulin/IGF-1 signaling (IIS) pathway is a conserved regulator of longevity, development, and metabolism. In Caenorhabditis elegans IIS involves activation of DAF-2 (insulin/IGF-1 receptor tyrosine kinase), AGE-1 (PI 3-kinase), and additional downstream serine/threonine kinases that ultimately phosphorylate and negatively regulate the single FOXO transcription factor homolog DAF-16. Phosphatases help to maintain cellular signaling homeostasis by counterbalancing kinase activity. However, few phosphatases have been identified that negatively regulate the IIS pathway. Here we identify and characterize pdp-1 as a novel negative modulator of the IIS pathway. We show that PDP-1 regulates multiple outputs of IIS such as longevity, fat storage, and dauer diapause. In addition, PDP-1 promotes DAF-16 nuclear localization and transcriptional activity. Interestingly, genetic epistasis analyses place PDP-1 in the DAF-7/TGF-β signaling pathway, at the level of the R-SMAD proteins DAF-14 and DAF-8. Further investigation into how a component of TGF-β signaling affects multiple outputs of IIS/DAF-16, revealed extensive crosstalk between these two well-conserved signaling pathways. We find that PDP-1 modulates the expression of several insulin genes that are likely to feed into the IIS pathway to regulate DAF-16 activity. Importantly, dysregulation of IIS and TGF-β signaling has been implicated in diseases such as Type 2 Diabetes, obesity, and cancer. Our results may provide a new perspective in understanding of the regulation of these pathways under normal conditions and in the context of disease.

The PTEN lipid phosphatase homolog DAF-18, which antagonizes signaling at the level of AGE-1/PI 3-kinase, is a major negative regulator of IIS. In contrast to the kinases in this pathway, loss-of-function mutations in daf-18 reduces lifespan, fat storage, dauer formation and stress resistance [32,[34][35][36][37][38][39]. Besides DAF-18, few negative modulators of the pathway have been identified. In particular, less is known about serine/threonine phosphatases that counterbalance kinase activity in the IIS pathway. We recently performed a directed RNA interference (RNAi) screen for serine/threonine phosphatases that regulate C. elegans IIS using dauer formation as an output [39]. We identified the PP2A regulatory subunit PPTR-1 as an important regulator of AKT-1 dephosphorylation as well as DAF-16-dependent phenotypes [39]. Here we characterize another candidate from this screen, pdp-1, as a positive regulator of dauer formation. PDP-1 is homologous to pyruvate dehydrogenase phosphatase (PDP) in higher organisms, an enzyme that positively regulates the pyruvate dehydrogenase enzyme complex (PDHc). RNAi of the other components of PDHc do not result in changes in dauer formation. Interestingly, we report that although PDP-1 is a robust modulator of multiple IIS-regulated processes as well as DAF-16 activity, genetic epistasis studies place pdp-1 in the DAF-7/TGF-b pathway. Through this study, we find that IIS and TGF-b signaling are more tightly connected than previously suggested, with distinct roles for the Co-SMAD DAF-3 in modulating the IIS pathway. Our data suggests that PDP-1 modulates the gene expression of several insulins, and that insulins may be a potential mediator of the crosstalk between these two pathways.

