Positive feedback between transcriptional and kinase suppression in nematodes with extraordinary longevity and stress resistance

Abstract

Insulin/IGF-1 signaling (IIS) regulates development and metabolism, and modulates aging, of Caenorhabditis elegans. In nematodes, as in mammals, IIS is understood to operate through a kinase-phosphorylation cascade that inactivates the DAF-16/FOXO transcription factor. Situated at the center of this pathway, phosphatidylinositol 3-kinase (PI3K) phosphorylates PIP(2) to form PIP(3), a phospholipid required for membrane tethering and activation of many signaling molecules. Nonsense mutants of age-1, the nematode gene encoding the class-I catalytic subunit of PI3K, produce only a truncated protein lacking the kinase domain, and yet confer 10-fold greater longevity on second-generation (F2) homozygotes, and comparable gains in stress resistance. Their F1 parents, like weaker age-1 mutants, are far less robust-implying that maternally contributed trace amounts of PI3K activity or of PIP(3) block the extreme age-1 phenotypes. We find that F2-mutant adults have <10% of wild-type kinase activity in vitro and <60% of normal phosphoprotein levels in vivo. Inactivation of PI3K not only disrupts PIP(3)-dependent kinase signaling, but surprisingly also attenuates transcripts of numerous IIS components, even upstream of PI3K, and those of signaling molecules that cross-talk with IIS. The age-1(mg44) nonsense mutation results, in F2 adults, in changes to kinase profiles and to expression levels of multiple transcripts that distinguish this mutant from F1 age-1 homozygotes, a weaker age-1 mutant, or wild-type adults. Most but not all of those changes are reversed by a second mutation to daf-16, implicating both DAF-16/ FOXO-dependent and -independent mechanisms. RNAi, silencing genes that are downregulated in long-lived worms, improves oxidative-stress resistance of wild-type adults. It is therefore plausible that attenuation of those genes in age-1(mg44)-F2 adults contributes to their exceptional survival. IIS in nematodes (and presumably in other species) thus involves transcriptional as well as kinase regulation in a positive-feedback circuit, favoring either survival or reproduction. Hyperlongevity of strong age-1(mg44) mutants may result from their inability to reset this molecular switch to the reproductive mode.

Figure 1. Both in vitro kinase activity for endogenous substrates and total phosphoprotein levels are diminished in age-1(mg44) homozygotes.

(A–C) In vitro kinase activity of post-gravid or sterile adult worms, 6 days after the L4/adult molt. Kinase activity of cleared, sonicated lysates was assessed as γ-³²P-ATP incorporation per 20-µg protein sample, in 1 min at 30°C. Samples, quenched on ice, were electrophoresed on 10% polyacrylamide-SDS gels (Invitrogen). (A) Gels stained with SYPRO Ruby (Invitrogen) to confirm constant protein loads. (B) Image of ³²P β-emissions (Molecular Dynamics Storm) from the gel in A, dried under vacuum. (C) Data summary from 2–3 independent expansions each of N2DRM, age-1(hx546), age-1(mg44) F1 homozygotes (labeled age-1 mg44^F1 ), age-1(mg44) F2 homozygotes (labeled age-1 mg44^F2), and daf-16(m26) double mutants with each age-1 allele. These quantitations covered the full length of each lane, not all of which is shown in A and B. (D–F) Phosphoproteins, resolved by polyacrylamide gel electrophoresis and visualized by Pro-Q Diamond staining. (D) Image of total protein stained with Coomassie blue. (E) Image of fluorescence after Pro-Q Diamond staining; note that only the two phosphoprotein markers were detected. (F) Summary of data from 3 independent expansions of each strain at adult day 6. Significance: (C, F), P values are shown for 2-tailed t-tests (unequal variance) comparing each age-1 strain to N2DRM (values directly over bars) or comparing two strains connected by brackets (values over brackets).

Figure 2. Most phosphoproteins are depleted in age-1(mg44) F2 adult worms.

A 2-dimensional polyacrylamide gel electrophoresis resolves dual-labeled phosphoproteins, enriched on phospho-affinity columns (Qiagen), from age-1(mg44) F2 worms collected 10 d after the L4/adult molt (red fluorescence) and from post-gravid, near-isogenic N2DRM controls 6 d after becoming adults (green fluorescence). A total of 1,669 spots were identified and quantified for each fluor after co-electrophoresis. Red/green ratios equal to or greater than 2 (i.e., at least 2-fold more abundant in age-1(mg44) than in N2DRM adults) were observed for 13 spots; 30 had ratios of 1.2–2.0; 122 were essentially constant (ratios of 0.8–1.2); 305 fell between 0.5 and 0.8; and 1,199 were reduced at least twofold in abundance in age-1(mg44) adults. A reduction of at least 3-fold was seen for 784 spots, and ≥5-fold for 287 spots.

Figure 3. Transcriptional suppression of IIS genes in age-1(mg44) F2 adults.

