C. elegans AMPKs promote survival and arrest germline development during nutrient stress

Juvenile animals coordinate development and growth with nutritional status in order to achieve reproductive maturation. C. elegans development shows a clear example of such coordination; when larvae hatch into nutritionally-deficient conditions, they suspend postembryonic development and can survive more than a week, resuming development if ample food is supplied (Johnson et al., 1984). This developmentally dormant state is called “L1 diapause” or “L1 arrest,” and it can be simply released by supplying nutrients. Once larval development is initiated, exposure to unfavorable conditions such as lack of nutrients, high population density or high temperature predispose larvae to arrest development in an alternative third larval (L3) stage, called “dauer diapause” (reviewed by Riddle and Albert, 1997). Progression to the L4 stage is then blocked until the environment becomes favorable for growth. Furthermore, once animals have reached the L4 stage, starvation triggers a response that delays reproductive onset and reduces viability of fertilized embryos (Angelo and Van Gilst, 2009; Seidel and Kimble, 2011). The starved animals molt into adults and their oogenic germlines shrink, but can regenerate if the animal survives until re-feeding, indicating the stem cells had entered a quiescent state. Although the significance of these various starvation responses to survival in the wild requires further study, their existence in the life cycle is likely to reflect the importance of coordinating development and growth with nutritional status for optimal fitness and reproduction.

Previous studies have shown that during L1 diapause germline stem cells arrest at the G2 phase of the cell cycle in a manner dependent on daf-18/pten, a negative regulator of the insulin/IGF signaling (IIS) pathway (Fig. 1A) (Fukuyama et al., 2006). Interestingly, daf-16/foxo, which antagonizes the IIS pathway downstream of daf-18, is not required for mitotic quiescence of germline stem cells during L1 diapause, suggesting that daf-18 suppresses germline proliferation by a daf-16-independent pathway (Fukuyama et al., 2006). In contrast, daf-16 does function downstream of daf-18 to control dauer larvae formation (Ogg and Ruvkun, 1998). Thus, these observations suggest that developmental arrest during L1 and dauer diapauses is regulated by partly distinct mechanisms.

Another component known to mediate cellular and organismal responses to starvation is AMP-activated kinase (AMPK). In mammals, this heterotrimeric serine/threonine kinase is stimulated by AMP and ADP as well as upstream kinases and hormones, and it is best known for its role in maintaining energy balance by directly activating metabolic pathways that generate ATP and inactivating those that consume ATP (reviewed by Hardie, 2011; Kahn et al., 2005). However, recent studies have shown that AMPK also regulates other starvation-responsive processes such as autophagy, mitochondria biogenesis, cell cycle arrest, and food intake (Wang et al., 2001; Zong et al., 2002; Minokoshi et al., 2004; Jones et al., 2005; Narbonne and Roy, 2006; Behrends et al., 2010; Egan et al., 2011; Kim et al., 2011).

Both AMPK and PTEN negatively regulate mechanistic Target of Rapamycin Complex 1 (mTORC1), which consists of several proteins including the atypical serine/threonine kinase mTOR and its scaffold protein Raptor (reviewed by Wullschleger et al., 2006; Shackelford and Shaw, 2009; Laplante and Sabatini, 2012). mTORC1 responds to signals from several environmental cues such as amino acids, growth factors and energy stresses to control diverse cellular processes including growth, autophagy and metabolism (Fig. 1A) (reviewed by Laplante and Sabatini, 2012). In response to energy stresses, AMPK directly and indirectly represses mTORC1 through both Rheb-independent and dependent manners (Inoki et al., 2003b; Gwinn et al., 2008). Growth factors stimulate mTORC1 activity through the PI3K (phosphatidylinositol 3-kinase)-AKT axis (Inoki et al., 2003a; Inoki et al., 2003b; Tee et al., 2003). Thus, deficiency of PTEN, which antagonizes PI3K as a phosphatidylinositol 3-phosphatase (Maehama and Dixon, 1998), results in constitutive activation of mTORC1 (Neshat et al., 2001; Podsypanina et al., 2001).

