Report

DAF-16 Target Genes That Control C. elegans Life-Span and Metabolism

See allHide authors and affiliations

Science  25 Apr 2003:
Vol. 300, Issue 5619, pp. 644-647
DOI: 10.1126/science.1083614

Abstract

Signaling from the DAF-2/insulin receptor to the DAF-16/FOXO transcription factor controls longevity, metabolism, and development in disparate phyla. To identify genes that mediate the conserved biological outputs of daf-2/insulin-like signaling, we usedcomparative genomics to identify 17 orthologous genes from Caenorhabditis and Drosophila, each of which bears a DAF-16 binding site in the promoter region. One-third of these DAF-16 downstream candidate genes were regulatedby daf-2/insulin-like signaling in C. elegans, and RNA interference inactivation of the candidates showed that many of these genes mediate distinct aspects of daf-16 function, including longevity, metabolism, anddevelopment.

The C. elegans daf-2 pathway controls longevity, metabolism, and development and is orthologous to the mammalian insulin and insulin-like growth factor 1 signaling cascade (1). Decreased daf-2 signaling causes up to threefold life-span extension, increased fat storage, and constitutive arrest at the dauer diapause stage (24). The daf-2 mutant phenotypes are suppressed by mutations in daf-16, indicating that daf-16 is negatively regulated by daf-2 signaling and is the major downstream effector. daf-16 encodes a fork-head transcription factor (5, 6), which translocates into the nucleus (7) and modulates transcription when daf-2 signaling is abrogated. Multiple daf-16 transcriptional targets are likely to mediate the diverse functions of daf-2/insulin-like signaling. Candidate gene and biochemical approaches revealed that genes encoding superoxide dismutase (sod-3), an FK506 binding protein, and a nucleolar protein are regulated by C. elegans daf-16 (8, 9). The mammalian DAF-16 orthologs (FOXO1, FOXO3, and FOXO4) regulate genes involved in growth control, apoptosis, DNA repair, and oxidative stress (10).

Because the pathway from DAF-2/insulin receptor to DAF-16/FOXO regulates both longevity and metabolism in C. elegans, Drosophila, and mammals (1, 1114), DAF-16/FOXO might control homologous target genes in different species to mediate conserved functions. DAF-16 and its mammalian homologs bind to an identical consensus DNA sequence (TTGTTTAC) in vitro (15), FOXO3 binds to this consensus site in the MnSod promoter in mammalian cells, and binding to this consensus site is required for FOXO3 transactivation of MnSod (16). We sought to identify DAF-16 transcriptional targets by searching for DAF-16 binding sites in the regulatory regions of genes. Given the high expected rate of detecting a DAF-16 binding site by chance alone [3700 sites expected by chance (17)], the search for such a site upstream of a C. elegans gene and upstream of its ortholog in a divergent animal species would highlight functional DAF-16 sites in conserved components of the DAF-16 transcriptional cascade. Because the Drosophila genome is relatively small and well assembled, we searched for DAF-16 binding sites in Drosophila and C. elegans orthologous genes.

We surveyed 1 kb upstream of the predicted ATG of 17,085 C. elegans and 14,148 Drosophila genes and identified 947 C. elegans and 1760 Drosophila genes that contain at least one perfect-match consensus DAF-16 binding site within the 1-kb promoter region. We then compared these DAF-16 binding site–containing worm and fly genes with a list of 3283 C. elegans and Drosophila genes that are orthologous to each other (17), and identified 17 genes that are orthologous between Drosophila and C. elegans and bear a DAF-16 binding site within 1 kb of their start codons in both species (Table 1). One Drosophila and one C. elegans candidate target gene had more than one DAF-16 binding site within the 1-kb region (Table 1).

Table 1.

daf-16 transcriptional target candidates predicted by the survey of 1 kb upstream of each ATG in both C. elegans and Drosophila genomes. n.c., no change from control; n.d., DAF-16 site was not searched for because no clear starting ATG of the C. briggsae homolog was determined; dash indicates no expression detected.

