daf-16: An HNF-3/forkhead Family Member That Can Function to Double the Life-Span of Caenorhabditis elegans

See allHide authors and affiliations

Science  14 Nov 1997:
Vol. 278, Issue 5341, pp. 1319-1322
DOI: 10.1126/science.278.5341.1319


The wild-type Caenorhabditis elegans nematode ages rapidly, undergoing development, senescence, and death in less than 3 weeks. In contrast, mutants with reduced activity of the genedaf-2, a homolog of the insulin and insulin-like growth factor receptors, age more slowly than normal and live more than twice as long. These mutants are active and fully fertile and have normal metabolic rates. The life-span extension caused by daf-2mutations requires the activity of the gene daf-16. daf-16appears to play a unique role in life-span regulation and encodes a member of the hepatocyte nuclear factor 3 (HNF-3)/forkhead family of transcriptional regulators. In humans, insulin down-regulates the expression of certain genes by antagonizing the activity of HNF-3, raising the possibility that aspects of this regulatory system have been conserved.

The identification of genes that regulate aging (1) is an important breakthrough because it provides a means of investigating this fundamental but poorly understood process. The nematode Caenorhabditis elegans has a very rapid rate of aging (2), which is due in part to the activity of the gene daf-2. Mutations that reduce the activity of daf-2, a homolog of the insulin and insulin-like growth factor (IGF) receptors (3), can slow the rate of aging and more than double the life-span of the animal (4) without substantially affecting its activity or fertility (4-6). The life-span extension caused bydaf-2 mutations requires the gene daf-16(4).

In addition to regulating the rate of aging, daf-2 anddaf-16 also regulate the decision to enter diapause (the dauer phase) (7-10). When food is limited, young animals become dauers instead of developing to adulthood (9). The dauer is a resilient, long-lived, juvenile form that remains small and reproductively immature. Wild-typedaf-2 activity promotes growth to adulthood and prevents dauer formation. Unlike partial loss of daf-2 function, which specifically affects life-span, more severe loss ofdaf-2 function causes the animals to become dauers even in the presence of food (7-11). Thus,daf-2(+) has two functions: It promotes growth to adulthood, and it shortens the life-spans of adult animals (12). In addition to its role in life-span extension, the wild-typedaf-16 gene is also required for dauer formation in both wild-type and daf-2(−) animals (7-11). Thus, daf-16 also has two functions: Under dauer-inducing conditions, it promotes dauer formation, and, under conditions that do not induce dauer formation, it allows fertile adults carrying weak daf-2 mutations to remain active for a much longer period and to live twice as long as normal (4).

Because so few genes are known to regulate aging in any organism, we asked how many other genes were likely to have functions similar to that of daf-16. To do this, we carried out a genetic screen for additional daf-16–like mutants. daf-16mutations, as well as mutations in functionally related genes required for both dauer formation and life-span, can be isolated easily as suppressors of daf-2 mutations (7-9). We mutagenized daf-2(e1370) hermaphrodites with trimethylpsoralen followed by ultraviolet (UV) irradiation, which causes both point mutations and deletions (13), and screened their descendants for rare individuals that did not become dauers but instead grew to adulthood. We found 24 independent daf-2suppressors, all of which proved to be new alleles of daf-16(14). This result suggests that there are not a large number of genes like daf-16. Instead, any genes that act along with, or downstream of, daf-16 to initiate dauer formation are likely to affect only certain aspects of dauer formation or else are functionally redundant or essential (15). Thus,daf-16 provides a unique starting point for understanding how aging can be regulated.

We cloned daf-16 by transposon tagging (16,17). We isolated a daf-16::Tc1insertion mutation, mu147, by looking for daf-2suppressors in a daf-2; mut-6 strain, in which the transposon Tc1 is active (Fig. 1, A and B) (16). We then cloned the genomic DNA containing this Tc1 element and used sequences flanking Tc1 as a probe in Southern blot analysis of DNA isolated from our trimethylpsoralen-induceddaf-16 mutants. In five mutants, we found changes in the mobility of the restriction fragments that hybridized to the probe (Fig. 1C).

Figure 1

Identification of adaf-16::Tc1 insertion mutant. (A andB) Cosegregation of a Tc1-containing fragment with thedaf-16(mu147) mutation. In (A), Xba I–digested genomic DNA was probed with Tc1. The arrow points to the 6.1-kb fragment, which was present in 20 of 20 daf-16(−) recombinants (three are shown in lanes 1 to 3) but absent in 15 of 15daf-16(+) recombinants (five are shown in lanes 4 to 8) and also absent in the wild-type strain N2 (not shown) (16). In (B), the same filter was probed with genomic sequences flanking this Tc1 element. All daf-16(+) recombinants as well as N2 (not shown) contained a single 4.5-kb hybridizing fragment (arrowhead); alldaf-16(−) recombinants lacked this band and instead contained a band 1.6 kb larger (arrow; Tc1 is 1.6 kb). (C) Genomic DNA from the trimethylpsoralen-induced mutants was probed with the sequence flanking Tc1. Five mutants exhibited changes in the mobility of these restriction fragments. Arrowheads indicate mobility shifts; the lack of a hybridizing band in mu86 is indicated by an asterisk. WT, wild type.

