Aging Cell (2003) 2, pp111–121

Transcriptional outputs of the Caenorhabditis elegans forkhead protein DAF-16

Blackwell Publishing Ltd.

Joshua McElwee,1* Kerry Bubb2 and James H. Thomas1,2 1

Molecular and Cellular Biology Program of the University of Washington and Fred Hutchinson Cancer Research Center, and 2 Department of Genome Sciences, University of Washington, Seattle WA 98195, USA

Summary In Caenorhabditis elegans , the forkhead protein DAF-16 transduces insulin-like signals that regulate larval development and adult lifespan. To identify DAF-16dependent transcriptional alterations that occur in a long-lived C. elegans strain, we used cDNA microarrays and genomic analysis to identify putative direct and indirect DAF-16 transcriptional target genes. Our analysis suggests that DAF-16 action regulates a wide range of physiological responses by altering the expression of genes involved in metabolism, energy generation and cellular stress responses. Furthermore, we observed a large overlap between DAF-16-dependent transcription and genes normally expressed in the long-lived dauer larval stage. Finally, we examined the in vivo role of 35 of these target genes by RNA-mediated interference and identified one gene encoding a putative protease that is necessary for the daf-2 Age phenotype. Key words: C. elegans; DAF-16; dauer; forkhead; insulin; microarray.

Introduction In C. elegans, lifespan is a dynamically controlled process involving the integration of diverse environmental and humoral stimuli. One of the principal components regulating lifespan in C. elegans is a conserved insulin-like signalling pathway (Hekimi et al., 2001). The insulin/insulin-like growth factor receptor homolog DAF-2 signals through a conserved PI3-kinase/Akt pathway, ultimately acting to inhibit the activity of the winged-helix transcription factor DAF-16 (Dorman et al., 1995; Lee et al., 2001). Many mutations within genes in the insulin-like signalling group have been shown to significantly extend both

Correspondence James H. Thomas, Department of Genome Sciences, University of Washington, Box 357730, Seattle, WA 98195, USA. Tel.: (206) 543 7877; fax: (206) 543 0754; e-mail: [email protected] * Present address: Department of Biology, Darwin building, University College London, Gower Street, WC1E 6BT, UK. Accepted for publication 7 January 2003 © Blackwell Publishing Ltd/Anatomical Society of Great Britain and Ireland 2003

mean and maximum lifespan (Dorman et al., 1995; Ailion et al., 1999; Paradis et al., 1999). This lifespan extension, the Age phenotype, is dependent on wild-type function of DAF-16 (Lin et al., 1997; Ogg et al., 1997). The C. elegans DAF-2 pathway also regulates entry into the dauer larval stage, a developmental diapause state specialized for long-term survival and dispersal (Cassada & Russell, 1975; Gottlieb & Ruvkun, 1994). While weak loss-of-function DAF-2 mutations result in the Age phenotype, strong DAF-2 mutations cause constitutive dauer formation (Daf-c phenotype) (Gems et al., 1998). Wild-type activity of DAF-16 is required for both daf-2 mutants and normal animals to become dauers. Analysis of C. elegans Age mutants has identified common characteristics that are strongly associated with lifespan extension. Typically, mutants with increased lifespan show increased thermotolerance, UV resistance, and resistance to reactive oxygen species (ROS) (Johnson et al., 2000). These characteristics are also observed in dauer larvae, which are exceptionally longlived (Klass & Hirsh, 1976). These physiological alterations are consistent with the oxidative damage theory of aging, which argues that senescence occurs because of cumulative cellular and systemic oxidative damage (Finkel & Holbrook, 2000). Longlived strains would be predicted either to decrease production of reactive oxygen species, or to increase the ability to detoxify ROS or mitigate their effects. In C. elegans, evidence indicates that Age mutants up-regulate systems that detoxify and protect from ROS. Biochemical analysis of mutants in daf-2 and the downstream PI3-kinase age-1 have shown that these mutations result in increases in activity and /or transcription of cytosolic catalase ctl-1, cytosolic super-oxide dismutase sod-1, and mitochondrial sod-3 in adult animals (Vanfleteren, 1993; Honda & Honda, 1999). Mutations in ctl-1 have been shown to be epistatic to the increase in lifespan observed for Age mutants, indicating that its transcription may be directly regulated by DAF-16 (Taub et al., 1999). To identify additional transcriptional changes that may be important for the Age phenotype of daf-2, we have used cDNA microarrays to analyse transcriptional alterations between daf-2 and daf-16;daf-2 strains. We observed a wide range of DAF-16-dependent transcriptional alterations, including down-regulation of metabolic genes, up-regulation of cellular stress response genes, and ectopic expression of genes normally expressed in the long-lived dauer stage. To identify candidate direct transcriptional targets of the DAF-16 transcription factor, we have identified a subset of genes that are both differentially expressed and contain known DAF-16 regulatory elements within their promoter regions. These putative direct targets include several genes known to be up-regulated in long-lived strains, including the mitochondrial sod-3 and a cytosolic sod. Finally, 111

112 DAF-16 transcriptional targets, J. McElwee et al.

we have used RNAi to examine the in vivo role of putative direct and indirect DAF-16 transcriptional targets. Using this approach, we identified a predicted protease necessary for the daf-2 Age phenotype.

Results Microarray analysis To identify potential transcriptional targets of DAF-16, we used cDNA microarrays, available through the Kim lab at Stanford University, to examine differential gene expression between daf-2(e1370) and daf-16(m27);daf-2(e1370) (Kimura et al., 1997; Lin et al., 1997). Both strains also contained the temperaturesensitive mutation glp-4(bn2), which prevents germline development at non-permissive temperature (Beanan & Strome, 1992). This mutation was used to prevent production of progeny and maternal mRNAs, both of which would probably increase non-specific background hybridization. The glp-4 mutation has no effect on daf-2 or daf-16;daf-2 adult lifespan (data not shown). We directly compared transcriptional differences between these two strains using DNA microarrays containing PCR fragments of genomic DNA corresponding to 17 871 of the 19 626 currently predicted genes in C. elegans + arrayed on a glass slide (Jiang et al., 2001). PolyA RNA was isolated from synchronously staged first-day adult worms in four independent experiments. mRNA samples from daf-16;daf-2 were used to synthesize Cy3-labelled cDNA, and mRNA samples from daf-2 were used to synthesize Cy5-labelled cDNA (see Experimental procedures). The two samples from each biological replicate were hybridized as described, and the entire dataset is available at http://calliope.gs.washington.edu/papers/ mcelwee2002a/ (Reinke et al., 2000). Because normal DAF-2 function acts to inhibit DAF-16 activity, these microarray comparisons can be conceptually simplified to examining transcriptional differences between a daf-16 active state (when daf-2 is mutant) and a daf-16 inactive state (when both daf-2 and daf-16 are mutant). Accordingly, we will refer to these states as daf-16(+) and daf-16(–).

