PRY-1/Axin signaling regulates lipid metabolism in Caenorhabditis elegans

The nematode Caenorhabditis elegans constitutes a leading animal model to study how signaling pathway components function in conserved biological processes. Here, we describe the role of an Axin family member, PRY-1, in lipid metabolism. Axins are scaffolding proteins that play crucial roles in signal transduction pathways by physically interacting with multiple factors and coordinating the assembly of protein complexes. Genome-wide transcriptome profiling of a pry-1 mutant revealed differentially regulated genes that are associated with lipid metabolism such as vitellogenins (yolk lipoproteins), fatty acid desaturases, lipases, and fatty acid transporters. Consistent with these categorizations, we found that pry-1 is crucial for the maintenance of lipid levels. Knockdowns of vit genes in a pry-1 mutant background restored lipid levels, suggesting that vitellogenins contribute to PRY-1 function in lipid metabolic processes. Additionally, lowered expression of desaturases and lipidomic analysis provided evidence that fatty acid synthesis is reduced in pry-1 mutants. Accordingly, an exogenous supply of oleic acid restored depleted lipids in somatic tissues of worms. Overall, our findings demonstrate that PRY-1/Axin signaling is essential for lipid metabolism and involves the regulation of yolk proteins.


INTRODUCTION
Axin was identified initially as a negative regulator of the Wnt-signaling pathway [1].
Subsequently, the role of Axin was studied in other processes as well, and shown to be essential in diverse developmental events, including embryogenesis, neuronal differentiation, and tissue homeostasis [2]. Axin homologs show functional conservation throughout metazoans [3]. As a scaffolding protein Axin plays a key role in the regulation of canonical Wnt pathway function. It contains multiple domains that facilitate homodimerization and interactions with destructioncomplex proteins, Dishevelled, APC, and GSK-3β [4,5]. The destruction complex initiates the phosphorylation and consequent proteolysis of the transcriptional regulator β-catenin, which promotes expression of Wnt target genes [4,5]. In an ON state, the Wnt-initiated signal inhibits the action of the destruction complex, thereby causing cytoplasmic accumulation of β-catenin, which subsequently translocate to the nucleus and promotes transcription of target genes [4,6].
Constitutive activation of β-catenin, due to the loss of destruction-complex function, is often associated with cancers and various other disorders affecting the lungs, heart, muscles, and bones [4]. Thus, Axin function is crucial to ensuring precise regulation of β-catenin-mediated Wnt signaling.
The nematode C. elegans (worm) is an excellent animal model to investigate the role of Axin and other Wnt pathway components. Similar to the mammalian Axin, PRY-1 in worms interacts with APR-1/APC and GSK-3/GSK-3β to regulate BAR-1/β-Catenin-mediated gene transcription [7].
pry-1 mutants show defects that are consistent with overactivation of Wnt signaling, such as Q cell migration and vulval induction [7,8]. As expected, these phenotypes are opposite to those observed in bar-1 mutants that inactivate Wnt signaling [8,9].
To understand the mechanism of pry-1/axin function a transcriptome profiling was carried out.
The analysis of the transcriptomic data uncovered genes associated with different biological processes including aging and lipid metabolism. In agreement with this, pry-1 mutants have a short lifespan, low brood size, and drastically reduced lipid levels. We focused on a set of differentially regulated genes, specifically yolk lipoprotein vitellogenins (VITs) that are distant homologue of human apolipoprotein B (ApoB) [10] and important for lipid distribution. Previous Page 4 studies showed that reducing VITs in wild-type animals increases both lifespan and lipid accumulation, with overexpression having an opposite effect in long-lived mutants [11,12]. Our results showed that the expression of vit genes was upregulated in pry-1 mutant animals during early larval stages. Moreover, vit-2 was overexpressed in adults as well. We also found that lowering VIT activity in pry-1 mutants restored lipids. These results suggest that VITs may be regulated by pry-1-mediated signaling to affect lipid levels.
Because lipid contents are affected by changes in the enzymatic activities of lipases and lipid desaturases [13][14][15], we examined lipid catabolism by measuring lipase activity but observed no detectable increase in lipid utilization. However, the expression of three conserved stearoyl-CoA desaturases, fat-5, fat-6, and fat-7, which are involved in the synthesis of monounsaturated fatty acids such as oleic acid (OA) [13][14][15], was reduced in animals lacking pry-1 function. In support of this, supplementing the bacterial diet with OA bypassed the requirements of desaturases and partially rescued the lipid phenotype of pry-1 mutants. Overall, our work demonstrates that the yolk proteins Vitellogenins contribute to PRY-1-mediated function in lipid metabolism.