Similar to its mammalian orthologues, C. elegans AMPK is also activated by AMP and energy stresses (Apfeld et al., 2004; Schulz et al., 2007), and it plays an important role in starvation responses, including extension of lifespan by nutritional restriction and maintenance of viability and germline quiescence during dauer diapause (Narbonne and Roy, 2006; Greer et al., 2007; Schulz et al., 2007). Here, we test the significance of AMPK in the regulation of L1 diapause. We find that aak-1 and aak-2, which encode the two catalytic α subunits of AMP-activated protein kinase, act redundantly to arrest germline stem cell proliferation and maintain viability during L1 diapause. Both aak-1; aak-2 double and daf-18 single mutants that experience L1 diapause develop into sterile adults when returned to food, indicating the importance of these genes for long term germline viability. Both daf-18 and the aak genes appear to regulate germline quiescence by suppressing activity of TORC1. Previous studies indicate that aak-2 acts in the hypodermis and excretory cell to maintain viability during dauer diapause (Narbonne and Roy, 2009). In contrast, we find that L1 diapause survival requires aak-2 in the intestine and neurons. Together, these results suggest that suppression of TORC1 activity by AMPK maintains germline stem cell quiescence and that AMPK promotes survival during L1 diapause in a manner distinct from that which controls dauer diapause.

Survival during L1 diapause depends on activity of aak-2 and daf-18/PTEN (Baugh and Sternberg, 2006; Sigmond et al., 2008), but the role of aak-1, the sole worm paralogue of aak-2 (Apfeld et al., 2004) has not been fully investigated. We tested the roles of the aak genes in L1 diapause using the putative null alleles, tm1944 and ok524, which result in deletion of a significant portion of the kinase domains of AAK-1 and AAK-2, respectively (Narbonne and Roy, 2006; Narbonne and Roy, 2009). Viability of aak-1(tm1944) mutants was reduced during L1 diapause relative to wild-type animals, although the animals survived better than did aak-2(ok524) mutants (Fig. 1B). However, animals lacking function of both aak genes (hereafter called “ampk mutants”) displayed a more severe reduction in viability than did either single mutant, indicating that aak-1 and aak-2 act redundantly to maintain survival during L1 diapause. The L1 diapause-induced lethality of ampk mutants is rescued by an aak-2 transgene (see below) demonstrating that loss of AMPK activity is responsible for this phenotype. Because L1 diapause is induced by nutritional stress, the observed lethality of ampk mutants may be a consequence of carbon and energy deprivation. Consistent with this possibility, supplementation of the starvation medium with glucose or ethanol, compounds known to promote population growth when added to chemically defined media, significantly increased viability of ampk mutants (Fig. 1C) (Lu et al., 1978; Castro et al., 2012).

AAK-2 mediates dietary restriction-induced longevity by acting through the FOXO transcription factor, DAF-16 (Greer et al., 2007). Similarly, daf-16 also acts downstream of daf-18 to control lifespan and dauer formation (Ogg and Ruvkun, 1998). These findings raise the possibility that the aak and daf-18 genes could maintain organismal survival during L1 diapause through a daf-16-mediated pathway. However, viability of daf-16 null mutant animals is much greater than that of ampk or daf-18 mutants (Fig. 1B), indicating that loss of daf-16 function is not solely responsible for the reduced survival in daf-18 and ampk mutants. Furthermore, viability of daf-18(nr2037); ampk triple mutants was reduced relative to daf-18 and ampk mutants (Fig. 1B). Because the three alleles used were null (Mihaylova et al., 1999), these findings suggest that the daf-18 and aak genes act in parallel pathways to maintain survival during L1 diapause (see also below).

