Influence of Steroid Hormone Signaling on Life Span Control by Caenorhabditis elegans Insulin-Like Signaling

Sterol-sensing nuclear receptors and insulin-like growth factor signaling play evolutionarily conserved roles in the control of aging. In the nematode Caenorhabditis elegans, bile acid-like steroid hormones known as dafachronic acids (DAs) influence longevity by binding to and regulating the activity of the conserved nuclear receptor DAF-12, and the insulin receptor (InsR) ortholog DAF-2 controls life span by inhibiting the FoxO transcription factor DAF-16. How the DA/DAF-12 pathway interacts with DAF-2/InsR signaling to control life span is poorly understood. Here we specifically investigated the roles of liganded and unliganded DAF-12 in life span control in the context of reduced DAF-2/InsR signaling. In animals with reduced daf-2/InsR activity, mutations that either reduce DA biosynthesis or fully abrogate DAF-12 activity shorten life span, suggesting that liganded DAF-12 promotes longevity. In animals with reduced DAF-2/InsR activity induced by daf-2/InsR RNAi, both liganded and unliganded DAF-12 promote longevity. However, in daf-2/InsR mutants, liganded and unliganded DAF-12 act in opposition to control life span. Thus, multiple DAF-12 activities influence life span in distinct ways in contexts of reduced DAF-2/InsR signaling. Our findings establish new roles for a conserved steroid signaling pathway in life span control and elucidate interactions among DA biosynthetic pathways, DAF-12, and DAF-2/InsR signaling in aging.


