RNAi screening implicates a SKN-1-dependent transcriptional response in stress resistance and longevity deriving from translation inhibition

Abstract

Caenorhabditis elegans SKN-1 (ortholog of mammalian Nrf1/2/3) is critical for oxidative stress resistance and promotes longevity under reduced insulin/IGF-1-like signaling (IIS), dietary restriction (DR), and normal conditions. SKN-1 inducibly activates genes involved in detoxification, protein homeostasis, and other functions in response to stress. Here we used genome-scale RNA interference (RNAi) screening to identify mechanisms that prevent inappropriate SKN-1 target gene expression under non-stressed conditions. We identified 41 genes for which knockdown leads to activation of a SKN-1 target gene (gcs-1) through skn-1-dependent or other mechanisms. These genes correspond to multiple cellular processes, including mRNA translation. Inhibition of translation is known to increase longevity and stress resistance and may be important for DR-induced lifespan extension. One model postulates that these effects derive from reduced energy needs, but various observations suggest that specific longevity pathways are involved. Here we show that translation initiation factor RNAi robustly induces SKN-1 target gene transcription and confers skn-1-dependent oxidative stress resistance. The accompanying increases in longevity are mediated largely through the activities of SKN-1 and the transcription factor DAF-16 (FOXO), which is required for longevity that derives from reduced IIS. Our results indicate that the SKN-1 detoxification gene network monitors various metabolic and regulatory processes. Interference with one of these processes, translation initiation, leads to a transcriptional response whereby SKN-1 promotes stress resistance and functions together with DAF-16 to extend lifespan. This stress response may be beneficial for coping with situations that are associated with reduced protein synthesis.

Figure 1. RNAi screening overview.

After overnight culture, RNAi bacteria were seeded onto 24-well RNAi feeding plates in triplicate. dsRNA synthesis was induced for 5–6 hours, then 3–7 L3 or L4 gcs-1p::GFP worms were deposited into each well. After four days growth at 20°C, the triplicate worm samples were transferred to 96 well plates for assessment of the GFP signal in the intestine. Approximately 300 candidates were identified in the first screening round, in which a worm population was assessed rapidly as to whether the intestinal GFP signal was elevated. Those cDNA clones were analyzed in a quantitative second round of screening, in which intestinal GFP signal was scored as High, Medium, or Low as described , , (Materials and Methods). Genes were scored as positive if gcs-1 upregulation was robust in all three trials in the second round (Figure 2A; Table 1). Four distinct RNAi clones are represented by different shading in individual wells, with the remainder of the plates being arbitrarily left blank.

Figure 2. Analysis of genes that prevent constitutive gcs-1 expression.

Confirmed RNAi screening positives and additional COP9 signalosome subunits were examined by RNAi knockdown for effects on the indicated GFP reporters in L4 stage C. elegans. Reporters were scored for levels of nuclear GFP localization (SKN-1B/C::GFP, SKN-1op::GFP, DAF-16::GFP) or GFP expression in the intestine as High, Medium, or Low (Figure 1; Materials and Methods). Percentages of worms in each group were plotted on the Y axis in each panel. In each case a representative example of at least three RNAi experiments is shown (n>30 for each experiment). (A) gcs-1p::GFP expression. (B) Expression of the gcs-1Δ2::GFP reporter, which lacks a SKN-1-independent pharyngeal regulatory sequence and serves as a control for (C) (Figure S1A) . (C) Expression of gcs-1Δ2mut3::GFP, in which an important SKN-1 binding site is mutated (Figure S1A) . (D) gcs-1p::GFP expression in skn-1(zu67) mutants. Two independent transgenic lines each gave similar results. (E) gcs-1p::GFP expression in the sek-1(km4) mutant, in which stress-induced p38 signaling is blocked . (F) Expression of the gst-4p::GFP promoter, a SKN-1 target , . (G) Levels of nuclear SKN-1 expressed from SKN-1B/C::GFP, which encodes SKN-1 isoforms b and c (Figure S1B) . (H) Nuclear accumulation of SKN-1 expressed from SKN-1op::GFP, which encodes all three SKN-1 isoforms (Figure S1B) . (I) Expression of gcs-1p::GFP in daf-16(mgDf47) animals. (J) Presence of DAF-16::GFP (Table S6) in intestinal nuclei. (K) Activity of the DAF-16 target sod-3 in the intestine. Black diamonds and asterisks indicate genes for which gcs-1 was induced independently of sek-1 or skn-1, respectively (summarized in Figure 3). Dots indicate genes that were associated with unambiguous accumulation of SKN-1::GFP in intestinal nuclei.

Figure 3. Pathways of gcs-1 activation in the intestine.

RNAi against 34 of the 41 genes we identified in this study resulted in gcs-1 promoter activation through a canonical mechanism that required both skn-1 and p38 MAPK signaling, as illustrated by requirement for the p38 MAPKK SEK-1 (black box)(Figure 2) . In many of these cases the levels of SKN-1::GFP in intestinal nuclei were not dramatically increased (Figure 2), implying that gcs-1 may be activated by SKN-1 through mechanisms besides increasing the overall levels of nuclear SKN-1 (see text). For three genes (C48B6.2: snoRNP component, phi-43 and wdr-23; blue box) RNAi-induced gcs-1 activation required SKN-1, but not p38 MAPK signaling (as revealed by sek-1-independence). For two genes (F30A10.9: nucleic acid binding protein, and Y71H10B.1: IMP-GMP specific 5′-nucleotidase), induction required SEK-1 but not SKN-1 (green box), implying that a different transcription factor was involved. In two cases (Y87G2A.1 and Y57E12AL.6), gcs-1 was activated independently of both SKN-1 and SEK-1 (red box).