DAF-16-dependent expression of genes involved in cellular stress responses There are many theories that attempt to explain the physiological causes of senescence. Some of these describe the accumulation of damaged or unfolded proteins, accumulated DNA damage or mitochondrial dysfunction (Gensler & Bernstein, 1981; Gershon, 1999; Friguet et al., 2000). The oxidative damage theory of aging postulates that oxygen free radicals produced by cellular metabolism cause systemic damage, and that gradual accumulation of this cellular damage leads to senescence (Finkel & Holbrook, 2000). In C. elegans, it appears that DAF-16-dependent alterations in physiology probably result in the extended lifespan observed for many Age mutants (Ogg et al., 1997; Henderson & Johnson, 2001). To attempt

to identify what physiological alterations may underlie the Age phenotype, we examined the expression of a variety of genes involved in specific cellular roles that may contribute to senescence (Fig. 1). First, we examined a large set of genes involved in protein synthesis and degradation, including tRNA synthetases, ribosomal subunits, translation factors and proteases. We observed no systematic difference in expression of genes involved in protein synthesis in daf-16(+) vs. daf-16(–) (Fig. 1A). For genes involved in protein degradation, we examined the expression of both proteasomal and non-proteasomal proteases (Fig. 1B). The expression of genes that encode proteasomal subunits was also not systematically different in daf-16(+) vs. daf-16(–). Taken together, there appear to be no differences in the expression of genes that mediate the general flux of cellular proteins. The expression of non-proteasomal proteases was more variable, which may indicate a role for specific protein degradation in the Age phenotype. Several genes encoding non-proteasomal proteases were expressed at lower levels in daf-16(+), and two were highly up-regulated in daf-16(+) (Fig. 1B). The role of these two up-regulated proteases is unknown, but they are among the 10 most up-regulated genes in daf-16(+), and one of them appears to be necessary for the daf-2 Age phenotype, as shown below. Second, to determine whether mitochondrial dysfunction could contribute to senescence, we examined the expression of mitochondrial genes (Fig. 1C). Again, there were no systematic differences in the expression of mitochondrial genes in daf16(+) vs. daf-16(–). This finding is interesting, as recent evidence has suggested that reducing the flow of electrons through the mitochondrial electron transport chain can result in decreased production of ROS and increased lifespan (Feng et al., 2001; Tsang et al., 2001). From our analysis, at least at a transcriptional level, there does not appear to be any reduction in mitochondrial function in daf-16(+). This implies that increased detoxification of ROS, rather than reduced production of ROS, may be necessary for the Age phenotype. In support of this finding, the only known mitochondrial gene that appears strongly up-regulated in daf-16(+) is the super-oxide dismutase sod3. sod-3 is one of two manganese-containing superoxide dismutases localized to the mitochondria, and is 10-fold upregulated in daf-16(+) (Hunter et al., 1997). This gene was previously known to be up-regulated in long-lived strains such as age-1 and daf-2, and it has been hypothesized that increased detoxification of damaging free-radicals by this enzyme contributes to the Age phenotype (Honda & Honda, 1999). Finally, we examined the expression of several classes of other genes previously implicated in senescence (Fig. 1D). These include genes involved in cellular stress responses and detoxification, such as heat-shock proteins and cytochrome p450s (Fernandes et al., 1990; Verbeke et al., 2001), apoptosis genes (Higami & Shimokawa, 2000), DNA repair genes (Gensler & Bernstein, 1981), and genes related to yeast SIR2, including sir-2.1 (Tissenbaum & Guarente, 2001). Among these classes of genes, we observed up-regulation of many heat-shock protein © Blackwell Publishing Ltd/Anatomical Society of Great Britain and Ireland 2003

DAF-16 transcriptional targets, J. McElwee et al. 113

Fig. 1 Expression of functional gene classes in daf-16(+) vs. daf-16(–). Each graph shows the log of the average daf-16(–) hybridization intensity (x-axis) compared to the log of the average daf-16(+) normalized hybridization intensity (y-axis) for four array experiments. Axes for all panels are identical to that shown for panel C. A 1 : 1 expression ratio (no-change) is depicted by the diagonal line. Genes expressed above the diagonal are up-regulated in daf-16(+) compared to daf-16(–), while genes below the diagonal are down-regulated. The functional annotations of these genes were identified by WormPD (Costanzo et al., 2000), or are functional classes identified by Kim et al. (Kim et al., 2001). (A) Genes involved in protein synthesis. (B) Genes involved in protein degradation. (C) Mitochondrial genes. (D) Genes previously implicated in lifespan determination.

mRNA’s in daf-16(+), including HSP-70 and several of the heat-shock inducible HSP16-1 family members. Expression of the SIR2 family of genes, DNA repair genes and apoptosis genes all showed no systematic differences in daf-16(+) vs. daf-16(–), while cytochrome p450 genes showed variable expression. The specific up-regulation of genes involved in protection from cellular stress responses, such as HSP and sod-3, may be responsible for many aspects of the Age phenotype, and supports the oxidative damage theory of senescence.

daf-2 expresses transcriptional programmes normally observed in the long-lived dauer stage The developmentally arrested dauer larval stage of C. elegans is specialized for long-term survival and dispersal in harsh environmental conditions. In addition to increased stress-resistance, dauer larvae are exceptionally long-lived, surviving roughly 7– 10-fold longer than the normal reproductive lifespan (Klass & Hirsh, 1976). Because dauer formation is reversible and has no effect on post-dauer lifespan, this developmental stage © Blackwell Publishing Ltd/Anatomical Society of Great Britain and Ireland 2003

presumably must have very efficient programmes for increasing lifespan. Recently, Jones et al. have examined differential gene expression between dauer larvae and a non-dauer wild-type population using serial analysis of gene expression (SAGE) (Jones et al., 2001). In this study, the authors identify 358 dauerspecific, and 533 non-dauer specific transcripts. To examine whether the daf-2 Age phenotype might be due to recapitulation of dauer-specific life-maintenance programmes, we analysed the expression of the stage-specific transcript sets identified by this SAGE analysis (Fig. 2A,B). Though the expression of specific genes varies substantially, the general trend clearly indicates that daf-16(+) expresses many dauer-specific transcripts at high levels while daf-16(–) tends to express non-dauer-specific tags. Additionally, three of the 20 most abundant dauer tags identified by Jones et al. (F38E11.2/Hsp-12.6, F36D3.9, and T25D10.3) were also among the 20 most up-regulated genes in daf-16(+) adults. This is consistent with the expression of dauer-like programmes in adult daf-16(+) animals. As further support of this finding, a recent microarray analysis examining differential gene expression during normal aging in C. elegans

