Dietary Restriction Induced Longevity Is Mediated by Nuclear Receptor NHR-62 in Caenorhabditis elegans

Dietary restriction (DR) extends lifespan in a wide variety of species, yet the underlying mechanisms are not well understood. Here we show that the Caenorhabditis elegans HNF4α-related nuclear hormone receptor NHR-62 is required for metabolic and physiologic responses associated with DR-induced longevity. nhr-62 mediates the longevity of eat-2 mutants, a genetic mimetic of dietary restriction, and blunts the longevity response of DR induced by bacterial food dilution at low nutrient levels. Metabolic changes associated with DR, including decreased Oil Red O staining, decreased triglyceride levels, and increased autophagy are partly reversed by mutation of nhr-62. Additionally, the DR fatty acid profile is altered in nhr-62 mutants. Expression profiles reveal that several hundred genes induced by DR depend on the activity of NHR-62, including a putative lipase required for the DR response. This study provides critical evidence of nuclear hormone receptor regulation of the DR longevity response, suggesting hormonal and metabolic control of life span.


Introduction
Genetic and environmental factors can cause profound changes in organism lifespan. Genetic alterations that stimulate robust longevity across taxa include reduced insulin/IGF and TOR signaling, reduced mitochondrial function, and reduced signaling from germline stem cells [1]. One of the most pervasive environmental alterations that impacts longevity is dietary restriction (DR), a reduction in caloric uptake without malnutrition, which can increase health and life span in different organisms, including yeast, worms, flies, and rodents [2]. Whether DR induces longevity in non-human primates is still under debate, however there are clear health benefits observed [3,4]. In humans, evidence indicates that DR lowers body temperature, insulin levels, and body fat [5,6]. Moreover, improved serum cholesterol and lipid levels suggest a decreased risk for cardiovascular disease [7]. Conversely, overnutrition may be a risk factor for age-related disease including obesity, diabetes, heart disease, neurodegeneration, and cancer [8].
In Caenorhabditis elegans several different DR regimens can induce longevity. The two most widely used are dilution of bacterial food and the genetic DR mimetic eat-2. Dilution of bacterial food in liquid culture (BDR) was first demonstrated by Klass in 1977 to extend C. elegans life span, and variations of this method have been shown to enhance longevity by 20-100% [9,10]. By this regimen, animals develop on bacterial plates ad libitum until adulthood, and then are shifted to liquid culture containing a dilution of bacterial food. The eat-2 mutation affects the function of a pharyngeal acetylcholine receptor subunit, which reduces pharyngeal pumping rate and subsequent food intake throughout the life of the animal, and extends life span by 15-40% [11]. Other ways of inducing DR in C. elegans adults include intermittent feeding (IF), in which worms are fed every two days, dietary deprivation (DD), in which adult worms are completely removed from food, and solid DR (sDR) where bacteria is diluted on solid agar plates [12][13][14]. Curiously, life extension by these regimens requires different sets of genes, indicating that DR is not a uniform process and could result from multiple responses [11][12][13][14][15][16].
From genetic studies in C. elegans, a few key transcriptional regulators of the DR response have started to emerge. PHA-4 is a FOXA homolog required for eat-2 and BDR induced longevity [10]. SKN-1 is an NF-E2 transcription factor required in a BDR model of DR longevity [17]. Additionally, heat shock factor and hypoxia inducible factor have been implicated in regimens resembling DD or IF [18,19]. Reduced signaling through the nutrient sensor TOR kinase, and processes downstream of TOR that increase autophagy, reduce protein synthesis, and alter energy homeostasis may contribute to the DR response [12,20,21]. Nevertheless, the regulatory networks and the underlying mechanisms promoting longevity still remain unclear.
In an effort to identify regulators of DR-induced longevity, we specifically focused on nuclear hormone receptors (NHRs).
Nuclear hormone receptors are transcription factors that respond to fat-soluble hormones, such as steroids and fatty acids, to directly regulate gene transcription. They are broadly implicated in the regulation of development, metabolism and homeostasis, and are well poised to coordinate events throughout the body in response to hormonal or nutritional signals [22,23]. We hypothesized that NHRs may mediate metabolic states associated with DR, and thus could be important for DR-induced longevity. In this work we identify the HNF4a-like nuclear hormone receptor nhr-62 as required for DR-induced metabolic and longevity responses.

