Given that endocrine signaling across tissues is a prominent feature of reproductive longevity, we examined whether wild-type PD_Stv adults shared any gene expression signatures with animals lacking a functional germ line. In glp-1 mutants, increased longevity is dependent on TOR (target of rapamycin) signaling, DAF-16/FOXO gene regulation, steroid hormone signaling, and fatty acid metabolism regulation (Lapierre and Hansen, 2012). With the exception of TOR signaling, we found significant gene expression changes in PD_Stv adults of key genes in each of these regulatory pathways (Supplementary file 1).

In the steroid signaling pathway, dafachronic acid (DA) biosynthesis requires the cytochrome P450 DAF-9, the Rieske-like oxygenase DAF-36, the short-chain dehydrogenase DHS-16, and the hydroxysteroid dehydrogenase HSD-1 (Figure 1A; Gerisch et al., 2001; Jia et al., 2002; Motola et al., 2006; Rottiers et al., 2006; Patel et al., 2008; Wollam et al., 2012; Mahanti et al., 2014). In animals lacking a functional germ line, the levels of daf-36 mRNA and Δ⁷-DA are significantly increased compared to wild type (Shen et al., 2012). We found that in wild-type PD_Stv adults with a germ line, daf-36 mRNA increased threefold (p = 3.25e-04; FDR = 0.01) compared to control adults (Supplementary file 1; Ow et al., 2018). To investigate whether DA signaling plays a role in mediating reproductive plasticity as a result of early-life experience, we asked whether mutations in DA biosynthesis genes would affect the reduced brood size observed in PD_Stv adults. We found that daf-9(rh50), daf-36(k114), and dhs-16(tm1890) mutants did not exhibit a significant decrease in brood size in PD_Stv adults compared to controls (CON), while hsd-1(mg433) brood sizes were similar to wild type. Interestingly, daf-9 and daf-36 mutants exhibited a significant increase in postdauer brood size compared to controls, opposite of what we observed in wild type (Figure 1B).

We next asked whether the steroid signaling that contributed to the reduced fecundity in postdauers acted through two related NHRs, DAF-12 and NHR-8. Since null mutants of daf-12 are dauer defective, we used two daf-12 mutant alleles, rh284 (Class 5) and rh285 (Class 4), with lesions in the ligand binding domain that affect steroid signaling activity but otherwise have a relatively normal dauer phenotype (Antebi et al., 2000). We found that daf-12(rh284) and daf-12(rh285) exhibited a significant increase in PD_Stv brood size compared to controls, similar to that observed for the daf-9 and daf-36 DA biosynthesis mutants (Figure 1B). The closest gene relative to daf-12, nhr-8, is also upregulated 2.5-fold in PD_Stv adults (Lindblom et al., 2001; Magner et al., 2013; Ow et al., 2018); however, nhr-8(ok186) adults exhibited a significant brood size reduction in PD_Stv animals compared to controls similar to wild type (Figure 1B). One possible explanation of this observation is that passage through dauer may partially rescue the reproductive phenotypes of the daf-12, daf-36, and daf-9 mutants, as has been previously observed for hypodermal and vulval precursor cell fates in postdauer heterochronic mutants (Liu and Ambros, 1991; Euling and Ambros, 1996), or that DAF-12 activity in the absence of DA is sufficient for reproduction in PD_Stv animals. Another explanation is that DA-dependent DAF-12 activity is required for the early-life starvation memory that programs a decrease in PD_Stv fecundity, and its loss results in a reproductive phenotype similar to what we have observed previously in pheromone-induced postdauers (Ow et al., 2018). Although we cannot distinguish between these possibilities with this data, we favor the latter explanation given the abrogation of the decreased fecundity phenotype in dhs-16 mutants, which lack significant reproductive defects as determined by the similarity of the brood size of dhs-16 control adults to wild type (Figure 1B).

DAF-16 and PQM-1 act in a mutually antagonistic manner to promote the expression of a group of stress response genes (Class I) or genes associated with growth and reproduction (Class II), respectively (Figure 2A; Tepper et al., 2013). We found that the set of genes with significant changes in mRNA levels between PD_Stv and controls was enriched for Class I and II targets (Figure 2—figure supplement 1A; Figure 2—figure supplement 1—source data 1). In addition, we found two genes that regulate DAF-16 cellular localization, pqm-1 and daf-18, were significantly up and downregulated, respectively, in PD_Stv adults compared to controls (Supplementary file 1; Ow et al., 2018). DAF-18 is the functional ortholog of the human PTEN tumor suppressor gene that promotes DAF-16 nuclear localization (Ogg and Ruvkun, 1998; Gil et al., 1999; Mihaylova et al., 1999; Solari et al., 2005). These observations prompted us to ask whether PQM-1 and DAF-18 contribute to the reduced fertility in PD_Stv adults by altering the regulation of genes involved in reproduction and sequestering DAF-16 in the cytoplasm. However, the brood size differences between control and PD_Stv adults in pqm-1(ok485) and daf-18(e1375) mutants were similar to wild type, indicating that gene regulation by PQM-1 is unlikely to contribute to the PD_Stv reduced fecundity (Figure 2B). Next, because daf-16 null mutants are dauer defective, we crossed a daf-16(mu86) null allele with a strain carrying a rescue transgene (daf-16a^AM::gfp) that constitutively localizes DAF-16 to the nucleus (Lin et al., 2001). Similar to what was observed for wild type, the daf-16(mu86); daf-16a^AM::gfp transgenic strain displayed a reduced brood size in PD_Stv compared to controls (Figure 2B). Since daf-18(e1375) is a hypomorph, we next tested the possibility that DAF-16 nuclear localization may play a role in regulating postdauer reproduction. First, we tested SMK-1/PPP4R3A, which promotes the nuclear localization of DAF-16 when animals are exposed to pathogenic bacteria, ultraviolet irradiation, or oxidative stress (Wolff et al., 2006), and found that smk-1(mn156) mutants continued to exhibit a decreased PD_Stv fertility compared to controls (Figure 2B). In addition, we examined the cellular localization of daf-16p::daf-16a/b::gfp transgene in two independent strains (CF1139 and TJ356) and observed diffuse cytoplasmic signal in intestinal cells of both control and PD_Stv adults (Figure 2—figure supplement 2). Together, these results do not support a role for DAF-16 per se in the diminished fertility phenotype in postdauers.