Identification of PRY-1 targets
To gain insights into the mechanism of pry-1-mediated signaling, a genome-wide transcriptome analysis was carried out to identify the potential downstream targets. Using RNA-Seq, we identified a total of 2,665 genes (767 upregulated and 1898 downregulated, False Discovery Rate (FDR) p-adj 0.05) that were differentially expressed in pry-1(mu38) animals during the L1 larval stage (Fig. 1A, Table S3, also see Methods). Of these, the transcription of 1,331 genes was altered twofold or more, (FDR, p-adj 0.05) (248 upregulated and 1083 downregulated) ( Table S3). The average and median fold changes in the expression were 2.2 and 2.0, respectively. Figure 1A shows a scattered plot of all expressed genes. A total of 20 genes were also tested by qPCR, which revealed an 85% validation rate (Fig. S1A, B).
We carried out gene ontology (GO) analysis (www.geneontology.org) to investigate the processes affected in pry-1(mu38) animals. Genes with altered expression were found to be enriched in GO Page 5 terms associated with determination of adult lifespan, aging, response to unfolded protein, oxidation-reduction process, metabolism, stress response and cell signaling, steroid hormone mediated signaling, lipid metabolic processes, and cellular response to lipids (Fig. 1B, complete list in Table S4). This indicates that pry-1 plays a role in stress response, lipid metabolism, and lifespan regulation. We also observed enrichment in neuron-related GO terms such as axon, synapse, synaptic transmission, and neuron development. This is expected from the requirements of pry-1 in neuronal development [7]. Other categories included molting cycle, regulation of transcription, DNA-template, and the reproductive process. In addition to these known categories, the dataset includes many non-annotated genes (Table S3) whose function remains uncharacterized.

Page 6
We noticed alterations in the expression of some of the Wnt pathway components as well. For example, pry-1 was upregulated 1.3-fold (Table S3, Fig. S1A). Although such upregulation of pry-1 was not reported earlier in pry-1 mutants, previous studies have shown that Axin is a target of Wnt signaling and its expression is increased in overactivated Wnt backgrounds [5,[18][19][20]. This, together with our finding, suggests that the positive regulation of Axin by the canonical Wnt signaling pathway is a conserved mechanism in eukaryotes. Other Wnt pathway components that are differentially expressed in pry-1(mu38) included mom-2/wnt (1.5-folds up), cfz-2/fz (1.7-folds down), lin-17/fz (1.6-folds up) and pop-1/tcf (1.7-folds up) levels (Table S3).