NHR-62 is Necessary for eat-2 DR-induced Longevity
To test if NHRs mediate DR-induced longevity, we performed RNAi knockdown of NHRs in animals carrying an eat-2 mutation (a genetic DR mimetic), and screened for a loss of eat-2-induced longevity. We screened 246 of the 284 C. elegans NHRs in a genetic background more sensitive to RNAi (eat-2;nre-1;lin-15b). Knockdown of most NHR genes had no effect on eat-2 longevity (e.g., nhr-35 Figure 1A). As expected, knockdown of a few NHRs substantially shortened eat-2 and wild-type life span (e.g., nhr-49) yet largely maintained DR-induced life span extension ( Figure S1). Interestingly, we found that only knockdown of nhr-62, an HNF4a related NHR, suppressed the longevity of eat-2;nre-1;lin-15b while having little effect on the longevity of nre-1;lin-15b control animals ( Figure 1B).
The nhr-62 locus encodes a predicted long isoform A (515 AA) consisting of DNA-and ligand-binding domains (DBD; LBD), and a short isoform B (353 AA) consisting of only the LBD ( Figure 1C). nhr-62 shares 26-28% overall identity with vertebrate and insect HNF4a nuclear receptors across both domains. Mammalian HNF4a proteins function in many processes including lipid and glucose metabolism, and HNF4a mutations in humans are associated with maturity onset diabetes of the young type 1 [24,25]. Similarly, the Drosophila melanogaster HNF4a homolog is important for lipid mobilization, fatty acid b-oxidation, and the starvation response [26]. However, 269 of the 284 NHRs in C. elegans represent a major expansion of the HNF4a family and are interrelated [27]. To determine specificity of the suppression of DR-induced longevity by nhr-62, we first compared the fraction of animals alive at day 15 between nre-1;lin-15b and eat-2;nre-1;lin-15b fed nhr-62's closest homologs, nhr-21, nhr-4, nhr-34, and nhr-100, and found that these genes were not required for longevity in eat-2;nre-1;lin-15b mutant animals ( Figure S1). Secondly, we measured the lifespan of dietarily restricted worms subjected to either empty vector RNAi or RNAi specific to the NHRs with closest homology, nhr-21 and nhr-4, and again found that these genes were not required for DR-induced longevity ( Figure 1D,E and Figure S1). These data suggest that nhr-62 uniquely regulates eat-2 DRinduced life span extension in C. elegans.
To validate our RNAi results, we obtained the deletion allele, nhr-62(tm1818), which removes a 658 bp region, including part of exon one, and results in an immediate stop codon, thus removing the DBD of the receptor. Although nhr-62(tm1818) is likely a strong loss-of-function allele, it may not be null because the predicted B isoform containing the LBD only is intact. We introduced this mutation into eat-2(ad465) animals and measured the lifespan of the double mutant. Similar to RNAi, nhr-62(tm1818) significantly suppressed the lifespan of eat-2(ad465) animals. Importantly, the nhr-62(tm1818) mutant had little or no effect on wild-type (N2), suggesting that nhr-62 does not suppress eat-2 longevity through general sickness ( Figure 1F). Though nhr-62(tm1818) consistently suppressed eat-2(ad465) lifespan, suppression was not always complete, suggesting other activities still contribute to the response.
To confirm the role of nhr-62 in DR-induced longevity, we generated an extra chromosomal line, dhEx627, expressing wildtype nhr-62, and introduced it into eat-2;nhr-62 mutant animals to measure whether complementing nhr-62 function was sufficient to restore longevity. As expected, the dhEx627 array restored longevity to eat-2;nhr-62 double mutants ( Figure 2A). Interestingly, when this array was crossed into the wild-type background, nhr-62 over-expressing worms displayed phenotypes similar to DR worms (e.g., small body size and reduced fat) without affecting pharyngeal pumping rate ( Figure S2). Furthermore, the dhEx627 array significantly extended the life span of wild-type animals in four of eight independent experiments, suggesting that nhr-62 can be sufficient to promote longevity ( Figure 2B and Table S1).
Reduced insulin/IGF receptor (IR) signaling or reduced mitochondrial function have both been shown to extend life span [28,29]. To test if nhr-62 also modulates longevity of these pathways, we measured the lifespan of nhr-62(tm1818) animals fed daf-2 (IR) or cco-1 (cytochrome c oxidase) RNAi. As expected, daf-2 RNAi and cco-1 RNAi robustly increased the life span of wild-type animals. Moreover, RNAi knockdown similarly increased longevity in the nhr-62(tm1818) background ( Figure 2C,D). These results reveal the nhr-62 mutation does not affect these pathways, but specifically modulates DR-induced longevity. Taken together, these data indicate that NHR-62 is a novel regulator of DRinduced longevity.

NHR-62 Mediates BDR Longevity Response at Lower Nutrient Levels
If NHR-62 is a robust mediator of DR-induced longevity, then it would be predicted to suppress longevity in a second DR regimen. To test whether nhr-62(tm1818) is also required for DRinduced longevity by BDR, we measured the lifespan of wild-type and nhr-62(tm1818) worms fed bacteria at ten different concentrations. We started with our ad libitum bacterial concentration of 2.5610 9 CFU/ml followed by nine subsequent two-fold serial dilutions. At the level of DR associated with the longest lifespan