114 DAF-16 transcriptional targets, J. McElwee et al.

Fig. 2 Expression of gene classes in daf-16(+) vs. daf-16(–). Graph format is identical to Fig. 1. The axes for all panels are identical to that shown for panel G. SAGE datasets (panels A and B) were originally identified by Jones et al. (Jones et al., 2001). Each gene expression ‘mountain’ (Panels C–G) represents genes clustered on the C. elegans gene expression map (Kim et al., 2001). The over-representation of differentially expressed genes within each mountain is highly significant (P << 0.0001), as determined by hypergeometric analysis (see Experimental procedures). (A) Expression of 358 dauer-specific SAGE transcripts. (B) Expression of 533 non-dauer-specific SAGE transcripts. (C) Expression of mount 15 genes, which are highly enriched for genes expressed in the dauer stage (Lund et al., 2002). (D) Expression of mount 8 genes, which are enriched for genes involved in intestinal function. (E) Expression of mount 19 genes, which are enriched for genes involved in amino acid, lipid and fatty acid metabolism. (F) Expression of mount 27 genes, which are enriched for amino acid metabolism and energy generation genes. (G) Expression of mount 22 genes, which is primarily made up of genes with no known function.

has identified a large number of genes specifically expressed in dauer larvae (Lund et al., 2002). The authors show that these dauer-specific genes are highly clustered in mount 15 of the C. elegans global gene expression map (described below), and we observe increased expression of many of the mount 15 genes in daf-16(+) vs. daf-16(–) (Fig. 2C).

Differentially expressed genes are over-represented in several C. elegans expression mountains The global gene expression map of C. elegans contains 44 gene clusters that represent 17 661 genes (Kim et al., 2001). These clusters were derived based on gene expression similarities from 553 different DNA microarray experiments testing a wide range of mutant strains and growth conditions, and are represented as ‘mountains’ in the gene expression map. Many of the moun-

tains are enriched for genes that have a common biological function, or that are expressed in specific tissues. To test whether any functionally related genes are specifically activated or repressed by DAF-16 function, we examined the overlap of genes within each of the expression mountains compared to genes that were differentially expressed in daf-16(+) vs. daf-16(–). To do this, we identified a set of 1646 genes that were differentially expressed by at least 1.5-fold in all four of our microarray experiments. This set contains 602 genes up-regulated and 1044 genes down-regulated in daf-16(+) compared to daf-16(–). We examined the overlap between this dataset and the expression mountains, and found statistically significant (P << 0.001) over-representation of differentially expressed genes in several mountains, including mount 8 (3.3-fold overrepresented), mount 19 (6.6-fold over-represented) and mount 22 (5.8-fold over-represented) (see website for full analysis). © Blackwell Publishing Ltd/Anatomical Society of Great Britain and Ireland 2003

DAF-16 transcriptional targets, J. McElwee et al. 115

Mount 8 is known to contain genes involved in digestion and intestinal function, including many proteases, carboxylesterases, lipases, antibacterial proteins and intestine-specific genes. The expression of mount 8 genes in daf-16(+) vs. daf-16(–) is highly variable, but generally down-regulated in daf-16(+) (Fig. 2D). It is known that life-long caloric restriction can significantly increase the lifespan of many organisms, including C. elegans (Lakowski & Hekimi, 1998; Merry, 2000). Perhaps decreased expression of genes involved in intestinal function results in a caloric restrictionlike effect, increasing lifespan in daf-2. Alternatively, both dauers and daf-2 animals are known to have increased intestinal lipid storage and altered metabolism, which may be reflected in mount 8 gene expression (Kimura et al., 1997; Ogg et al., 1997). Mount 19 contains several classes of genes involved in metabolism. These include genes involved in both amino acid and lipid metabolism, as well as many cytochrome p450 genes. The differentially expressed genes in mount 19 are generally strongly down-regulated in daf-16(+) vs. daf-16(–) (Fig. 2E). Additionally, we observe significant over-representation of differentially expressed genes in several other mountains enriched for metabolic genes, including mount 24 (amino acid, lipid and fatty acid metabolic genes), mount 27 (amino acid and energy generation genes) and mount 21 (lipid metabolism) (Fig. 2F and data not shown). The majority of differentially expressed genes in these three mountains are also down-regulated, indicating a tendency for reduced expression of genes involved in metabolism in daf-16(+) compared to daf-16(–). Finally, many of the genes in mount 22 are down-regulated in daf-16(+) compared to daf-16(–) (Fig. 2G). The only class of genes mount 22 is known to be enriched for is collagens; however, further examination showed that none of the collagens within mount 22 is differentially expressed. The majority of differentially expressed genes within this mountain have no known function, but our analysis indicates they are probably coordinately regulated by DAF-16 and may be important for the daf-2 Age phenotype.

Identification of putative direct transcriptional targets of DAF-16 Current genetic evidence indicates that DAF-16 is necessary to transduce signals in the insulin-like pathway in C. elegans (Paradis & Ruvkun, 1998). Mutations in other members of the insulin-like signalling branch that result in a Daf or Age phenotype are completely suppressed by mutations in daf-16, indicating that it is probably the principal transcriptional output of this signalling cascade. Whereas several genes, such as sod-3 and the stress-inducible receptor-tyrosine kinase old-1, have been shown to be dependent on DAF-16 activity for transcription, no direct transcriptional targets of DAF-16 have been identified (Honda & Honda, 1999; Murakami & Johnson, 2001). Furuyama et al. (2000) have identified a core consensus DNA binding site for DAF-16 using in vitro binding site selection. This consensus site has been termed the DAF-16 binding element, or DBE. © Blackwell Publishing Ltd/Anatomical Society of Great Britain and Ireland 2003

Fig. 3 Analysis of DAF-16 binding element (DBE) positioning. Significant clustering of DBE sequences was identified between 0 and 500 base pairs upstream of the start site of genes (P = 0.0006). This clustering was not observed when the DBE sequence was scrambled (two example sequences are shown).

To identify candidate direct transcriptional targets of DAF-16, we used sequence search algorithms to identify genes that contain DBEs within their promoter regions (see Experimental procedures). We searched for the occurrence of DBEs between 0 and 3000 base pairs upstream of the known or predicted start codon (ATG) of C. elegans genes. In these regions, we identified 1492 separate DBE occurrences. To further refine this list of possible direct transcriptional targets, we examined the distribution of DBE sites to see if there were specific regions of the promoters that appeared enriched for DBE sequences (Fig. 3). We found that there is a statistically significant (P = 0.0006) overrepresentation of DBEs near the start site of genes (0 –500 base pairs upstream), compared to the rest of the promoter (500 – 3000 base pairs upstream). To test whether this enrichment of DBE sites is due to the sequence of the DBE, as opposed to a different base composition in gene promoters, we performed searches using scrambled DBE sequences. We observed no significant enrichment of the scrambled sites within specific regions of promoters (two examples are shown in Fig. 3), indicating that enrichment of DBE sites between 0 and 500 base pairs is sequence specific. There are 316 genes that have at least one DBE within 500 base pairs of their initiating codon (for full list, see website). To test whether transcription of any of these genes is dependent on DAF-16 function, we examined the expression of these genes in our microarray datasets. Of these 316 genes, 35 are at least 1.5-fold differentially expressed in daf-16(+) vs. daf-16(–) (Table 1), making them good candidates for direct transcriptional targets of DAF-16.