pry-1 mutants are short-lived, and exhibit altered lipid metabolism
The presence of several aging-related genes in the pry-1(mu38) transcriptome led us to examine lifespan phenotype. We found that pry-1(mu38) animals are severely short lived with an 80% decrease in the mean lifespan ( Table 1, Fig. 2A). Although pry-1 transcription is higher in these animals (see above), the aging defect is likely caused by the loss of pry-1 function due to the nonsense nature of mu38 allele. A similar phenotype was observed in a CRISPR mutant pry-1(gk3682) as well ( Table 1, Fig. 2A). The RNAi experiments also caused a reduction in lifespan.
Specifically, mean lifespans were reduced by 18-29% when treated from the embryonic stage and 22-31% when treated during adulthood ( Table 1, Fig. 2B, also see Methods). It should be noted that RNAi effects are less severe compared to mutants, possibly due to the partial KD of pry-1 activity and negative autoregulation of pry-1 transcription (see Fig. S1A). Therefore, we conclude that pry-1 is important for the normal lifespan of C. elegans.
Throughout the lifespan of an animal, lipids are persistently mobilized to afford energy demands for growth, cellular maintenance, tissue repair, and reproduction [25,26]. Changes in lipid levels affect an organism's ability to survive in stressful conditions. For example, exposure of animals to oxidative stress causes mobilization of somatic fat to germline as a mechanism to balance survival and reproduction [27,28]. Many genes that are involved in the synthesis, breakdown and transport of lipids are differentially expressed in pry-1 mutants (Fig. S2 - (Table S2). The expression of vit genes and desaturase was measured by qPCR and all but fat-4 were successfully validated (Fig. S1A, B). The fat-4 levels were down by 20%, unlike the 1.5-fold increase observed by RNA-Seq. We also tested another desaturase, fat-7, that functions redundantly with fat-6 [29] but was not present in our dataset. fat-7 mRNA levels were Page 8 below the limit of detection (Fig. S1A). Thus, all four fat desaturase genes are downregulated in pry-1(mu38) animals. Enrichment of several lipid metabolism genes in the pry-1 transcriptome led us to examine lipid accumulation in worms. Staining with Oil Red O revealed that the lipid content was less than half in pry-1(mu38) one-day young adults compared with controls (Fig. 3A, B). Examination of total fat at each larval stage revealed that pry-1 mutants have lower somatic lipid stores (25-80%) at all stages except for L2 (Fig. S3A). In addition, the lipid distribution was altered such that the staining was mostly restricted to gonadal tissue (Fig. 3C). These results suggest that pry-1 plays a role in lipid metabolism. Consistent with this we found that pry-1(mu38) animals lay fewer fertilized eggs and have poor survival upon starvation-induced L1 diapause (Fig. 3E, F).
One explanation for reduced lipid phenotype could be that lipids are being rapidly utilized. This is unlikely because several lipases (lips family members, mentioned above) are downregulated. We also measured total lipase activity in one-day-old adults from whole worm lysates. As expected, the total lipase activity was 34% lower in the mutant compared with the N2 control (Fig. 3D).
Next, we examined lipids in pry-1(mu38) animals following knock-down of lipl-4 or lips-7, lipase genes that regulate the gonad dependent somatic lipid levels [25,26,30] but observed no change in the pattern of lipid distribution (Fig. 3G, H). We conclude that lower somatic lipids in animals lacking pry-1 function are not due to increased utilization, raising the possibility of the involvement of other metabolic processes.

Vitellogenins contribute to lipid metabolism defects in pry-1 mutants
To understand the molecular basis of low lipid levels in pry-1(mu38) worms we focused on the vitellogenin family of genes whose expression is repressed by pry-1. VITs are the major yolk proteins in C. elegans that are synthesized in the intestine and mediate lipid transport from the intestine to the gonad during the reproductive period [31]. Examination of vit levels in pry-1(mu38) animals revealed abnormal expression at all developmental stages. Thus, starting with the L1 stage where all six vit genes were upregulated, the number of overexpressed genes was five in L2, two in L3 and zero in L4 stage (Fig. 4).

Page 9
Next, lipid contents in pry-1(mu38) worms were examined following RNAi-mediated KDs of individual vit genes. Previously, vit RNAi in wildtype animals was shown to cause accumulation of lipids in the intestine [12]. We found that pry-1 mutants had altered lipid distribution in all cases such that lipids accumulated at higher levels in somatic tissues (Fig. 5A, B, Fig. S3B). A quantification of overall lipids revealed a significantly higher accumulation in mutants (2.5 -3 fold) compared to wildtype N2 (1.3 -1.4 fold) (Fig. 5A, B, Fig. S3B). We also compared vit genomic sequences used to perform KD experiments, which revealed that vit-1 RNAi also targets vit-2 due to significant identity (Table S6, Fig S4). Likewise, any one of the vit-3, 4 or 5 RNAi can simultaneously KD the other two (Table S6). Thus, multiple VITs play roles in the regulation of both lipid levels as well as its distribution and their misregulation contributes to lipid metabolism defects in pry-1 mutants.