A key feature of L1 diapause is maintenance of quiescent germline stem cells. Intriguingly, loss of daf-18, but not daf-16, can uncouple these processes (Fukuyama et al., 2006). We tested whether the aak genes are also required for maintaining germline quiescence (see Materials and Methods). Animals lacking aak-1, but not aak-2, occasionally showed ectopic germline proliferation during L1 diapause (Fig. 2A). In contrast, analysis of ampk double mutants revealed a striking enhancement; germ cells in all such animals examined (n = 150) failed to arrest proliferation during L1 diapause, indicating that the aak genes act redundantly to maintain quiescence in the germline (Fig. 2A,B).

When ampk and daf-18 mutants in L1 diapause are cultured in M9 medium containing ethanol, many surviving animals lack the overt tissue degeneration observed in the absence of ethanol. This prompted us to ask whether the germline hyperplasia observed in these L1 animals would have a consequence on subsequent germline development. After 5 days of L1 diapause in M9 containing ethanol, animals were re-fed to allow postembryonic development, and animals that developed to the adult stage were examined. In contrast to wild type, aak-1 and aak-2 single mutants, most ampk double mutants were sterile when grown to adulthood (Fig. 2C). Microscopic examination revealed variable, disorganized gonad morphologies, ranging from a failure of gonad extension to well-extended gonads containing oocytes (supplementary material Fig. S1). The observed sterility appears to be a consequence of L1 diapause because all continuously fed ampk double mutant larvae were fertile (Fig. 2D). Similar but less severe effects of L1 diapause on fertility were also detected in daf-18 mutants (Fig. 2C,D; supplementary material Fig. S1). Although the cause of the sterility in these mutants requires additional analysis, these observations further illustrate the requirements for aak and daf-18 in ensuring proper germline development in animals that have experienced nutritional stress.

Although the aak and daf-18 genes are important for maintaining survival and germline quiescence during L1 diapause (Baugh and Sternberg, 2006; Fukuyama et al., 2006; Sigmond et al., 2008; this study), the tissue(s) in which they function to control these processes have not been elucidated. We used a series of gfp-tagged reporter genes to identify these tissues (supplementary material Table S1). We first generated reporter fusions for each aak gene that contain genomic DNA, spanning the promoter and extending through the full coding region, fused to gfp and the unc-86 3′-UTR (Paak-1::aak-1::gfp and Paak-2::aak-2::gfp). Because Paak-1::aak-1::gfp expression patterns were weak and variable when expressed from its native promoter (data not shown), aak-2 was used for these studies. During L1 diapause, Paak-2::aak-2::gfp was highly expressed in most tissues including the intestine, excretory canal, pharynx, somatic gonad, neurons, hypodermis, and body wall muscle (supplementary material Fig. S2; data not shown), a pattern similar to that observed during continuous development (Lee et al., 2008) and dauer diapause (Narbonne and Roy, 2009). The Paak-2::aak-2::gfp transgene rescued the lethality and, albeit less effectively, the ectopic germline proliferation phenotype observed in ampk mutants during L1 diapause (Fig. 3A,B). Although the construct contains a heterologous 3′-UTR, the significant rescuing activity conferred by Paak-2::aak-2::gfp suggests that its expression largely recapitulates that of the endogenous gene.

To determine the tissue(s) in which aak-2 acts to maintain survival and germline quiescence during L1 diapause, gfp::aak-2 was expressed in ampk mutant animals under the control of various tissue-specific promoters. Intestinal (Ppgp-1) and pan-neuronal (Prgef-1) expression of gfp::aak-2 significantly restored the survival of ampk mutant animals (Fig. 3A), identifying these tissues as foci of aak activity relative to survival and indicating that the N-terminal GFP fusion also retains activity. In contrast to their ability to rescue the lethality phenotype, neither of these constructs suppressed the abnormal germline proliferation phenotype of ampk mutants during L1 diapause (Fig. 3B). These results suggest that the aak genes control germline quiescence in a manner distinct from that which maintains energy homeostasis during L1 diapause.

