Regulation of Aging and Age-Related Disease by DAF-16 and Heat-Shock Factor

See allHide authors and affiliations

Science  16 May 2003:
Vol. 300, Issue 5622, pp. 1142-1145
DOI: 10.1126/science.1083701

This article has a correction. Please see:


The Caenorhabditis elegans transcription factor HSF-1, which regulates the heat-shock response, also influences aging. Reducing hsf-1 activity accelerates tissue aging and shortens life-span, and we show that hsf-1 overexpression extends lifespan. We find that HSF-1, like the transcription factor DAF-16, is required for daf-2–insulin/IGF-1 receptor mutations to extend life-span. Our findings suggest this is because HSF-1 and DAF-16 together activate expression of specific genes, including genes encoding small heat-shock proteins, which in turn promote longevity. The small heat-shock proteins also delay the onset of polyglutamine-expansion protein aggregation, suggesting that these proteins couple the normal aging process to this type of age-related disease.

Heat-shock factor activates transcription of heat-shock genes, which encode chaperones and proteases, in response to heat and other forms of stress. Previous studies have implicated heat-shock proteins (HSPs) in aging. For example, mild heat stress can cause a period of decreased mortality rate in Drosophila, and hsp70 has been implicated in this effect (1). In addition, expression of genes encoding small heat-shock proteins (sHSPs) is increased in Drosophila lines selected for increased life-span (2), and overexpression of hsp70F increases the life-span of C. elegans (3).

Previously, we showed that reducing the activity of C. elegans HSF-1 causes a rapid-aging phenotype and shortens life-span (4). Conversely, we found that animals carrying additional hsf-1 gene copies (5), which were resistant to heat and oxidative stress (fig. S1, A and B), lived approximately 40% longer than normal (Fig. 1A). Thus HSF-1 activity promotes longevity.

Fig. 1.

HSF-1 promotes longevity. (A) Blue, survival of wild-type animals grown on control bacteria containing vector alone; green, animals grown on bacteria expressing hsf-1 dsRNA; red, animals carrying additional copies of hsf-1 (CF1824). Two additional HSF-1 overexpressing lines were obtained and were found to extend life-span (table S1). (B) hsf-1 overexpression extends life-span in a daf-16–dependent manner. Adultlife-spans of wild-type and hsf-1–overexpressing (CF1824) animals grown on daf-16 RNAi bacteria (green and blue lines). In addition, hsf-1 overexpression extends the life-span of animals treated with daf-2 RNAi (red and orange lines). (C and D) RNAi of hsf-1 completely prevents the daf-2(e1370) mutation, but not the isp-1(qm150) or eat-2(ad1116) mutations, from extending life-span. Animals were grown on hsf-1 RNAi bacteria from the time of hatching. Adult life-spans of wild-type (N2) (blue), daf-2(e1370) (red), eat-2(ad1116) (orange), and isp-1(qm150) (green) animals grown on (C) control bacteria or (D) hsf-1 RNAi bacteria. All experiments have been repeated more than once with similar effects. For statistical data, see table S1.

The FOXO transcription factor DAF-16, which functions in the C. elegans insulin/IGF-1 signaling pathway (6, 7), also promotes longevity (810). As with hsf-1, inhibiting daf-16 activity shortens life-span, and elevating daf-16 activity increases life-span.

We found that the longevity phenotypes produced by hsf-1(RNAi) were similar to those produced by daf-16(–) mutations. DAF-16 is required for daf-2–insulin/IGF-1 receptor mutations to increase life-span; in the short-lived daf-16 mutants, daf-2 mutations are unable to increase life-span (6, 7, 11). HSF-1 appeared to be required as well, because neither of two daf-2 mutations we examined, e1370 and mu150, was able to extend the short life-spans of hsf-1 RNAi-treated animals (Fig. 1, C and D; table S1). In contrast, the life-spans of both daf-16(–) mutants (12, 13) and hsf-1 RNAi-treated animals (Fig. 1, C and D) can be extended by eat-2(ad1116) mutations [which inhibit pharyngeal pumping and may cause caloric restriction (12)] and by isp-1(qm150) mutations [which inhibit mitochondrial respiration (13)]. [Control experiments indicated that hsf-1 RNAi sharply decreased hsf-1 mRNA levels in all of these strains (fig. S1C).] Together these similarities suggested the hypothesis that HSF-1, like DAF-16, might function in the insulin/IGF-1 system. We also found that daf-16 was required for hsf-1 overexpression to extend lifespan (Fig. 1B). This, too, suggested that DAF-16 and HSF-1 might act together to promote longevity.