Results
C. elegans PDP-1 regulates daf-2 dauer formation independent of PDH Our RNAi screen was designed to identify serine/threonine phosphatases that modulated dauer formation of daf-2(e1370), a non-null, temperature-sensitive mutant of the C. elegans insulin/ IGF-1 receptor gene, daf-2 [39]. We were particularly interested in phosphatases that would negatively regulate IIS similar to DAF-18/PTEN, and for all RNAi based assays described below, daf-18 RNAi was used as a positive control [39]. From this screen, we identified pdp-1 as a modulator of daf-2(e1370) dauer formation ( Figure 1A and Figure S2). BLAST analyses using amino acid sequence revealed that PDP-1 is homologous to fly and mammalian PDP (,52% positive and ,38% identical). pdp-1 RNAi significantly reduces dauer formation of daf-2(e1370) worms, similar to daf-18 RNAi ( Figure 1A and Figure S2). This phenotype is not allele-specific, as pdp-1 RNAi results in suppression of dauer formation in a second allele of daf-2, daf-2(e1368) ( Figure 1B and Figure S2). Similar to the results with the RNAi, a mutation in pdp-1 also affects dauer formation -pdp-1(tm3734); daf-2(e1370) double mutants form significantly fewer dauers when compared to the daf-2(e1370) parental strain ( Figure S2).
Given its homology to PDP in higher organisms, we wondered whether the effect of pdp-1 knockdown on daf-2 dauer formation was a consequence of modulating the activity of the PDHc. The PDHc is a multi-subunit enzyme complex consisting of three major enzymes: E1 pyruvate dehydrogenase, E2 dihydrolipoyl acetyltransferase and E3 dihydrolipoyl dehydrogenase that regulate energy metabolism [40]. PDHc converts pyruvate to acetyl-coA, which can either enter the TCA cycle or be used for fatty acid synthesis. In mammals, regulation of PDHc activity is primarily achieved through reversible phosphorylation/dephosphorylation of the E1a subunit by pyruvate dehydrogenase kinase (PDHK) and PDP, with phosphorylation inactivating the enzyme complex [40]. All of the components of the PDH complex have conserved C. elegans homologs, encoded by the genes T05H10.6 (E1a), C04C3.3 (E1b), F23B12.5 (E2), LLC1.3 (E3), pdhk-2 (PDHK) and pdp-1 (PDP).
To test whether modulation of PDHc activity affects daf-2 dauer formation, we grew daf-2(e1370) worms on PDHc RNAi. Quantification the RNAi efficiency of the PDHc components revealed that we achieved 60-90% knockdown ( Figure S1). To our surprise, RNAi of the E1a subunit had no effect on daf-2 dauer formation, while pdp-1 RNAi resulted in dauer suppression ( Figure 1C and Figure S2). In addition, RNAi of either the other E1 subunit E1b, or the E2 subunit, did not affect daf-2 dauer formation ( Figure 1C and Figure S2). Knockdown of the E3 subunit resulted in lethality (data not shown). Interestingly, pdhk-2 RNAi resulted in slight suppression daf-2(e1370) dauer formation but had no effect on dauer formation of daf-2(e1368) mutants ( Figure 1C and Figure S2). Therefore pdhk-2 modulates the IIS pathway in an allele-specific manner and we did not perform further characterization of this gene.
To further evaluate the components of the PDH complex, we examined their expression patterns. The expression pattern of PDP-1 does not completely overlap with that of the E1 and E2 subunits of PDHc ( Figure S3). PDP-1 expression was enriched in the head and tail neurons, head muscle and the intestine. We did not observe any expression in the pharynx. In contrast, the expression of the E1 and E2 subunits, was observed throughout the body of the worm and was significantly enriched in the pharynx. Taken together, PDP-1 modulates daf-2 dauer formation and this function is likely to be independent of its role in regulating the PDHc.

PDP-1 regulates multiple outputs of the IIS pathway
In addition to dauer formation, the IIS pathway also regulates longevity, stress resistance and fat storage [17,18]. Mutations in daf-2 and age-1 result in a significant extension in lifespan, enhanced resistance to various stresses and increased fat storage [7,35,[41][42][43][44]. These phenotypes are suppressed by loss-of-function mutations in daf-18 and daf-16 [32,34,35,39]. We therefore investigated whether dosage modulation of pdp-1 would affect additional outputs of the pathway. We first tested the role of PDP-1 in regulating lifespan ( Figure 2 and Figure S4). The lifespan of wild-type worms was not affected by pdp-1 RNAi and slightly reduced by a mutation in pdp-1 (Figure 2A and 2D). In contrast, the mean and maximal lifespan of long-lived daf-2(e1370) and age-