In contrast to the rescuing activity observed with the intestine and neuronal promoters, gfp::aak-2 expression in the hypodermis, body wall muscles, or excretory (Exc) cell did not significantly rescue either of the ampk mutant phenotypes during L1 diapause (Fig. 3A,B). This result is interesting in light of a previous study that found that aak-2 acts in the hypodermis and Exc to regulate survival during dauer diapause (Narbonne and Roy, 2009). The lack of rescue observed in our studies was not likely to be due to insufficient expression of the transgenes, since gfp::aak-2 expression was sufficient to allow transgenic animals to be identified. Thus, these findings illustrate the distinct modes by which aak-2 functions to control viability during L1 and dauer diapauses.

We next sought to determine in which tissue(s) daf-18 functions to maintain survival during L1 diapause. As observed with the aak genes, intestinal expression of daf-18 significantly rescued survival (Fig. 3C). However, in contrast to aak function, the hypodermal expression of daf-18 also rescued survival, whereas neuronal expression did not (Fig. 3C). These observations, together with the genetic analysis already described (Fig. 1B), indicate that the aak and daf-18 genes act by at least partially distinct mechanisms to maintain survival during L1 diapause.

Germ cells arrest in the G2 phase of the cell cycle during L1 diapause (Fukuyama et al., 2006). Although AMPK has been proposed to cause G1 cell cycle arrest through regulation of the tumor suppressor p53 and cyclin-dependent kinase inhibitor p27 in mammalian cultured cells (Imamura et al., 2001; Jones et al., 2005; Liang et al., 2007), the mechanism by which AMPK induces G2 arrest is currently unknown. However, studies in multiple organisms including C. elegans suggest that TORC1 promotes G2 progression (Nakashima et al., 2008; LaFever et al., 2010; Gaur et al., 2011; Kapoor et al., 2012; Korta et al., 2012). Given that mammalian mTORC1 is negatively regulated by both AMPK and PTEN (reviewed by Wullschleger et al., 2006; Shackelford and Shaw, 2009; also see Introduction), we considered the possibility that the germline proliferation phenotype observed in ampk and daf-18 mutant animals during L1 diapause may be caused by inappropriate activation TORC1. To test this possibility, C. elegans TORC1 activity was depleted in ampk and daf-18 mutant animals by RNAi. Key components of TORC1 are the protein kinase TOR and its scaffold protein Raptor, proteins whose worm orthologues are encoded by let-363 (Long et al., 2002) and daf-15 (Jia et al., 2004), respectively. RNAi directed against each of these genes significantly suppressed the germline phenotype of ampk and daf-18 mutant L1 larvae, but did not alter germ cell number during L1 diapause of wild type (Fig. 4A,B; supplementary material Fig. S3), suggesting that TORC1 functions either in parallel to, or downstream of, the aak and daf-18 genes. Recently, loss of aak activity has been shown to partially suppress male tail defects caused by plx-1 mutation, which significantly reduces TORC1 activity (Nukazuka et al., 2011). Furthermore, the inhibitory AMPK phosphorylation site is conserved between mammalian Raptor and C. elegans DAF-15 (Gwinn et al., 2008). Thus, our RNAi results are consistent with a model in which abnormal germline proliferation of ampk and daf-18 mutants during L1 diapause is caused by derepressed activity of TORC1.

Interestingly, neither let-363(RNAi) nor daf-15(RNAi) significantly restored the viability of ampk or daf-18 mutants during L1 diapause (supplementary material Fig. S3). These observations suggest that aak and daf-18 maintain survival during L1 diapause through TORC1-independent pathways. Alternatively, RNAi against let-363 and daf-15 may not be effective enough to confer detectable suppression of this phenotype.

