daf-2, an Insulin Receptor-Like Gene That Regulates Longevity and Diapause in Caenorhabditis elegans

See allHide authors and affiliations

Science  15 Aug 1997:
Vol. 277, Issue 5328, pp. 942-946
DOI: 10.1126/science.277.5328.942


A C. elegans neurosecretory signaling system regulates whether animals enter the reproductive life cycle or arrest development at the long-lived dauer diapause stage. daf-2, a key gene in the genetic pathway that mediates this endocrine signaling, encodes an insulin receptor family member. Decreases in DAF-2 signaling induce metabolic and developmental changes, as in mammalian metabolic control by the insulin receptor. Decreased DAF-2 signaling also causes an increase in life-span. Life-span regulation by insulin-like metabolic control is analogous to mammalian longevity enhancement induced by caloric restriction, suggesting a general link between metabolism, diapause, and longevity.

A reversible arrest ofC. elegans development at the metabolically less active dauer stage is triggered by a dauer-inducing pheromone (1). This pheromone is detected by sensory neurons, which then signal by a complex pathway to target tissues, such as the germ line, pharynx, intestine, and ectoderm, that are remodeled and metabolically shifted (1). Genetic analysis of daf mutants, which arrest at the dauer stage or enter the reproductive life cycle independent of pheromone regulation, has revealed that parallel genetic pathways regulate distinct aspects of the dauer metamorphosis (2). The pathway that includes daf-2 controls both reproductive development and normal senescence: daf-2mutant animals arrest development at the dauer larval stage and exhibit a marked increase in longevity (1-3).

Genetic mapping, using both visible genetic markers and restriction fragment length polymorphism (RFLP) markers, placed daf-2between mgP34 and mgP44 (Fig.1). Yeast artificial chromosome (YAC) Y53G8 carries the genomic region that includes mgP34and mgP44, which flank daf-2 (Fig. 1). Analysis of 570 random Y53G8 M13 subclone DNA sequences from the Genome Sequencing Center revealed that four sequences (00667, 00622, 00318, and 00706) were homologous to regions of the mammalian insulin receptor family. The detection of multiple daf-2mutations in the gene (see below), and the coincidence of the genetic location of this insulin receptor homolog with daf-2(Fig. 1) establish that this insulin receptor homolog corresponds todaf-2.

Figure 1

Genetic and physical map of daf-2. daf-2maps between mgP34 and mgP44 in a region not covered by cosmid clones but covered by YAC Y53G8. Cosmids from the region detect RFLPs between C. elegans strains Bristol N2 and Bergerac RC301. Using 33 recombinants in the dpy-1 daf-2 interval and 24 recombinants in the daf-2 ben-1interval, we were able to map daf-2 0.7 map units (m.u.) to the right of mgP34 and 0.7 map units to the left ofmgP44. Y53G8 YAC DNA hybridizes to multiple restriction fragments from cosmids bearing mgP34 and mgP44. A probe from the insulin receptor homolog on Y53G8 detects themgP48 RFLP. daf-2 could not be separated from this insulin receptor gene in 47 recombination events in a 13–map unit interval. The initial insulin receptor homolog shotgun sequences (00667, 00622, 00318, and 00706) are shown as thin bars above a box corresponding to daf-2 cDNA. A probable full-lengthdaf-2 cDNA bears an SL1 transspliced leader (25), a 120-base 5′ untranslated region (UTR), a 5538-base open reading frame, and a 159-base 3′ UTR. SP and TM denote putative signal peptide and transmembrane region, respectively. Intron positions (triangles) and lengths were detected by genomic and cDNA sequence and PCR amplification comparisons. The minimum daf-2 gene size, based on this analysis, is 33 kb.

The daf-2 gene structure and DNA sequence were determined with the use of polymerase chain reaction (PCR) primers derived fromdaf-2 genomic subclone sequences to amplify daf-2genomic and cDNA regions (Fig. 1, GenBank accession number AF012437). The predicted DAF-2 protein is 35% identical to the human insulin receptor, 34% identical to the human insulin-like growth factor–I (IGF-I) receptor, and 33% identical to the human insulin receptor–related receptor (4). DAF-2 is the only member of the insulin receptor family in the 90-Mb C. elegans genome sequence (∼90% complete) or in the 10-Mb C. elegansexpressed sequence tag database. Because it is equally distant from the human insulin, IGF-I, and insulin receptor–related receptors, DAF-2 is probably the homolog of the ancestor of these duplicated and diverged receptors, and thus may subserve any or all of their functions. Like these receptors, DAF-2 has a putative signal peptide, a cysteine-rich region in the putative ligand-binding domain, a probable proteolysis site, a transmembrane domain, and a tyrosine kinase domain (Fig. 1); in addition, DAF-2 has a COOH-terminal region that may serve a function similar to that of mammalian insulin receptor substrate–1 (IRS-1) (Fig. 2) (5).