Author Summary
Cells in the body respond to a variety of on/off signals that are relayed in a defined spatial and temporal manner. These signals influence several processes such as growth, fat storage, and the repair of damaged molecules. As humans age, the onset of diseases such as Type 2 Diabetes, obesity, and cancer often results from an imbalance in the levels of on/off signals in the cell. The insulin/IGF-1 signaling pathway is an important regulator of longevity, development, and metabolism across phylogeny. While the protein kinases that activate this pathway have been well studied, less is known about the protein phosphatases that tune down the signals. The roundworm C. elegans has been an excellent model system to study the role of insulin/IGF-1 signaling in the aging process. Here, we identify a new phosphatase that negatively regulates the insulin/IGF-1 pathway to enhance longevity and stress-resistance. Interestingly, the phosphatase achieves this function by tuning down the activity of a conserved TGF-b pathway, a pathway important for development. By reducing TGF-b pathway activity, this phosphatase decreases expression of insulin molecules that may stimulate the insulin/IGF-1 pathway. Our studies not only unravel a new regulator of these pathways, but also point to how they are more linked than previously thought. Both insulin/IGF-1 and TGF-b signaling have been implicated in age-associated diseases, and understanding their connection will provide us with potential therapeutic avenues.
To examine the effect of increased dosage of pdp-1, we generated transgenic worms bearing a translational fusion containing pdp-1 fused to gfp and driven by its own promoter (pdp-1::gfp). In addition, we also crossed the pdp-1::gfp worms to daf-2(e1370) mutants to generate the daf-2(e1370); pdp-1::gfp strain. Overexpression of pdp-1 results in a significant extension in lifespan compared to wild-type worms ( Figure 2D and Figure S4). Interestingly, pdp-1 overexpression further extends the lifespan of daf-2(e1370) mutants ( Figure 2B and Figure S4). In both of these cases, the increased lifespan was suppressed by daf-16 RNAi ( Figure S5). Therefore, dosage modulation of pdp-1 regulates lifespan in a DAF-16 dependent manner.
Next, we asked if PDP-1 modulated additional outputs of the IIS signaling pathway. We first tested whether PDP-1 regulates stress resistance by assaying the survival of pdp-1 mutants and transgenic animals when exposed to heat stress at 37uC ( Figure 2E and Figure S7). Dosage modulation of pdp-1 affects the response to heat stress, with a pdp-1 mutation decreasing and pdp-1 overexpression slightly enhancing thermotolerance ( Figure 2E). Importantly a pdp-1 mutation drastically reduced the thermotolerance of daf-2 mutants ( Figure 2E).
To confirm these results, we investigated whether pdp-1 RNAi could suppress dauer formation of daf-2(e1370); daf-3(mgDf90) double mutants (Table 2 and Figure S10). In this strain, input from the TGF-b pathway is removed due to the daf-3 null mutation, and dauer formation is presumably mediated through activated DAF-16 [39]. Therefore, if pdp-1 was indeed acting in the TGF-b pathway, we would not see any effect of pdp-1 RNAi on daf-2(e1370); daf-3(mgDf90) double mutants. Expectedly, pdp-1 RNAi had no effect on daf-2(e1370); daf-3(mgDf90) double mutants ( Table 2 and Figure S10). DAF-3 itself is unlikely to be a substrate for PDP-1, as similar to mammalian Co-SMADs, it lacks the SMAD phosphorylation motif [28]. Therefore, our genetic epistasis analysis supports a model whereby pdp-1 acts in the DAF-7 TGF-b pathway at the level of daf-8 and daf-14.