Analysis of putative direct and indirect DAF-16 transcriptional targets To search among the DAF-16 transcriptional candidates for genes that might have an important role in aging or dauer formation, we used RNA-mediated interference (RNAi) to examine a subset

116 DAF-16 transcriptional targets, J. McElwee et al.

Gene

Description

Fold change (daf-2 vs. daf-16;daf-2)

C08A9.1 ZC412.7 H12D21.1 F57B1.4 F49E11.9 F48G7.8 W02D3.6 F48G7.5 ZK742.4 M60.3 F17B5.1 W01C9.2 ZC412.6 K07B1.4 C50B8.2 F54C8.7 C26H9A.1 C33H5.10 K03H1.6 C17B7.7 Y52B11A.9 R08E5.2 D2096.4 F40F4.3 B0222.9 F36G9.14 K03H1.4 F43E2.5 F29D10.4 T13F2.1 F46C3.2 C34B4.2 C05C10.4 F55G11.4 F35E8.11

SOD-3: Manganese containing superoxide dismutase Protein of unknown function Protein of unknown function Putative collagen, similar to human COL5A2 Member of the PRY protein family Member of an uncharacterized protein family Mitochondrial carrier protein family member Member of an uncharacterized protein family Similar to NAPDH dehydrogenase Protein of unknown function Member of the thioredoxin gene family Protein of unknown function Protein of unknown function Member of an uncharacterized protein family Putative ortholog of C. elegans bir-1 Similar to human arfaptin-1 Member of the vacuaolar H+ ATPase family Protein of unknown function Member of an uncharacterized family Member of the polypeptide chain release family Similar to human Kin17 Cystathionine beta-synthase protein family Similar to TDPGD, dTDP-D-glucose 4,6-dehydratase Fatty acid binding protein homolog 2 Xanthine dehydrogenase protein family Member of an uncharacterized protein family Member of an uncharacterized protein family Similar to TDPGD, dTDP-D-glucose 4,6-dehydratase Mysoin I, class I unconventional heavy chain Delta5-fatty acid desaturase Protein of unknown function Protein of unknown function Acid phosphatase protein family Member of an uncharacterized protein family Glutathione S-transferase protein family

10.2 7.6 7.0 4.1 3.8 2.8 3.7 3.4 3.2 2.9 2.9 2.7 2.4 2.3 2.2 0.6 0.5 0.5 0.5 0.5 0.5 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.3 0.2 0.2 0.1

of possible DAF-16 transcriptional targets (Fig. 4 and supplementary materials). We chose a sample of 35 genes, which represents 16 direct candidate transcriptional targets (differentially expressed genes with a DBE) and 19 indirect candidate transcriptional targets (differentially expressed genes with no DBE). To analyse what effect target inhibition had on dauer formation and lifespan in both N2 and daf-2(e1370), we used the RNAi by feeding method described by Kamath et al. (2001). During all of the RNAi experiments, unc-22 RNAi was performed in parallel as an induction control, which gives a well-characterized and easily discernible locomotion defect (Waterston et al., 1980). To verify the ability of RNAi to suppress the Age and Daf-c phenotypes, we performed RNAi of R13H8.1/daf-16 and saw slight suppression of the Daf-c phenotype of daf-2(e1370) and complete suppression of the Age phenotype. It is possible that dauer formation is less sensitive to reduced DAF-16 activity than lifespan, or that RNAi by feeding is more effective at suppressing the Age phenotype. Because daf-2 is exceptionally long-lived, large-scale RNAi experiments for all 35 candidate genes would have been

Table 1 Putative direct transcriptional targets of DAF-16. Genes were identified as putative direct transcriptional targets if they contained at least one DBE between 0 and 500 base pairs upstream of their transcriptional start site and were at least 1.5fold differentially regulated in all four array experiments

prohibitively time and labour intensive. Thus, we initially tested a small number of animals to survey possible effects on lifespan. These data are available at the supplementary data website. If a suggestive difference between the experimental RNAi lifespan and empty vector control was observed, we performed further experiments to verify the result. Because many of the lifespan experiments used a small number of animals, the results should not be considered an exhaustive analysis. From this survey we found three genes that affect lifespan when inhibited (Fig. 4). RNAi of two genes, ZC334.2/ins-7 and ZK430.3/sod, caused a small but significant increase in the lifespan of daf-2. ZC334.2 encodes an insulin-like peptide upregulated 5.6 fold in daf-16 (+) vs. daf-16 (−), while ZK430.3 encodes a predicted cytosolic Cu/ Zn superoxide dismutase which is upregulated 12.9 fold. RNAi of ZK384.3, which encodes a predicted protease of unknown function, significantly reduced the lifespan of daf-2 by approximately 35% from 44.5 days to 29.5 days (P < 0.001). Inhibition of ZK384.3 resulted in increased mortality rates at all stages during daf-2 lifespan, while having no effect on mortality rates in N2 (Fig. 4A,B). RNAi of ZK384.3 had no © Blackwell Publishing Ltd/Anatomical Society of Great Britain and Ireland 2003

DAF-16 transcriptional targets, J. McElwee et al. 117

Fig. 4 RNAi analysis of DAF-16 transcriptional targets. Shown are five of 35 putative DAF-16 target genes that had a significant effect on dauer formation or lifespan determination when targeted by RNAi (see website for full analysis). Gene/RNAi refers to the gene targeted by RNAi using the feeding method described (see Experimental procedures). L4440 is the empty vector control feeding strain. (A) Mortality rate of N2 grown on RNAi bacterial strains. (B) Mortality rate of daf-2(e1370) grown on RNAi bacterial strains. (C) Mean lifespan and dauer formation frequencies for N2 and daf-2(e1370) grown on RNAi bacterial strains. Sample size (N) is shown in parentheses. (a) Mean lifespan ± standard error at 20 °C as described. Assays with N > 15 represent at least two separate experiments. (b) Per cent dauer formation at 25 °C as described. (c) The reduced dauer formation frequency is due to larval arrest. (d) P < 0.0001.

effect on daf-2 dauer formation, and did not significantly affect N2 lifespan. We also observed changes in larval development when we inhibited C08A9.1/sod-3. This sod is the most up-regulated candidate direct transcriptional target for DAF-16. Inhibition of C08A9.1 resulted in an increased frequency of non-dauer larval arrest in daf-2 at 25 °C, but did not have any effect on lifespan or mortality rates in either daf-2 or N2. While this result does not necessarily indicate a direct role in dauer formation, high levels of sod-3 may be necessary to survive the stressful conditions encountered at higher temperatures.