KEYWORDS
Caenorhabditis elegans steroid hormones insulin signaling aging longevity Steroid hormones have critical functions in development and maintenance of homeostasis throughout metazoan phylogeny. They exert their effects largely by binding to and regulating the activity of transcription factors of the nuclear receptor superfamily (Wollam and Antebi 2011). In the nematode Caenorhabditis elegans, bile acid-like steroid hormones known as dafachronic acids (DAs) are nuclear receptor ligands that control development and life span by binding to and regulating the activity of the nuclear receptor DAF-12 . Two structurally related DAs, D 4 -and D 7 -DA, differ in potency but appear to have similar functions in regulating larval development (Sharma et al. 2009).
Genetic analyses and rescue experiments with presumed DA biosynthetic intermediates are consistent with a model whereby D 4and D 7 -DA are synthesized from cholesterol via distinct pathways ( Figure 1A) (Wollam et al. 2012). The Rieske oxygenase family member DAF-36 catalyzes the first step of D 7 -DA biosynthesis by synthesizing 7-dehydrocholesterol [7-DHC; Yoshiyama-Yanagawa et al. 2011)]. 7-DHC is thought to be converted into lathosterol, the 3-OH group of which is subsequently oxidized by the 3-hydroxysteroid dehydrogenase DHS-16 to create lathosterone Wollam et al. 2012). Lathosterone is a direct D 7 -DA precursor and a substrate for the cytochrome P450 family member DAF-9 . The enzyme that catalyzes the conversion of 7-DHC into lathosterol has not been identified.
DAF-9 catalyzes the final common step of DA biosynthesis, converting lathosterone into D 7 -DA and 4-cholesten-3-one into D 4 -DA . Whereas D 7 -DA is detectable in lipid extracts from wild-type C. elegans, it is not detectable in extracts from daf-36 or daf-9 mutants, indicating that both DAF-36 and DAF-9 are required for D 7 -DA synthesis in vivo . D 4 -DA has not been unequivocally identified in C. elegans extracts.
DAs and DAF-12 have multiple functions during larval development. Under conditions of high population density, food scarcity, and high temperature, wild-type C. elegans larvae undergo developmental arrest in an alternative third larval stage known as dauer. Dauer larvae are long-lived and resistant to environmental insults (Hu 2007). daf-9 mutants, which lack endogenous DAs , arrest as dauer larvae constitutively, even when ambient conditions favor reproductive development (Gerisch et al. 2001;Jia et al. 2002). This dauer-constitutive phenotype is fully suppressed by exogenous DA (Giroux et al. 2008;Motola et al. 2006;Sharma et al. 2009) as well as by null mutations in daf-12 (Gerisch et al. 2001). daf-12 ligand binding domain mutants also have a dauer-constitutive phenotype (Antebi et al. 1998;Antebi et al. 2000). Therefore, unliganded DAF-12 promotes dauer arrest. The dauer-constitutive phenotype of daf-9 mutants and daf-12 ligand binding domain mutants is also suppressed by mutations in din-1S, which encodes a transcriptional coregulator that binds to DAF-12 (Ludewig et al. 2004). Taken together, these results support a model whereby unliganded DAF-12 acts together with DIN-1S to promote dauer arrest; DAs permit reproductive development by binding to DAF-12, thereby preventing its interaction with DIN-1S ( Figure 1B) . DAs are also required during larval development for proper gonadal migration (Gerisch et al. 2001;Motola et al. 2006) and expression of let-7-family microRNAs that coordinate the timing of cell divisions (Bethke et al. 2009;Hammell et al. 2009). In adult males, DAs are required for normal mate searching behavior (Kleemann et al. 2008).