Author Summary
Dietary restriction extends the life span of diverse species across taxa, yet the underlying mechanisms are poorly understood. In humans there are clear health benefits associated with DR such as improved serum cholesterol and lipid levels. In Caenorhabditis elegans, genes implicated in the TOR pathway, autophagy, protein synthesis and energy homeostasis have been shown to modulate the dietary restriction response; however their mechanism of action is still unclear. In this work, we find that the C. elegans nuclear hormone receptor, nhr-62, is required for longevity in multiple DR regimens, providing the first evidence of a nuclear receptor required for DR-induced longevity. Additionally, nhr-62 is required for physiologic changes associated with DR, including increased autophagy and decreased levels of triglycerides, possibly through lipolysis. Moreover, nhr-62 is responsible for regulating hundreds of genes under DR, as measured by qPCR and RNA-seq. Importantly, this work is the first to report transcriptome analysis of DR in C. elegans and the first to provide functional evidence that nuclear receptors are key regulators of the DR longevity response, which imply hormonal and metabolic control of longevity, possibly through alterations in fat metabolism, lipolysis, and autophagy. (optimal; 3.23610 8 CFU/ml), wild-type worms exhibited a 188% increase in median lifespan compared to ad libitum conditions ( Figure 2E). When bacteria were diluted past this optimal DR condition, the median lifespan of wild-type worms decreased, resulting in a characteristic tent-shaped DR response curve. Unexpectedly, across the first two dilutions nhr-62(tm1818) worms exhibited a 40% increase in median lifespan compared to ad libitum conditions, identical to the wild-type DR response curve at these , eat-2(ad465), nhr-62(tm1818), eat-2;nhr-62, and eat-2;nhr-62(dhEx627). Extra chromosomal array dhEx627 (carrying a wild-type nhr-62) in eat-2;nhr-62 resulted in a significant increase in lifespan compared to both nhr-62(tm1818) and eat-2;nhr-62 mutants (p,0.001). (B) Wild-type (N2) worms with dhEx627 exhibited a significant increase in lifespan compared to wild-type (p,0.001). See Figure S2 for physiologic traits of overexpression. (C,D) Both wild-type (N2) and nhr-62(tm1818) worms fed daf-2 RNAi or cco-1 RNAi had a significant increase in lifespan when compared to controls (p,0.001). (E) BDR curve for wildtype (N2) and nhr-62(tm1818) worms. The lifespan of wild-type and nhr-62(tm1818) were not different at the three most concentrated food dilutions, but were significantly different at the remaining 7 dilutions (p,0.001). (F) Survivorship curves for wild-type (N2) and nhr-62(tm1818) fed either 6.45610 8  high nutrient conditions. However, at the optimal DR condition, the DR response curve for nhr-62(tm1818) mutants diverged from the wild-type curve; these animals were unable to respond to DR as effectively, with only a partial increase in the median lifespan at all subsequent concentrations ( Figure 2E,F and Table S2). These results suggest that nhr-62 blunts the BDR response at lower nutrient concentrations.

NHR-62 Is Widely Expressed
We obtained a transgenic strain from the Hope laboratory containing 2 kb of the nhr-62 promoter fused to gfp (pnhr-62::gfp). As previously described, transgene expression was observed from embryo to adult in the pharynx and intestine. We also observed clear expression in various neurons including sensory neurons, motor neurons, hermaphrodite specific neuron, and pharyngeal neurons (data not shown). A similar pattern of expression was seen in the eat-2(ad465) background. A full-length integrated low copy nhr-62::gfp expressing strain from the TransgeneOme Project [30] was weakly expressed in many tissues including the nuclei of pharynx, sensory neurons, intestine, spermatheca, hypodermis, and excretory cell in both wild-type ( Figure 3A-F) and eat-2 backgrounds. Again no obvious change in nhr-62::gfp intensity or localization was observed in the eat-2 background, nor were changes seen in nhr-62 mRNA levels as measured by qPCR. Notably several tissues that express nhr-62::gfp, such as the intestine, ASI sensory neurons, and hypodermis, are major endocrine tissues that coordinate metabolic states and have been implicated in aging [10,17,20,31,32].