Discussion Using microarray analysis, we have identified a large number of DAF-16-dependent transcriptional alterations. Wild-type © Blackwell Publishing Ltd/Anatomical Society of Great Britain and Ireland 2003

DAF-16 appears necessary for up-regulation of many genes involved in cellular stress responses, including heat-shock proteins, super-oxide dismutases, and many cytochrome p450s. These results agree with several recent studies examining the transcriptional output of a mammalian relative of DAF-16, FKHRL1 (Kops et al., 2002; Nemoto & Finkel, 2002). In these studies, FKHRL1 activation was shown to increase cellular stress responses by increasing hydrogen peroxide scavenging and directly activating transcription of both Mn-SOD and catalase. Transcriptional targets of FKHRL1 and another DAF-16 relative, AFX, also play a role in cell cycle progression in both G1 and G2 (Medema et al., 2000; Tran et al., 2002). In C. elegans, DAF16 probably does not have these cell cycle functions, since we find no significant overlap between DAF-16-regulated genes and genes involved in cell cycle progression or DNA damage repair (Fig. 1 and data not shown). This is not surprising, since daf-16 mutations have not been reported to affect cell cycle progression during larval development and our studies were conducted on germ-line defective adults, which are entirely post-mitotic. It is possible that the ancestral DAF-16 functionally diversified during mammalian evolution to co-ordinate additional biological processes. In our attempts to identify additional transcriptional targets of DAF-16, we examined the in vivo role of 35 candidate DAF16 direct and indirect transcriptional targets using RNAi. In this analysis, we saw very little effect of these genes on dauer formation in either wild-type or daf-2(e1370). This is unsurprising, since numerous genetic screens have been performed to identify strong Daf-c mutations and to identify strong suppressors of the daf-2 Daf-c phenotype, making it likely that most genes of this nature have been identified. However, it is possible that there is functional redundancy or compensation for many genes involved in dauer formation and lifespan determination, in which case, single-gene RNAi analysis would fail to identify them. Ideally, simultaneous loss-of-function for several genes could be examined to address this issue; however, the effectiveness of RNAi by feeding is drastically reduced when it targets more than one gene (A. Frasier, personal communication). In contrast to dauer formation, we did identify several genes that affected the daf-2 Age phenotype when targeted by RNAi. Inhibition of the predicted aspartyl protease ZK384.3 significantly reduced the mean lifespan of daf-2 from 44.5 days to 29.5 days, and resulted in increased mortality rates throughout the lifespan of daf-2 (Fig. 4). Interestingly, ZK384.3 is one of only two non-proteasomal proteases that are highly upregulated in daf-16(+). This protease may play a specific role in degradation or proteolytic activation of proteins important in determining lifespan. Additionally, we find that inhibition of ZC334.2/ins-7 and ZK430.3/sod enhances the daf-2 Age phenotype. Initially, this seems paradoxical, as expression of both ZC334.2/ins-7 and ZK430.3/sod is dramatically increased in daf-16(+). An analysis of mortality rate indicates that the increases in lifespan may be due to harmful effects mediated by these two genes at specific time periods during the lifespan of daf2. RNAi of ZC334.2/ins-7 appeared to decrease the mortality

118 DAF-16 transcriptional targets, J. McElwee et al.

of daf-2 during mid-life, while RNAi of ZK430.3/sod decreased the mortality of daf-2 very early in life (Fig. 4B). It is possible that the daf-16(+) state may be physiologically similar to insulin-resistant type II diabetes, which could reconcile our data for ZC334.2/ins-7. A common occurrence in type II diabetes is increased levels of insulin production, which generally follows the initiation of insulin resistance. Insulin-resistance, and subsequent hyperglycaemia, has been linked to other debilitating diseases such as cardiovascular disease (LeRoith, 2002). In C. elegans, decreased insulin-receptor activity may lead to an insulin-resistant state, resulting in increased levels of insulin production and harmful pleiotropic effects from elevated insulin levels. In this case, RNAi of ZC334.2/ins-7 could lead to extended lifespan by mitigating the harmful effects of this insulin. Finally, the DAF-16-dependent physiological alterations we observed in daf-2 have substantial overlap with physiological changes observed in long-lived dauer larva. Dauer larvae have increased expression or activity of protective enzymes such as HSP-90, HSP-70, SOD-3, Cu/ Zn SOD, and catalase (Dalley & Golomb, 1992; Vanfleteren, 1993; Vanfleteren & De Vreese, 1995; Honda & Honda, 1999). Additionally, dauer larvae have a radically altered metabolism specialized for long-term survival, efficient glycogen utilization and increased lipid storage (O’Riordan & Burnell, 1989, 1990). It has been hypothesized that the Age phenotype of certain dauer formation mutants may be due to the inappropriate expression of dauer-specific life-extending transcriptional programmes in the adult (Kenyon et al., 1993). In agreement with this hypothesis, we find that transcriptional alterations in daf-16(+) vs. daf-16(–) partially mirror transcriptional alterations between dauer larvae and non-dauer animals. While these findings are consistent with the idea of ectopic expression of a life-extending transcriptional programme normally unique to the dauer stage, they could also result from universal molecular mechanisms underlying lifespan extension that are shared by dauer larva and Age mutants. One way to address this question would be to compare transcriptional alterations in other models of aging to identify shared mechanisms of lifespan determination. Caloric restriction has been shown to increase lifespan in taxonomically diverse organisms including yeast, fruit flies, rodents and nematodes (Yu et al., 1985; Chapman & Partridge, 1996; Jiang et al., 2000; Braeckman et al., 2001). Recently, Pletcher et al. (2002) have examined genome-wide transcription in fruit flies under caloric restriction, and identified many classes of genes that are altered under these conditions. Initial comparisons between our daf-16 data and the caloric restriction data suggest that several classes of genes involved in metabolism are down-regulated in a similar fashion. This may indicate that metabolic suppression is a universal change that is necessary for lifespan extension. It is commonly believed that many different processes contribute to lifespan determination and senescence. Our analysis of the transcriptional outputs of DAF-16, which seems to act as a major molecular switch for the Age phenotype, supports this view. Cellular stress responses, global metabolism and developmental regulation all seem to act co-ordinately under

the control of DAF-16, and probably function together to control the Age phenotype. Many other organismal models for lifespan extension exist, including environmental, pharmacological and genetic interventions across several taxonomically diverse species. Microarray analysis is a powerful tool to begin to ask what processes may be shared between all of these diverse models, allowing the identification of aging-specific pathways.