In mammals and Drosophila, two G proteins, Rheb and Rag, mediate growth factor and amino acid signaling in response to TORC1 activation (Fig. 1A) (reviewed by Zoncu et al., 2011). There is a sole C. elegans Rheb orthologue, rheb-1 (Li et al., 2004). Rag proteins are subdivided into two subfamilies, which associate with each other to form heterodimers (Nakashima et al., 1999; Sekiguchi et al., 2001). One subfamily includes S. cerevisiae Gtr1p and mammalian RagA and RagB, and the other includes S. cerevisiae Gtr2p and mammalian RagC and RagD. Similar to yeast, C. elegans possesses only one gene, called raga-1, whose translational product belongs to the Gtr1p/RagA/B subfamily (Schreiber et al., 2010). We identified the ragc-1 gene, which encodes a sole orthologue of Gtr2p, RagC and RagD (supplementary material Fig. S5). As in other organisms (Garami et al., 2003; Saucedo et al., 2003; Stocker et al., 2003; Kim et al., 2008; Sancak et al., 2008), C. elegans rheb-1, raga-1, and ragc-1 orthologues are likely to be required for full activation of TORC1. Consistent with this view, inhibition of TORC1 and ragc-1 results in similar phenotypes including increased autophagy, decreased overall mRNA translation, and enhanced stress tolerance (Robida-Stubbs et al., 2012). If exit from germline quiescence in ampk and daf-18 mutant animals is triggered by derepression of TORC1 activity, then reducing or eliminating the activity of rheb-1, raga-1 or ragc-1 should suppress these defects. Consistent with this idea, the ectopic germline proliferation observed in both ampk and daf-18 mutants was significantly suppressed by RNAi of rheb-1 or by making double mutants between these genes and probable null alleles of raga-1 or ragc-1 (Fig. 4C,D). One could argue that the observed genetic suppression may simply reflect a requirement for rheb-1 and the Rag genes during larval cell divisions. However, this is clearly not the case for the Rag genes; in contrast to let-363 and daf-15, neither the raga-1 nor ragc-1 mutations result in fully penetrant larval arrest or sterility; both mutants can be maintained as homozygotes (Schreiber et al., 2010; data not shown).

For L1 diapause to be advantageous to an animal, the organism must be able to survive an extended period of starvation and resume development when conditions improve. Our studies reveal a role for AMP-activated kinase genes aak-1 and aak-2 in two key aspects of L1 diapause induced by nutrient deprivation: survival and arrest of germ cell proliferation.

Loss of AMPK activity dramatically impairs survival during L1 diapause. We speculate that energy deprivation contributes to this lethality for two main reasons. First, the lethality can be partially rescued by supplementation of the medium with glucose or ethanol. Previous studies have shown that glucose and ethanol, as well as other carbohydrates, promote population growth when included in chemically defined media, leading to the conclusion that these compounds were utilized as energy source (Braeckman et al., 2009). Second, tissue-specific expression of aak-2 in the intestine, an energy storing organ, significantly increases survival of ampk mutants. A large body of literature has established that AMPK upregulates catabolic pathways that generate ATP, and downregulates anabolic pathways that consume ATP in energy storing organs such as muscle, liver, and adipose tissue of mammals (Kahn et al., 2005); thus, it is reasonable to hypothesize that multiple metabolic pathways controlled by AMPK contribute to maintenance of survival during L1 diapause. Narbonne and Roy have shown that regulation of a triglyceride lipase, ATGL-1, by AMPK in the hypodermis plays a key role in maintaining lifespan during dauer diapause (Narbonne and Roy, 2009). Because an atgl-1::gfp reporter is expressed in the intestine (Zhang et al., 2010), regulation of ATGL-1 by AMPK in this tissue may provide an additional contribution to survival during L1 diapause.

We found that neuronal AMPK activity also plays an important role in maintaining survival during L1 diapause. Recent studies have shown that rodent hypothalamic AMPK affects expression of neuropeptides such as AgRP (Agouti-related peptide), NPY (neuropeptide Y) and POMC (proopioidmelanocortin) to control food intake (Minokoshi et al., 2004). Although C. elegans does not possess clear orthologues of these neuropeptides, multiple genes encoding members of neuropeptide Y/RFamide receptor family and RFamides are present in its genome (Li et al., 1999; Hewes and Taghert, 2001). Therefore, there might be an RFamide-receptor pair that functions downstream of AMPK in the C. elegans nervous system. Whether loss of AMPKs in C. elegans results in abnormal control of food intake, as it does in mice, remains to be examined. However, such a defect is not likely to make a significant contribution to survival during the L1 diapause, which occurs in the absence of food. Thus, roles of AMPKs in the nervous system may not simply be limited to regulation of feeding.