In the reproductive longevity pathway, two genes, kri-1 (ortholog of the human intestinal ankyrin-repeat protein KRIT1/CCM1) and tcer-1 (homolog of the human transcription elongation factor TCERG1), are required for DAF-16 nuclear localization and increased longevity in germ line-less animals (Berman and Kenyon, 2006; Ghazi et al., 2009). TCER-1 regulates target genes in both a DAF-16-dependent and independent manner (Amrit et al., 2016). Since we determined that DAF-16 itself does not contribute to the PD_Stv reproductive phenotype, we tested whether TCER-1 and KRI-1 regulate PD_Stv reproduction independent of DAF-16. Interestingly, we found that the decreased fecundity in PD_Stv was abrogated in kri-1(ok1251) and tcer-1(tm1452) strains (Figure 2B), indicating that KRI-1 and TCER-1 are required for the reproductive plasticity in PD_Stv animals in a DAF-16-independent manner.

In animals lacking a germ line, DAF-16 and TCER-1 are required to bolster the expression of lipid biosynthesis, storage, and hydrolysis genes to promote adult longevity (Amrit et al., 2016). Similar to germ line-less glp-1 mutants, PD_Stv adults exhibited a significantly altered expression of ~26% (33 of 127 genes) of all the fatty acid metabolic genes, including ~46% (18 of 39 genes) also targeted by DAF-16 and TCER-1 (Figure 2—figure supplement 1B; Figure 2—figure supplement 1—source data 2). To investigate whether reduced fecundity of PD_Stv is modulated by upregulating fatty acid metabolism, we asked if mutations in known regulators of lipid metabolism would exhibit changes in brood size in PD_Stv adults when compared to controls. One of the genes jointly upregulated by DAF-16 and TCER-1 is nhr-49, a nuclear hormone receptor homologous to the mammalian HNF4α lipid sensing nuclear receptor involved in the regulation of fatty acid metabolism and the oxidative stress response (Ratnappan et al., 2014; Amrit et al., 2016; Moreno-Arriola et al., 2016; Goh et al., 2018; Hu et al., 2018). Additional nuclear hormone receptors, NHR-80, NHR-13, and NHR-66, and the Mediator complex subunit MDT-15, partner with NHR-49 and co-regulate the expression of genes involved in various aspects of lipid metabolism such as fatty acid β-oxidation, transport, remodeling, and desaturation (Gilst et al., 2005; Van Gilst et al., 2005; Taubert et al., 2006; Nomura et al., 2010; Pathare et al., 2012; Ratnappan et al., 2014; Folick et al., 2015; Amrit et al., 2016). In addition, SBP-1 (homolog of mammalian SREBP) and NHR-49 are co-regulated by MDT-15 as part of a transcriptional network coordinating the expression of delta-9 (Δ9) fatty acid desaturase genes (Figure 3—figure supplement 1A; Yang et al., 2006; Taubert et al., 2006).

When we examined the control and PD_Stv brood sizes of strains carrying mutations in fatty acid metabolism genes, we found that the reduced fecundity of PD_Stv characteristic of wild-type animals was also observed in nhr-80(tm1011), nhr-13(gk796), and in the nhr-49(ok2165) allele (Figure 3A). Postdauers expressing nhr-49 gain-of-function (gf) alleles et7 or et13 also exhibited reduced brood size (Figure 3A). However, nhr-66(ok940), mdt-15(tm2182), and sbp-1(ep79) strains, in addition to three nhr-49 alleles, nr2041, gk405, and the et8 gf, failed to exhibit the decreased fecundity in PD_Stv adults compared to CON (Figure 3A). The gf nhr-49 alleles, et7, et8, and et13, harbor missense lesions located at or near the ligand-binding domain (Svensk et al., 2013; Lee et al., 2016). The nature of et7 and et13 could modify NHR-49 activity following the dauer experience, resulting in a significant decrease in postdauer brood size (Figure 3A). Because the six nhr-49 mutant alleles differ in the nature and the location of their lesions, their physiological function could vary and result in disparate reproductive phenotypes.