NHR-62 Modulates DR-Induced Fat Metabolism
Animals under DR generally display dramatic changes in stored fat compared to animals in ad libitum conditions. Since nhr-62 shares homology to HNF4a-like NHRs known to function in fat metabolism such as nhr-49 and nhr-80 [33,34], we hypothesized that nhr-62 might be necessary for proper fat regulation under DR. To test this hypothesis, we treated wild-type, nhr-62(tm1818), eat-2(ad465), and eat-2;nhr-62 animals with the lysochrome dye Oil Red O, which stains neutral triglycerides and lipids ( Figure 4A). The intensity of Oil Red O has been reported to correlate positively with triglyceride levels as measured by biochemical assays [35]. We found Oil Red O staining intensity in eat-2 mutant animals was noticeably less than wild-type. Notably, mutation of nhr-62 modestly increased staining intensity relative to wild-type or eat-2 controls. eat-2;nhr-62 mutants had a moderate but significant 14.1% increase in staining intensity compared to eat-2 animals ( Figure 4B). This difference was not due to increased feeding as pumping rates of eat-2 and eat-2;nhr-62 were identical, as were body size and progeny production ( Figure S3). We also measured the levels of triglycerides (TAGs) and similarly observed that eat-2 had a significant decrease in TAG/protein compared to both wildtype and nhr-62 mutants. Furthermore, nhr-62 partially reversed this effect, giving a significant 15.6% increase in total TAG/ protein in eat-2;nhr-62 worms compared to eat-2 worms ( Figure 4C). These data indicate a general role for nhr-62 in regulating lipid levels.
Longevity in C. elegans has been associated with altered fatty acid composition. For example, loss of germline stem cells, which increases life span, results in increased monounsaturated fatty acid content, and this longevity depends on production of oleic acid (C18:1n9) [36]. To determine if nhr-62 regulates fatty acid composition under DR, we quantified individual fats by gas chromatography. eat-2;nhr-62 double mutants exhibited increased levels of saturated fatty acids compared to wild-type. Moreover, eat-2;nhr-62 double mutants showed a significant reduction of various monounsaturated fatty acids (MUFAs) and polyunsaturated fatty acids (PUFAs) compared to eat-2 mutants alone ( Figure 4D,E). These observations suggest that nhr-62 modulates changes in fatty acid composition under DR. Interestingly, we observed that the mRNA level of fat-2, a key enzyme in the production of PUFAs, was up-regulated 1.6-fold in eat-2 worms compared to wild-type as measured by qPCR. Mutation of nhr-62 dampened this up-regulation, although this effect did not reach statistical significance ( Figure S4), suggesting that post-transcriptional regulation might also be at work.
Lipases are involved in the liberation of fatty acids from triglyceride stores. Given the changes in fatty acid composition and previous studies implicating lipases in modulating longevity [37,38], we hypothesized that lipases might modulate DR-induced longevity. To test this hypothesis we screened through 34 predicted lipases in C. elegans for suppression of longevity in the eat-2;nre-1;lin-15b background (Table S3). Interestingly, we found that RNAi targeting one predicted lipase, C40H1.8, partially reduced longevity of dietary restricted animals, reaching significance in 4 out of 6 experiments ( Figure 5A and Figure S5, Table  S1). One explanation for a partial response is that C40H1.8 RNAi does not result in complete knockdown ( Figure S6). Additionally, C40H1.8 has two closely related homologs, C40H1.7 and C40H1.9, which were unaffected by C40H1.8 RNAi knockdown and might compensate for its reduction of function ( Figure S6). Consistent with a unified pathway, C40H1.8 knockdown did not further reduce the life span of eat-2;nhr-62 double mutants ( Figure 5B). By qPCR analysis, we found that C40H1.8 was upregulated five-fold under DR. In eat-2;nhr-62 animals, this upregulation was attenuated by approximately 50%, suggesting some regulation of C40H1.8 expression by nhr-62 ( Figure 5C). These data indicate that C40H1.8 may play at least a partial role in mediating eat-2-induced longevity. C40H1.8 contains a class 3 lipase domain, found in a variety of proteins predicted to have triacylglycerol lipase activity. In particular, C40H1.8 contains the catalytic triad comprised of serine, aspartate, and histidine (S198, D256, H316), a nucleophilic elbow, and other features typical of lipases ( Figure S7) [39,40]. Although C40H1.8 does not have a clear mammalian ortholog, class 3 lipases from other species include Atg15p, which is involved in autophagy in Saccharomyces cerevisiae and is required for DR-mediated longevity [38], and the Sn-1 specific diacylglycerol lipase implicated in endocannabinoid signaling in mice [41].

NHR-62 Regulates DR-Induced Autophagy
Autophagy, the catabolic process involving the degradation of cellular components which are recycled to serve as a source of energy and precursors for biosynthesis, has also been linked to DR, as well as to fatty acid metabolism and lipolysis [20,42,43]. To determine if nhr-62 influenced DR induced autophagy, we utilized a GFP::LGG-1 reporter strain. LGG-1 is the worm ortholog of S. cerevisiae ATG8p, which is incorporated into pre-autophagosomal membranes, and is necessary for proper cellular degradation during autophagy [44]. Under normal conditions GFP::LGG-1 is cytoplasmic and diffuse in the worm's hypodermal seam cells, but under DR conditions, forms puncta which are associated with an increase in autophagy [20]. We observed a significant increase in the average number of GFP::LGG-1 puncta per seam cell in the eat-2 background, and found that nhr-62 RNAi suppressed puncta formation back to wild-type levels, suggesting that nhr-62 modulates autophagy up-regulation under DR ( Figure 5D). Previous studies have shown that lipases are required for autophagy induction in the context of the gonadal longevity pathway [43]. We wondered whether the C40H1.8 lipase implicated here in eat-2 longevity played any role in autophagy, and found that C40H1.8 RNAi knockdown reduced the number of GFP::LGG1 puncta in the eat-2 background ( Figure 5E). Additionally, levels of lgg-1 mRNA were likewise induced in eat-2 mutants in an nhr-62 dependent manner ( Figure 5F). These results suggest that nhr-62 regulates DR-mediated longevity, possibly through lipolysis and autophagy.
A major regulator of the DR response is the target of rapamycin (TOR) kinase. Down-regulation of let-363/TOR induces autophagy and extends life span through DR pathways [21]. To see how nhr-62 interacts with TOR signaling, we reduced let-363/TOR by RNAi in the nhr-62(tm1818) mutants. We found that let-363 knockdown induced longevity in both wild-type and nhr-62 mutants ( Figure 5G and Table S1). These results suggest that let-363/TOR acts downstream or parallel to nhr-62.

