Protein synthesis.

Sorting by the lifespan increase, inactivation of genes involved in protein synthesis caused the most potent lifespan increase. We identified several RNAi clones that target components of the translation initiation factor (eIF) complex; egl-45 (eIF3, 52%), eif-3.F (eIF3, 32%), eif-3.B (eIF3, 51%), ifg-1 (eIF4G, 55%), T27F7.3 (eIF1, 25%), and two clones targeting inf-1 (eIF4A, 46% and 28%). Inhibition of these genes increased lifespan up to 50% longer than with vector control RNAi. In mammals, insulin stimulates protein synthesis in vivo and in vitro by increasing the rate of translation initiation by regulating the eIF4F complex consisting of eIF4G, eIF4E, eIF4A, and the 40S ribosome [44]. eIF4F then recruits eIF3 to generate the 43S ribosome. Protein synthesis may be closely monitored in the cell to allow energy equivalents to be redirected to other processes such as genomic maintenance and stress adaptation. Mice lacking the translational inhibitor 4E-BP1 are lean and have increased metabolism and mitochondrial biogenesis, which are opposing phenotypes found in long-lived mutants [45]. We found that three clones that target components of the 40S subunit of the ribosome, rps-3 (32%), rps-8 (18%), and rps-11 (28%) also increased mean lifespan. Although components of the 80S ribosome complex were represented in the RNAi library that we screened, only components of the 43S complex, predominantly translation initiation factors emerged as regulators of lifespan.

Under conditions of ER stress a transient block in protein synthesis occurs. Inactivation of the ER oxidoreductase ero-1 (32%) also increased lifespan (Table 1). In mammals, C. elegans, and yeast, the ATF4/atf-5/GCN4 gene contains an extensive 5′UTR with multiple upstream open reading frames. Under normal conditions the upstream open reading frames are preferentially utilized for translation initiation yielding truncated proteins, whereas under low nutrient conditions the proper ATG is used to synthesize ATF4/atf-5/GCN4. eIF1 has been shown to regulate initiation codon selection [46]. In support of our findings two recent studies also identified ifg-1 and rps-11 as regulators of longevity [47,48]. Therefore, although complex, translation can be an effective regulator of cellular stress and adaptation.

Signaling.

Genes in the insulin-signaling pathway are well-established and potent regulators of longevity in the worm [43]. Our screen identified many signaling molecules that extend adult lifespan (Figure S2). For example, we identify SEM-5 (24%), a SH2- and SH3-like protein that negatively regulates RAS-MAP and -IP3 signal transduction. sem-5 was also identified as an enhancer of dauer formation (S.S. Lee, unpublished data). In addition, inactivation of four annotated kinases—the serine/threonine protein kinases tpa-1 (28%) and tag-181 (21%), an ARK kinase family member sel-5 (29%), and a previously identified integrin-linked kinase pat-4 (13%) [38]—increased lifespan. daf-16(mgDf47) epistasis (discussed below) reveals that some of these kinases may act in the insulin pathway.

The C. elegans genome contains one insulin-like receptor, daf-2, but 39 insulin-like peptides of which only a few have known functions [49]. Loss-of-function mutations in daf-2 are of the strongest enhancers of lifespan in C. elegans, and in our screen RNAi of daf-2 increased lifespan by ∼79%. In humans, the relaxin family of peptides are structurally similar to insulin and bind to seven transmembrane receptors to initiate signaling cascades [50]. We identified two serpentine receptors, str-49 (15%) and sre-25 (23%), that act in lifespan control. Since the lifespan extension phenotype of these gene inactivations requires DAF-16 (Figure 2), it is conceivable that these seven transmembrane receptors utilize a subset of the insulin-like peptides, in analogy to relaxin in humans.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Identification of RNAi Clones that Act through DAF-16

(A) Shading indicates that daf-16(mgDf47); eri-1(mg366) is epistatic to the lifespan extension observed in eri-1(mg366) (Table S2).

(B) DAF-16 nuclear localization after feeding RNAi clone is shown. Light shading indicates some nuclear localization, and medium shading indicates more nuclear localization (Figure S4).

(C) sod-3p::gfp expression after RNAi feeding is shown. Light shading indicates weak expression, medium shading indicated modest expression, and dark shading indicates strong expression.

https://doi.org/10.1371/journal.pgen.0030056.g002

Signals from proliferating germ cells negatively regulate C. elegans aging, via the insulin-signaling pathway [43]. We identified glp-1 (33%), which codes for a DSL-family ligand receptor in the germline that controls germ cell proliferation and negatively regulates lifespan [22]. The design of our screen was fortunate in that the somatic gonad and the establishment of the initial pool of proliferating germ cells at adulthood was not perturbed. Down-regulating genes such as glp-1 or genes that function in the reproductive system (Table S1) could disrupt signaling from the germline to increase lifespan.