Amongst the various nhr-49 alleles that eliminated the PD_Stv reproductive phenotype, we chose to use the well-characterized nr2041 for further experiments owing that it is a complete loss-of-function mutant whose lesion encompasses a deletion in its DNA binding domain as well as over half of its ligand-binding domain (Gilst et al., 2005; Van Gilst et al., 2005; Pathare et al., 2012). Because of the interaction between NHR-49, NHR-80, NHR-13, and NHR-66, we also examined if double and triple mutants of these NHRs would have any reproductive plasticity phenotypes. Strains carrying mutations in nhr-80, nhr-13, or nhr-66 in addition to nhr-49 showed an abrogated phenotype compared to wild type (Figure 3B). A triple mutant strain, nhr-49; nhr-80; nhr-13, showed a similar abrogated phenotype (Figure 3B). In contrast, the nhr-80; nhr-13 double mutant exhibited a wild-type phenotype, indicating the importance of NHR-49 in regulating PD_Stv brood size (Figure 3B). Together, these data suggest that SBP-1, MDT-15, NHR-49, and interacting NHR, NHR-66, are important in the postdauer reproduction program, likely by upregulating fat metabolism genes.

NHR-49, MDT-15, and SBP-1 upregulate the expression of genes involved in fatty acid biosynthesis, including the Δ9 desaturases, fat-5, fat-6, fat-7, and the delta-12 (Δ12) desaturase fat-2 (Yang et al., 2006; Nomura et al., 2010; Han et al., 2017). FAT-5, FAT-6, and FAT-7 convert saturated fatty acids (SFAs) to mono-unsaturated fatty acids (MUFAs), while FAT-2 catalyzes the conversion of MUFAs to poly-unsaturated fatty acids (PUFAs) (Figure 3—figure supplement 1B; Watts and Ristow, 2017). Our previous mRNA-Seq results showed that the expression of fat-5, fat-6, fat-7, and fat-2 increased significantly between 3.8- and 26.6-fold in wild-type PD_Stv adults compared to controls (Supplementary file 1; Ow et al., 2018). When we compared the PD_Stv brood size to controls for these mutant strains, fat-6 and fat-2 exhibited an abrogated phenotype, while fat-5 and fat-7 strains retained the decreased brood size phenotype similar to wild type (Figure 3C). Furthermore, the double mutant strains with combinations of mutations in fat-5, fat-6, and fat-7 genes all exhibited an elimination of the decreased brood size phenotype (Figure 3C). These results suggest: (1) a functional redundancy between the Δ9 fatty acid desaturases in modulating lipid homeostasis of PD_Stv adults, with FAT-6 playing a more principal role than FAT-5 and FAT-7; and (2) MUFA and PUFA levels may be upregulated to promote the decreased fertility phenotype in PD_Stv adults compared to controls.

In C. elegans, MUFAs are essential for viability as a fat-5; fat-6; fat-7 triple mutant is lethal (Brock et al., 2006). MUFAs, such as oleic acid (OA), can be remodeled to become PUFAs, phospholipids, and neutral lipids such as triacylglycerols (TAG), which serve as energy storage molecules in the intestine, hypodermis, and germ line (Figure 3—figure supplement 1B; Watts and Ristow, 2017). In addition to acting as key regulators of fat metabolism, FAT-5, FAT-6, and FAT-7 are also essential in promoting the long lifespan of adult worms lacking a germ line (Gilst et al., 2005; Brock et al., 2006; Goudeau et al., 2011; Ratnappan et al., 2014). Given that MUFAs may be required for the reduced fecundity of PD_Stv adults, we asked whether the dietary addition of OA to PD_Stv animals would further reduce their brood size. To test this, we compared the brood size of PD_Stv adults fed E. coli OP50 grown with OA and PD_Stv adults whose bacterial diet was not pre-loaded with OA. We tested the N2 wild type, nhr-49, and Δ9 desaturase double mutant strains. We found that for wild type and the strains that included a mutation in nhr-49, the supplementation of OA significantly increased PD_Stv adult fecundity compared to the control diet (Figure 4A). In addition, the fat-6; fat-5 double mutant strain continued to exhibit a significant increase in brood size when fed food supplemented with OA. However, the Δ9 desaturase double mutant strains carrying a mutation in fat-7, fat-7; fat-5 and fat-6; fat-7, did not exhibit an OA-induced increase in brood size (Figure 4A). These results suggest that OA is not required for decreased fecundity but may rather be a limiting factor for reproduction after passage through the starvation-induced dauer stage, whether for nutrition or as a signaling molecule across tissues to regulate physiology (Schmeisser et al., 2019; Starich et al., 2020). These results also indicate that fat-7 is required for the OA-induced increase in brood size, which was unexpected given that it acts upstream of OA in fatty acid synthesis and suggests that FAT-7 has additional roles in fatty acid metabolism (Figure 3—figure supplement 1B; Watts, 2009). Altogether, these results suggest that the upregulation of fatty acid desaturases are critical for the decreased fertility in PD_Stv adults by mediating and promoting the synthesis of sufficient levels of lipids needed for reproduction after the animals experienced starvation-induced dauer formation.

In long-lived glp-1 mutants lacking a germ line, fat stores are increased relative to wild type (Steinbaugh et al., 2015; Amrit et al., 2016). Given that PD_Stv adults have an increased expression of lipid metabolism genes similar to glp-1, but have limited quantities of OA for reproduction, we questioned what the status of fat stores would be in wild-type PD_Stv adults with an intact germ line. Using Oil Red O (ORO) staining, we compared the amounts of neutral triglycerides and lipids (O'Rourke et al., 2009) in PD_Stv one-day-old adults and developmentally matched controls. Despite having a significant upregulation in fatty acid metabolism genes, we found that PD_Stv adults have a significantly reduced amount of stored lipids relative to controls in their intestine (Figure 4B and C; Figure 4—figure supplement 1A and B). These results are consistent with a model that PD_Stv adults have increased expression of lipid metabolism genes for reproduction rather than somatic lipid storage.