NHR-62 Regulates the DR-specific Transcriptome
Because nhr-62 promotes DR longevity and associated physiological processes, we speculated that it could modulate the overall transcriptional change upon DR. To test this hypothesis, we performed RNA-seq on wild-type, nhr-62(tm1818), eat-2(ad465), and eat-2;nhr-62 animals. This approach not only enabled us to identify genes regulated under DR, but also which genes showed nhr-62 dependence. Our criteria for differential gene regulation included a greater than 1.5-fold change in expression compared to wild-type, a false discovery rate less than 0.05, and greater than 10 reads when normalized to the base mean. For each sample we typically obtained 25-37 million unambiguous reads, and could assign these to approximately 15,000 genes. Heat maps of the mean Euclidean distance between samples showed that nhr-62 was most similar to wild-type, while eat-2 was most different, with eat-2;nhr-62 in between ( Figure S8). qPCR on 12 of 13 regulated genes showed consistency with RNA-seq data, validating this approach for our analysis ( Figure S8).
Global profiles comparing wild-type and eat-2 gave rise to over 3,000 genes with significant gene expression changes out of 17,788 examined genes, depicting an altered transcriptome for longevity induced by the eat-2 mutation ( Figure 6A and Table S4). DAVID analysis [45,46] revealed an enrichment of biological processes involved in oxidation/reduction, phosphorus metabolism, unsaturated fatty acid, eicosanoid and ceramide metabolism, lipid modification and transport, amino acid, amine and chitin metabolism, and neuropeptide signaling among others ( Figure 6D and Table S5).