TGF-b signaling can modulate the IIS pathway
How does a phosphatase in the TGF-b signaling pathway have such robust effects on the outputs of the IIS pathway and DAF-16?
A number of studies have previously identified roles for the TGF-b pathway in lifespan and fat storage [7,54,55]. However, genetic epistasis analysis on dauer formation placed DAF-7 TGF-b signaling and IIS as two parallel pathways where components of one pathway did not affect the other [14,56,57]. Yet in our studies, PDP-1 was able to regulate multiple outputs of IIS. Therefore, we decided to further investigate the potential crosstalk between the IIS and TGF-b signaling pathways. First, we focused on DAF-3 and DAF-5, which are positive regulators of dauer formation similar to PDP-1, and asked whether mutations in daf-3 or daf-5 could also affect phenotypes of the IIS pathway [14,28,29].
To further explore the crosstalk between both pathways, we next asked whether DAF-18 and DAF-16, which are components of the IIS pathway, affect TGF-b pathway signaling. For this, we assayed dauer formation of TGF-b pathway mutants on daf-18 and daf-16 RNAi (Table 3 and Figure S10). Interestingly, dauer formation of daf-7(e1372), daf-14(m77) and daf-8(m8 5) worms was robustly suppressed by daf-16 RNAi. We observed similar results for dauer formation daf-7(e1372) and daf-14(m77) mutants on daf- 18 RNAi. However, in the case of daf-8(m85) mutants, daf-18 RNAi had no effect on dauer formation of (Figure S10), suggesting a complex crosstalk between both pathways. The enhanced dauer formation of daf-2(e1370); daf-3(mgDf90) is suppressed by both daf-18 and daf-16 RNAi but not pdp-1 RNAi (Table 3 and Figure S10). Therefore, we not only observe DAF-3 and DAF-5 affecting various phenotypes of the IIS pathway, but also the converse, where DAF-16 and DAF-18 robustly regulates TGF-b dauer formation. These results unravel a more complex interaction between the two pathways, where DAF-16 is likely to be the major downstream effector regulating longevity, dauer formation and other physiological outputs.
Insulins are a possible connection between TGF-b signaling and IIS How can these two pathways, once considered to be parallel to each other, be mechanistically linked? Thus far our data suggests that PDP-1, a component of the TGF-b pathway can modulate multiple phenotypes of IIS by positively regulating DAF-16. In addition, we observe extensive crosstalk between the two pathways at multiple levels. A feed-forward model that has been proposed to connect TGF-b signaling to the IIS pathway suggests insulins as a possible link [55,58]. The C. elegans genome encodes 40 insulin genes [59,60] (WormBase 215: www.wormbase.org). Studies using mutants and RNAi have characterized some of the insulins as agonists or antagonists of the IIS pathway [13,[59][60][61]. Importantly, microarray studies have identified several insulin genes that are regulated by TGF-b signaling, including ins-1, ins-4, ins-5, ins-6, ins-7, ins-17, ins-18, ins-30, ins-33, ins-35 and daf-28 [55,57]. We tested changes in the levels of these insulins using Q-PCR in TGFb pathway mutants such as daf-3(mgDf90), daf-14(m77) as well as pdp-1(tm3734) and compared them to wild-type worms ( Figure 5A-5C, Figure S14, Tables S2 and S3). Interestingly, both pdp-1(tm3734) and daf-3(mgDf90) showed elevated levels of several insulins as compared to wild-type worms ( Figure 5A and Figure  S14). In contrast, expression of these insulins was markedly reduced in daf-14(m77) mutants ( Figure 5B and Figure S14). We next looked at the effects of overexpressing DAF-3 and PDP-1 on insulin gene expression ( Figure 5C and Figure S14). The levels of several insulins are markedly reduced in daf-3::gfp and pdp-1::gfp animals when compared to wild-type worms. Therefore, dosage modulation of DAF-3 and PDP-1 modulates insulin gene expression. INS-4, for example, has been reported as a positive regulator TGF-b pathway and a suppressor of dauer formation of daf-7 and daf-8 mutants [62]. ins-4 transcript levels were elevated in pdp-1 and daf-3 mutants but reduced in daf-14.
To investigate insulin gene expression regulated by DAF-16, we tested daf-2(e1370), pdp-1(tm3734); daf-2(e1370) and daf-16(mgDf50); daf-2(e1370) mutants. Several insulins were changed relative to daf-2(e1370) worms, with the trend between pdp-1(tm3734); daf-2(e1370) and daf-16(mgDf50); daf-2(e1370) being quite similar ( Figure 5D and Figure S14). Interestingly, ins-7 levels were elevated both double mutants ( Figure 5E and Figure S14). Previous studies have shown ins-7 to be an agonist of the IIS pathway as well as a DAF-16 target gene [13,63]. In contrast, ins-1 levels were drastically reduced, and INS-1 has been characterized as a potential antagonist of IIS [59]. We did not observe a significant change in ins-18, another potential DAF-16 target [13]. We also did not detect any appreciable differences in insulin gene expression in daf-16(mgDf50) single mutants ( Figure S14). In addition, we were unable to detect ins-33 and ins-35 transcripts in all the strains tested, and the trend observed with daf-28 was inconclusive (Table S2 and S3). Taken together, our results suggest the possibility that insulins downstream of TGF-b signaling mediate at least part of the cross talk between the two pathways. Therefore, PDP-1 would modulate to regulate expression of several insulins that can potentially feed into or antagonize the IIS pathway to regulate DAF-16 and its associated phenotypes.