Figure 2

Comparison of DAF-2 ligand-binding and kinase domains to insulin receptor family members. The amino acid comparison of DAF-2 to the human insulin receptor (InR) and IGF-I receptor (IGF-IR) and to the Drosophila insulin receptor homolog (DInR) shown. The mutations found in human patients are indicated at the top of the rows, and daf-2 allele substitutions are indicated below with allele names. Sixdaf-2 mutations map in the ligand-binding domain:sa187 (C469S, TGT to AGT), e1368(S573L, TCA to TTA), e1365 (A580T, GCT to ACT), sa229 (D648N, GAT to AAT), and two mutations in mg43 (C401Y, TGT to TAT; P470L, CCC to CTC). Three daf-2 mutations substitute conserved amino acid residues in the insulin receptor kinase domain: sa219(D1374N, GAT to AAT), e1391(P1434L, CCC to CTC), and e1370 (P1465S, CCA to TCA). Residues with black and shaded backgrounds indicate amino acid identity and similarity, respectively. The percentages under each domain represent the percentage identity between DAF-2 and each receptor. Conserved cysteine residues in the ligand-binding domain (top) are marked with dots. In the kinase domain, active-site residues that mediate insulin receptor kinase specificity are marked with stars. All of these residues are homologous in DAF-2. 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.

In the ∼500–amino acid ligand-binding domain, DAF-2 is 36% identical to the insulin receptor and 35% identical to the IGF-I receptor. Of 23 phylogenetically conserved cysteine residues in this domain, 21 are conserved in DAF-2 (Fig. 2). The DAF-2 Cys-rich region is 34% identical to the insulin receptor and 28% identical to the IGF-I receptor. Six daf-2 mutations map in this domain (Fig.2 and Table 1). The mg43 andsa187 mutations substitute conserved residues in the Cys-rich region (Fig. 2). Other substitutions at nonconserved residues cause less severe phenotypes (Table 1).

Table 1

Percentage dauer formation of daf-2 alleles (N, number of animals). Eggs from animals grown at 15°C (day 0) were incubated at 15°, 20°, or 25°C. Numbers of wild-type animals and dauers were counted on day 3 (20° and 25°C) or day 5 (15°C). The dauers marked with asterisks recovered by day 4 (sa229 at 25°C) or by day 8.sa187 and sa229 were also studied (27). Amino acid abbreviations are as in Fig. 2. ND, not determined.

View this table:

The 275–amino acid DAF-2 tyrosine kinase domain is 70% similar and 50% identical to the human insulin receptor kinase domain. Eight amino acid residues mediate target site specificity of the insulin receptor kinase (6); in DAF-2, six of these residues are invariant and two are replaced with similar amino acids (Fig. 2), which suggests that DAF-2 phosphorylates similar target tyrosine motifs to the insulin receptor kinase. Multiple DAF-2 tyrosine residues in conserved sequence contexts and analogous positions are likely autophosphorylation targets (7) (Table2). Three daf-2 missense mutations substitute conserved amino acid residues in the kinase domain (Fig. 2 and Table 1). All three mutations cause moderate to strong dauer-constitutive phenotype, but none are as strong as the nonconditional alleles, for example, mg43 (Table 1). Remarkably, a human diabetic insulin-resistant patient bears the same amino acid substitution (Pro1178 → Leu) as daf-2 (e1391) (Fig. 2) (8). This patient was reported to be heterozygous for this substitution. daf-2(e1391) is not dominant (9). It is possible that the patient was a compound heterozygote or carries mutations in other genes (for example, other complex type II diabetes loci) that enhance the dominance of Pro1178 → Leu (10).

Table 2

Sequence motifs in the DAF-2 amino acid sequence. Amino acid abbreviations are as in Fig. 2. PTB, phosphotyrosine-binding domain.