Gene Homology DAF-16 siteView inline (C. elegans) DAF-16 SiteView inline (Drosophila) DAF-16 siteView inline (C. briggsae) mRNA in daf-2(—) RNAi inactivation phenotype
Life-span Dauer Fat storage
C08B11.8 Similar to yeast glucosyltransferase 48 324 ∼500View inline n.c. n.c. n.c. n.c.
C10G11.5 (pnk-1) Pantothenate kinase 389 334 ∼300 5X Shortened n.c. Reduced
C39F7.5 Cytochrome c heme binding site 375 299 ∼350View inline 3X n.c. n.c. n.c.
E04F6.2 Unknown 240 150 ∼400View inline n.c. n.c. n.c.
F14F4.3 (mrp-5) ABC transporter 111,920 567 ∼900View inline n.c. Extended Enhanced n.c.
F27C8.1 Amino acid transporter 915 828 ∼2500View inline n.c. n.c. n.c.
F43G9.5 Subunit of pre-mRNA cleavage factor 1 371 609 ∼350 0.4X n.c. n.c. n.c.
F52H3.5 Similar to yeast stress-induced protein 763 982,400 ∼2200View inline 7X n.c. n.c. n.c.
F54D5.7 Acyl-CoA dehydrogenase 513 825 n.d. n.c. n.c. n.c. n.c.
K07B1.3 Mitochondrial carrier 895 69 n.d. n.c. n.c. n.c. n.c.
T20B3.1 Carnitate acyltransferase 536 96 n.d. n.c. n.c. n.c.
T20B5.3 Hyaluronoglucosaminidase 588 507 ∼500, ∼600 n.c. n.c. n.c. n.c.
T21C12.2 (hpd-1) Hydroxyphenylpyruvate dioxygenase 983 175 ∼1700View inline 0.5X Extended Enhanced n.c.
T23B12.4 Similar to yeast glucose repressible protein MAK10 90 633 ∼100 n.c. n.c. n.c. n.c.
Y106G6H.7 Mitochondrial energy transfer protein signature 71 343 ∼70 n.c. n.c. n.c. n.c.
ZC506.3 Phosphatidylserine synthase 1 702 358 ∼630View inline n.c. n.c. n.c. n.c.
ZK593.4 (rbp-2) Similar to retinoblastoma binding protein 2 27 716 ∼2700View inline 0.5X Extended n.c. n.c.
  • View inline * Nucleotide position upstream of the predicted ATG.

  • View inline These binding sites contain one mismatch from the consensus that retains DAF-16 binding in vitro.

  • To examine whether the predicted DAF-16 downstream genes are regulated by insulin signaling through DAF-16, we compared the RNA expression level of each candidate in wild-type, daf-2(e1370), and daf-2(e1370); daf-16(mgDf47) animals (Fig. 1). Under conditions in which sod-3 was robustly induced in the daf-2 mutant (18), we found that 6 of the 17 (∼35%) predicted DAF-16 down-stream genes were differentially expressed in daf-2 and daf-2;daf-16 mutant animals (Fig. 1), indicating that their expression was regulated by insulin signaling through DAF-16. Three of the six genes were expressed at levels three to seven times higher in a daf-2 mutant than in the wild type or the daf-2;daf-16 double mutant. This fraction of genes, robustly regulated by the daf-2 pathway, is much higher than the fraction expected to occur by chance; data from a microarray analysis indicate that 1% of the 16,721 C. elegans genes tested were regulated by three-fold or more (19).

    Fig. 1.

    The expression of seven DAF-16 target candidate genes is regulated by daf-2/insulin-like signaling in a daf-16–dependent manner. RNA from wild-type, daf-2(e1370), and daf-2(e1370);daf-16(mgDf47) animals was tested. Fold differences in expression levels are shown below each band.

    The expression of ZK593.4, T21C12.2, and F43G9.5 was down-regulated and that of C10G11.5, F52H3.5, and C39F7.5 was up-regulated in the daf-2 mutant in a daf-16–dependent manner (Fig. 1 and Table 1). Because the positively and negatively regulated genes bear conserved DAF-16 binding sites and are likely to be direct targets of DAF-16, these results suggest that DAF-16 acts as both a transcriptional activator and a transcriptional repressor, depending on gene context, similar to the fork-head transcription factor LIN-31 (20). We failed to detect the expression of three of the DAF-16 downstream gene candidates (E04F6.2, F27C8.1, and T20B3.1), probably because of low endogenous expression. For the remaining eight candidates, we did not detect a noticeable change of expression under the conditions tested. These genes may represent false positives predicted by informatics. Alternatively, some of these genes may be regulated by daf-2 signaling in a tissue- or stage-specific manner, so that their differential expression was not detected in RNA that was isolated from whole adult animals. Because neuronal daf-2 signaling is sufficient to regulate C. elegans longevity (21), analysis based on changes of mRNA levels in whole animals might miss regulatory genes acting in particular tissues, such as neurons. Such regulatory genes would be identified by the informatic search for DAF-16 binding sites. Green fluorescent protein fusions to these candidate genes might reveal whether they are expressed in particular tissues and whether their expression is regulated by daf-2 signaling.