Discussion
We identified pdp-1 from a RNAi screen for serine/threonine phosphatases that modulate daf-2 dauer formation. C. elegans PDP-1 is homologous to mammalian pyruvate dehydrogenase phophatase (PDP), a metabolic enzyme that is a positive regulator of the pyruvate dehydrogenase enzyme complex (PDHc). Remarkably, other components of the PDHc in C. elegans do not affect daf-2 dauer formation. Microarray and SAGE studies on dauers have indicated that genes involved in anaerobic metabolism are upregulated while genes involved in the TCA cycle and mitochondrial oxidative phosphorylation are downregulated, suggesting that PDHc activity may not be critical for dauer diapause [64][65][66]. Further, annotations indicate that the C. elegans genome encodes approximately 60 serine/threonine phosphatases, in contrast to the 400 plus protein kinases, suggesting that phosphatases are likely to have a number of cellular substrates [39,67]. We find that PDP-1 also regulates longevity, fat storage and stress resistance in addition to dauer formation. Interestingly, these phenotypes are more severe in mutants such as daf-2 and age-1, where IIS is reduced. Further, PDP-1 positively regulates DAF-16 activity. We reason that PDP-1 function is critical under conditions of stress or low food availability, when DAF-16 activation is required [39].
Intriguingly, genetic epistasis analyses place PDP-1 in the DAF-7/TGF-b pathway, at the level of the R-SMAD proteins DAF-14 and DAF-8. A recent functional RNAi screen for serine/threonine phosphatases that modulate BMP signaling identified PDP as a SMAD1 phosphatase in Drosophila S2 cells and mammalian 293T cells [68]. Our study complements these findings and reveals a molecular conservation in the role of PDP-1 in regulating TGF-b signaling. Early genetic epistasis studies had suggested that TGF-b signaling and IIS pathways are parallel signaling pathways that modulate dauer diapause [31]. Importantly, in these studies, the conclusion was that both these pathways acted independently, and it was the IIS pathway that regulated longevity and stress resistance [31,32].
However, the effect of PDP-1 on DAF-16 activity led us to reinvestigate the interaction between the IIS and TGF-b signaling. Previous studies have shown that DAF-3 and DAF-5 are negatively regulated by TGF-b signaling, and function similarly as repressors of gene expression to ultimately promote dauer formation [28,29,69,70]. We find that under conditions of reduced IIS, DAF-3 and DAF-5 affect various outputs of the IIS pathway in opposite ways. DAF-3 in particular regulates IIS depending upon the level of signaling through the pathway ( Figure 6). In our hands, mutants of the TGF-b signaling pathway do not exhibit a pronounced increase in lifespan. However, components of this pathway are important for the long lifespan of mutants in the IIS pathway, as well as other phenotypes such as dauer formation, fat storage and stress resistance. Our epistasis studies reveal that daf-18 and daf-16 RNAi can strongly suppress dauer and fat storage of TGF-b pathway mutants. Together, these results point to a feedforward model where signals through the TGF-b pathway are relayed to modulate activity of the IIS pathway as well as DAF-16. Indeed, recent studies have suggested that TGF-b pathway regulates the expression of insulins, leading to a feed-forward model, where signals from the TGF-b pathway are relayed to modulate activity of the IIS pathway as well as DAF-16 [55,58].
In support of this model, we find TGF-b signaling regulates the expression of several insulin genes with DAF-3 and PDP-1 negatively modulating insulin gene expression. This is in agreement with previous studies that identify DAF-3 as a repressor of gene expression [69,70]. The expression of several insulins is also modulated by DAF-16, with pdp-1(tm3734); daf-2(e1370) and daf-16(mgDf50); daf-2(e1370) worms showing similar trends in insulin levels. Therefore, in the absence of PDP-1, increased levels of agonists or reduced levels of antagonists hyperactivate the DAF-2 pathway to negatively regulate DAF-16, thereby affecting the enhanced lifespan, stress resistance, dauer formation and fat storage of daf-2 mutants.