The roles of DAs and DAF-12 in the control of adult life span are complex. daf-9 mutants are long-lived when cultured at 15° (Gerisch et al. 2007;Jia et al. 2002) but short-lived when cultured at temperatures between 20°and 25° (Gerisch et al. 2007;Gerisch et al. 2001;Jia et al. 2002;Lee and Kenyon 2009). These temperature-dependent phenotypes are suppressed by daf-12 loss-of-function mutations (Jia et al. 2002;Lee and Kenyon 2009) and exogenous DA (Gerisch et al. 2007), suggesting that unliganded DAF-12 promotes longevity at low temperatures but shortens life span at higher temperatures ( Figure  1C). din-1S mutation suppresses the life span extension conferred by daf-9 mutation at low temperatures (Ludewig et al. 2004), indicating that at 15°, unliganded DAF-12 and DIN-1S act together to extend life span ( Figure 1C).
How DAs and DAF-12 influence life span in the context of reduced DAF-2/InsR signaling is poorly understood. The daf-9 missense allele rh50 has distinct effects on life span in the context of specific daf-2 mutant alleles. At 15°, daf-9(rh50) shortens the life span of both daf-2(e1368) (harboring a missense mutation in the DAF-2 ligand binding domain) and daf-2(e1370) (harboring a missense mutation in the tyrosine kinase domain) mutant animals. However, at 22.5°, daf-9(rh50) shortens daf-2(e1368) life span but lengthens daf-2(e1370) life span (Gerisch et al. 2001). Accordingly, exogenous D 4 -DA prolongs the life span of daf-2(e1368) animals but does not significantly influence the life span of daf-2(e1370) animals (Gerisch et al. 2007). Furthermore, daf-12 mutant alleles influence the life span of daf-2/ InsR mutants in an allele-specific manner. For example, the non-null allele daf-12(m20) (see Supporting Information, Figure S1) (Antebi et al. 2000;Snow and Larsen 2000) suppresses the extended life span phenotype of daf-2(e1368) harboring a mutation in the ligand binding domain (Patel et al. 2008) at all temperatures tested (Gems et al. 1998), whereas it enhances daf-2(e1370) life span extension at high temperatures (Gems et al. 1998;Larsen et al. 1995). In aggregate, these data underscore the need for further investigation into how steroid hormone signaling and DAF-2/InsR signaling interact in life span control. Specifically, the relative contributions of liganded and unliganded DAF-12 to life span control have not been defined. Prior studies on the interactions of daf-12 and daf-2/InsR mutants in life span control were performed with non-null alleles of daf-12 (Gems et al. 1998;Larsen et al. 1995), complicating the interpretation of these experiments.
Here we used null alleles of daf-36 and daf-12 to explore the relationship between DA pathways and DAF-2/InsR signaling in life span regulation. Our results are consistent with a model whereby both liganded and unliganded DAF-12 influence life span. Liganded DAF-12 promotes longevity in animals with reduced DAF-2/InsR signaling. Unliganded DAF-12 also extends life span in animals subjected to daf-2/InsR RNA interference (RNAi) but shortens life span in daf-2/ InsR mutants and in animals lacking a germline. These findings establish that distinct DAF-12 activities interact with DAF-2/InsR signaling to control life span.

C. elegans strains
The wild-type N2 Bristol strain was used. Mutant alleles used are described in Table S1. Compound mutants were constructed using standard techniques.

Dauer arrest assays
Dauer arrest assays were performed at the indicated temperatures in I-36NL model incubators (Percival Scientific, Inc., Perry, IA) as described previously (Hu et al. 2006). P values were calculated using the Student t-test. Statistical analysis of all data is presented in Table S2.