    To examine whether the candidate DAF-16 downstream genes are biologically important targets of daf-2 signaling, we used RNA interference (RNAi) (22) in wild-type or rrf-3(pk1426) strains and daf-2(e1370) or age-1(hx546) strains to reduce the expression of each gene and to determine whether life-span, dauer arrest, and fat storage were affected. rrf-3(pk1426) animals are hypersensitive to RNAi (23) but are otherwise wild type in our functional assays (18). age-1(hx546) animals live long but do not arrest as dauer constitutively at 25°C (24), and they represent a sensitized genetic background with a slight reduction of daf-2 pathway signaling. We expected RNAi inactivation of the genes that are down-regulated in the daf-2 mutant to promote daf-2 mutant phenotypes, including life-span extension, dauer arrest, and increased fat storage, and we expected RNAi inactivation of the genes upregulated in the daf-2 mutant to suppress the daf-2 mutant phenotypes.

    RNAi of ZK593.4 (rbp-2) and T21C12.2 (hpd-1), genes that are down-regulated in the daf-2 mutant, caused rrf-3(pk1426) animals to live considerably longer than those undergoing control RNAi or RNAi of an unrelated gene (Fig. 2, A and B) (18). The life-span extension was modest compared to that of RNAi inactivation of daf-2 (a 30% increase in mean life-span for rbp-2 or hpd-1 RNAi as compared with a 100% increase for daf-2 RNAi). rbp-2 and hpd-1 might constitute a fraction of the DAF-16 transcriptional cascade. RNAi of hpd-1 also promoted dauer arrest under sensitized conditions (Table 2), whereas RNAi of rbp-2 did not. Although RNAi inactivation of hpd-1 or rbp-2 in wild-type animals did not induce dauer arrest, hpd-1 RNAi inhibited dauer recovery of daf-2(e1370) at 22°C, compared with control or rbp-2 RNAi (Table 2) (18). rbp-2 might specifically regulate life-span, whereas hpd-1 might have a broader role in daf-16 regulation of both dauer arrest and longevity.

    Fig. 2.

    Longevity after RNAi of DAF-16 transcriptional targets. Life-span was determined in (A to C) rrf-3(pk1426),(D) age-1(hx546),or (E) wild-type animals undergoing the indicated RNAi. The mean life-span of rrf-3(pk1426) animals undergoing control RNAi was 11.7 ± 3 days, for ZK593.4(rbp-2) RNAi it was 15.3 ± 4 days (P < .0001), for T21C12.2 (hpd-1) RNAi it was 15.3 ± 4days (P < .0001), and for F14F4.3 (mrp-5) RNAi it was 16.1 ± 2 days (P < .0001). The mean life-span of age-1(hx546) animals undergoing control RNAi was 16.9 ± 3 days and for C25E10.12 RNAi it was 14.4 ± 2 days (P = 0.0009). The mean life-span of wild-type animals undergoing control RNAi was 12.1 ± 2 days and for C25E10.12 RNAi it was 11.3 ± 3 days (P = 0.24). Student's t test P values are shown in parentheses.

    Table 2.

    Dauer formation of daf-2(e1370) animals at 22°C under the indicated RNAi conditions.

    Day 4 at 22°C Control RNAi daf-2 RNAi T21C12.2 RNAi F14F4.3 RNAi
    daf-2(e1370) adult 100% 0 10% 2%
    daf-2(e1370) dauer 0 100% 90% 98%

    rbp-2 encodes a homolog of the mammalian RB binding protein 2 (RBP2), which is implicated in gene expression control and chromatin remodeling (25, 26). sir-2, which modulates longevity in yeast and in C. elegans (27, 28), encodes a histone deacetylase, also highlighting a role for chromatin remodeling in longevity control. rbp-2 might be regulated by DAF-16 to further modify chromatin when daf-2 signaling is decreased. hpd-1 encodes the enzyme 4-hydroxyphenylpyruvate dioxygenase involved in the catabolism of phenylalanine and tyrosine to fumarate and acetoacetate. Insulin signaling might regulate amino acid degradation and contribute to the coupling of nutritional status and amino acid turnover. In Drosophila, reduced function of the Indy transporter, which carries metabolic intermediates including fumarate, markedly extends life-span (29, 30). hpd-1 might also affect the balance of metabolic intermediates such as fumarate and influence longevity through a mechanism similar to that of Indy in Drosophila. Alternatively, hpd-1 encodes a dioxygenase in a degradation pathway from tyrosine; mutations in this dioxygenase could affect tyrosine pools and in turn affect dopaminergic signaling, or they could affect free radical production, an expected byproduct of dioxygenases. pnk-1 (C10G11.5), a gene up-regulated in the daf-2 mutant, encodes one of the two pantothenate kinases in C. elegans, the rate-limiting enzymes in coenzyme A synthesis. Because coenzyme A is key to fat metabolism, we examined fat storage in pnk-1 RNAi animals, using Nile Red staining (31). RNAi of pnk-1 caused dramatic reduction of fat storage in the intestine of wild-type or daf-2 mutant animals (Fig. 3). Thus, increased fat storage in daf-2 mutants might be partly a result of pnk-1 upregulation. RNAi of pnk-1 also dramatically shortened wild-type and daf-2 mutant adult lifespan (23), suggesting that inactivation of pnk-1 compromises the health of animals.