Experimental procedures Full datasets and supplementary material are available at http:// calliope.gs.washington.edu/papers/mcelwee2002a/

RNA isolation, cDNA synthesis and microarray hybridization Strains used: wild-type is C. elegans variety Bristol strain N2. Linkage group III: daf-2(e1370) (Kimura et al., 1997); Linkage group I: daf-16(m27) (Lin et al., 1997) and glp-4(bn2) (Beanan & Strome, 1992). Strains were grown at 15 °C in liquid culture as described (Reinke et al., 2000). Worms were isolated by floating on sucrose, and large numbers of eggs were prepared by sodium hypochlorite treatment. Worms were synchronized at the L1 stage by hatching these eggs in liquid culture without food. The staged worms were then grown at 20 °C on peptone plates seeded with a lawn of bacteria. Strains were allowed to develop until L4 stage, then shifted to 25 °C to prevent progeny production. By visual inspection, glp-4; daf-2 cultures contained ∼5% (one culture) or << 1% (three cultures) dauers, and daf16 glp-4; daf-2 cultures contained no dauers. Worms were harvested at the first day of adulthood and total RNA was isolated as described (Burdine & Stern, 1996). Poly(A)+ RNA was isolated using the Messagemaker mRNA isolation kit (Gibco BRL) and quantified by UV spectroscopy. A total of four separate pairs of worm cultures (glp-4; daf-2 and daf-16 glp-4; daf-2) were prepared, and RNA from each culture was prepared independently. The RNA representing these four biological replicates were used for the four arrays described. Labelled cDNA probe for DNA microarray hybridization was prepared and hybridized by K. Duke and S. Kim at Stanford University as described (Reinke et al., 2000), except that no wild-type reference was used. glp-4; daf-2 cDNAs labelled with Cy5 and daf-16 glp-4; daf-2 cDNAs labelled with Cy3 from each biological replicate were hybridized together directly, without the use of a reference RNA. While this approach limits the ability to compare our data directly to other published C. elegans arrays, direct hybridizations allow substantial increases in measurement precision (Kerr & Churchill, 2001). We used two additional independent tests to verify the validity of our microarray data. First, we compared the data we generated to another recent study by Yu and Larsen, which used differential-display PCR to identify downstream target genes that were transcriptionally over-expressed or repressed in daf-2 © Blackwell Publishing Ltd/Anatomical Society of Great Britain and Ireland 2003

DAF-16 transcriptional targets, J. McElwee et al. 119

mutant animals compared to wild-type (Yu & Larsen, 2001). In their study, the authors identified nine genes that showed differential expression between daf-2(m41) and daf-16(m26); daf-2(m41). When we examined the expression of eight of these genes in our microarray dataset (one gene was not represented on the array), we observed similar expression profiles for 6/8 genes (data not shown). Second, we performed mRNA slot blot analysis on a random sample of 18 putative DAF-16 transcriptional targets identified in our study (see Supplementary materials). In this analysis, 17 of the 18 genes we examined showed differential expression in the direction predicted by the array data, verifying our methodologies.

Identification of DAF-16 binding elements We implemented a program that searches the C. elegans genome for annotated genes or EST sequence matches downstream of instances of the DBE motif. The program first finds all sequences that match exactly with the DBE sequence. It then uses cross_match (P. Green, unpublished) to compare a 3-kb window downstream of each match to (i) a file containing the DNA sequence for all annotated genes, and (ii) a file containing the assembled EST sequences. Cross_match parameters-minmatch and -minscore were set to 40. All other parameters were kept at default values. Regions in which genes or ESTs begin within 500 nucleotides from the end of the promoter sequence match were flagged for further investigation. To obtain a copy of our program, contact [email protected]. Cross_match must be obtained separately. See www.phrap.com for access/ download information. To determine whether there was a significant over-representation of the DBE within specific promoter regions, we defined the promoter as 0–3000 bp upstream of the initiating ATG of genes. Using 1492 genes that had at least one DBE occurrence in this region, we calculated the mean of the number of occurrences of DBE sites within defined subregions and compared this to the mean of the number of DBE occurrences within the rest of the promoter. A Mann–Whitney U-test was used to compare 500-bp regions throughout the defined promoter region. The most significant over-representation of DBEs occurred between 0 and 500 bp upstream of the initiating ATG.

RNAi analysis The loss-of-function phenotype of putative DAF-16 transcriptional targets was examined using an RNAi feeding protocol (Kamath et al., 2001). It is important to keep in mind that RNAi in C. elegans is often ineffective at suppressing neuronal gene expression (Timmons et al., 2001). Because of this, it is possible that our analysis might miss relevant genes that act in neurones to control lifespan or dauer formation. We used primers originally designed by Reinke et al. that amplify at least 700 bp of predicted coding sequence for all of the genes tested (Reinke et al., 2000). PCR was used to amplify target genes, and all reactions yielded a single band of the © Blackwell Publishing Ltd/Anatomical Society of Great Britain and Ireland 2003

predicted size. PCR products were gel-purified and subcloned using the TopoTA cloning kit (Invitrogen). These were then subcloned into the L4440 RNAi feeding vector using compatible restriction sites for the gene of interest and transformed into the HT115(DE3) RNAi feeding bacterial strain (Fire et al., 1998). Bacteria expressing RNAi constructs were induced overnight on NG agar plates containing 1 mM IPTG as described. All assays were performed on C. elegans strains that had been maintained on induced bacterial cultures for at least three generations. For the dauer formation assays, 5 –10 adult animals were allowed to lay eggs overnight on RNAi plates. Parents were removed the next day and plates were shifted to 25 °C. Three days later, the number of dauer and non-dauer progeny was counted. For the lifespan assays, 5 –10 adult animals were allowed to lay eggs overnight on RNAi plates. Parents were removed the next day and eggs were allowed to develop at 20 °C. L4 larvae were transferred to new plates, and this time point was used as day 0 to account for differences in developmental time between strains. Assays were then performed as described (Dorman et al., 1995), using RNAi plates throughout the experiment. A Mann–Whitney U-test was performed to compare mean lifespan of each RNAi experiment to the empty L4440 vector control.

Data analysis Instat version 2.01 for Macintosh was used to perform Mann– Whitney U analyses. Hypergeometric analysis: to determine whether there was significant over-representation of differentially expressed genes within any of the C. elegans expression map mountains, we used a hypergeometric test for overlap between the two sets of genes. This test returns the probability of a given number of sample successes, given the sample size, population successes and population size. We used a total population size of 19 282 corresponding to the number of predicted genes in the genome, and a total sample size of 1646, corresponding to 1.5fold differentially expressed genes. For full analysis, see website.