Lipid defect in pry-1 mutants does not involve the lipoprotein receptors rme-2 and lrp-2
We wanted to understand how pry-1 and vit genes might work to regulate lipid levels. Because VITs are transported via the RME-2 receptor [31], we examined the involvement of rme-2 in the pry-1-mediated pathway to regulate lipid accumulation. The knock-down of rme-2 by RNAi led to intestinal accumulation and ectopic deposition of lipids (Fig. 5C, D) due to blockage of yolk protein transport to the developing oocytes [31]. Specifically, rme-2(RNAi) animals showed an approximately 45% increase in total lipid content such that the gonad-to-somatic ratio was roughly 30% lower compared with controls. However, this phenotype was not observed in pry-1(mu38) due to a reduction in lipid levels both in somatic and gonadal tissues (Fig. 5C, D). These results allow us to suggest that VITs act independently of the RME-2 transport mechanism to regulate lipid metabolism in response to pry-1 signaling. Moreover, since lipid levels are further reduced in pry-1(mu38) animals following rme-2 KD, it may be that rme-2 has an unknown nonvitellogenin-mediated role in lipid accumulation. There may also be other possibilities for such a phenotype.
We also examined the possibility of other VIT-interacting factors that might be involved in PRY-1-mediated regulation of lipid levels. Our transcriptome dataset contained one LDL-like receptor Page 10 gene lrp-2 that was overexpressed in pry-1 mutants (Supplementary Table S3). It was shown previously that lrp-2 positively regulates yolk protein synthesis [32]. To test whether lipid levels are affected by lrp-2, RNAi KD experiments were performed, which showed a small but significant rescue of lipid phenotype in pry-1(mu38) animals (Fig. 5E, F). However, since lrp-2 KD in wildtype animals also caused a similar increase in lipids, it is unclear whether PRY-1mediated signaling involves LRP-2 function to affect lipid levels.
Because the expression of all three fat genes depends on nuclear hormone receptors nhr-49 and nhr-80 [15,33], we determined levels of both these NHR transcripts in pry-1 mutant animals.
Although RNA-Seq transcriptome data showed no change in nhr-49 and nhr-80, qPCR experiments revealed that nhr-80 transcription showed a subtle but significant upregulation whereas nhr-49 was unchanged (Fig. S1B). Thus, transcriptional regulation of these two nuclear hormone receptors is unlikely to be a mechanism affecting pry-1-mediated expression of Δ9desaturase genes, although we cannot rule out the possibility that activities of one or both may be regulated post-transcriptionally in response to PRY-1 function.
The changes in the expression of Δ9-desaturases can lead to reduced FA synthesis in pry- 1(mu38) animals. To investigate this, we quantified lipid levels by gas chromatography-mass spectrometry (GC-MS) approach. The results showed that while the relative ratios of fatty acids in pry-1 mutants Page 11 are normal, the absolute level of each species is significantly reduced (Fig. 6A, Fig. S5). The result agrees with the overall low lipid levels in pry-1(mu38) animals and together supports the important role of PRY-1 signaling in lipid metabolism in C. elegans.

mutants
One of the fatty acid species that showed 50% reduced levels in our GC-MS analysis is OA. OA is required for fatty acid metabolism and is synthesized by C. elegans as it cannot be obtained through the normal E. coli (OP50) diet. OA acts as a precursor for the synthesis of polyunsaturated fatty acids and triacylglycerides, which are used for fat storage [13]. The addition of exogenous OA as a fat source has been shown to rescue several fat-deficient mutants, including fat-6 and fat-7 by restoring their fat storage, resulting in improved fertility and increased locomotion [29].
Moreover, the addition of OA in sbp-1, fat-6, and fat-7 animals fully rescued defects in satiety quiescence [14,15,34]. In the context of lifespan OA has been shown recently to impart beneficial effects [35], which may involve hormesis mechanism due to elevated activities of genes such as glutathione peroxidase, catalase, and superoxide dismutase [28]. We, therefore, reasoned that supplementation of OA may improve lipid levels in pry-1(mu38) mutants. Treatment with 1 mM OA resulted in the restoration of lipids in animals lacking pry-1 function (up to twofold higher compared with the untreated control, Fig. 6B, C). No significant changes were seen in the gonadal lipid levels, suggesting that lipid metabolism in the gonads was unaffected. In addition to restored lipid levels, OA-treated pry-1(mu38) mutants showed a 1.3-fold increase in lifespan (  6D).