    Fig. 3.

    RNAi of pnk-1 reduced lipid storage. Nile Red staining of wild-type or daf-2(e1370) animals undergoing the indicated RNAi is shown. (A and C) Nile Red staining showing intestinal fat droplets in wild-type or daf-2(e1370) animals. (B and D) Reduced Nile Red staining in wild-type or daf-2(e1370) animals undergoing RNAi against pnk-1.

    RNAi inactivation of F43G9.5, C39F7.5, and F52H3.5 did not affect dauer arrest, lifespan, or fat storage under the conditions tested (Table 1). It is possible that RNAi did not reduce their expression to a level necessary to produce a phenotype. Alternatively, these genes might have more subtle functions in daf-2 regulation of metabolism or longevity, or other genes might provide redundant functions to compensate for their inhibition.

    RNAi inactivation of F14F4.3 (mrp-5) promoted life-span extension and dauer arrest (Fig. 2C and Table 2). Although we did not detect differential expression of mrp-5 in daf-2 as compared with daf-2;daf-16, it is possible that daf-2 signaling regulates mrp-5 expression in specific tissues or at specific times, and this was not detected under our experimental conditions. mrp-5 encodes an adenosine triphosphate–binding cassette, subfamily C transporter. Members of this subclass are implicated in modulating insulin secretion and in transport of nucleoside analogs and glutathione (32). mrp-5 might act as a feedback regulator of insulin secretion to influence life-span and dauer arrest. Alternatively, mrp-5 might also affect life-span by regulating glutathione transport and antioxidant defense.

    The genome of the nematode C. briggsae has been sequenced. Because C. elegans and C. briggsae are more closely related than C. elegans and Drosophila (33), we examined whether the DAF-16 binding site that is conserved between orthologous C. elegans and Drosophila genes is also conserved in the promoters of the C. briggsae homologs. Among the 14 C. elegans DAF-16 downstream gene candidates that have a close C. briggsae homolog, 5 genes have a DAF-16 binding site within 1 kb of the predicted ATG, and 5 genes have a DAF-16 binding site containing one mismatch, with specific substitutions that would retain DAF-16 binding (15) (Table 1). For the remaining four DAF-16 downstream gene candidates, we found DAF-16 binding sites only when intergenic regions further upstream were surveyed (up to 2.7 kb) (Table 1). It is possible that DAF-16 binding sites drift and relocate frequently, and for some of the C. elegans and Drosophila genes that bear DAF-16 binding sites within 1 kb of the ATG, the counterparts in C. briggsae might have relocated the binding site away from the 1-kb promoter region.

    This informatic search for DAF-16 sites within the 1 kb upstream of the ATG is not yet saturating. A more complete search would cover the intergenic regions that are located upstream of the worm and fly genes, as well as large introns near the ATG. This would make the C. elegans search space about five times larger and the Drosophila search space about six times larger (34). In addition, allowed mismatches in the consensus that retain DAF-16 binding could also be searched. However, because enhancer elements are highly enriched in the region proximal to the start codon, our 1-kb search is a reasonable first stage of the analysis.

    We have thus far expanded the informatic search to cover 1.5 kb of the worm promoter and 5 kb of the fly promoter, and this yielded 66 additional DAF-16 downstream gene candidates (table S1). Inspection of the molecular identity of the predicted candidates led us to focus on candidate C25E10.12, which encodes a serine/threonine phosphatase. The expression of C25E10.12 was up-regulated in the daf-2 mutant in a daf-16–dependent manner (Fig. 1). When C25E10.12 was RNAi-inactivated, it shortened the life-span of age-1(hx546) animals (Fig. 2D) but did not alter the life-span of wild-type animals (Fig. 2E), indicating that C25E10.12 RNAi specifically suppressed the life-span extension caused by reduced daf-2/insulin signaling.

    Continued characterization of DAF-16 targets conserved between disparate animal taxa will identify additional key mediators of the conserved longevity and metabolism functions of insulin signaling.

    Note added in proof: We searched C. elegans and Drosophila intergenic regions and detected 115 orthologous genes that each contain at least one DAF-16 site in the region between the start codon and the next gene upstream (table S3).

    Supporting Online Material

    www.sciencemag.org/cgi/content/full/1083614/DC1

    Materials and Methods

    Tables S1 to S3

    References and Notes

    View Abstract

    Navigate This Article