Next, we investigated whether the decreased lipid stores in PD_Stv adults was dependent upon DAF-12 and TCER-1. DA-dependent DAF-12 activity was shown previously to promote fat utilization for reproduction (Wang et al., 2015). In contrast to previous results, we found that control adults of both daf-12(rh284) and daf-12(rh285) strains displayed low levels of lipid storage, and postdauer adults exhibited a significant increase in lipid storage relative to controls (Figure 4B and C; Figure 4—figure supplement 1B). Interestingly, the levels of intestinal ORO staining positively correlated with control and PD_Stv brood sizes in wild type and the daf-12 mutants, and the daf-12 mutant postdauers have statistically similar lipid stores compared to wild-type postdauers, further supporting the conclusion that DA-dependent DAF-12 activity is not required in postdauers to regulate lipid storage and reproduction (Figures 1B, 4B and C; Figure 4—figure supplement 1B). In contrast, tcer-1(tm1452) lipid staining was diminished in PD_Stv compared to controls, similar to wild type (Figure 4B and C; Figure 4—figure supplement 1B). However, since both tcer-1(tm1452) control and PD_Stv adults have reduced fat stores compared to their wild-type counterparts (Figure 4B and C), and TCER-1 is known to positively regulate NHR-49 and fatty acid metabolism, this result may be due to the inability of these animals to store fat and not because TCER-1 regulates the levels of stored fat in PD_Stv adults. These results suggest that fine-tuning the balance of somatic lipid stores between the CON and PD_Stv life histories may be correlated with reproductive output.

Given that fatty acid metabolism is important for regulating fecundity in postdauer animals, we profiled fatty acids in wild-type control and PD_Stv adults. Quantification of the level of SFAs, MUFAs, and PUFAs revealed that most of these fatty acids, including oleic acid, remained unchanged in control and PD_Stv adults. Only two PUFAs, α-linolenic acid (ALA or C18:3n3) and dihomo-γ-linolenic (DGLA or C20:3n6) were significantly downregulated and upregulated, respectively, in PD_Stv adults (Figure 4—figure supplement 1C). ALA is an omega-3 fatty acid whose level is augmented in glp-1 animals and is reported to extend adult lifespan in a manner that is dependent on NHR-49 and the SKN-1/Nrf2 transcription factor (Ratnappan et al., 2014; Amrit et al., 2016; Qi et al., 2017). Dietary supplementation of omega-6 fatty acid, DGLA, has been shown to trigger sterility through ferroptosis, an iron-dependent germ line cell death resulting from the production of toxic lipid metabolites (Deline et al., 2015; Perez et al., 2020). Through the activities of the fat-2, fat-1, fat-3, elo-1/2 and/or let-767 genes, oleic acid serves as the substrate for the production of ALA and DGLA (Figure 3—figure supplement 1B). Interestingly, with the exception of fat-2, we found that the mRNA levels of these genes were significantly increased in postdauers that experienced starvation but not crowded conditions (Ow et al., 2018). Thus, the fatty acid profiling suggests that a complex interplay between various PUFAs and their biosynthesis, and not the levels of oleic acid per se, may play a role in modulating fecundity in postdauers adults that experienced starvation.

Next, we investigated how PD_Stv animals prioritize fat utilization for reproduction rather than intestinal storage. During C. elegans reproduction, intestinal fat stores are reallocated into low-density lipoproteins (LDL)-like particles (yolk lipoproteins or vitellogenins) that are incorporated into oocytes through receptor-mediated endocytosis in a process called vitellogenesis to supply nutrients to the developing embryos (Kimble and Sharrock, 1983; Grant and Hirsh, 1999). Six vitellogenins homologous to the human ApoB proteins are encoded in the C. elegans genome, vit-1 through vit-6, and concomitant multiple RNAi knockdown of the vit genes increases adult lifespan in a process that requires NHR-49 and NHR-80 (Spieth et al., 1991; Seah et al., 2016). Because vitellogenesis mobilizes intestinal fat resources for reproduction and depletes somatic lipid stores (Kimble and Sharrock, 1983), we hypothesized that PD_Stv adults have reduced fat reservoirs because they prioritize vitellogenesis as a reproductive investment over intestinal storage. To test this, we first examined the lipid content of PD_Stv and CON adult embryos in utero using ORO staining. In contrast to the intestinal fat stores, we observed that ORO staining of PD_Stv embryos was significantly increased compared to CON embryos (Figure 4D and E). Next, we examined the lipid content of CON and PD_Stv embryos of daf-12 mutant strains. If intestinal lipid storage is indeed negatively correlated with the amount of vitellogenesis, we would predict that daf-12 adults would exhibit the opposite phenotype compared to wild type, with daf-12 PD_Stv embryos having less fat than CON embryos. Instead, we observed that daf-12 PD_Stv and CON embryos have similar levels of ORO staining, all of which were significantly higher than the wild-type levels (Figure 4D and E). In contrast to our previous results in Figure 4B, these results indicate that DAF-12 does play a role in lipid allocation in control and postdauer animals, potentially by regulating vitellogenesis.