Vacuolar H⁺-ATPases were another potent lifespan regulator to emerge from the screen. These proteins acidify intracellular compartments and act in synaptic transmission and cell death signaling cascades [51]. UNC-32, a subunit of the vacuolar ATPase, regulates male longevity [52]. The vacuolar ATPase is comprised of a V0/V1 heteromultimer. We identified one of each type of subunit from our screen; vha-6 (V0, 24%) and tag-300 (V1, 23%). In addition, we identified two other candidate longevity genes linked to cell-death pathways; the apoptotic cysteine protease ced-3 (19%) and the cell death related nuclease crn-5 (41%). We were surprised to identify an RNAi clone targeting ced-3 from our screen since ced-3(0) mutants fail to undergo programmed cell death but are not lethal [53]. However, we noted synthetic lethality between ced-3(717) and daf-2(e1370) indicating a genetic interaction between these longevity-promoting pathways (unpublished data). In support of this finding, the increased lifespan phenotype of ced-3 (RNAi) was dependent upon daf-16 (Figure 2).

Mitochondria.

We were surprised to find that late inactivation of mitochondrial genes also increased lifespan. RNAi clones targeting F26E4.6 (28%) and atp-3 (21%) were previously identified to increase lifespan [37,38]. However, previous work has suggested that perturbations that disable mitochondria to increase lifespan must occur during larval development, rather than after the L4 stage as we observe here [34]; these differences may be attributed to the enhanced RNA interference phenotype of the eri-1(mg366) strain. In addition, we identified seven other mitochondrial RNAi clones that significantly increased adult lifespan when the gene inactivation occurred postdevelopmentally: two subunits of respiratory complex V (atp-2 [43%] and phi-37 [27%]); three subunits of the NADH:ubiquinone oxidoreductase complex (F59C6.5 [19%], nuo-1 [23%], and Y56A3A.19 [21%]); a predicted mitochondrial carrier protein (F43E2.7, 21%); and a mitochondrial AAA-protease (spg-7, 22%). In support of a role for nuo-1 in negative regulation of lifespan, overexpression of nuo-1 causes decreased fecundity and shortened lifespan [40].

Gene expression.

RNA binding/processing and chromatin-associated factors were among the largest class of regulators; comprising over 20% of the genes identified in our screen and yielding some of the most potent extension of lifespan (Figures S2 and S3; Table 1). The strong enrichment for RNA factors was intriguing because few have been identified in the previous genome-wide RNAi screens for longevity [37,38]. We identified three clones targeting predicted RNA helicases, dic-1 (22%), B0511.6 (50%), and ZK686.2 (24%) that scored as among the strongest gene inactivations that increase lifespan (Table 1). We also found two clones with predicted conserved RNA binding domains; hrp-1 (49%) contains an RRM domain, and C56G2.1 (20%) has a KH and a Tudor domain that may facilitate RNA and chromatin interactions, respectively. In addition, the protein products of a large number of genes identified are predicted to modify and/or process RNA; these include a predicted mRNA splicing factor D1054.14 (36%); two ribonucleases crn-5 (41%) and Y54E10BR.4 (15%); mRNA cleavage and polyadenylation factor F25C6.2 (24%); RNA cyclase ZK1127.5 (24%); and two RNA methylases, T07A9.8 (24%) and W07E6.1 (31%).

Chromatin-associated factors are conserved players in altering gene expression, coordinating cell fate, and modulating small RNA mechanisms of gene regulation in eukaryotes [54–56]. Inactivation of these chromatin-associated factors caused increased lifespan: C56G2.1 (20%), sas-5 (30%), htp-3 (19%), and the DNA replication licensing factor mcm-2 (15%). In support of our findings, mcm-2 was also identified by Serial Analysis of Gene Expression analysis to be transcriptionally repressed in daf-2/insulin-receptor mutants and may contribute to their longevity phenotype [57].

Two transcription factors, unc-62/homothorax/MEIS (38%) and Y65B4BR.5/NAC/TS-N (17%) were found to negatively regulate lifespan. Recently, the Dubcovsky laboratory reported that mutations in the NAC transcription factor NAM-B1 delay senescence by more than three weeks in Triticum turgidum, and the ancestral wild wheat allele functions to accelerate senescence and increase nutrient remobilization [58], suggesting this may be an evolutionary conserved mechanism regulating lifespan. Although a role for unc-62 has not previously been described in an adult animal, the unc-62 promoter element is active in adult neuronal tissues (Figure S1). While a mechanism for RNA and chromatin-associated factors in controlling lifespan is unknown, there are significant alterations in the transcriptional profile of aging populations [59]; these genes may regulate such changes.