Our results suggest a model where under favorable growth conditions, signals through the TGF-b pathway activate the SMAD transcriptional complex to regulate the expression of insulins that activate the IIS pathway to phosphorylate and inhibit DAF-16 activity, thereby promoting growth, reproduction and normal lifespan ( Figure 6, top panel). However, when food is limiting or under harsh survival conditions, TGF-b signaling is downregulated by PDP-1 to activate DAF-3 and DAF-5, to regulate the repression of insulin genes that may feed into the IIS pathway ( Figure 6, middle panel). DAF-3 has also been reported to negatively regulate daf-7 and daf-8 gene expression in a feedback loop [24]. We find that pdp-1 expression is elevated in daf-3(mgDf90) mutants, suggesting a similar feedback regulation The trend in Insulin levels are similar between pdp-1(tm3734); daf-2(e1370) and daf-16(mgDf50); daf-2(e1370) double mutants compared to the daf-2(e1370) parental strain. *p,0.03, **p,0.01, # a significant reduction in ins-18 levels is observed in daf-16(mgDf50; daf-2(e1370) mutants this set but not in others, p,0.001. (E) ins-7 levels are drastically elevated in daf-16(mgDf50); daf-2(e1370) and pdp-1(tm3734); daf-2(e1370) worms as compared daf-2(e1370) worms. *p,0.01, **p,0.004. doi:10.1371/journal.pgen.1001377.g005 ( Figure S15). Repression of TGF-b and insulin gene expression by DAF-3 results in a reduction in signaling through the IIS pathway, and promotes DAF-16 nuclear localization. DAF-16 then regulates the transcription of hundreds of target genes that ultimately modulate longevity, stress resistance, dauer formation and fat storage. Under low TGF-b signaling and IIS conditions, DAF-3 and DAF-5 regulate these outputs in an opposite manner, with DAF-5 synergizing and DAF-3 antagonizing DAF-16 function (Figure 6 lower panel). With our Q-PCR data, we found that PDP-1 affected only a subset of the DAF-16 target genes tested. These could represent genes that are regulated by DAF-16 and SMAD proteins. SMAD proteins have low affinity for binding DNA, and the orchestration of cellular signals into defined outputs requires their association with additional co-factors [71]. Mammalian SMAD proteins can bind several co-activators and corepressor proteins to modulate gene transcription [23]. Specifically, a synergy between mammalian FOXO (FOXO1, FOXO3a and FOXO4) and SMAD2/3 was identified for the regulation of several genes involved in cell cycle regulation and the response to stress [72]. Importantly, these interactions required the function of the co-SMAD protein SMAD-4, which is homologous to DAF-3 [72]. Therefore, DAF-3 and DAF-5 could also directly modulate the IIS pathway at the transcriptional level.
A clear interpretation of our results is complicated by three main factors. First, the sheer number of insulins in the worm makes it difficult to assess whether they are functionally distinct. Secondly, the role of temperature in modulating the readouts of the pathway has not been closely explored. For example, we observe the effects of pdp-1 RNAi on daf-2 lifespan at 15uC but the effect decreases at a higher temperature, as the pathway gets more inactive. It is therefore likely that a certain level of signaling through the pathway is required to activate and target PDP-1 to its substrate(s). At higher temperatures such as 20uC or 25uC, there may be extremely low levels of phosphorylated substrate available for PDP-1. Similarly, the effect of a daf-3 null mutation on daf-2 phenotypes is more pronounced at higher temperatures but not at 15uC. Third, the lack of null alleles may provide an incomplete picture of the phenotypes observed. For example, previous studies using non-null alleles of daf-16 only partially suppressed dauer formation of TGF-b pathway mutants and therefore DAF-16 was thought to only affect the IIS pathway [31]. Therefore, temperature, level of signaling and the kind of mutants used (null versus weak) are important additional inputs that need to be considered to better understand the crosstalk between the IIS and the TGF-b pathways.
In conclusion, our studies show that PDP-1 acts through the TGF-b pathway to negatively regulate IIS and promote DAF-16 activity. PDP-1 may mediate this function in part by negatively regulating TGF-b signaling to repress expression of several insulins that feed into the IIS pathway. In humans, dysregulation of TGFb signaling and the insulin/IGF-1 signaling axis have been implicated in the onset of age-associated diseases such as Type 2 Diabetes and cancer [73][74][75][76][77]. Future studies exploring the interactions between these two pathways as well as the factors that modulate these interactions may ultimately provide a better understanding of the pathophysiology of these diseases.