DISCUSSION
Wnt pathway components are involved in cellular senescence, tissue aging, and nutrient metabolic processes. However, the mechanism by which the pathway affects these various processes is not well understood. Here, we investigated the role of pry-1/Axin, a negative regulator of Wnt signaling, and provide evidence for its important role in the maintenance of lipid metabolism through regulation of vitellogenesis.

pry-1 regulates expression of genes involved in lipid metabolism
The transcriptome profiling of pry-1 mutants revealed altered expression of many genes including those that affect hypodermis, stress-response, aging, and lipid metabolism. The hypodermalrelated genes include collages, cuticulins, and hedgehogs. Previously, expression of some of the hedgehog genes was found to be altered in bar-1 mutants [20,21]. Considering that cuticular defects are observed in bar-1 [18,20] and pry-1 mutants (Mallick et al., manuscript in preparation), and that hedgehog family members play roles in cuticle shedding and formation of alae [17], these results lead us to suggest that a genetic pathway involving pry-1 and bar-1 interacts with hedgehogs for normal cuticle development.
One of the key findings of our pry-1(mu38) transcriptome analysis is the enrichment of genes related to lipid metabolism. We found that multiple lipogenic and lipolytic genes had altered expression. For example, all four fatty acid desaturases (∆5 and ∆9 desaturases) were downregulated in pry-1 mutants. While single fat gene mutants affect fatty acid composition without altering overall lipid levels, double mutants have a low lipid level [14,29], suggesting that pry-1 positively regulates fatty acid synthesis. With regards to lipolytic genes, such as those involved in beta-oxidation, changes in gene expression between peroxisomal and mitochondrial beta-oxidation genes had an opposite trend (4 of 5 upregulated and 8 of 11 downregulated, respectively). This may indicate selective utilization of long-chain fatty acids over short-chain fatty acids by the pry-1 pathway. We also observed that all four lipases (lips family) are downregulated, including lips-7 which was earlier shown to be involved in lifespan extension and the maintenance of lipid levels [30]. Although lips-7 did not alter the pry-1(mu38) phenotype, it remains to be seen whether pry-1 regulates any, or all, of the remaining three lips gene(s) to modulate lipids.
In addition to lipogenic and lipolytic genes, several lipid transporters are also present in the pry-1(mu38) transcriptome, including two lipid-binding proteins (lbp-5 and lbp-8; both downregulated), six lipoproteins (vit-1 to -6; all upregulated) and a LDL-like receptor protein (lrp-2). Knock-downs of lbp-5 and VITs negatively affect lipid storage [36], which further emphasizes the important role of pry-1 in the maintenance of lipids and suggests that the pry-1-mediated signaling is involved in utilization of lipids for energetics as well as signaling mechanisms.

PRY-1-mediated lipid metabolism involves Vitellogenins
Reduced lipids may affect tissue function and physiology in different ways, for example, due to altered membrane structure and compartmentalization, altered signaling, reduced energy demands, and impact on autophagy. The Oil Red O staining of pry-1(mu38) showed a severe reduction in lipid content with a marked decline in the somatic lipid storage. This phenotype was also observed in dNT-bar-1 animals that carry a constitutively active form of BAR-1 (Fig. S6A, B). Together these findings suggest that pry-1-bar-1 pathway is involved in lipid metabolism. Earlier, a somatic lipid depletion phenotype was reported in the skn-1 gain of function mutant and H2O2-treated wildtype animals [27]. However, unlike pry-1 and bar-1, defects in these cases were observed near the end of the reproductive senescence (termed age-dependent somatic depletion of fat or asdf).
One possibility for reduced lipids in pry-1 and bar-1 mutants may be due to elevated breakdown of lipids. We investigated this issue in pry-1 mutants and found no increase in total lipase activity.
Moreover, KDs of lipl-4 (lysosomal lipase) and lips-7 (cytosolic lipase) in pry-1(mu38) -both of which negatively regulate lipid levels [25,26] -had no observable effect. Thus, selective and rapid lipid catabolism does not appear to be a factor in lipid depletion in the absence of pry-1 function. We then investigated the role of VIT proteins in maintaining lipid levels. As major yolk proteins, VITs are involved in somatic mobilization of lipids to the developing germline. pry-1 mutants show misregulation of all six vit genes such that they are overexpressed in L1 but decline thereafter with all being downregulated by the L4 stage. Interestingly, vit-2 levels were found to be significantly higher in adults. As expected, the low lipid phenotype of pry-1 mutant animals was suppressed by knocking down vit genes (vit-1/2 and vit-3/4/5) providing evidence that VITs play an important role in PRY-1-mediated lipid metabolism. We have also shown that such a role of VITs may not utilize lipoprotein receptors RME-2 (VIT transporter) and LRP-2 (VIT synthesis).
Overall, these findings along with the role of VITs in regulating lipid levels [12], allow us to propose that PRY-1-mediated signaling involves VITs to regulate processes that depend on energy metabolism and lipid signaling [25,26].