The Tc1-tagged DNA was sequenced and found to be present on the cosmid R13H8, which was sequenced by the C. eleganssequencing project (18). We obtained the corresponding cDNA sequences by performing reverse transcription polymerase chain reaction (RT-PCR) and by analyzing clones from the C. elegansexpressed sequence tag (EST) database (19) (Fig.2). Both methods identified a cDNA species consisting of 10 exons and containing a single open reading frame predicted to encode a protein of 510 amino acids (Fig. 2, A and B).

Figure 2

daf-16 encodes an HNF-3–forkhead homolog. (A) The structure of thedaf-16 gene. The genomic DNA sequence spans at least 14 kb and contains 10 exons (boxes). The predicted coding regions are filled in black, and the regions homologous to the forkhead–winged helix domain are hatched. The changes in seven daf-16 mutations,m26, m27, mu147, mu100,mu92, mu86, and mu84, are shown [four additional mutations (mu85, mu87,mu89, and mu91) contained deletions that were detected by Southern blot analysis or PCR but were not sequenced].m26 harbors a single GC to AT transition at the 5′ splice junction (indicated by a vertical bar) of intron 2. The positions of the Xba I (X) fragment containing the Tc1 insertion inmu147 and those of the Eco R1 (E) and Hind III (H) restriction fragments that were altered in the daf-16mutants (Fig. 1C) are shown. The use of two alternative 3′ splice sites for intron 2 is indicated by solid and dashed lines. (B) The daf-16 cDNA and predicted amino acid sequence (GenBank accession number AF032112). The daf-16 message we analyzed was trans-spliced with an SL1 leader (33). Open triangles indicate intron positions inferred from the cDNA sequences. The six base pairs that are missing from one of the alternatively spliced forms are boxed; the predicted amino acid residue at position 121 is a Glu in the shorter (508 amino acids) form (not shown) instead of an Asp in the longer (510 amino acid) form. A polyadenylation [poly(A)] signal is found 11 base pairs upstream of the site of poly(A) in RT-PCR products containing the longest 3′ untranslated region (bold underline). The forkhead domain is indicated by brackets, and the position of the Tc1 insertion in mu147 is indicated by a vertical arrow. Primers used in RT-PCR are indicated by horizontal arrows over the sequences. The 3′ RACE primers (QT, QO, and QI) used for isolating 3′ cDNA ends are also shown (34). The m27 GC to AT amber mutation is indicated by an asterisk over the site. The m26 mRNA analysis is not shown, but RT-PCR detected two abnormally spliced mRNA species from this mutant. The major form contained unspliced intron 2 and therefore is 645 bp longer than the wild-type longer form; the minor form (∼10% of the product) lacked the last two nucleotides of exon 2 because of the use of a noncanonical 5′ splice site GA instead of GU. Both forms predicted early stop codons shortly after exon 2. A downstream in-frame ATG was found in both cases, however,with the potential to encode a 400–amino acid protein in the major form bearing an altered NH2-terminal region and the intact forkhead domain and COOH-terminal region as well as a 353–amino acid protein in the minor form with the NH2-terminal region (including part of the forkhead domain) truncated. No correctly spliced form was detected in the m26 mRNA. Also not shown is themu100 mRNA analysis,which resulted in an abnormally spliced mRNA bearing an early stop codon predicted to remove most of the forkhead domain and all COOH-terminal regions. Abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.

The sequence of this gene was found to be homologous to members of the HNF-3/forkhead family, a large class of transcription factors characterized by the presence of a forkhead domain, an ∼110–amino acid domain that forms a winged helix structure and mediates DNA binding (20). Members of this family have many different roles in embryogenesis, tumorigenesis, and differentiation and have been found to function at downstream positions in several types of signaling pathways (20), including insulin pathways (described below). The daf-16 forkhead domain was most similar to those of human FKHR and AFX proteins (67% and 64% identity, respectively), both of which were identified as human oncogenic fusion proteins (20) (Fig.3). As with other HNF-3/forkhead family members, little similarity was present outside of this region.