Life span assays
Life span assays were performed in I-36NL incubators (Percival) at the indicated temperatures. After alkaline hypochlorite treatment and two generations of growth, young adult animals were placed onto nematode growth media (NGM) plates containing 25 mg/ml (100 mM) 5-fluoro-29-deoxyuridine (FUDR; Sigma) and 10 mg/ml nystatin (Sigma) that had been seeded with 20· concentrated Escherichia coli OP50. For life span assays of strains carrying glp-1(e2141), animals were raised at 25°, and sterile young adult animals were placed onto NGM plates containing nystatin but lacking FUDR as described above. Assays were conducted at 20°unless otherwise noted. Viability was assessed visually or with gentle prodding. Prism software (GraphPad Software, La Jolla, CA) was used for data representation and statistical analysis. P values were calculated using the log-rank test. Statistical analysis of all data is presented in Table S2.

RNAi
Feeding RNAi was performed using variations of standard procedures (Boulton et al. 2002). For dauer assays, NGM plates containing 5 mM isopropyl beta-D-1-thiogalactopyranoside (IPTG) and 25 mg/ml carbenicillin were seeded with 500 ml of overnight culture of E. coli HT115 harboring either control L4440 vector or daf-2 RNAi plasmid. Gravid animals cultured on control or daf-2 RNAi plates were picked to assay plates for 6-hr egg lays. Dauer larvae were scored after progeny had been incubated at 25°for 48-60 hr. For life span assays, NGM plates containing 5 mM IPTG, 25 mg/ml carbenicillin, 25 mg/ml FUDR, and 10 mg/ml nystatin were seeded with 500 ml of 5· concentrated overnight culture of E. coli HT115 harboring either control L4440 vector or daf-2 RNAi plasmid. Young adult animals cultured on standard NGM plates seeded with E. coli OP50 were picked to RNAi plates and scored for viability as described above.
Notably, daf-12(m20) is a nonsense mutation that specifically affects DAF-12A isoforms; it is predicted to truncate DAF-12A upstream of the C-terminal ligand binding domain, potentially resulting in a DAF-12A polypeptide that contains an intact zinc finger in the N-terminal DNA binding domain. The DAF-12B isoform, which contains the ligand binding domain but lacks the DNA binding domain, is not affected by m20 ( Figure S1) (Antebi et al. 2000;Snow and Larsen 2000). The influence of a daf-12 null allele on the dauer-constitutive phenotype of daf-2 mutants has not been explored.
daf-12(m20) also has disparate effects on the longevity of Class 1 and Class 2 daf-2 mutants (Gems et al. 1998;Larsen et al. 1995). To gain insight into how DAF-12 influences life span in animals with reduced DAF-2/InsR signaling, we measured life spans of daf-12(null) animals in three contexts of reduced DAF-2/InsR activity. First, we performed daf-2 RNAi in wild-type and daf-12(null) animals. daf-2 RNAi does not induce dauer arrest in wild-type animals at 25°but enhances dauer arrest at 27° (Dillin et al. 2002), suggesting that the extent to which RNAi reduces DAF-2 activity is less than that caused by Class 1 and Class 2 daf-2 mutant alleles, which all have strong dauer-constitutive phenotypes at 25° (Gems et al. 1998). As previously observed (Dillin et al. 2002), daf-2 RNAi extended life span to a degree comparable to daf-2 mutation (Figure 2, A-C). Life span extension induced by daf-2 RNAi was significantly attenuated in daf-12(null) animals (Figure 2, A and D; Table S2); daf-12(null) animals subjected to daf-2 RNAi exhibited a 34.5% decrease in median survival compared to wild-type animals on daf-2 RNAi (P , 0.0001, log-rank test), suggesting that DAF-12 is required for life span extension in animals with reduced DAF-2/InsR activity. At 25°, daf-12(null) mutants live approximately as long as wild-type animals do (Figure 2A, P = 0.0018; Figure 2B, P = 0.1787; Figure 2C, P = 0.0678; Table S2), indicating that the effect of daf-12(null) on the life span of animals subjected to daf-2 RNAi is unlikely to be due to general frailty. Furthermore, RNAi of three unrelated genes in wild-type and daf-12(null) animals revealed that daf-12(null) animals do not have an Rde (RNAi-defective) phenotype ( Figure S3). This indicates that the relative reduction in life span extension caused by daf-2 RNAi in daf-12(null) animals is unlikely to be due to reduced inactivation of daf-2.
A possible explanation for the differences in the influence of DAF-12 on life span between the contexts of daf-2 RNAi and mutational reduction of DAF-2/InsR activity is the distinct food sources employed in each experimental condition. RNAi by feeding, as used to reduce daf-2/InsR activity, involves the use of an E. coli strain, HT115, which is distinct from the standard lab food source, E. coli OP50. It has been shown that using HT115 in place of OP50 as a food source is sufficient to impact C. elegans longevity (Maier et al. 2010). In the context of DAF-2/InsR activity reduction by daf-2 RNAi, the daf-12(null) mutation shortened median life span [daf-12(null) vs. that of the wild-type, P , 0.0001]. (B) In the context of the Class 1 daf-2(e1368) allele, daf-12(null) shortened the median life span [daf-2(e1368);daf-12(null) vs. daf-2(e1368), P , 0.0001]. (C) The daf-12(null) mutation did not shorten the median life span of animals with DAF-2/InsR activity reduction via the Class 2 daf-2(e1370) allele [daf-2 (e1370);daf-12(null) vs. daf-2(e1370), P = 0.4275]. (D) Scatter plot of median survival of daf-12(null) animals normalized to that of daf-12 wild-type animals in the three contexts of reduced DAF-2/InsR activity, separated by assay food source. Error bars indicate SEM. (E and F) Food source control experiments described in (B) and (C), respectively. E. coli HT115 expressing vector control RNAi was used as the assay food source as opposed to E. coli OP50. For each experiment, more than 60 animals were scored per genotype, and at least two experimental replicates were performed. See Table S2 for all raw data and statistics.
To test whether E. coli strain differences influence the effect of daf-12 (null) on life span in the context of reduced DAF-2/InsR activity, we performed life span assays with daf-2(e1368) and daf-2(e1370) mutant animals grown on E. coli HT115 (expressing empty vector control RNAi) as the food source. Under these conditions, daf-2(e1368);daf-12 double mutants were not shorter lived than daf-2(e1368) single mutants [Figures 2,D and E,and Table S2: daf-2(e1368);daf-12 animals exhibited a 0% change in median life span compared to daf-2 (e1368), P = 0.1989]. daf-2(e1370);daf-12 double mutants grown on E. coli HT115 were shorter lived than daf-2(e1370) single mutants, but the difference in median life span was not statistically significant [Figures 2,D and F,and Table S2: 12.1% decrease in median life span compared to daf-2(e1370), P = 0.1439]. These results suggest that the food source does not account for the differential effects of daf-12(null) on longevity in the three contexts of reduced daf-2/InsR activity that we examined.