Abnormal germline proliferation during L1 diapause in ampk and daf-18 mutants was significantly suppressed by knockdown or elimination of TORC1 activity or its conserved activators (i.e. rheb-1, raga-1 and ragc-1). These results suggest that aak and daf-18 maintain quiescence by suppressing TORC1 activity. Similarly, rapamycin, a potent inhibitor of TORC1, can induce G2 arrest in some types of mammalian cultured cells (Gaur et al., 2011; Kapoor et al., 2012). Genetic inhibition of TOR also results in G2 delay in Saccharomyces cerevisiae and adult germline stem cells in Drosophila melanogaster and C. elegans (Nakashima et al., 2008; LaFever et al., 2010; Korta et al., 2012). Although how TORC1 affects activity of the core cell cycle machinery such as cyclins, cdk (cyclin-dependent kinase), and cdk inhibitors, remains unknown in multicellular animals (Russell et al., 2011), studies in yeast suggest that polo-like kinase CDC5, which is a regulator of the mitotic cdk, Cdc28, mediates the TOR signaling pathway that promotes G2/M transition (Nakashima et al., 2008). Similar mechanisms may function in higher organisms.

One unresolved question is, in which tissue(s) do aak and daf-18 act to regulate germline quiescence? Immunostaining studies showed that cells expressing DAF-18 include the germline precursors Z2 and Z3 (Brisbin et al., 2009), and previous studies have suggested that the aak and daf-18 genes act cell autonomously to regulate germline quiescence during diapause. The germline-cell autonomy conclusions were based on RNAi experiments carried out in rrf-1 mutants, in which RNAi was thought to be active only in the germline (Sijen et al., 2001; Narbonne and Roy, 2006; Watanabe et al., 2008). Indeed, we also found that when aak-1; rrf-1 and aak-2; rrf-1 double mutants were subjected to aak-2 and aak-1 RNAi, respectively, each strain exhibited a significant increase in failure to arrest germline proliferation during L1 diapause (data not shown). However, rrf-1 mutants were recently found to maintain RNAi activity in some somatic tissues, including the gut and hypodermis (Kumsta and Hansen, 2012), leaving the question as yet unanswered. aak-2::gfp expressed under the control of its endogenous promoter can weakly rescue the defects in germline quiescence (Fig. 3B), but its expression was not detected in Z2 or Z3 (not shown), typical for expression from extrachromosomal arrays which are often silenced in the germline. Extrachromosomally expressed daf-18 also significantly suppressed the germline phenotype during L1 diapause (Fukuyama et al., 2006). Thus, one could argue that aak-2 and daf-18 act in the soma to regulate germline quiescence. Alternatively, low levels of germline expression from extrachromosomal arrays could be responsible for the rescue; low levels of maternally expressed lin-35 from the arrays results in substantial suppression of developmental phenotypes in lin-35; fzr-1 double mutants (Fay et al., 2002). Therefore, it remains possible that aak and daf-18 act in the germline to regulate its quiescence.

Recent studies using lineage-specific pten knockout mice have demonstrated that pten is required for quiescence of haematopoietic stem cells as well as premature oocytes (Yilmaz et al., 2006; Zhang et al., 2006; Reddy et al., 2008). Thus, one of the physiological roles of pten conserved between mammals and worms may be to prevent unwanted exit from quiescence in germline and stem cells. In this respect, it would be of great interest to test whether AMPKs are also essential for maintaining quiescence of stem cells and premature oocytes in mammals.

Worm culture and genetics were conducted according to standard procedures (Brenner, 1974). Cultures were grown at 20°C unless otherwise noted. All strains used in this study were outcrossed against wild-type (Bristol N2) worms at least twice.