Strains
All strains were maintained at 15uC using standard C. elegans techniques [78]. For all RNAi assays, worms were maintained on the RNAi bacteria for two generations except for the assays on the PDHc RNAi. Strains used in this manuscript are listed in Table  S4.

RNAi-based assays
RNAi plates were prepared as previously described [39]. All RNAi clones were sequenced and verified before any assays were carried out. L4 worms were picked onto fresh RNAi plates and maintained for two generations prior to the assay, with the exception PDHc RNAi plates. Worms exhibit lethality when maintained on the following RNAi clones: T05H10.6 (E1a), C04C3.3 (E1b), F23B12.5 (E2), or LLC1.3 (E3) [79]. To circumvent this problem, strains were maintained on vector RNAi for two generations and transferred to E1a, E1b, E2 or E3 plates prior to the assay.

Strain construction
For the pdp-1(tm3734);daf-2(e1370) double mutant, daf-2(e1370) males were mated to pdp-1(tm3734) hermaphrodites at 15uC. A total of 30 F1 progeny were picked onto individual plates and allowed to have progeny at 25uC. From the F2 progeny on each plate, dauers were selected and transferred to fresh plates and incubated for an additional 24 hours at 25uC. The next day, the dauers were allowed to recover at 15uC until they reached adulthood. Subsequently, adult worms were picked onto individual plates and transferred to 25uC and allowed to have progeny. Among the F3 progeny, we observed that some plates had 100% dauers at 25uC, while worms in some of the plates exhibited a developmental delay and could not form complete dauers even after 5-6 days at 25uC. Worms from both sets of plates were recovered, picked to individual plates and allowed to self at 15uC. Parents were then tested for pdp-1(tm3734) deletion by PCR. As anticipated, the pdp-1(tm3734);daf-2(e1370) double mutants are unable to form 100% dauers at 25uC.
The daf-2(e1370);pdp-1::gfp strain was made by crossing daf-2(e1370) males to pdp-1::gfp hermaphrodites at 15uC. About 30 F1 animals were transferred to individual plates and allowed to have progeny at 25uC. From the progeny, F2 dauers were selected from each plate and allowed to recover at 15uC. The recovered adult worms were then checked for the presence of GFP, and Under these conditions, DAF-3 and DAF-5 repress the transcription of agonistic insulins as well as expression of the daf-7 TGF-b ligand and daf-8, leading to further downregulation of the TGF-b pathway. Alternatively, DAF-3 and DAF-5 may promote transcription of potential antagonistic insulins. This results in reduced signaling through the IIS pathway, enhancing DAF-16 nuclear localization. Lower panel: Under low IIS conditions, DAF-16 localization is predominantly nuclear, where it regulates the transcription of hundreds of target genes that act in combination to regulate longevity, stress resistance, dauer formation and the response to stress. Paradoxically, under low IIS conditions, DAF-3 and DAF-5 play opposite roles. DAF-5 is likely to synergize with DAF-16 and modulate the activity of its target genes. DAF-3 acts to antagonize DAF-16, either directly or through suppression of DAF-16 target genes. Therefore, the role of DAF-3 in modulating IIS depends upon the level of signaling through the pathway. doi:10.1371/journal.pgen.1001377.g006 GFP-positive worms were transferred to individual plates and incubated at 25uC. Plates where 100% of the progeny were dauers and GFP positive were selected and established as the strain for the assays.

Dauer assays
Strains were maintained on RNAi plates for two generations or regular OP50 plates at 15uC. Dauer assays were performed by picking approximately 100 eggs onto 2 fresh plates and incubated at the appropriate temperature. The pdk-1(sa680), daf-7(e1372) and daf-14(m77) worms have a strong Egl phenotype. For dauer assays on these strains, gravid adult worms growing on the RNAi plates were washed off the plate with sterile PBS onto a 1.5 mL eppendorf tube. After 2 washes at 2000 g for 30 seconds, the adults were vortexed for 5 mins in 5 mL of 1 N sodium hydroxide and 3% sodium hypochlorite (final concentration). The samples were then washed twice with sterile PBS and eggs were aspirated with a glass pipette onto fresh RNAi plates. For all dauer assays, plates were scored for the presence of dauers or non-dauers after 3.5-5.5 days, depending upon the strain. Dauer assays were performed at the temperature indicated. Significance was determined by Student's t-test.

Lifespan assays
Strains were maintained at 15uC and synchronized by picking eggs onto fresh RNAi or OP50 plates. Approximately 60 young adult worms were transferred per plate to a total of three fresh RNAi or regular OP-50 plates containing 5-fluorodeoxyuridine (FUDR) at final concentration of 0.1 mg/mL [80]. All RNAibased lifespan assays were carried out at 15uC. Lifespans on OP50 plates were performed at the temperature indicated. Survival was scored by tapping with a platinum wire every 2-3 days. Worms that died from vulval bursting were censored from the analysis. Statistical analyses for survival were conducted using the standard chi-squared-based log rank test.

Heat stress assay
Strains were maintained on RNAi or regular OP50 bacteria at 15uC, as described above. From these plates, approximately 30 young adult worms were picked onto fresh RNAi or regular plates and upshifted to 20uC for 6 hrs. The plates were then transferred to 37uC and heat stress-induced mortality was determined every few hours till all the animals died. Statistical analyses for survival were conducted using the standard chi-squared-based log rank test.

Fat staining
Strains maintained RNAi or on regular OP50 plates were synchronized by picking eggs on to fresh plates and grown synchronously at 15uC. The plates were then upshifted to 20uC for 8 hours, at the L2 stage to get L3 worms and at the L4 stage to get young adult worms. Worms were then washed off the plates into microcentrifuge tubes and incubated in 16 PBS buffer for 20 minutes on a shaker at RT. After 2 washes at 3000 rpm for 30 seconds with 16 PBS, the strains were fixed according to the type of staining performed. Oil Red O and Sudan black staining was performed as previously described [39,45,81,82]. After incubation overnight at RT, worms were mounted on slides and visualized using the Zeiss Axioscope 2+ microscope.