Transcript levels were assayed by real-time polymerase chain reaction (RT-PCR). Expression histograms are shown superimposed on a simplified schematic of the IIS pathway. Yellow arrows indicate protein phosphorylations (symbolized by circled P's), and orange arrows indicate binding/activation by phosphatidylinositol 3,4,5-triphosphate (PIP₃, red “structural” symbols). Transcriptional outputs are indicated by open block arrows. Within each histogram, means±SEM are shown on a log(2) scale for steady-state transcript level, comparing wild-type to four age-1 mutant populations and to dauer larvae. For each strain indicated, fold changes are shown (e.g., “3×”), of age-1(mg44)-F2 relative to N2DRM. The age-1(mg44) worms are post-gravid F1 homozygotes at days 8–9 of adulthood, or F2 homozygotes at day 10 after the L4/adult molt; other strains (N2DRM, age-1(hx546), and daf-16(mu86); age-1(mg44) double mutants) were harvested as post-gravid adults (day 6 of adulthood), or as dauer larvae (N2DRM only, from starved, dense cultures 1 day after >98% of worms had become resistant to 1% sodium dodecyl sulfate). Transcript levels were assayed for 3–8 independent biological replicates, with two cDNA syntheses and two RT–PCRs for each. All results were normalized to the mean of three control genes (β-actin, T08G5.3, and Y71D11.3) that did not differ significantly among strains. Significant differences, relative to N2DRM controls, are indicated by asterisks in Table 1.

Figure 4. Most transcriptional changes in age-1(mg44) F2 adults remain even after scaling for lifespan.

Transcript levels were assayed by real-time polymerase chain reaction (RT–PCR) as described in the legends to Figure 3 and Table 1. Within each histogram, means±SEM are shown on a log(2) scale for steady-state transcript level, comparing wild-type N2DRM adults at three ages to age-1(mg44) F2 homozygotes at one or two adult ages. All transcript values were first normalized to β-actin mRNA (to adjust for variation in RNA inputs), and then each biological group was normalized to the mean value for N2DRM day-6 adults (to correct for run-to-run variation in C(t) values). Ages are measured from the L4/adult molt; a table (inset) indicates the percent of adult lifespan represented by each age. To monitor longitudinal changes, two biological preparations of each group were assessed (fewer than used for Figure 3 or Table 1). Significance of differences was ascertained by two-tailed Behrens-Fisher t-tests, appropriate to samples of unequal or unknown variance. The transcript levels observed in this experiment differ from the means seen in previous experiments (e.g., Table 1), but those differences are not significant.

Figure 5. Survival of oxidative and electrophilic stresses.

N2DRM worms were exposed to (A) 5-mM hydrogen peroxide (H₂O₂), or (B) 10-mM 4-hydroxynonenal (4-HNE). Stress survivals began on adult day 3–4 and worms were maintained in liquid medium, 20±0.3°C, without bacteria. Results shown were replicated in independent experiments. Statistics (log-rank test, each n = 40–50, both A and B): age-1(mg44) at F2 vs. F1 generation: P<10⁻⁷; age-1(mg44) F1 vs. any other strain: P<10⁻³; age-1(mg44) F2 vs. any other strain or group: P<10⁻⁹; age-1(hx546) vs. N2DRM: P<0.004; daf-16(mu86); age-1(mg44) vs. N2DRM: P<0.005. (C) Functional consequences of suppressing signaling genes downregulated in age-1(mg44) F2 adults. N2DRM worms were picked on day 1 of adulthood, and maintained on dsRNA-expressing E. coli, at 20°C, for 3 days. RNAi extension of 50^th-percentile H₂O₂ survival, as percent gain over N2DRM fed on bacteria carrying empty expression vector (significance, by Gehans log-rank test), is as follows: vps-34, 26% (P<0.03); let-60/RAS, 40% (P<0.001); daf-3, 49% (P<0.0003); aak-1, 49% (P<0.00003); daf-4, 57% (P<0.0001). All significant effects except vps-34 were confirmed (each P<0.001) in independent experiments. Peroxide survival of F2 age-1(mg44) day-62 adults, without RNAi, exceeded the N2DRM controls by 5.6-fold (P<10⁻⁶), at the 60^th percentile (the lowest survival observed for the F2 group).

Figure 6. Model of the IIS “molecular switch.”

Three proposed system states are depicted, reflecting the balance between kinase signaling and transcriptional feedback suppression: reproductive mode, in which kinase signaling predominates, activated via the insulin/IGF-1 receptor (DAF-2, here labeled Ins-R); longevity mode, wherein FOXO (DAF-16) prevails and suppresses kinase transcription; and hyperlongevity mode—in which switching between the first two modes is blocked, thus strongly favoring survival but with no possibility of resuming reproduction. State transitions are normally triggered by signaling modulators (e.g., insulin-like peptides or SIR-2/14-3-3 complex [77]), to which age-1(mg44)-F2 mutants (“hyperlongevity” state) cannot respond.