Discussion
Nuclear hormone receptors are critical regulators of animal metabolism, development and homeostasis, and are particularly well suited to couple nutrient and lipid availability to transcriptional cascades. By screening through the set of NHR genes in C. elegans we identified the HNF4a homolog, nhr-62, as an important mediator of the DR response. Several lines of evidence are consistent with the notion that nhr-62 promotes aspects of DRinduced metabolism and longevity. RNAi knockdown or mutation of nhr-62 specifically suppressed eat-2 DR-induced longevity, but not that of reduced insulin/IGF receptor (daf-2) or mitochondrial function (cco-1) arguing for a DR specific function. Importantly, nhr-62 mutation prevented animals from fully responding to the longevity-inducing effects of bacterial food dilution (BDR), thereby making it one of only a handful of DR-regulating genes that clearly modulates multiple forms of DR. By comparing eat-2 to eat-2;nhr-62 mutants we found a reversal of Oil Red O staining intensity, triglyceride levels, and autophagy induction, without seeing a difference in body size or pumping rate. Additionally, we found that nhr-62 mutation decreased the mono-and polyunsaturated fatty acid content of dietary restricted worms. Finally, we found a reversal of the DR-induced transcriptional regulation of a predicted lipase required for DR, as well as a plethora of genes through RNA-seq analysis. These results point to a model in which a dietary restricted state promotes activation of nhr-62, the lipase C40H1.8, and numerous other genes to modulate DR-induced longevity, possibly through fat metabolism, autophagy, and other processes ( Figure 6F). Although nhr-62 was the most visible candidate that emerged from our screens, it is possible that other NHRs may also play a role in the DR response, and were missed due to either incomplete knockdown or functional redundancy.
It was unexpected to see a partial BDR response by nhr-62(tm1818) mutants. One potential explanation for this observation is that by mutating nhr-62, the optimum DR threshold has been shifted to a new concentration. However, nhr-62(tm1818) mutants do not exhibit a globally shifted BDR response curve since the highest three food concentrations give median life spans that align precisely with wild-type. At further food dilution, nhr-62(tm1818) mutation significantly suppressed life span extension compared to wild-type, which results in a parallel response but with lower median life spans compared to wild-type. These results are consistent with a blunted response to DR in the lower nutrient range. In contrast to a ''master regulator'' of DR, which might completely abrogate life span extension at all food concentrations, this type of response curve could be indicative of a gene that plays a role in modulating or fine tuning the BDR longevity response at specific nutrient levels. It is also possible that nhr-62 mutation disrupts metabolism in a non-specific manner that is incompatible with a proper DR response. What argues against this is that nhr-62 mutation on its own had very little effect on gene expression or physiology under ad libitum conditions, suggesting nhr-62 is engaged in a regulatory role primarily under DR states.
nhr-62 exhibits several phenotypes suggesting it functions in mobilizing fat to partition energetic and biosynthetic demands under nutrient limitation. Notably nhr-62 mutation prevents some of the reduction of TAG and Oil Red O staining stores seen under DR longevity. Several dynamic changes in fatty acid metabolism have been observed upon DR shifts in other species. In flies there is a shift toward increasing fatty acid synthesis and breakdown, and inhibiting this suppresses DR longevity [47]. Mice initially increase endogenous fatty acid synthesis followed by prolonged fatty acid oxidation [48]. We observed that deletion of nhr-62 alters fatty acid composition in the eat-2 background, increasing the amount of the saturated fatty acid palmitate, and reducing the amount of various MUFAs and PUFAs, including C18:1n7, C20:3n6, C20:4n6, C20:4n3, and C20:5n3 fatty acids. It is possible that these changes could contribute towards the suppression of life span. Interestingly, a handful of these fatty acids have been recently implicated in longevity. The omega-6 fatty acid, DGLA (C20:3n6) stimulates autophagy and extends C. elegans life span [49]. Oleic acid (C18;1n7) has been shown to be important for life span extension upon germline removal [36]. It seems plausible that some of these lipids or their derivatives function in DR-mediated longevity. Additionally germline removal also results in up-regulation of the lipase, lipl-4, which stimulates autophagy [43]. Our results suggest a model in which lipolysis by C40H1.8 lipase could stimulate autophagy, since RNAi against this gene diminished DR longevity and GFP::LGG-1 punctual formation. Alternately, lipolysis could liberate fatty acids that serve a signaling role, binding to NHR-62 or related receptors. Future studies should further clarify the complex relationships between fatty acid composition, lipolysis, autophagy and longevity in the DR response.
Dietary restriction is associated with reduced TOR signaling, reduced protein synthesis, and increased autophagy [20,21,51]. Within the TOR pathway, the amino acid sensing G-proteins RAGA-1, RAGC-1 and TOR itself were down regulated about 20% but was significant for TOR only. Consistent with a possible downstream or parallel role, let-363/TOR RNAi induced longevity in the nhr-62 mutant background. We also found that the autophagy gene lgg-1 was up-regulated 1.5-fold, in a manner dependent on nhr-62. Consistently, we observed that autophagic vesicles visualized with GFP::LGG-1 were increased under DR in an nhr-62 dependent manner. Previous studies have shown that lgg-1 and other autophagy genes are required for DR and various longevity pathways [52]. However, other autophagy genes did not show significant regulation under DR, suggesting that posttranscriptional mechanisms contribute to this response.
Reduced protein translation extends life span and may be one of the important outputs of the DR response [53]. We observed significant down regulation of several translation initiation and release factors including inf-1, whose down-regulation has been shown to extend life span [54]. Additionally, a systematic, albeit small (approximately 20%) down-regulation of scores of ribosomal proteins was evident under DR; however none of these showed definitive nhr-62 dependence. It is possible that these minor transcriptional changes collectively recapitulate the physiologic effect of DR. Alternately, modest regulation of a critical enzyme or key regulator in the pathway might be sufficient. nhr-62 might also regulate unidentified post-transcriptional modifiers of these processes, especially given the many kinases and phosphatases affected (Tables S6,S7). Many of the genes regulated by nhr-62 also include nonconserved genes that fall into large families including F-box proteins, lectins, activated in blocked unfolded protein response (abu) and fungal induced proteins, which are speculated to be involved in innate immunity. In the future it will be important to understand how the diverse transcriptional output of nhr-62 relates to longevity and to identify direct targets by ChIP-seq.
It is currently unclear which mammalian homolog is most functionally analogous to nhr-62. The C. elegans genome has undergone a radical expansion of the HNF4 family of nuclear receptors, making an assignment of homologous function for nhr-62 challenging. One such HNF4-like nematode receptor, nhr-49, has been proposed as PPARa-like because it is involved in the starvation response, turns on genes implicated in b-oxidation, fatty acid desaturation, binding and transport. However, in our hands, nhr-49 knockdown did not abrogate the DR response. Conceivably, nhr-62 functions similar to the PPARs as well, since these nuclear receptors are known lipid sensors that regulate metabolism under different nutritional states. Corton et al originally observed a substantial overlap in the expression profile of DR and PPAR nuclear receptor regulation in mice [55]. Furthermore, in nonhuman primates, regulation of PPARs may be pivotal for the effects of DR [56]. Similar meta-analysis of expression profiles implicates PPARa in the DR response [57]. It is also possible that nhr-62 functions like HNF4, which regulates apolipoprotein levels and lipid and glucose homeostasis. Notably, both PPARs and HNF4 nuclear receptors can bind to fatty acids, which may be regulating their activity [58,59]. The finding that nhr-62 is involved in fat metabolism raises the possibility it too is regulated by fatty acid like ligands.
Interestingly, the C. elegans transcription factor SKN-1/NF-E2 regulates DR-induced longevity from the pair of ASI sensory neurons [17]. This implies that cell non-autonomous signals downstream of SKN-1/NF-E2 mediate a systemic physiological response to DR. Conceivably, this DR response could be communicated through an endocrine mechanism, and suggest that specific hormones and hormone receptors such as NHR-62 are required for DR-induced longevity. If so, it would be exciting to identify the ligand for NHR-62 and determine if it could promote longevity under replete conditions. Alternately, NHR-62 may not be ligand regulated, but instead function as a competency factor instructed by other regulatory molecules that control its response to nutrient availability and facilitate metabolic remodeling. In the future it will be interesting to distinguish these possibilities. To our knowledge, this study is the first to find a nuclear hormone receptor specifically required for DR-induced longevity, potentially though regulation of fat metabolism and autophagy.