Quantification of fat staining
For Sudan Black Staining, we used Image J software to measure the average pixel intensity for a 84-pixel radius below the pharynx of each animal in the anterior intestine area. Next, an 84-pixel radius of the background was measured, and subtracted from the values obtained for the staining. At least 10 animals were measured for each RNAi clone. Significance was determined by Student's t-test.
For Oil Red O Staining, Image J was used to separate out each color image into its RGB channel components. As previously described [45], Oil Red O absorbs light at 510 nm and therefore, the green channel was used for further analysis. We measured the average pixel intensity for a 84-pixel radius below the pharynx of each animal in the anterior pharynx area. We next measured a 84pixel radius of the background, which was later subtracted from the values obtained from the staining. At least 10 animals was measured for each RNAi clone. Significance was determined by Student's t-test. DAF-16::GFP localization assay DAF-16 localization assays were performed as previously described [39,52]. daf-2(e1370); daf-16::gfp worms were maintained on RNAi plates at 15uC similar to the dauer assays. Approximately 30 L4 worms were transferred to fresh RNAi bacteria and the plates were shifted to 20uC for 1 hr. The worms were visualized under a fluorescence microscope (Zeiss Axioscope 2+ microscope). Worms were classified into four categories based on the extent of DAF-16::GFP nuclear-cytoplasmic distribution: completely cytosolic, more cytosolic than nuclear in most tissues, more nuclear than cytosolic in most tissues and completely nuclear.

Psod-3::gfp expression
Quantification of Psod-3::gfp was performed as previously described [39]. daf-2(e1370);sod-3::gfp worms were grown at 15uC on RNAi as described above. Approximately 30 L4 animals were transferred to fresh RNAi bacteria and shifted to 25uC for 1 hr. The expression of sod-3::gfp was visualized using Zeiss Axioscope 2+ microscope. GFP expression was categorized as follows: High: GFP expression seen throughout the worm Medium: Weak expression detected in the body of the worm along with the head and the tail Low: Low GFP expression only detected in the head and tail

Transgenic worms
Promoter and ORF entry clones of pdp-1 obtained from the promoterome and ORFeome were combined using multisite Gateway cloning (Invitrogen) into the pDEST-DD03 or the R4-R2 GFP destination vectors to create the Ppdp-1::gfp or Ppdp-1::pdp-1 ORF :: gfp constructs [83,84]. All constructs contain the unc-119 minigene. The vectors were verified by sequencing as well as restriction digestion. Transgenic worms were generated by ballistic transformation into unc-119(ed3) mutant worms as previously reported (Biorad, USA) [83]. Integrated lines that were obtained were used for further analyses. For the pdp-1::gfp translational fusion strain, additional lines were generated by integration of extrachromosomal array lines by UV irradiation as previously described [85]. All translational fusion lines were backcrossed 46 to wild-type prior to analysis.

RT-PCR experiments
For all RT-PCR experiments, strains were maintained at 15uC. Eggs were obtained from gravid adult worms by hypochlorite treatment described earlier. The eggs were seeded onto large plates maintained at 15uC until the worms entered the L4 stage. The plates were then upshifted to 20uC for 8 hours until they became young adults. Worms were then collected with sterile 16PBS and washed twice at 2000 g for 30 seconds. The supernatant was removed, and 0.5 mL of AE buffer (50 mM acetic acid, 10 mM EDTA), 0.1 mL of 10% SDS, and 0.5 mL of phenol was added to the worm pellet and the mixture was vortexed vigorously for 1 min, followed by incubation at 65uC for 4 min. Total RNA was purified by phenol:chloroform extraction and ethanol precipitation. The quality of the RNA isolated was determined by checking the 28 S and 18 S RNA on an agarose gel. 2 ug of total RNA was used for making cDNA using the SuperScript cDNA synthesis kit (Invitrogen, USA). The expression of the DAF-16 target and insulin genes was checked by RT-PCR using the SYBR Green PCR Master Mix and 7000 Real-Time PCR System (Applied Biosystems, USA). The relative expression of the genes tested was compared to actin as an internal loading control. Significance was determined by Student's t-test. Primers used for the RT-PCR experiments are listed in Table S5.

Locomotion assay
Young adult wild-type and pdp-1(tm3734) worms were picked onto 6 individual plates each. After 5 minutes, the worms were picked off the plate. The average distance covered was calculated by measuring the traces on the bacterial lawn using ImageJ. Significance was determined by Student's t-test.

Brood size measurements
Wild type, daf-2(e1370), pdp-1(tm3734) and pdp-1(tm3734); daf-2(e1370) worms were maintained at 15uC. 5 L4 worms were picked onto individual plates and allowed to lay eggs at 22.5uC. Worms were transferred to a new plate every 12 hours. After 22.5 hours, the parental worms were picked off the plates, and the total number of eggs laid was scored. The number of progeny from these eggs was scored again after 38 hours. The % hatched eggs was calculated as a percentage of the average number of progeny over the average number of eggs laid. Significance was determined by Student's t-test.