Furthermore, we examined the effects on the intestinal fat stores in control and PD_Stv adults when vitellogenesis is disrupted by RNAi knockdown of vit-1, which also results in the knockdown of vit-2/3/4/5 due to the high-sequence homology amongst the vit genes (Seah et al., 2016; Figure 5—figure supplement 1). In animals treated with the empty vector (EV) negative control, PD_Stv adults continued to exhibit decreased fat stores compared to control adults. However, PD_Stv adults treated with vit-1 RNAi have significantly greater intestinal fat deposits than PD_Stv negative controls (Figure 5A and B), indicating that increased vitellogenesis in PD_Stv adults may be a contributing factor to the lack of stored intestinal fat. Altogether, these results suggest that PD_Stv adults utilize fat accumulated after diapause for reproduction and not somatic storage.

C. elegans lifespan is dependent on nutrition and intestinal fat stores. Both animals with increased intestinal fat storage, such as glp-1 animals, and animals with decreased fat storage, such as dietary restricted animals, exhibit prolonged lifespan (Kenyon, 2010a; Kenyon, 2010b). To examine whether the decrease in fat stores in postdauers would affect their longevity, we measured adult lifespan in wild-type controls and postdauers and found that postdauers have a significantly increased longevity compared to controls (Figure 5C). Interestingly, we previously reported that crowding-induced postdauers also exhibited increased mean lifespan, suggesting that the increased longevity of PD_Stv adults may be due to passage through dauer and not their specific early-life stress (Hall et al., 2010). We next asked if a disruption in PD_Stv vitellogenesis resulting in increased level of intestinal fat would affect their lifespan by performing vit-1 knock-down. First, we again observed postdauers fed with EV RNAi lived longer than controls; however, this lifespan differential between controls and postdauers was eliminated when animals were fed with vit-1 RNAi (Figure 5D). Both control and PD_Stv animals treated with vit-1 RNAi exhibited a significant increase in lifespan when compared to cognate animals fed with empty vector (EV) control RNAi, consistent with previous reports of vit-1 RNAi prolonging lifespan (Figure 5D; Murphy et al., 2003; Seah et al., 2016). However, inhibiting vitellogenesis appears to result in a particular threshold for increased longevity instead of an additive effect, resulting in a similar median lifespan regardless of life history (Figure 5D). We next asked whether fecundity was compromised in PD_Stv animals as a result of vit-1 knock-down. Consistent with our previous results, postdauer animals fed with EV RNAi showed a decreased in brood size compared to controls; however, this brood size difference was eliminated following vit-1 RNAi (Figure 5—figure supplement 2). Together, these results support the model that the complex crosstalk between the intestine and germ line shown to regulate somatic aging is also mediating the physiology of postdauer adults (see Discussion).

Our results suggest that PD_Stv animals upregulate their fatty acid metabolism to increase lipid transport to embryos. In humans, nutritional stress in utero not only promotes metabolic syndrome in adulthood of the affected individuals, but also promotes obesity in subsequent generations (Painter et al., 2008; Veenendaal et al., 2013). Therefore, we investigated whether subsequent generations inherit the starvation memory by examining if they exhibit any postdauer aging, reproduction, or lipid storage phenotypes. First, we tested if PD_Stv F₁ progeny had significantly increased longevity compared to CON F₁ progeny, but we found no significant differences in lifespan between the two populations (Figure 5C). Next, we assessed whether increased fat content in PD_Stv embryos affected the reproduction of F₁ adults by measuring the brood sizes of CON and PD_Stv F₁ and F₂ progeny, but we found no significant differences beyond the parental generation (Figure 5—figure supplement 3). Finally, we examined whether PD_Stv progeny exhibited altered lipid content by quantitating intestinal fat storage in control and postdauer F₁ and F₂ adults using ORO staining. We observed that adult F₁ progeny of PD_Stv adults had an increased level of stored fat compared to F₁ progeny of control adults, but the difference in lipid storage was abolished in the F₂ generation (Figure 6; Figure 6—figure supplement 1). These results indicate that the F₁ progeny of PD_Stv adults inherit a starvation memory that results in metabolic reprogramming to increase their stored fat reserves.

In C. elegans, small RNA pathways often mediate the inheritance of gene expression states (Feng and Guang, 2013; Rechavi and Lev, 2017). The germline-specific, nuclear Argonaute HRDE-1/WAGO-9 promotes transgenerational silencing via the formation of heterochromatin at targeted genomic loci (Buckley et al., 2012; Feng and Guang, 2013; Rechavi and Lev, 2017). In addition, starvation-induced L1 diapause alters the small RNA populations of subsequent generations in a HRDE-1-dependent manner (Rechavi et al., 2014). We therefore asked whether HRDE-1 was required for the generational inheritance of starvation memory by quantifying fat storage in hrde-1(tm1200) control and PD_Stv adults and their F₁ and F₂ progeny. Similar to the wild-type P₀ generation, the stored fat levels in hrde-1 P₀ PD_Stv adults was significantly reduced when compared to hrde-1 controls (Figure 6; Figure 6—figure supplement 1). However, the ORO staining between control and PD_Stv adults in both the hrde-1 F₁ and F₂ progeny was statistically similar (Figure 6; Figure 6—figure supplement 1), indicating that HRDE-1 is required for the F₁ inheritance of the parental starvation memory that promotes increased lipid storage in wild-type animals.