View this table:

Like the Drosophila insulin receptor homolog (11), DAF-2 has a long COOH-terminal extension that may function analogously to mammalian IRS-1, which is phosphorylated by the insulin receptor to recruit signaling proteins bearing Src homology 2 (SH2) domains (5). Many of the DAF-2 COOH-terminal extension tyrosines bear flanking sequence motifs (7) that suggest they are autophosphorylated (Table 2). On the basis of precedents from IRS-1 interactions with mammalian phosphatidylinositol (PI) 3-kinases (5), a Tyr-X-X-Met motif at DAF-2 Tyr1626 is likely to mediate interaction with the AGE-1 PI 3-kinase, which acts in the same genetic pathway as daf-2 to signal reproductive development and normal senescence (Fig.3) (2, 3, 5, 12, 13). The insulin receptor also couples to other signaling pathways (5); analogous DAF-2 phosphotyrosine residues may mediate these interactions (Table 2) (13).

Figure 3

A model of insulin-like signaling in the C. elegans dauer formation pathway. In the absence of dauer pheromone, an insulin-like ligand activates DAF-2, and a DAF-7 TGF-β–like signal activates the DAF-1 and DAF-4 receptors. Activated DAF-2 recruits AGE-1, which produces the second messenger PIP3. This second messenger may regulate glucose transport (5), metabolic kinase cascades that include AKT and GSK-3 (16), and transcription and translation of metabolic genes (5). DAF-16 acts downstream of DAF-2 and AGE-1 in this pathway and is negatively regulated by them (2).

Insulin and its receptor families play key roles in vertebrate metabolic and growth control (5). Upon stimulation of pancreatic insulin release by increasing blood glucose and autonomic inputs, insulin receptor engagement causes shifts in the activities of key metabolic enzymes, as well as changes in the transcription and translation of metabolic regulators in fat, liver, and muscle cells, all of which lead to assimilation of glucose into glycogen and fat and inhibition of glucose synthesis (5).

Diapause arrest in general and dauer arrest in particular are also associated with major metabolic changes (14), consistent with a model in which daf-2 acts in a metabolic regulatory pathway related to insulin signaling. In wild-type animals, DAF-2 signaling allows nondauer reproductive growth, which is associated with the use of food (for growth in cell number and size) and with small stores of fat (Fig. 4). Indaf-2 mutant animals, metabolism is shifted to the production of fat (Fig. 4) and glycogen (9) in intestinal and hypodermal cells. Even when a temperature-sensitivedaf-2 mutant allele is shifted to the nonpermissive temperature at the L4 or adult stage (after the critical period fordaf-2 control of dauer formation) (1), metabolism is shifted toward storage of fat (Fig. 4). Similar metabolic shifts are seen in wild-type pheromone-induced dauers, age-1, anddaf-7 mutants (Fig. 4) (9). In support of this metabolic shift, in dauer larvae, enzymes that regulate glycolysis are down-regulated while those that regulate glycogen and fat synthesis are up-regulated, and there is ultrastructural evidence for increased lipid and glycogen (14).

Figure 4

Metabolic control bydaf-2 and daf-7. (A) Low accumulation of fat in a wild-type L3 animal grown at 25°C stained with sudan black (26). (B) Much more extensive fat accumulation in daf-2(e1370) grown at the nonpermissive temperature of 25°C. These animals accumulate fat in both intestinal and hypodermal cells. daf-2(e1370)animals grown at 15°C, the permissive temperature, accumulate small amounts of fat, like the wild type (9). (C) Extensive fat accumulation in the intestine and hypodermis of daf-7(e1372) animals grown at 25°C. (D) Extensive fat accumulation indaf-2(e1370) animals grown at the permissive temperature until the L4 stage and then shifted to the nonpermissive temperature. This shows that daf-2 regulates metabolism without entry into the dauer stage.

Even though DAF-2 and the mammalian insulin receptor both regulate metabolism, the metabolic defects associated with mutations in these receptors appear to be different. Complete loss of mammalian insulin receptor activity causes growth arrest at birth (leprechaunism in humans) and a metabolic shift to runaway lipolysis and ketoacidosis (5), rather than the fat accumulation we observe indaf-2 mutants (Fig. 4). This distinction between insulin receptor and daf-2 mutants may reflect distinct metabolic responses to this signaling, or a difference between complete loss and declines in insulin signaling. Because none of the daf-2mutations we have so far identified are clear null mutations, it is possible that daf-2 dauer-constitutive alleles are more analogous to non-null human insulin receptor mutations (5). In fact, the 14-year-old patient with the same insulin receptor mutation as daf-2(e1391) was morbidly obese (8), suggesting that the metabolic effects of decreased insulin signaling may be similar to those of daf-2 mutants.