Figure 4. Effects of gcs-1-regulatory genes on stress resistance.

(A) Effects on tert-butyl hydrogen peroxide (TBHP) resistance in wild-type (N2) animals. L4 worms were placed on RNAi or control bacteria for three days at 20°C, then transferred to assay plates containing a lawn of OP50 and 9.125 mM TBHP. Survival was then scored over a time-course. A bar graph shows the percent change in mean survival time for each gene compared to control RNAi. Representative experiments are shown here and plotted in Figure S2. Where only 2 experiments were performed, the experiment in which RNAi gave the less robust effect is graphed. Results of individual experiments, numbers of worms analyzed, and statistical analyses are presented in Table S1. (B, C) Effects on TBHP resistance in skn-1(zu67) animals. RNAi assays of selected genes were performed and analyzed as in (A), but using 15.4 mM TBHP. Representative experiments are presented as plots of proportional survival over time, with results and statistical analysis of individual experiments provided in Table S3. Note that in each case, the increase in stress resistance deriving from RNAi of these genes was almost completely dependent upon skn-1.

Figure 5. Induction of SKN-1–dependent target gene expression and stress resistance in response to translation initiation factor RNAi.

(A) Translation initiation factors that were examined in this study. The eIF4F complex stabilizes capped mRNAs and activates them for translation by interacting with their 5′ cap and poly-A-binding protein (PABP) . This interaction promotes binding of these mRNAs by the translation pre-initiation complex (PIC), which includes the 40 S ribosome subunit and the initiator tRNA. Subsequent steps in initiation follow this binding event. Initiation factors that we examined in this study are shown in green. (B) Activation of the gcs-1p::GFP reporter. N2 or skn-1(zu67) worms that carry the gcs-1p::GFP reporter were exposed to the indicated RNAi or control bacteria beginning at the L2 stage. They were scored for GFP fluorescence at day one of adulthood as in Figure 2, at which time the worms appeared normal and were laying eggs that hatched. p-values indicated above individual bars correspond to comparison with control RNAi. Similar reporter induction was observed after three days of initiation factor RNAi that began at adulthood day one, and no reporter activity was observed when control RNAi was performed in skn-1(zu67) animals (not shown). p values were calculated by the Chi² method. (C) Activation of the gst-4 reporter, scored as in (B). (D) Induction of endogenous SKN-1 target gene expression in response to translation initiation factor RNAi, analyzed by quantitative RT-PCR (qRT-PCR) performed in triplicate. RNAi was performed as in (B). Each gene assayed is upregulated under stress conditions . A representative experiment is shown, in which Fold Change and p-values above individual bars refer to comparison to control RNAi. Additional qRT-PCR experiments and statistical analyses are described in Table S4. (E) Induction of skn-1-dependent stress resistance. After exposure to the indicated RNAi bacteria as in Figure 4, N2 or skn-1(zu67) worms were placed on plates containing 15.4 mM TBHP, then scored for survival over time. In each case, the worms appeared normal and were laying eggs when they were transferred to TBHP plates. In N2 but not skn-1 mutant worms, stress resistance was dramatically enhanced by prior exposure to translation initiation factor RNAi. All experiments and statistics are provided in Table S3. (F) Comparison of TBHP resistance in N2 and daf-16 mutant worms, performed and analyzed as in (E). daf-16 was not required for the increases in oxidative resistance that derive from translation initiation factor RNAi.

Figure 6. Importance of SKN-1 for lifespan extension deriving from translation initiation factor RNAi.

(A) Survival plot showing effects of ife-2(eIF4F) RNAi. This lifespan extension was greatly reduced by the skn-1(zu135) mutation, which was used in all experiments in this figure. (B) Longevity extension by ifg-1(eIF4F) RNAi. Survival of ifg-1(RNAi) worms was not substantially decreased by skn-1 mutation. (C,D) Longevity extension by eif-1(PIC) and eif-1A(PIC) RNAi. The longevity associated with inhibiting these genes was decreased but not eliminated by skn-1 mutation. Note that overall survival of these RNAi animals is nevertheless significantly impaired in the skn-1 background compared to N2. All longevity analyses were performed at 20°C, with lifespans measured from hatching and RNAi initiated at day 1 of adulthood. Each panel shows a composite of multiple experiments in which the populations shown were analyzed in parallel, with the proportion surviving indicated on the y-axis. These data are summarized in Table 2, with individual experiments described in Table S5.

Figure 7. Lifespan extension in response to translation initiation factor RNAi is mediated primarily by DAF-16 and SKN-1.

(A) Survival after exposure to translation initiation factor RNAi and 2% glucose. Glucose feeding increases IIS, which inhibits both DAF-16 and SKN-1 , . Lifespan extension by translation factor inhibition is eliminated under these conditions. (B) ife-2(eIF4F) RNAi fails to extend lifespan in a daf-16(mgDf47) mutant. (C, D) Modest lifespan extension in response to ifg-1(eIF4F) or eif-1(PIC) RNAi in daf-16(mgDf47) animals. Note that survival of these RNAi animals is impaired in the daf-16 background compared to N2 (Figure 6B and 6C). (E, F) Survival of daf-16(mgDf47); skn-1(zu67) double mutants exposed to ifg-1(eIF4F) or eif-1(PIC) RNAi. In this genetic background ifg-1(eIF4F) RNAi extends lifespan exclusively among the longest-lived worms, and eif-1(PIC) RNAi has only a negligible effect. Experiments were performed and plotted as in Figure 6 and are summarized in Table 2, with individual experiments described in Table S5.