Role of unliganded DAF-12 in life span control by DAF-2/InsR signaling
To elucidate the relative contributions of liganded and unliganded DAF-12 to life span control in animals with reduced DAF-2/InsR signaling, we determined the influence of daf-12(null) mutation on the life spans of daf-36(null) animals with reduced DAF-2/InsR activity. Since daf-36(null) animals do not make D 4 -or D 7 -DA , DAF-12 activity in the context of daf-36(null) is largely attributable to unliganded DAF-12.
In aggregate, our results (summarized in Table 1) support roles for both liganded and unliganded DAF-12 in life span control in animals with reduced DAF-2/InsR signaling. Liganded DAF-12 promotes longevity in all contexts tested, whereas unliganded DAF-12 modulates life span in a context-dependent manner; in the context of reduced DAF-2/InsR signaling via daf-2 RNAi, unliganded DAF-12 promotes longevity. In contrast, in the context of DAF-2/InsR signaling reduction via daf-2(e1368) mutation, unliganded DAF-12 is detrimental to life span.

DISCUSSION
Although the interface between C. elegans hormone signaling and the DAF-2/InsR pathway has been explored previously (Gems et al. 1998;Larsen et al. 1995), how these pathways interact to influence longevity remains obscure. Our work provides novel insights into the genetic interactions of liganded and unliganded DAF-12 with DAF-2/InsR signaling in life span control.
Liganded DAF-12 promotes longevity in animals with reduced DAF-2/InsR activity Ambiguity about the role of DAF-12 in determining longevity is due at least in part to the use of the non-null daf-12(m20) allele in previous investigations (Gems et al. 1998;Larsen et al. 1995). We now show that the daf-12(rh61rh411) null allele and the non-null daf-12(m20) allele have distinct effects on the life spans of animals with reduced DAF-2/InsR signaling ( Figure 2) (Gems et al. 1998;Larsen et al. 1995;McCulloch and Gems 2007). Our results indicate that at high temperatures, DAF-12 promotes longevity in animals with reduced DAF-2/ InsR signaling (Figure 2). The magnitude of this life-span-extending effect of DAF-12 is greater in animals subjected to daf-2 RNAi than in animals harboring daf-2 mutation, indicating that the specific context of reduced DAF-2/InsR activity influences the role of DAF-12 in life span control (Table 1). The disparity between our results and those obtained with the non-null daf-12(m20) allele (Gems et al. 1998;Larsen et al. 1995) suggests that the longevity-promoting effect of daf-12(m20) and other non-null daf-12 mutations (that specifically affect DAF-12A isoforms) on the life span of daf-2(e1370) and other Class 2 daf-2 mutants (Antebi et al. 2000;Gems et al. 1998;Larsen et al. 1995;McCulloch and Gems 2007) may be attributable to a lifespan-extending activity of either the DAF-12B isoform, which contains a ligand binding domain but no DNA binding domain, or truncated DAF-12A polypeptides containing most of the DNA binding domain but lacking the ligand binding domain (Antebi et al. 2000;Snow and Larsen 2000). These DAF-12 polypeptides do not play a significant role in dauer regulation by DAF-2/InsR, as daf-12 (null) and daf-12(m20) have similar effects on the dauer-constitutive phenotypes of daf-2 mutants ( Figure S1).   The observation that mutations in either daf-12 ( Figure 2) or genes encoding DA biosynthetic components ( Figure 3) reduce life span in animals with reduced DAF-2/InsR signaling is consistent with a model whereby liganded DAF-12 promotes longevity when DAF-2/InsR signaling is reduced ( Figure 5A). Similar results indicate that liganded DAF-12 also promotes longevity in germline-ablated animals ( Figure  4D) (Gerisch et al. 2007;Gerisch et al. 2001;Hsin and Kenyon 1999;Yamawaki et al. 2010). The magnitude of the effect of reducing the activity of DA biosynthetic components or DAF-12 on life span is greater in animals lacking a germline than in animals with reduced DAF-2/InsR activity (Figures 2-4) (Gerisch et al. 2007;Gerisch et al. 2001;Hsin and Kenyon 1999;Yamawaki et al. 2010). The molecular basis for this observation is not known.
Unliganded DAF-12 has context-dependent influences on life span in animals with reduced DAF-2/InsR activity Unliganded DAF-12 promotes longevity in animals cultured at low temperatures (Gerisch et al. 2001;Jia et al. 2002) but shortens life span in animals that are cultured at high temperatures (Lee and Kenyon 2009). Here we show that in the context of reduced DAF-2/InsR signaling, unliganded DAF-12 can either extend or shorten life span. In daf-36(null) animals, which lack both D 4 -and D 7 -DA , DAF-12 extends life span in the context of daf-2 RNAi ( Figure 4A) but shortens life span in the contexts of the Class 1 daf-2 (e1368) allele ( Figure 4B) and germline ablation ( Figure 4D). Since DAF-16/FoxO is a major target of both DAF-2/InsR signaling and germline signaling in life span control (Hsin and Kenyon 1999;Kenyon et al. 1993), it is likely that the impact of unliganded DAF-12 on longevity is strongly influenced by relative levels of DAF-16/FoxO activity. This notion is supported by a recent report demonstrating that DAF-12 and DAF-16/FoxO mutually influence target gene expression in animals lacking a germline (McCormick et al. 2011).
Transcriptional coregulator DIN-1S shortens life span in animals lacking a germline DIN-1S acts together with unliganded DAF-12 at 15°to promote longevity (Ludewig et al. 2004). Here we show for the first time that the DAF-12 coregulator DIN-1S plays a major role in life span control , respectively. E. coli HT115 expressing vector control RNAi was used as the assay food source as opposed to E. coli OP50. For each experiment, more than 60 animals were assayed per genotype, and at least two experimental replicates were performed. Raw data and statistics are presented in Table S2. in germline-ablated animals. Both daf-12(null) and din-1S(null) suppressed the life-span-shortening effect of daf-36(null) on animals lacking a germline (Figure 4, D and E). This is consistent with a model whereby unliganded DAF-12 and DIN-1S act together to shorten life span. The role of DIN-1S in life span control by unliganded DAF-12 in the context of reduced DAF-2/InsR signaling is not known.

Context-dependent life span control by DAF-12 complexes
Our results define new functions for DAF-12 complexes in life span control and underscore the context-dependence of these activities ( Figure 5). As summarized in Table 1, liganded DAF-12 promotes longevity both in animals with reduced DAF-2/InsR activity (Figures 2, A-C, E and F, and 3, C-H) as well as in animals lacking a germline (Figure 4, D and E) (Gerisch et al. 2007;Yamawaki et al. 2010). Unliganded DAF-12 also promotes longevity in animals subjected to daf-2 RNAi ( Figure 4A) but shortens life span in animals harboring daf-2 mutations or lacking a germline (Figure 4, B-E). The basis for the context-dependent influence of unliganded DAF-12 on life span may involve context-specific proteins and/or undiscovered DAF-12 ligands present in daf-36(null) animals that influence the transcriptional regulatory activity of DAF-12 complexes.