Acknowledgments We would like to thank S. Kim and K. Duke for preparing and performing the microarray hybridizations. Thanks to T. Johnson, P. Swoboda and members of the Thomas lab for helpful discussion and commentary. This work was supported by NIA grant T32 AG00057.

References Ailion M, Inoue T, Weaver CI, Holdcraft RW, Thomas JH (1999) Neurosecretory control of aging in Caenorhabditis elegans. Proc. Natl Acad. Sci. USA 96, 7394–7397. Beanan MJ, Strome S (1992) Characterization of a germ-line proliferation mutation in C. elegans. Development 116, 755 –766.

120 DAF-16 transcriptional targets, J. McElwee et al.

Braeckman BP, Houthoofd K, Vanfleteren JR (2001) Insulin-like signaling, metabolism, stress resistance and aging in Caenorhabditis elegans. Mech. Ageing Dev. 122, 673– 693. Burdine R, Stern M (1996) Easy RNA isolation from C. elegans: a TRIZOL based method. Worm Breeder’s Gazette 14, 10. Cassada RC, Russell RL (1975) The dauerlarva, a post-embryonic developmental variant of the nematode Caenorhabditis elegans. Dev. Biol. 46, 326–342. Chapman T, Partridge L (1996) Female fitness in Drosophila melanogaster: an interaction between the effect of nutrition and of encounter rate with males. Proc. R. Soc. Lond. B Biol. Sci. 263, 755 –759. Costanzo MC, Hogan JD, Cusick ME, Davis BP, Fancher AM, Hodges PE, Kondu P, Lengieza C, Lew-Smith JE, Lingner C, Roberg-Pere ZKJ, Tillberg M, Brooks JE, Garrels JI (2000) The yeast proteome database (YPD) and Caenorhabditis elegans proteome database (WormPD): comprehensive resources for the organization and comparison of model organism protein information. Nucl Acids Res. 28, 73 –76. Dalley BK, Golomb M (1992) Gene expression in the Caenorhabditis elegans dauer larva: developmental regulation of Hsp90 and other genes. Dev. Biol. 151, 80 – 90. Dorman JB, Albinder B, Shroyer T, Kenyon C (1995) The age-1 and daf-2 genes function in a common pathway to control the lifespan of Caenorhabditis elegans. Genetics 141, 1399 –1406. Feng J, Bussiere F, Hekimi S (2001) Mitochondrial electron transport is a key determinant of lifespan in Caenorhabditis elegans. Dev. Cell 1, 633–644. Fernandes G, Venkatraman J, Khare A, Horbach GJ, Friedrichs W (1990) Modulation of gene expression in autoimmune disease and aging by food restriction and dietary lipids. Proc. Soc. Exp. Biol. Med. 193, 16 –22. Finkel T, Holbrook NJ (2000) Oxidants, oxidative stress and the biology of ageing. Nature 408, 239 –247. Fire A, Xu S, Montgomery MK, Kostas SA, Driver SE, Mello CC (1998) Potent and specific genetic interference by double-stranded RNA in Caenorhabditis elegans. Nature 391, 806 – 811. Friguet B, Bulteau AL, Chondrogianni N, Conconi M, Petropoulos I (2000) Protein degradation by the proteasome and its implications in aging. Ann. NY Acad. Sci. 908, 143 –154. Furuyama T, Nakazawa T, Nakano I, Mori N (2000) Identification of the differential distribution patterns of mRNAs and consensus binding sequences for mouse DAF-16 homologues. Biochem. J. 349, 629 – 634. Gems D, Sutton AJ, Sundermeyer ML, Albert PS, King KV, Edgley ML, Larsen PL, Riddle DL (1998) Two pleiotropic classes of daf-2 mutation affect larval arrest, adult behavior, reproduction and longevity in Caenorhabditis elegans. Genetics 150, 129 –155. Gensler HL, Bernstein H (1981) DNA damage as the primary cause of aging. Q. Rev. Biol. 56, 279 –303. Gershon D (1999) The mitochondrial theory of aging: is the culprit a faulty disposal system rather than indigenous mitochondrial alterations? Exp. Gerontol. 34, 613 – 619. Gottlieb S, Ruvkun G (1994) daf-2, daf-16 and daf-23: genetically interacting genes controlling Dauer formation in Caenorhabditis elegans. Genetics 137, 107–120. Hekimi S, Burgess J, Bussiere F, Meng Y, Benard C (2001) Genetics of lifespan in C. elegans: molecular diversity, physiological complexity, mechanistic simplicity. Trends Genet. 17, 712–718. Henderson ST, Johnson TE (2001) daf-16 integrates developmental and environmental inputs to mediate aging in the nematode Caenorhabditis elegans. Curr. Biol. 11, 1975 –1980. Higami Y, Shimokawa I (2000) Apoptosis in the aging process. Cell Tissue Res. 301, 125–132. Honda Y, Honda S (1999) The daf-2 gene network for longevity regulates oxidative stress resistance and Mn-superoxide dismutase gene expression in Caenorhabditis elegans. FASEB J. 13, 1385 –1393.