Lifespan Assays
Lifespans were recorded as previously described [62] unless otherwise noted. A log-rank p value of less than 0.001 was used for establishing significance. Experimental worms were grown at 20uC for three generations without starving and were staged as eggs on experimental plates using a timed egg lay. L4s were moved into the appropriate conditions for the start of the aging experiment at a density of 10-12 worms per 6 cm plate. At the L4 stage, or day 0 of aging experiment, worms were moved to fresh experimental plates at density of 10-12 worms per plate. Worms were scored every other day across their lifespan for movement. Worms that did not move on their own were gently touched with a sterilized platinum wire pick and monitored for movement. If there was no movement worms were scored as dead. During these experiments worms were transferred to fresh experimental plates every other day until the end of the reproductive period and then moved once a week. Animals that crawled off the plate, exploded due to a ruptured vulva, or became Egl (egg laying defective), in which unlaid eggs hatch inside the mothers, were censored from the aging experiment. The number of aging experiments performed, mean, median, and maximum lifespans, log-rank analysis, number of worms used in each experiment and worms censored were recorded (Table S1). For the high-throughput aging screen, worms were grown to gravid adult and bleached to collect staged eggs. To increase the efficacy of RNAi, RNAi sensitive strain nre-1(hd20);lin-15b(hd126) was used. Eggs were pipetted at a target density of 15 eggs per well into 12 well cell culture plates with RNAi NGM seeded with RNAi clone of interest. At L4, or day 0 of adult, 15 ml of FUdR (5-Fluoro-29-deoxyuridine) (50 mM) was pipetted into each well to inhibit progeny production. On day 7 of adulthood, the number of worms alive was determined for each well. On day 15 of adulthood, the number of worms alive was again determined and the percent surviving was calculated. For RNAi experiments, worms were maintained on NGM plates with 20 mg/ml carbenicillin, 10 mg/ml tetracycline, and 1 mM IPTG. Worms were transferred to fresh RNAi plates every other day. For BDR longevity experiments, lifespans were recorded as previously described [10]. Worms were grown for three generations without starving on NGM plates seeded with OP50 and then staged using a timed egg lay. At L4, or day 0 of aging experiment, worms were transferred to NGM plates with FUdR (100 mg/mL). After 48 hours, worms were then washed for one hour in BDR liquid media (5.85 g NaCl, 1 g K 2 HPO 4 , 6 g KH 2 PO 4 , 1 ml cholesterol at 5 mg/mL in ethanol, and MilliQ H 2 O to 1 L) with carbenicillin (50 mg/ml), kanamycin (10 mg/ml), and tetracycline (1 mg/ml) added to inhibit bacteria replication. 90 worms were then moved into defined bacteria concentrations and scored every 3-4 days for survival at a density of 15 worms per well. On every score day worms were transferred to freshly made bacteria condition. FUdR was added to the BDR liquid media for the first 14 days of adulthood to inhibit progeny production. Bacteria concentration was determined through serial dilution, plating, and counting of colony forming units (CFUs). The number of aging experiments performed, mean, median, and maximum lifespans, log-rank analysis, number of worms used in each experiment and worms censored were recorded (Table S2).

Fat Analysis
Oil Red O protocol was adapted from [35]. 200-300 day 1 adult worms synchronized through a timed egg lay were washed with M9 and fixed with 50% isopropanol for 15 minutes. Oil Red O stock solution (0.5 g Oil Red O in 100 ml isopropanol) was diluted in water to 60% Oil Red O working solution and used to stain the worms overnight. Worms were then washed with PBS and 0.01% TritonX-100 (diluted in PBS) was added prior to mounting the worms. Images were captured using a DIC microscope and Zeiss AxioCam MRc5 camera. ImageJ was used to invert the color images and a 100 pixel diameter circle was drawn to quantify the intensity of the first intestinal cell and background was subtracted out. Statistics were done using GraphPad Prism software.
Biochemical triglyceride assays were performed using approximately 5,000 day 1 adult worms staged through timed egg lay. Worms were from NGM plates and washed 3 times in PBS then ground up using a pestle and homogenized using a Branson Sonifier Cell Disruptor. Triglyceride and protein levels were determined using a Triglyceride Colorimetric Assay Kit (Cayman Chemical Company) and the Pierce BCA Protein Assay Kit (Thermo Scientific) and analyzed on a Biotek Power Wave XS plate reader according to manufacturer's instructions. Three technical replicates were preformed for every biological replicate. Standards were also done in triplicate. Statistics were done using GraphPad Prism software.