To investigate this hypothesis, we first asked whether DAF-16's activities required HSF-1. We found that hsf-1(RNAi) did not prevent DAF-16 from accumulating in the nuclei of daf-2 mutants (911) (fig. S2) or activating expression of two known downstream genes, the metallothionein gene mtl-1 (14) and the super-oxide dismutase gene sod-3 (15) (Fig. 2A). This indicates that DAF-16 can function independently of HSF-1. In addition, it implies that, without adequate HSF-1 activity, the increased expression of sod-3 and mtl-1 activity are not sufficient to increase the life-span of daf-2 mutants [though these and other hsf-1–independent genes could certainly contribute to longevity (16)]. Next, we asked whether DAF-16 was required for HSF-1 to activate gene expression after heat-shock. This seemed plausible, because daf-2 mutants are thermotolerant (17, 18), and we found that their thermotolerance required daf-16 (table S2). However, the heat-inducibility of the genes aip-1, unc-33, and the HSP-70 homolog F44E5.4 (19) was not diminished by daf-16(–) mutations (Fig. 2B). Thus, HSF-1 can function independently of DAF-16.

Fig. 2.

HSF-1 is required for the expression of a subset of DAF-16 targets in daf-2 mutants. Likewise, DAF-16 is required for the expression of a subset of heat-shock response genes after heat shock. (A) RT-PCR analysis of hsf-1, mtl-1, sod-3, and four shsps in wild-type (N2) (lanes 1 to 3 and 10 to 12), daf-16(mu86) (lanes 4 to 6) and daf-2(e1370) (lanes 7 to 9 and 13 to 15) animals grown on either control or hsf-1 RNAi bacteria. Shown are RT-PCR products from serial dilutions of total RNA isolated from the animals after RNAi treatment. act-1 (β-actin) served as an internal control. (B) RT-PCR analysis of hsf-1, hsp-70, aip-1, unc-33, and four shsps before (lanes 1 to 3 and 7 to 9) or after (lanes 4 to 6, 10 to 12, and 13 to 15) heat shock. RNA was harvested on day 1 of adulthood from wild-type (N2) animals grown on control bacteria (lanes 1 to 6) or bacteria expressing hsf-1 dsRNA (lanes 13 to 15) as well as daf-16(mu86) animals grown on control bacteria (lanes 7 to 12). The heat-shock treatment was carried out less than 5 min before the RNA was harvested. Only slight variation was observed among replicates.

The finding that neither DAF-16 nor HSF-1 was absolutely required for the other's activity suggested that the two proteins might function together to turn on a specific subset of genes. Using DNA microarrays (16) we asked whether any known heat-shock genes (19) were regulated by the DAF-2 pathway. Others have found (20), and we confirmed (16), that expression of hsp-70 and hsp-90 did not change in daf-2(–) adults. In addition, hsf-1 mRNA levels were not affected by daf-2 mutations (16) (Fig. 2A). However, the expression of several heat-inducible genes, including four small heat-shock protein (shsp) genes, hsp-16.1, hsp-16.49, hsp-12.6, and sip-1, was sharply increased in animals with reduced daf-2 activity and decreased in animals with reduced daf-16 activity (16).

We investigated the shsp genes in more detail. First, we confirmed the microarray data using reverse transcription polymerase chain reaction (RT-PCR) (Fig. 2A). We then asked whether HSF-1 was required for increased shsp gene expression in daf-2 mutants. We found that it was (Fig. 2A). Thus, HSF-1 functions in the insulin/IGF-1 system. In addition, DAF-16 was required along with HSF-1 to activate shsp expression after heat shock (Fig. 2B). Thus, DAF-16 functions in the heat-shock response.

A simple model to explain these findings is that in daf-2 mutants and in normal animals subjected to heat shock, DAF-16 and HSF-1 both bind directly to regulatory sequences in the shsp (and possibly other) genes (Fig. 3A). We favor this model because we found sequences identical to consensus DAF-16 (16, 21) and HSF-1 (19) binding sites upstream of all of these shsp genes (Fig. 3B). Heat shock triggers DAF-16 nuclear localization (9, 11), suggesting that it may increase the ability of DAF-16 to activate shsp expression. Thus it is possible that DAF-16 functions as a heat-shock factor to regulate part of the heat-shock response. Likewise, DAF-2 pathway mutations could potentially increase the ability of HSF-1 to activate shsp expression (Fig. 3A).

Fig. 3.

(A) Model. In DAF-2 pathway mutants and in normal animals subjected to heatshock, DAF-16 and HSF-1 bind the regulatory regions of common target genes, including shsp (and possibly other) genes and activate their expression. It is possible that DAF-2 pathway mutations or heat shock somehow increase the ability of HSF-1 or DAF-16 (respectively) to activate, in a specific fashion, the common target genes (dashed lines), although this need not be the case. Only a subsetof DAF-16's or HSF-1's downstream targets require both proteins for their expression. These common target genes are required for the increased longevity of daf-2 mutants and hsf-1–overexpressing animals. Other genes (such as sod-3 or mtl-1) may also contribute to longevity; however, our findings suggest that without sufficient hsf-1 activity, their increased expression is not sufficient to extend life-span. (B) Potential DAF-16 (blue) and HSF-1 (red) binding sites, GTAAAc/tA (16, 21) and T TCTa/cGAA (19), respectively, located upstream of the shsp genes.