To further investigate the role of the generational inheritance of a starvation memory, we examined the effect of the prg-1(tm872) mutation on the fat stores of control and PD_Stv adults and their F₁ and F₂ progeny. PRG-1 is a Piwi-class Argonaute that acts upstream of HRDE-1 to perpetuate transgenerational epigenetic memory in the germ line (Ashe et al., 2012; Shirayama et al., 2012). We found that in the P₀ generation, PD_Stv prg-1(tm872) adults exhibited a significant decrease in stored fats compared to controls, similar to wild type (Figure 6; Figure 6—figure supplement 1). Also similar to wild type, the F₁ progeny of prg-1 PD_Stv animals exhibited increased intestinal lipid storage. However, the F₂ progeny of CON and PD_Stv prg-1 mutants continued to exhibit the increased PD_Stv fat stores phenotype instead of ‘resetting’ like in the wild type (Figure 6; Figure 6—figure supplement 1), suggesting that PRG-1 plays a role in erasing the starvation memory inherited from PD_Stv adults in the F₂ generation. Although PRG-1 and HRDE-1 work in the same nuclear RNAi pathway required for transgenerational inheritance, our ORO staining show that these proteins play different roles in the transmission of an ancestral starvation memory. Namely, HRDE-1 promotes the inheritance of the starvation memory to the next generation, and PRG-1 halts the propagation of an ‘expired’ memory to the grand-progeny. In addition, the P₀ CON of adults of hrde-1 and prg-1 mutant strains have significantly increased stored fat compared to wild type, indicating these pathways also seem to play a role in regulating lipid storage in continuously developing animals (Figure 6; Figure 6—figure supplement 1; Figure 7—figure supplement 2). Altogether, our results show that C. elegans, like humans, inherit the disposition for increased adiposity from parents that experienced early-life starvation.

Thus far, our results indicate that the fatty acid metabolism and reproductive longevity pathways are required for the reduced fecundity phenotype in PD_Stv adults. While our data suggests a role for DA-dependent DAF-12 activity in regulating vitellogenesis, its contribution to regulating the reduced fecundity of postdauer adults is less certain given the multiple possible interpretations of our results. Moreover, how these pathways are potentially interacting to regulate PD_Stv phenotypes is unclear. The daf-12, nhr-49, and tcer-1 PD_Stv mutant phenotypes are distinct, suggesting the possibility that they may act in different pathways, tissues, or developmental time points to regulate PD_Stv fecundity. Furthermore, strains carrying double mutations in daf-12 and either tcer-1 or kri-1 have control and postdauer brood sizes similar to the daf-12 mutations alone, suggesting that daf-12 alleles are acting downstream and masking the phenotypes of the reproductive longevity mutants (Appendix 1; Figure 1—figure supplement 1B).

To further investigate the developmental mechanism of steroid signaling, reproductive longevity, and fatty acid metabolism pathways in the regulation of reproductive plasticity, we examined if DAF-12, TCER-1, and NHR-49 play a direct role in the timing of germline proliferation in postdauer larvae. We previously demonstrated that wild-type PD_Stv animals delay the onset of germline proliferation compared to control animals, contributing to a reduction in brood size (Ow et al., 2018). In C. elegans hermaphrodites, undifferentiated germ cells initiate spermatogenesis during the L4 larval stage, followed by a transition to oogenesis at the adult stage (L'Hernault, 2006). Therefore, reproduction in C. elegans hermaphrodites is sperm-limited (Byerly et al., 1976; Ward and Carrel, 1979; Kimble and Ward, 1988). In previous results, when control and PD_Stv somatic development was synchronized using vulva morphology, we observed significantly fewer germ cell rows in PD_Stv larvae compared to control larvae, correlating with fewer sperm available for self-fertilization in adulthood (Ow et al., 2018). To determine if DAF-12, TCER-1, and NHR-49 play a direct role in germline proliferation as a mechanism to regulate reproductive plasticity, we counted the number of germ cell rows in control and PD_Stv mutant larvae that were developmentally synchronized by their somatic morphology. Because daf-12(rh284) and daf-12(rh285) mutants have altered gonad morphologies that prevent accurate synchronization, we used daf-36(k114) to disrupt the steroid signaling pathway. First, we recapitulated our previous results showing that wild-type PD_Stv larvae have significantly fewer total germ cell rows compared to control larvae due to significantly reduced cell rows in the meiotic transition zone (Figure 7; Figure 7—figure supplement 1). In contrast, daf-36(k114) control and PD_Stv larvae did not exhibit a significant difference in total, mitotic, or meiotic cell rows, indicating that DAF-36-dependent DA is required during early germline development for the delay in PD_Stv germ cell proliferation (Figure 7; Figure 7—figure supplement 1). This result is consistent with the increased brood size of PD_Stv adults in daf-12(rh284) and daf-12(rh285) mutants, which express DAF-12 proteins unable to bind to DA.