Page 14
While it remains unclear how VITs participate in pry-1 signaling, one possibility may involve downregulation of the autophagy pathway. Manipulating VIT levels has been shown to affect the lifespan by altering autophagy [12]. Autophagy is a complex process that involves multiple enzymes to recycle cellular contents by converting them into usable metabolites. Although the pry-1 transcriptome did not contain known autophagy-related candidates, the pathway may still be involved. This could be tested by examining the roles of specific autophagosome genes in lipoprotein synthesis and autophagy, which in turn should reveal a link between pry-1-mediated lipid metabolism.
Another possibility for low lipid levels in pry-1 mutants may be due to reduced fatty acid synthesis.
To this end, our GC-MS analysis of fatty acid composition revealed that pry-1 is needed to maintain normal levels of every fatty acid species analyzed. A global reduction in fatty acids due to the loss of pry-1 function may affect processes that require utilization of lipids such as aging that is supported by short lifespan of pry-1 mutants. However, it should be noted that previous studies examining lipid levels in lifespan defective mutants have found no clear relationship between fat content and longevity. For example, daf-2 and nhr-49 mutants have high fat but their lifespan phenotypes are opposite (daf-2 are long lived and nhr-49 short lived) (reviewed in [25,26]). The lifespan extension phenotype is also observed in mutants with reduced fat content, such as dietary restricted eat-2 [37]. Thus, instead of absolute levels, the quality of lipids may be more important [28]. We investigated this using OA, one of the species involved in fatty acid signaling.
Exogenous treatment with OA restored lipid levels as well as partially rescued lifespan defect in animals that lack pry-1 function. Thus, pry-1 may play a role in maintaining the levels of beneficial fatty acids. How then might the pry-1-vit-mediated pathway affect lipid levels? It may be that lipid synthesis and storage processes are compromised. This is in part supported by reduced expression of desaturases, however, additional mechanisms are also likely to be involved, such as reduced conversion of acetyl-CoA to saturated fatty acid (palmitate), lower synthesis of diglycerides, and increased peroxisomal beta oxidation (Fig. S2). It would be interesting to examine these possibilities in the future.

Page 15
Our study provides the first evidence of PRY-1/Axin function in lipid metabolism. The involvement of lipids in age-related disorders in humans, as well as animal models, is well documented. Genetic and acquired lipid diseases are associated with loss of subcutaneous fat, accumulation of visceral and ectopic fat, and metabolic syndromes such as insulin resistance, glucose intolerance, dyslipidemia, and hypertension [38]. Yang et al. showed that Axin expression in mice contributes to an age-related increase in adiposity in thymic stromal cells [39] (Yang et al. 2009). Although our data shows that PRY-1/Axin is not likely to affect fat storage, whether Axin family members play roles in any of these lipid-related diseases is unknown. Therefore, findings that PRY-1/Axin is necessary for the maintenance of lipid levels provide a unique opportunity to investigate the role of Axin signaling in age-related lipid metabolism.

Strains
Worms were grown on standard NG-agar media plates using procedure described previously [40].