Hunter T, Bannister WH, Hunter GJ (1997) Cloning, expression, and characterization of two manganese superoxide dismutases from Caenorhabditis elegans. J. Biol. Chem. 272, 28652–28659. Jiang JC, Jaruga E, Repnevskaya MV, Jazwinski SM (2000) An intervention resembling caloric restriction prolongs lifespan and retards aging in yeast. FASEB J. 14, 2135 –2137. Jiang M, Ryu J, Kiraly M, Duke K, Reinke V, Kim SK (2001) Genomewide analysis of developmental and sex-regulated gene expression profiles in Caenorhabditis elegans. Proc. Natl Acad. Sci. USA 98, 218 –223. Johnson TE, Cypser J, de Castro E, de Castro S, Henderson S, Murakami S, Rikke B, Tedesco P, Link C (2000) Gerontogenes mediate health and longevity in nematodes through increasing resistance to environmental toxins and stressors. Exp. Gerontol. 35, 687– 694. Jones SJ, Riddle DL, Pouzyrev AT, Velculescu VE, Hillier L, Eddy SR, Stricklin SL, Baillie DL, Waterston R., Marra MA (2001) Changes in gene expression associated with developmental arrest and longevity in Caenorhabditis elegans. Genome Res. 11, 1346 –1352. Kamath RS, Martinez-Campos M, Zipperlen P, Fraser AG, Ahringer J (2001) Effectiveness of specific RNA-mediated interference through ingested double-stranded RNA in Caenorhabditis elegans. Genome Biol. 2, RESEARCH0002. Kenyon C, Chang J, Gensch E, Rudner A, Tabtiang R (1993) A C. elegans mutant that lives twice as long as wild-type. Nature 366, 461– 464. Kerr MK, Churchill GA (2001) Statistical design and the analysis of gene expression microarray data. Genet. Res. 77, 123–128. Kim SK, Lund J, Kiraly M, Duke K, Jiang M, Stuart JM, Eizinger A, Wylie BN, Davidson GS (2001) A gene expression map for Caenorhabditis elegans. Science 293, 2087–2092. Kimura KD, Tissenbaum HA, Liu Y, Ruvkun G (1997) daf-2, an insulin receptor-like gene that regulates longevity and diapause in Caenorhabditis elegans. Science 277, 942– 946. Klass M, Hirsh D (1976) Non-aging developmental variant of Caenorhabditis elegans. Nature 260, 523 –525. Kops GJPL, Dansen TB, Polderman PE, Saarloos I, Wirtz KWA, Coffer PJ, Huang TT, Bos JL, Medema RH, Burgering BMT (2002) Forkhead transcription factor FOXO3a protects quiescent cells from oxidative stress. Nature 419, 316 –321. Lakowski B, Hekimi S (1998) The genetics of caloric restriction in Caenorhabditis elegans. Proc. Natl Acad. Sci. USA 95, 13091–13096. Lee RY, Hench J, Ruvkun G (2001) Regulation of C. elegans DAF-16 and its human ortholog FKHRL1 by the daf-2 insulin-like signaling pathway. Curr. Biol. 11, 1950 –1957. LeRoith D (2002) β-cell dysfunction and insulin resistance in type 2 diabetes: Role of metabolic and genetic abnormalities. Am. J. Med. 113 (6A), 3S –11S. Lin K, Dorman JB, Rodan A, Kenyon C (1997) daf-16: An HNF-3/forkhead family member that can function to double the life-span of Caenorhabditis elegans. Science 278, 1319 –1322. Lund J, Tedesco P, Duke K, Wang J, Kim SK, Johnson TE (2002) Transcriptional profile of aging in C. elegans. Current Biol. 12, 1566 –1573. Medema RH, Kops GJPL, Bos JL, Burgering BMT (2000) AFX-like forkhead transcription factors mediate cell-cycle regulation by Ras and PKB kip1 through p27 . Nature 404, 782–787. Merry BJ (2000) Calorie restriction and age-related oxidative stress. Ann. NY Acad. Sci. 908, 180 –198. Murakami S, Johnson TE (2001) The OLD-1 positive regulator of longevity and stress resistance is under DAF-16 regulation in Caenorhabditis elegans. Curr. Biol. 11, 1517–1523. Nemoto S, Finkel T (2002) Redox regulation of forkhead proteins through a p66shc-dependent signaling pathway. Science 295, 2450 –2452. O’Riordan V, Burnell A (1989) Intermediary metabolism in the dauer larva of the nematode Caenorhabditis elegans-I. Glycolysis, gluconeogenesis, © Blackwell Publishing Ltd/Anatomical Society of Great Britain and Ireland 2003

DAF-16 transcriptional targets, J. McElwee et al. 121

oxidative phosphorylation and the tricarboxylic acid cycle. Comp. Biochem. Physiol. 92B, 233 –238. O’Riordan V, Burnell A (1990) Intermediary metabolism in the dauer larva of the nematode Caenorhabditis elegans-II. The glyoxylate cycle and fatty-acid oxidation. Comp. Biochem. Physiol. 95B, 125 –130. Ogg S, Paradis S, Gottlieb S, Patterson GI, Lee L, Tissenbaum HA, Ruvkun G (1997) The Fork head transcription factor DAF-16 transduces insulinlike metabolic and longevity signals in C. elegans. Nature 389, 994 – 999. Paradis S, Ailion M, Toker A, Thomas JH, Ruvkun G (1999) A PDK1 homolog is necessary and sufficient to transduce AGE-1 PI3 kinase signals that regulate diapause in Caenorhabditis elegans. Genes Dev. 13, 1438–1452. Paradis S, Ruvkun G (1998) Caenorhabditis elegans Akt/ PKB transduces insulin receptor-like signals from AGE-1 PI3 kinase to the DAF-16 transcription factor. Genes Dev. 12, 2488 –2498. Pletcher SD, Macdonald SJ, Marguerie R, Certa U, Stearns SC, Goldstein DB, Partridge L (2002) Genome-Wide Transcript Profiles in Aging and Calorically Restricted Drosophila melanogaster. Curr. Biol. 12, 712–723. Reinke V, Smith HE, Nance J, Wang J, Van Doren C, Begley R, Jones SJ, Davis EB, Scherer S, Ward S, Kim SK (2000) A global profile of germline gene expression in C. elegans. Mol. Cell 6, 605 – 616. Taub J, Lau JF, Ma C, Hahn JH, Hoque R, Rothblatt J, Chalfie M (1999) A cytosolic catalase is needed to extend adult lifespan in C. elegans Daf-C and clk-1 mutants. Nature 399, 162–166. Timmons L, Court DL, Fire A (2001) Ingestion of bacterially expressed

© Blackwell Publishing Ltd/Anatomical Society of Great Britain and Ireland 2003

dsRNAs can produce specific and potent genetic interference in Caenorhabditis elegans. Gene 263, 103 –112. Tissenbaum HA, Guarente L (2001) Increased dosage of a sir-2 gene extends lifespan in Caenorhabditis elegans. Nature 410, 227–230. Tran H, Brunet A, Grenier JM, Datta SR, FornaceAJ, JrDiStefano PS, Chiang LW, Greenberg ME (2002) DNA repair pathway stimulated by the forkhead transcription factor FOXO3a through the Gadd45 protein. Science 296, 530 –534. Tsang WY, Sayles LC, Grad LI, Pilgrim DB, Lemire BD (2001) Mitochondrial respiratory chain deficiency in Caenorhabditis elegans results in developmental arrest and increased lifespan. J. Biol. Chem. 276, 32240 –32246. Vanfleteren JR (1993) Oxidative stress and ageing in Caenorhabditis elegans. Biochem. J. 292, 605 – 608. Vanfleteren JR, De Vreese A (1995) The gerontogenes age-1 and daf-2 determine metabolic rate potential in aging Caenorhabditis elegans. FASEB J. 9, 1355 –1361. Verbeke P, Fonager J, Clark BF, Rattan SI (2001) Heat shock response and ageing: mechanisms and applications. Cell Biol. Int. 25, 845 –857. Waterston RH, Thomson JN, Brenner S (1980) Mutants with altered muscle structure in Caenorhabditis elegans. Dev. Biol. 77, 271–302. Yu H, Larsen PL (2001) DAF-16-dependent and independent expression targets of DAF-2 insulin receptor-like pathway in Caenorhabditis elegans include FKBP’s. J. Mol. Biol. 314, 1017–1028. . Yu BP, Masoro EJ, McMahan CA (1985) Nutritional influences on aging of Fischer 344 rats. I. Physical, metabolic, and longevity characteristics. J. Gerontol. 40, 657– 670.