Because HSF-1 and DAF-16 are both required for the longevity of daf-2 mutants, we next asked whether their common targets, the shsp genes, influenced life-span. RNAi of each gene shortened the life-span of daf-2(e1370) mutants by approximately 25% (Fig. 4B) and that of daf-2(mu150) mutants by a similar extent (16). shsp RNAi also decreased the longevity of animals overexpressing HSF-1 (Fig. 4C). In addition, like hsf-1 RNAi, shsp RNAi decreased the life-span of wild type, though to a lesser extent than for daf-2 mutants (Fig. 4A). Together these findings suggest that DAF-16 and HSF-1 increase longevity, at least in part, by increasing shsp expression. Moreover, because at least some of these sHSPs are likely to be functionally redundant, together they may make a substantial contribution to longevity.

Fig. 4.

sHSP activity extends life-span and delays polyglutamine protein aggregation Adultlife-spans of (A) wild type, (B) daf-2(e1370) mutants, and (C) hsf-1 overexpressing animals (CF1824) grown on control bacteria (green) or bacteria expressing dsRNA of hsp-16.1 (violet), hsp-16.49 (lightblue), hsp-12.6 (yellow), sip-1 (orange), or hsf-1 (red). Dark blue, survival of wild-type (N2) animals grown on control bacteria containing the vector alone. (D) Fluoresence micrographs of late L4/young adult transgenic animals expressing a fusion of polyQ (40 repeats) to yellow fluorescence protein (Q40-YFP) in muscle cells (26). Animals were classified into separate groups according to the numbers of aggregates they contained. (Top) Animals with less than 20 aggregates. (Bottom) Animals with more than 60 aggregates. (E) Quantitation of number of aggregates for animals grown on control bacteria or bacteria expressing daf-16, hsf-1, four shsps, hsp-70, sod-3, ctl-1 (catalase), and mtl-1 dsRNA. Q40-YFP expressing animals grown on RNAi bacteria from the time of hatching were examined as 1-day-old young adults. Shown in the graph are the percentages of the animals containing less than 20 aggregates. daf-16, hsf-1, and four shsp RNAi all significantly accelerated the onset of polyglutamine aggregates, whereas the other stress-response genes we tested did not. [Unexpectedly, RNAi of two other hsp-70 homologs we tested, hsp-1 and hsp-70F, delayed aggregate formation (table S3). This may be due to reduced negative feedback regulation of HSF-1 by these HSP-70s (29, 30).] At least 200 animals were examined in each experiment. Asterisk, P < 0.0001; pound sign, P < 0.01 (Chi-square test). Additional data and statistics are shown in table S3.

The failure of previous workers to observe increased shsp-gfp expression in DAF-2 pathway mutants (22) was probably due to low GFP expression levels. However, this group found that when such mutants are subjected to transient heat shock, both shsp expression later in life and life-span increase substantially (22). This has suggested the hypothesis that shsp expression extends life-span (22). Curiously, treating wild-type animals in the same way does not increase shsp expression later in life, and produces only small increases in life-span (22). One possible explanation is that DAF-16 activity, which is required for shsp expression, is elevated in daf-2 mutants. Heat-shocking these animals provides additional HSF-1 activity, which then further increases shsp expression and life-span. Consistent with this, we found that overexpression of HSF-1 further increased the life-spans of daf-2(–) animals (Fig. 1B).

How might sHSPs protect cells against aging? The sHSPs are known to form large oligomers that bind to unfolded proteins and prevent them from aggregating (2325). In aging animals, this activity may prevent oxidized or otherwise damaged proteins from aggregating before they can be refolded or degraded (although sHSPs could conceivably influence longevity in a different way).

A fundamental mystery in biology is how the normal aging process is coupled to the diseases of aging. At least part of the answer appears to involve the insulin/IGF-1 pathway. For example, Huntington's-like polyglutamine-repeat proteins expressed in C. elegans form aggregates as the animals age, and this aggregation is delayed in long-lived insulin/IGF-1 pathway mutants (26). Because sHSPs are known to inhibit protein aggregation (27), we asked whether shsp RNAi might accelerate the onset of polyglutamine aggregates in C. elegans. We found that it did (Fig. 4, D and E), whereas RNAi of the other stress-response genes we tested did not (Fig. 4). As predicted, daf-16 or hsf-1 RNAi accelerated aggregation formation to an even greater extent (Fig. 4E) (28). Thus, possibly by functioning as molecular chaperones, sHSPs may influence the rates of aging and polyglutamine aggregation coordinately. In this model, mutations in the DAF-2 pathway delay both aging and susceptibility to this agerelated disease, at least in part, by increasing shsp gene expression.

Supporting Online Material

Materials and Methods

Figs. S1 to S3

Tables S1 to S3


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article