Molecular Biology and transgenics
Primers used in this study are listed in Table S1. The pry-1p::pry-1::GFP transgenic strain DY596 was created by injecting 100 ng/ml of the pDC10 plasmid into the pry-1(mu38) animals. The plasmid was kindly provided by Korswagen lab and has been described earlier [7]. Briefly, it contains the full coding sequence of C37A5.9/pry-1 along with 3.

RNAi
For RNAi experiments, E. coli HT115 expressing target specific dsRNA were grown on plates containing β-lactose [41]. Worms were bleach synchronized and seeded onto plates. After becoming young adult, worms were transferred to fresh plates every other day and numbers of dead worms recorded. For adult specific RNAi, synchronized worms were cultivated on NGM/OP50 plates until the young adult stage and then transferred to the RNAi plates.
pry-1 RNAi was carried out using two different plasmids, one of which was from Ahringer library (termed plasmid #1) and the other was created as part of this study (termed plasmid #2, see Molecular Biology section above). Except for data presented in Fig. 2B, for which plasmid #2 was used, all other pry-1 KD experiments were performed using plasmid #1.

Microscopy and quantification
Animals were paralyzed in 10 mM Sodium Azide and mounted on glass slide with 2% agar pads and covered with glass coverslips for immediate image acquisition using Image NIS Element software (Nikon, USA) with a Hamamastu Camera mounted on a Nikon 80i upright microscope.
For GFP and autofluorescent age-pigment analysis excitation 470/40 nm and for emission 525/50 nm wavelengths were used. Each experiment was repeated at least two times with similar outcome.

Page 17
For the lipofuscin assay, fluorescence intensity of individual worms was determined using ImageJ software. For acquisition of PRY-1::GFP fluorescence micrographs Zeiss Apotome was used.
Quantification of pixel densities for GFP reporters was performed with Image J™. Fluorescence of hsp-6::gfp and hsp-16::gfp was examined by analyzing the degree of GFP intensity. The animals were categorized into three arbitrary states, i.e., low, medium, or high, based on GFP fluorescence and the proportion of animals for each of these was plotted.

Lifespan assay
Lifespan analyses were carried out as prescribed [11] at 20 °C unless otherwise indicated on NGmedia plate devoid of Fluorodeoxyuridine (FUDR). Young adult worms (day zero for lifespan) were plated and transferred to new seeded plates every alternate day during the reproductive period to avoid over-crowding due to progenies and following transfers were done when necessary.
Animals were scored every alternate day and animals that did not respond to repeated prodding with the pick were counted as dead. Results are expressed as mean life span ± SE. Differences were considered statistically significant at p < 0.05. Shown is one of three replicate trials with similar outcomes.

Oil Red O staining
Oil Red O staining was performed as previously reported [42]. In short, worms were collected from NGM plates, washed with 1X PBS buffer and re-suspended in 60 µl of 1× phosphate-buffered ImageJ software was used to quantify Oil Red O intensities [42]. 15 to 30 worms were randomly selected from each category in at least two separate batches.

Brood Assay
Worms were bleach synchronized and allowed to grow to L4 stage for determining the progression of egg-laying and the brood size. Individual worm was picked onto a separate NGM plate with OP50 bacteria and allowed to grow for several days. Worms were repeatedly transferred to a freshly seeded NGM plate and progeny was counted every 24 hours. Data from escaping or dying mothers were omitted from the analyses [32].

Oleic acid supplementation assay
To make OA supplemented NGM agar plates, a 0.1 M water based stock solution of OA sodium salts (NuCheck Prep, USA) was prepared and stored at -20 °C in the dark. The OA solution was added continuously to the NGM and promptly poured into the plates. The plates covered with aluminum foil and kept at temperature overnight to dry. The E. coli OP50 strain was seeded to each plate and allowed to further dry for one to two days in the dark. Aging assays and Oil Red O staining was performed as described above [43].