The structural and functional homology of DAF-2 to the mammalian insulin receptor suggests that components of the insulin signaling pathway may act in metabolic and diapause control in C. elegans. There is precedent for invertebrate metabolic and growth control by insulin superfamily signaling (15). We hypothesize that an insulin-like signal is up-regulated during C. elegans reproductive development and stimulates DAF-2 receptor autophosphorylation and recruitment of the AGE-1 PI 3-kinase to produce the second messenger phosphatidylinositol-3,4,5-trisphosphate (PIP3). AGE-1 is likely to be a major signaling output of DAF-2 because of the similarity of the age-1 and daf-2 mutant phenotypes and because of their similar placement in the epistasis pathway (2, 12) (Fig. 3). Precedent from insulin signaling suggests that PIP3 activates membrane fusion of vesicles bearing sugar transporters, an AKT–GSK-3 kinase cascade (16) to regulate the activities of glycogen and fat synthetic and lytic enzymes, and the transcription and translation of metabolic genes such as those that encode PEPCK and lipases (5).

Conversely, C. elegans daf genes may identify components of the mammalian insulin signaling pathway or converging pathways. The major target of DAF-2–AGE-1 signaling in C. elegans is DAF-16 (2). Mammalian homologs of DAF-16 may represent the output of insulin signaling. The dauer genetic pathway also suggests that DAF-7 transforming growth factor–β (TGF-β) neuroendocrine signals act in addition to DAF-2 insulin-like regulation of metabolism (Fig. 4) (17). Mammalian homologs of DAF-7 and its downstream signaling cascade (18) may represent signals that converge with insulin signals. Human homologs of the dafgenes may be defective in diabetic pedigrees or inactivated in obesity-onset diabetes (19).

The DAF-2 receptor may act in the hypodermal and intestinal target tissues, where a change in metabolism is triggered by the dauer regulatory cascade (Fig. 4). It is also possible that DAF-2 regulates the metabolism and remodeling of tissues indirectly, for example, by controlling the production of other hormones (20). Expression and genetic mosaic analysis of daf-2 is essential to distinguish these models.

In addition to these metabolic changes, the DAF-2 signaling cascade also controls the reproductive maturation of the germ line as well as morphogenetic aspects of the pharynx and hypodermis (1). This growth control by daf-2 may be homologous to that mediated by the mammalian IGF-I receptor, which acts downstream of growth hormone (21). Like daf-2 mutants, life-span is markedly increased in dwarf mice with defects in growth hormone signaling and, presumably, decreased IGF-I signaling (21).

Weak daf-2 and age-1 mutants that do not arrest at the dauer stage nevertheless live much longer than the wild type (3, 12). This connection between longevity and diapause control may not be unique to C. elegans. Diapause arrest is an essential feature of many vertebrate and invertebrate life cycles, especially in regions with seasonal temperature and humidity extremes (14). Animals in diapause slow their metabolism and rates of aging and can survive for periods much longer than their reproductive life-span (14). Thus, partial induction of diapause by environmental conditions or genetic predispositions may underlie variations in life-span within and between species.

The increase in longevity associated with decreased DAF-2 signaling is analogous to mammalian longevity increases associated with caloric restriction (22). It is possible that caloric restriction causes a decline in insulin signaling to induce a partial diapause state, like that induced in weak daf-2 and age-1mutants. The induction of diapause-like states, or changes in the mode and tempo of metabolism itself, may affect postreproductive longevity (22), as in C. elegans. This association of metabolic rate with longevity is also consistent with the correlation of free radical generation with aging (22).

Just as environmental extremes can select for variation in the genetic pathways that regulate C. elegans dauer formation, famines and droughts in human history may have selected for analogous variants in the human homologs of the daf genes. In fact, heterozygous mice carrying either the db or obrecessive diabetes genes survive fasting about 20% longer than do wild-type controls (23). The high frequency of type II diabetes in many human populations (19) may be the legacy of such selections.

  • * These authors contributed equally to this report.

  • To whom correspondence should be addressed. E-mail: ruvkun{at}


View Abstract

Navigate This Article