Figure 3

Alignment of the putative DAF-16 forkhead domain with forkhead domains from human FKHR (GenBank accession number U02310), human AFX (GenBank accession number X93996), murine HNF-3γ (GenBank accession number P35584), and Drosophilaforkhead (GenBank accession number A32380) proteins (20). Positions of the α helices (H1, H2, and H3), β strands (S1, S2, and S3), and wing (W1 and W2) domains as determined by x-ray crystallography of the HNF-3γ DNA binding domain (35) are indicated. Residues with greater than 50% identity and similarity are shaded in black and gray, respectively. Asterisks refer to residues that are identical in all five proteins; dots refer to residues that are similar in more than three of the proteins.

We verified that this gene was daf-16 by identifying sequence changes in the Tc1 insertion mutant, two previously identifieddaf-16 mutants, m26 and m27, and eight trimethylpsoralen-induced mutations that we analyzed (Fig. 2, A and B). Several mutations affected the conserved forkhead domain and were predicted to prevent DNA binding. The mu147 Tc1 insertion was located within the forkhead domain, m26 altered a splice donor sequence upstream of the forkhead domain, mu84 deleted part of the forkhead domain and some of the COOH-terminal region, andmu100 and mu92 both affected splice junctions of exon 3, which encodes part of the forkhead domain. In contrast, one mutation, m27, created a stop codon ∼100 amino acids COOH-terminal to the forkhead domain. Semiquantitative RT-PCR analysis indicated that this mutation did not affect daf-16 mRNA levels, implying that it may encode a truncated DAF-16 protein containing the DNA binding domain but lacking the COOH-terminal region (21). This finding suggests that sequences downstream of the forkhead domain may be required for DAF-16 function. The COOH-terminal region of the forkhead family member FAST-1 has been found to mediate protein-protein interactions (22); whether a similar function exists for this region of DAF-16 awaits further analysis.

The dauer-defective phenotype of daf-16 is semidominant (8); therefore, it was particularly important to confirm that the Daf-16 mutant phenotype resulted from reduced rather than altered or novel gene activity. Our findings indicated that this is the case, because most of these mutations would be predicted to reduce or eliminate daf-16 activity. To determine whether any of these mutations were null alleles, we analyzed mRNA expression in the mutants. For the majority of alleles, it was not possible to completely rule out the possibility that residual daf-16 function might still exist (Fig. 2B). However, one mutation, mu86, was likely a null allele. This mutation was a large deletion that removed most of the coding sequence, including all of the forkhead domain. This mutant, like all other known daf-16 alleles, grew to become an active, fertile adult. This finding suggests that daf-16functions primarily to regulate life-span and dauer formation and does not have essential activities.

The finding that daf-16 encodes an HNF-3/forkhead family member is important because it implies that mutations in thedaf-2 insulin/IGF receptor homolog exert their effects not simply by changing the activities of preexisting enzymes (23) but instead by initiating a new genetic regulatory program that extends youthfulness and postpones death. There are several intriguing parallels between this C. elegans pathway and the human insulin/IGF pathways (3). In humans, insulin and IFG regulate food utilization pathways and promote growth (23). Similarly, daf-2 activity promotes growth to adulthood when food is abundant; conversely, lack ofdaf-2 activity maintains the dauer state, in which the animals use stored food sources. In animals that lack daf-16activity, daf-2 mutations have little or no effect, which raises the possibility that the primary role of daf-2 is to prevent daf-16 function (7-9). In humans, insulin appears to mediate some of its effects by blocking the activity of HNF-3 (24). Of four insulin-repressed genes that have been studied extensively, three, phosphoenolpyruvate carboxykinase, tyrosine amino transferase, and IGF binding protein–1, appear to be up-regulated by HNF-3 in the absence of insulin. Each gene contains a similar insulin-response sequence (IRS) that can act as a binding site for HNF-3 and that is required for repression by insulin. It has been proposed that insulin signaling acts by preventing HNF-3 from binding to the IRS (24, 25).

So far, little is known about how daf-2 might affectdaf-16 activity in C. elegans. daf-2 may exert its effects by activating the phosphatidylinositol 3-kinaseage-1, because mutations in this gene also extend life-span in a daf-16–dependent fashion (5,26-30).

In vertebrates, caloric restriction, which affects insulin levels, has an effect analogous to that of weak daf-2 mutations: It also extends life-span without decreasing the rate of metabolism (31). Thus, life-span in both C. elegans and vertebrates may be regulated by an evolutionarily conserved mechanism involving a forkhead homolog that promotes longevity when food is scarce and an insulin family member that counteracts it. In addition, it is possible that the different aging rates of different individuals within a species, as well as the markedly different aging rates exhibited by members of different species, are due in part to intrinsic differences in the resting levels of this signaling pathway.

Note added in proof: Working independently, Ogg et al. have also cloned and molecularly analyzed daf-16(36).

  • * Present address: Molecular and Cellular Biology Program, University of Washington, Seattle, WA 98195– 7275, USA.


View Abstract

Navigate This Article