Sterilized embryos were hatched in 10 ml of sterile M9 with or without 0.08% ethanol in a 15 ml polypropylene tube, and hatched larvae were continuously cultured in the medium to assess survival or germline proliferation, respectively. L1 larvae were prepared as follows: about 40 to 50 gravid young adults were placed in a 10 cm “4× peptone” plate (standard NGM plus four times more peptone) seeded with OP50 E. coli. For experiments presented in supplementary material Fig. S4, 10 cm 4× peptone plates containing 25 mg/ml carbenicillin and 1 mM IPTG seeded with HT115 harboring RNAi construct were used. Adult F1 animals were washed into a 15 ml tube in M9, gravid adults were allowed to settle, and the supernatant was removed. Embryos were purified by the addition of 5 ml of alkali/bleach solution (0.5 M KOH; 10% sodium hypochlorite solution [Wako Inc, Japan]), washed three times with M9, and incubated in 10 ml of M9 to assess viability or M9 plus 0.08% ethanol to determine the average number of germline precursors. Incubation was at 20°C with 30 rpm rotation (Rotamix, ATR, Inc.). Germline stem cells and somatic gonadal precursors were identified by inspection (Kimble and Hirsh, 1979) using a Zeiss Axio Imager M1.

To assess the viability for Fig. 1B,C and supplementary material Fig. S4, about 50 µl of culture were placed on a slide, and animals exhibiting any movement were scored as alive, as in previous studies (Baugh and Sternberg, 2006). Animals lacking movement were reassessed with prodding. For the viability assay in Fig. 3A,B, starved animals were mounted on an agar pad, and the transgenic animals were identified by GFP expression. Motility was used as a proxy for viability. Motility and tissue integrity, which would be lost quickly once the animal were starved to death, were well correlated.

About 50 gravid adults were plated onto a 10 cm 4× peptone plates. After 4 days embryos were isolated and incubated 10 ml M9 plus 0.08% ethanol for 5 days. About 1,000 larvae were then plated onto a 100 mm 4× peptone plate and 3 to 4 days later young adults and L4 animals were randomly singled onto seeded NGM plates and fertility and viability were assessed 5 days later. To assess fertility of the continuously-fed animals, L4 larvae were randomly singled onto seeded NGM plates and scored 5 days later.

Feeding RNAi was conducted according to the standard protocols (Kamath et al., 2003). Sterilized embryos from RNAi-treated animals were treated as already described except that NGM plates seeded with HT115 bacteria harboring the appropriate RNAi vector were used. For RNAi targeting of rheb-1, approximately 1000 sterilized embryos were plated onto three 10 cm RNAi plates. For let-363(RNAi) and daf-15(RNAi), this resulted in many animals arresting during larval stages. Thus, for these experiments, larvae were synchronously grown on OP50-seeded NGM plates until the L3 to L4 stage and then were transferred to RNAi plates.

We thank Keiko Ando, Cori Bargmann, Andy Fire, Oliver Hobert, Yuji Kohara, Morris Maduro, Shohei Mitani, and Susan Strome for reagents and strains. Some strains used in this study were provided by the Caenorhabditis Genetics Center, which is funded by the National Institutes of Health, and the Mitani Laboratory at the Tokyo Women's Medical University School of Medicine (National Bioresource Project for the Experimental Animal “Nematode C. elegans”). This work was supported by Grant-in-Aid for Young Scientists (B) from Japan Society for the Promotion of Science (JSPS), and Grants-in-Aid for Scientific Research on Innovative Areas, ‘Regulatory Mechanism of Gamete Stem Cells’ and Scientific Research on Priority Areas, ‘G Protein Signaling’ from The Ministry of Education, Culture, Sports, Science and Technology (MEXT). Work in the Rougvie lab was supported by grants from the National Science Foundation (IOB-0515682; A.R.) and American Heart Association Midwest Affiliate (M.F.).