In contrast to the daf-36 mutant, the nhr-49(nr2041) mutant exhibited a significant decrease in the number of total germ cell rows in PD_Stv larvae relative to control larvae, similar to wild-type animals (Figure 7), although the numbers of germ cell rows in the mitotic and meiotic zones of nhr-49 control and PD_Stv animals were not significantly different (Figure 7—figure supplement 1). These results suggest that NHR-49 acts later in development to regulate the PD_Stv brood size, perhaps by modulating fatty acid metabolism in the intestine to support vitellogenesis in adults. Like NHR-49, TCER-1 plays an intestinal role in upregulating fatty acid metabolism genes in animals lacking germ cells, but also functions redundantly with PUF-8, a member of an evolutionarily conserved stem cell proliferation modulatory family, to potentiate germ cell proliferation in animals undergoing continuous development (Pushpa et al., 2013; Amrit et al., 2016). We found that a mutation in tcer-1 resulted in a similar number of total germ cell rows in control and PD_Stv larvae, indicating that TCER-1 plays a role in PD_Stv reproductive plasticity by regulating germ cell proliferation in addition to its intestinal role (Figure 7). Interestingly, tcer-1 PD_Stv larvae also appear to have a defect in the onset of meiosis, since PD_Stv larvae have significantly greater number of mitotic germ cell rows compared to controls (Figure 7—figure supplement 1). Together, these results indicate that DAF-12 and TCER-1 act early in germline development to delay the onset of germ cell proliferation in PD_Stv animals, while NHR-49 primarily functions later in development to promote the reduced fertility of PD_Stv adults.

N2 Bristol wild-type strain was used as the reference strain. Worms were grown at 20°C unless otherwise indicated in Nematode Growth Medium (NGM) seeded with Escherichia coli OP50 using standard methods (Brenner, 1974; Stiernagle, 2006). Mutants that were not previously backcrossed were backcrossed at least four times to our laboratory N2 wild type before use. Control and starvation-induced postdauer animals were obtain in a similar manner as described before (Ow et al., 2018). Briefly, to gather PD_Stv animals, well-fed worms grown on seeded NGM plates and monitored until the bacteria food was depleted and dauers were visible (about 1 week). Dauers were selected with 1% SDS, followed by recovery by feeding on seeded NGM plates. One-day-old PD_Stv adults were collected on day two following recovery (first day of adulthood). Control adults were obtained by collecting embryos from hypochlorite-treated well-fed gravid adults that did not experience dauer. Embryos were grown on seeded NGM plates until the first day of adulthood. All strains used in this study are listed in Supplementary file 2.

Ten L4 larvae were placed individually onto 35 mm NGM plates seeded with E. coli OP50 and incubated at 20°C. Animals were transferred daily to fresh 35 mm NGM plates until egg laying ceased. Live progeny from each egg laying plate were counted. Assays were performed from at least three biological independent replicates.

Animals were induced into dauer by starvation as well as recovered on peptone-less NGM plates seeded with E. coli OP50 pre-loaded with oleic acid (NuChek Prep, Inc.; Elysian, Minnesota) as described by Devkota et al., 2017. OP50 was grown overnight at 37°C in liquid LB supplemented with 600 μM of oleic acid or with an equivalent volume of ethanol (the oleic acid solvent) to serve as the control. Cultures were pelleted and washed several times with M9 buffer (Stiernagle, 2006) and resuspended at a 10x concentration. The 10x OP50 was seeded onto peptone-less NGM plates and allowed to dry overnight before use. At least three independent replicates were performed.

Fat stores were stained using ORO dye as described by O'Rourke et al., 2009. Age matched one-day-old adults were washed from 60 mm seeded NGM plates with 1x PBS pH 7.4 and rinsed 3–4 times until they were cleared of bacteria. Worms were permeabilized in 1x PBS pH 7.4 with an equal volume of 2x MRWB buffer (160 mM KCl, 40 mM NaCl, 14 mM Na₂EGTA, 1 mM spermidine-HCl, 0.4 mM spermine, 30 mM PIPES pH 7.4, 0.2% β-mercaptoethanol) and supplemented with 2% paraformaldehyde. Samples were rocked for 1 hr at room temperature. Following fixation, worm samples were washed with 1x PBS pH 7.4, resuspended in 60% isopropanol, and incubated at room temperature for 15 min. An ORO stock solution (prepared beforehand as a 0.5 g/100 mL in isopropanol and equilibrated for several days) was diluted to 60% with dH₂O and rocked for at least one day to be used as the working stock. The ORO working stock was filtered through a 0.22 or 0.45 μm filter immediately before use. Fixed worms were incubated in filtered ORO working stock and rocked overnight at room temperature. Next day, worm samples were allowed to settle and the ORO stain was removed. Worm pellets were washed once with 1x PBS pH 7.4 and resuspended in 200 μL of 1x PBS with 0.01% Triton X-100. Aliquots of worm samples were mounted onto microscope glass slides and imaged. Quantification of embryo ORO staining was done by singling out individual embryos in utero from one-day-old adults. Images were captured with a Nikon Eclipse Ci with Spot 5.2 software, an iPhone through iDu Optics equipped with a LabCam adapter (New York, NY), or with a Leica DM5500 B microscope with the LAS X Core Workstation fitted with a MC170 Color HD camera. All images from parallel experiments were captured using the same microscope platform. Color images were separated into their RGB channel components and the intensity of staining in the anterior intestine was measured on the green channel as previously described (Yen et al., 2010) using ImageJ (NIH). Because the unstained pharynx immediately above the anterior intestine was used as the ORO staining subtraction background, negative ORO staining values (Figures 5A and 6A) are a result of a high background in specific worm samples.