Identification of clones that act in or converge upon the insulin-signaling pathway.

The DAF-16/FOXO transcription factor is a key player in many biological processes that regulate lifespan including stress adaptation and metabolism [10,11,39,60]. Because loss of function daf-16 alleles are epistatic to many longevity promoting mutations, we inactivated the 64 candidate longevity genes in an eri-1(mg366) strain also carrying a null daf-16(mgDf47) mutation and scored for lifespan extension (Figure 2; Table S2). In accord with previous studies, most RNAi clones targeting mitochondrial genes that increase adult lifespan did not depend on DAF-16 [34,35]. The gene inactivations targeting the protein synthesis machinery also increase lifespan in the absence of DAF-16. However, RNAi of a few representative genes from most classes could increase the lifespan of the daf-16(mgDf47);eri-1(mg366) strain and thus represent DAF-16 independent pathways of lifespan extension (Figure 2). We classified the remaining RNAi clones as DAF-16 dependent because they failed to extend mean lifespan in the absence of DAF-16 by at least 10% compared to vector controls. Among the DAF-16-dependent genes were most of the signaling molecules and RNA-related genes including mcm-2, hrp-1, pos-1, sas-5, crn-5, T07A9.8, ZK686.2, and ZK1127.5. These genes may act in or converge onto the insulin-signaling pathway for lifespan control.

Although the role of DAF-16 in insulin-signaling is the best understood, DAF-16 receives inputs from multiple cellular pathways and has functions outside of adult lifespan extension. DAF-16 transcriptional activity is negatively regulated by cytoplasmic sequestration. Upon loss of inhibition, DAF-16 translocates to the nucleus. Therefore, we tested whether any of the gene inactivations identified in our screen could induce DAF-16 nuclear localization, regardless of the necessity for DAF-16 in adult lifespan extension. Despite being independent of daf-16 gene activity for increasing adult lifespan, inactivation of atp-3, ifg-1, vha-6, F25G6.2, B0511.6, eif-3.B, inf-1, egl-45, rps-8, rps-3, eif-3.F, or rps-11 induced localization of DAF-16 to the nucleus and consequently drove expression of the DAF-16 target gene sod-3 (Figure 2). Since these genes do not function through DAF-16 to extend lifespan, the translocation of DAF-16 to the nucleus may be due to a general stress or may represent a new regulator of DAF-16 activity.

Larval arrest and survival phenotypes of RNAi clones during development.

Loss of DAF-2 during development causes constitutive arrest in the dauer stage. daf-16(lf) mutations are defective for dauer arrest induced by low insulin signaling and also fail to arrest as L1 stage larvae in the absence of food [60,61]. Many of the 2,700 gene inactivations initiated by feeding RNAi from the L1 stage cause highly penetrant developmental arrest at stages ranging from early larval to sterile adult stage (Table S1). A subset of these gene inactivations caused arrested larvae to survive longer than control animals and facilitated DAF-16 nuclear localization at the arrested stage (Figure 2; Table 2). We tested if the larval arrest and increased survival of the arrested larvae induced by these gene inactivations depend on DAF-16 activity by feeding the RNAi clones that cause larval arrest in eri-1(mg366) to a daf-16(mgDf47);eri-1(mg366) double mutant. The absence of DAF-16 abrogated the long-lived phenotype for most of the arrested larvae (Table 2). For most gene inactivations, daf-16 mutant animals still arrested when fed the RNAi clone; however they did not survive as long in the arrested state. The daf-16(mgDf47) allele did, however, weakly suppress the larval arrest of inf-1 or spg-7 inactivation (Table 2). These data suggest that at almost any stage, stress pathways requiring daf-16 may be triggered by gene inactivations to ensure the survival of the arrested larvae.

We noted that many of the gene inactivations that cause increased survival of arrested larvae encode translation factors. Translation is a major target of antibiotics, which are produced by a wide range of fungi and microbes that nematodes encounter in the environment. As a larvae or adult enters an environment with an antibiotic, there may be signaling pathways that detect ribosomal deficiency to trigger developmental arrest as well as xenobiotic protective pathways. The induced stress adaptation and survival pathways would ensure that the animal could live long enough to escape the antibiotic and resume reproductive development. Inhibition of translation by RNAi of ribosomal and other translation factors may mimic the ribosomal deficiencies induced by antibiotics in the normal C. elegans ecosystem, and trigger this physiological response, developmental arrest, and cessation of aging. Because it is a prereproductive arrest, this response may be under natural selection to trigger longevity enhancing pathways [62]. It may be significant that other gene inactivations with potent arrested larval lifespan increases, a vacuolar ATPase and the mitochondrial ATP synthase, are also targets of natural antibiotics [63,64].