GC Analysis of FAMEs
Day 1 adults animals were collected in 100 ml M9 buffer, washed 2-times, and run through 10-freeze/thaw cycles before being sonicated to lyse the cuticle and tissue. Protein concentration was determined using the Pierce BCA Protein Assay Kit (Thermo Scientific) according to manufacturer's instructions. 50 ml was removed to a 1.5 ml gas chromatography (GC) vial for derivatization. To each vial, 200 ng each of methyl C11:0, C13:0, and C23:0 (NuChek Prep, Inc.; 20 mg/ml stock in methanol) were added as internal reference standards. To convert fatty acids into their fatty acid methyl ester (FAME) derivatives, 0.5 ml of 2.5% H 2 SO 4 (Sigma) in methanol (Biosolve) was added to each sample and incubated at 80uC for 20 minutes with shaking. 0.75 ml of water (Biosolve, ULC Grade) and 350 ml of hexane (Sigma) were next added and the sample was incubated with shaking for 10 min to extract FAMEs into the hexane layer. The hexane layer was removed, concentrated by evaporation, and used directly for analysis. FAME analysis was carried out using an Agilent 7890A GC equipped with a flame ionization detector (FID) (Agilent Technologies, Inc) and a DB-23 column (30 m60.25 mm I.D., 0.25 mm, Agilent) using helium as a carrier gas at a flow rate of 3 ml/min. 1 ml per sample was injected in pulsed-splitless mode. The initial oven temperature was set to 80uC, held for 1 minute; increased to 170uC at 6.5uC/min; then increased to 215uC at 2.75uC/min. FAMEs were identified and compared against known reference standards for quantification.

qRT-PCR
Synchronized worms were collected into TRIzol (Invitrogen) when reaching adulthood but bearing no embryos. Total RNA and cDNA was prepared using RNeasy Mini kit (QIAGEN) and iScript cDNA Synthesis Kit (Bio-Rad) respectively. qRT-PCR was performed with Power SYBR Green master mix (Applied Biosystems) on a ViiA7 384 Real-Time PCR System (Applied Biosystems). A combination of ama-1 and cdc-42 was used as control. 4 biological replicates containing 300-400 worms each and four technical replicates were tested for each experiment. qPCR primers sequences are in Table S8.
Autophagy Marker GFP::LGG-1 Counting of puncta formation was done as previously described [20]. Briefly, young adults were transferred to RNAi plates targeting the gene of interest. L3 animals of the next generation were scored. The number of GFP puntca in each seam cell was recorded using a Zeiss Axioskop2 Plus microscope.

RNA-seq and Bioinformatic Analysis
Worms of indicated genotypes were synchronized by two rounds of bleaching and approximately 300 worms for each genotype were handpicked into TRIzol (Invitrogen) when they just reached adulthood without carrying embryos. Total RNA was prepared with RNeasy Mini Kit (QIAGEN). cDNA library was subsequently constructed by TruSeq RNA Sample Preparation Kit (Illumina). Illumina sequencing (100 bp single end) was performed on 3 biological replicates for each of the 4 treatments eat-2, nhr-62;eat-2, nhr-62, N2, giving a total of 12 samples. Sequencing reads were clipped to remove adapter sequence using fastx_clipper (http://hannonlab.cshl.edu/fastx_toolkit/), clipped reads were aligned to the C. elegans reference genome (WBcel215.67) using tophat version 1.3 (Trapnell, Pachter & Salzberg 2009) with option 2g1 to ensure unique mapping and option 2G to pass the Caenorhabditis_elegans.WBcel215.67.gtf. The number of reads mapping to each Ensemble gene was counted with htseq-count (http://www-huber.embl.de/users/ anders/HTSeq/doc/count.html#count). Statistical analysis and drawing of plots was performed in R (http://www.r-project.org/) using the bioconductor package Deseq (Anders, Huber 2010) and R function gplots (http://cran.r-project.org/web/packages/ gplots/index.html). Scatterplots of the RNA-seq data were drawn to illustrate differential expression between samples. A heatmap showing Euclidean distances between samples (calculated from variance stabilized transformed data) was plotted in R. For each data comparison a list of ''significant differentially expressed genes'' was compiled from the Deseq-result as follows: FDR,0.05, mean normalized count .10 and absolute Foldchange (FC) .1.5. Furthermore a list of genes, which are up or down regulated by eat-2, compared to the control (eat-2 vs N2), but show that they are rescued in the nhr-62;eat-2 phenotype (nhr-62;eat-2 vs eat-2) was compiled and referred to as Rescued-Genes. Further analysis of ''significant differentially expressed genes'' and ''Rescued-Genes'' was performed as follows: 1) A Venn diagram was plotted in R for ''significant differentially expressed genes'' to show the overlap of the significant differentially expressed genes. 2) Annotation and Gene-enrichment analysis of ''significant differentially expressed genes'' and ''Rescued-Genes'' was performed with David [45,46]. 3) ReviGO analysis [63] was conducted on the DAVID GO-BP gene-enrichment result of ''Rescued-Genes''. ReviGO summarizes GO-BP-IDs. The resultant network was uploaded into Cytoscape [64] for further editing and the image exported.  Figure S3 Physiologic traits of nhr-62 mutation. (A) Pumping rates of day 1 adult eat-2(ad465) and eat-2;nhr-62 mutants are reduced relative to wild-type (N2) and nhr-62(tm1818) mutants (combined data from 2 independent experiments). (B) Relative body length of day 1 adult eat-2(ad465) and eat-2;nhr-62 mutants are smaller than ad libitum controls. (C) Day 1 adult eat-2(ad465) and eat-2;nhr-62 worms have reduced total progeny compared to ad libitum controls. nhr-62(tm1818) mutants have a modest but significant decrease in total hatched progeny compared to wildtype (N2) (combined data from 3 independent experiments).