To obtain control animals, small agar chunks (approximately 1 cm x 1 cm) from a well-fed mixed population of worms grown on NGM plates at 20°C were transferred to 100 mm NGM plates freshly seeded with 10x concentrated OP50 (10x NGM plates). After 3 days of propagation, embryos were harvested by standard methods using sodium hypochlorite (Stiernagle, 2006) and transferred to 10x NGM plates. One-day-old adults were collected three days later and washed with Milli-Q water until the supernatant was clear. Excess water was removed by centrifugation (3000 rpm for 30 s) and worm pellets (0.25 to 1.09 g) were flash frozen in a dry ice and ethanol slurry and stored at −80°C until analysis. To obtain postdauer animals, agar chunks from worms grown in a similar manner as control animals were transferred to 10x NGM plates and incubated for 2 weeks at 20°C for starvation-induced dauer formation. Starved worms were collected with Milli-Q water and dauers were selected by treatment with 1% SDS (Stiernagle, 2006). Dauers were transferred to 10x NGM plates and fed for 2 days. Postdauer one-day-old adults were harvested with Milli-Q water and washed until the supernatant was cleared. Excess water was removed by centrifugation and worm pellets (0.86 to 1.35 grams) were flash frozen and stored at −80°C until analysis. Total fatty acids were quantitatively measured by Creative Proteomics (Shirley, NY) using gas chromatography (GC) with flame ionization detection as follows: to extract fatty acids, worm samples were weighed into a screw-cap glass vial containing tritricosanoin as an internal standard (tri-C23:0 TG) (NuCheck Prep, Elysian, MN). Samples were homogenized and extracted with a modified Folch extraction. A portion of the organic layer was transferred to a screw-cap glass vial and dried in a speed vac. After samples were dried, BTM (methanol containing 14% boron trifluoride, toluene, methanol; 35:30:35 v/v/v) (Sigma-Aldrich, St. Louis, MO) was added. The vial was vortexed briefly and heated in a hot bath at 100°C for 45 min. Following cooling, hexane (EMD Chemicals, USA) and HPLC grade water were added, tubes were recapped, vortexed, and centrifuged to help in the separation of layers. An aliquot of the hexane layer was transferred to a GC vial. GC was performed using a GC-2010 Gas Chromatograph (Shimadzu Corporation, Columbia, MD) equipped with a SP-2560, 100 m fused silica capillary column (0.25 mm internal diameter, 0.2 μm film thickness; Supelco, Bellefonte, PA). Fatty acids were identified by comparison with a standard mixture of fatty acids (GLC OQ-A, NuCheck Prep), which was also used to determine the individual fatty acid calibration curves. Fatty acid composition was expressed as a percent of total identified fatty acids and concentrations as µg/mg of worms.

Gravid adults were treated with hypochlorite to obtain embryos using standard methods (Stiernagle, 2006). Embryos were placed on NGM plates supplemented with 1 mM IPTG and 50 μg/ml carbenicillin seeded with a 10x concentrated bacterial culture expressing the k09f5.2 (vit-1) RNAi clone obtained from the Ahringer library (Kamath et al., 2001). Embryos were allowed to grow until adulthood at which time they were treated again with hypochlorite to obtain embryos. The recovered embryos were grown until day 1 of adulthood under the same conditions and collected for ORO staining.

For control animals, ten L4 larvae (P₀ generation) grown in a mixed population cultured on 60 mm NGM plates at 20°C were placed onto each of 3–4 60 mm NGM plates (30–40 worms per replicate) seeded with OP50 and supplemented with 50 μM of 5-fluoro-2’-deoxyuridine (FUDR; Sigma Aldrich) to prevent reproduction. Worm survival was assessed every two days and was deemed dead if no movement was detected after gentle prodding with a worm pick. Animals that crawled to the side of the assay plate and died due to desiccation were censored from the experiment. To obtain F₁ animals, 2–3 P₀ L4 larva were placed onto one seeded 60 mm NGM plate lacking FUDR and grown at 20°C. Next day, following P₀ egg playing, the parents were removed and their F₁ progeny allowed to grow at 20°C until they reached the L4 larval stage. Ten L4 F1 larvae were placed onto each of 3–4 OP50-seeded 60 mm NGM plates supplemented with FUDR. F₁ worm lifespan was assessed in the same manner as the parental generation. Lifespan assays for RNAi treated animals were done in the same manner except that assay plates were supplemented with 50 μM FUDR, 1 mM IPTG and 50 μg/ml carbenicillin and seeded with a 10x concentrated bacterial culture expressing the k09f5.2 (vit-1) RNAi clone or an empty RNAi vector.

All worm strains used for germ cell row counting have the integrated transgene gld-1::gfp::3xFLAG from the BS1080 strain in their genetic background. Worms with L3 vulva morphology were identified as described by using Nomarski DIC microscopy at 630x (Seydoux et al., 1993) and DAPI stained using the standard whole worm DAPI staining protocol. The stained worms were imaged using a Leica DM5500B microscope with a Hamamatsu camera controller C10600 ORCA-R2. When performing the germ cell row counts, the start of the transition was identified when at least two cells in a row exhibited the crescent-shaped nuclei morphology (Shakes et al., 2009).