Lipase assay
Lipase activity was estimated using commercially available QuantiChrom™ Lipase Assay Kit (BioAssay Systems, USA, Catalog number DLPS-100) and processed according to the manufacturer's instructions. 1 unit of Lipase catalyzes the cleavage of 1 µmol substrate per minute.
Three independent samples of one-day-adult worms were prepared by homogenizing in a 20% glycerol, 0.1 M KCl, 20 mM HEPES (pH 7.6) buffer for further measurements as described earlier [44].

L1 survival assay
Worms were bleach synchronized and kept in 1.5 ml centrifuge tube. Worms were seeded onto NGM plates approximately 24 hours afterwards regularly for 12 days, and numbers of seeded worms counted. Worms were grown to young adult stage before survivors were counted. The L1 diapause data were statistically compared using an analysis of covariance (ANCOVA) model.

RNA-Seq and data analysis
pry-1 targets were examined in synchronized L1 stage animals. At this stage Wnt ligands, receptors, and targets are highly expressed as revealed by microarray studies from SPELL database [45,46] (Fig. S1C-E). Also, our qRT-PCR experiments showed significant upregulation of three of the Wnt targets, lin-39, egl-5 and mab-5, in pry-1 animals at L1 stage (Fig. S1F). The pry-1 transcriptome profile can be found in the GEO archive with accession number GSE94412. For RNA-Seq experiments synchronized L1 stage animals were obtained by two successive bleach treatments and RNA was isolated using Trizol-reagent (Sigma, USA, Catalog Number T9424) [47]. The quality of total purified RNA was confirmed using bioanalyzer (Agilent 2100 and In total, 30 to 38 million reads were obtained for each sample analyzed for differential gene expression. The adapters were trimmed using cutadapt/trimgalore, reads with QC values (Phred score) lower than 30 bases were discarded after trimming process [48]. Later, processed sequencing reads were mapped to the reference genome (ce6) (UCSC 2013) using the software package Bowtie 1.0.0 [49]. 92-95% of total sequenced fragments could be mapped to the genome (Table S2). Transcriptlevel abundance estimation was performed using eXpress 1.5 software package [50]. Among all genes analyzed, 18867 were mapped to known transcripts by at least one sequencing fragment in C. elegans. To avoid biases between samples, the gene counts were quantile normalized [48,51].
Using a negative binomial distribution model of DESeq package in R, differentially-expressed genes were called at a false discovery rate (FDR) of 0.05% [52].
A GO-term containing at least three genes with a p-value adjusted for multiple comparisons and < 0.05 (Benjamini-Hochberg method) was counted significant [53]. Tissue enrichment analysis was performed using Wormbase online TEA tool that employs a tissue ontology previously developed by WormBase [54].

Gas Chromatography Mass Spectrophotometry (GC-MS)
Fatty Acid analysis protocol was modified from a previously published method [13,33].

Statistical Analysis
Statistic tests for lifespan and stress resistance assays were performed using SigmaPlot software 11. The survival curves were estimated using the Kaplan-Meier test and differences among groups were assessed using the log-rank test. Survival data are expressed relative to the control group.
Other statistics were performed using Microsoft Office Excel 2016. If not specifically mentioned, p values for the fertility, motility, fat content, fluorescence intensity, L1 survival and enzyme activity assays were calculated by Student's t test after testing for equal distribution of the data and equal variances within the data set. Experiments were performed in triplicate except where stated otherwise. Differences were considered statistically significant at p < 0.05, thereby indicating a probability of error lower than 5%. Hypergeometric probability tests and statistical significance of the overlap between two gene sets were done using an online program (http://nemates.org/MA/progs/overlap_stats.html).

ACKNOWLEDGMENTS
We are grateful to Brian Golding for providing server space, Paul Sternberg lab for help with RNA-Seq computational analysis, Hendrik Korswagen for pry-1p::PRY-1::GFP plasmid, and Don Moerman lab for pry-1 CRISPR alleles. Some of the strains were obtained from CGC, which is funded by the NIH Office of Research Infrastructure Programs (P40OD010440). We thank Lesley Macneil and anonymous reviewers for comments on previous versions of the manuscript, Jessica Knox for assistance with initial qRT-PCR experiments, and Gupta lab members for helpful discussions. This work was supported by NSERC Discovery grant to BG.  Table S4 for a detailed list).