Life span extension via eIF4G inhibition is mediated by posttranscriptional remodeling of stress response gene expression in C. elegans

Abstract

Reducing protein synthesis slows growth and development but can increase adult life span. We demonstrate that knockdown of eukaryotic translation initiation factor 4G (eIF4G), which is downregulated during starvation and dauer state, results in differential translation of genes important for growth and longevity in C. elegans. Genome-wide mRNA translation state analysis showed that inhibition of IFG-1, the C. elegans ortholog of eIF4G, results in a relative increase in ribosomal loading and translation of stress response genes. Some of these genes are required for life span extension when IFG-1 is inhibited. Furthermore, enhanced ribosomal loading of certain mRNAs upon IFG-1 inhibition was correlated with increased mRNA length. This association was supported by changes in the proteome assayed via quantitative mass spectrometry. Our results suggest that IFG-1 mediates the antagonistic effects on growth and somatic maintenance by regulating mRNA translation of particular mRNAs based, in part, on transcript length.

Figure 1. Inhibition of ifg-1 Differentially Regulates Gene Expression at the Level of Translation

(A) Genome-wide comparison of relative expression in low (L) versus high (H) translation fractions (left panel). Coordinates correspond to the log2 of the ratio of gene expression between conditions (ifg-1 RNAi: control RNAi) for the translated fractions. The right panel shows the relative expression in low (L) versus high (H) translation fractions for genes selected on the basis of significantly altered TI. (B) Cluster image map of genes with altered TI versus their inclusion among overrepresented GO categories. Red indicates upregulation when ifg-1 is inhibited, and blue indicates downregulation. The plot represents 250 genes with altered TI present at the greatest frequency among the 100 most significantly overrepresented GO categories. (C) Overrepresented biological GO categories showing the number of genes in each category and the statistical significance. The p value indicates significantly overrepresented categories in yellow, while orange boxes indicate the parent terms (i.e., the ontological lineage). A total of 8179 genes were tagged withGOterms falling under the heading of “Biological process.” (D) Cluster analysis of TI genes, representing the level of coregulation for each gene with every other gene (722 plotted along both the x axis and y axis). Red, yellow, green, blue, and purple colors indicate high-to-low level of coregulation among genes with significantly altered TI, respectively.

Figure 2. Translationally Upregulated Stress Response Genes Are Important for Life Span Extension upon ifg-1 Inhibition

(A) Mean life span for the ifg-1(cxTi9279) mutant was 29.0 days (n = 110) on control RNAi; 16.1 days (n = 69) on sca-1 RNAi (p < 0.0001). Mean life span for N2 was 23.6 days (n = 274) on control RNAi; 21.4 days (n = 143) on sca-1 RNAi (p < 0.0001). (B) Mean life span for the ifg-1(cxTi9279) mutant was 29.0 days (n = 110) on control RNAi; 25.3 days (n = 106) on unc-70 RNAi (p < 0.0001). Mean life span for N2 was 23.6 days (n = 274) on control RNAi; 22.1 days (n = 142) on unc-70 RNAi (p < 0.0001). (C) Mean life span for the ifg-1(cxTi9279) mutant was 28.3 days (n = 107) on control RNAi; 19.8 days (n = 119) on let-711 RNAi (p < 0.0001). Mean life span for N2 was 22.9 days (n = 140) on control RNAi; 15.3 days (n = 160) on let-711 RNAi (p < 0.0001). (D) Mean life span for the ifg-1(cxTi9279) mutant was 29.0 days (n = 110) on control RNAi; 23.4 days (n = 132) on fasn-1 RNAi (p < 0.0001). Mean life span for N2 was 23.6 days (n = 274) on control RNAi; 21.9 days (n = 232) on fasn-1 RNAi (p < 0.0001). (E) Mean life span for the ifg-1(cxTi9279) mutant was 27.2 days (n = 73) on control RNAi; 18.5 days (n = 119) on mdt-15 RNAi (p < 0.0001). Mean life span for N2 was 22.6 days (n = 144) on control RNAi; 17.7 days (n = 160) on mdt-15 RNAi (p < 0.0001). (F) Mean life span for the ifg-1(cxTi9279) mutant was 27.2 days (n = 73) on control RNAi; 23.7 days (n = 112) on chc-1 RNAi (p < 0.0001). Mean life span for N2 was 22.6 days (n = 144) on control RNAi; 21.0 days (n = 83) on chc-1 RNAi (p < 0.0001). (G) Mean life span for the aak-2 (ok524) mutant on control RNAi was 15.7 days (n = 72); 17.3 days (n = 61) on ifg-1 RNAi (p = 0.0014). Mean life span for N2 was 19.7 days (n = 81) on control RNAi; 24.8 days (n = 80) on ifg-1 RNAi (p < 0.0001). (H) Mean life span for the daf-16 (mu86) mutant on control RNAi was 15.6 days (n = 126); 16.9 days (n = 137) on ifg-1 RNAi (p < 0.0001). Mean life span for N2 was 25.4 days (n = 98) on control RNAi; 28.7 days (n = 98) on ifg-1 RNAi (p < 0.0001). (I) Life span pathway dependence on stress response genes was determined for ifg-1, rsks-1, daf-2, isp-1, eat-2, daf-16, and aak-2 mutants. Results are given as the percent change in life span for each RNAi treatment (corresponding to Table S4). All life spans were repeated at least one time. Complete statistical analysis can be found in Table S4.

Figure 3. High-Resolution Polysomal Profiles of Individual Stress Response Genes upon ifg-1 Inhibition

Individual expression profiles for stress response genes important for life span are shown for the polysome region after 48 hr of control (gray line) or ifg-1 (black line highlighted in red) RNAi feeding. The average relative expression ratio between these two conditions (ifg-1 RNAi: control RNAi) and standard error of the mean (SEM) are shown for three independent experiments. Fractions were standardized according to Coleoptera luciferase control RNA spiking. Gene expression values for each fraction were then normalized according to the average expression value for that fraction. The ratio of TI values (ifg-1 RNAi to control RNAi conditions) associated with each gene is listed directly below the gene name, with values greater than one indicating an increase in translation state upon ifg-1 inhibition. Characteristic profiles for worms on control RNAi and ifg-1 RNAi are shown for one experiment to illustrate where relative expression is measured.

Figure 4. Differentially Translated Genes Are Important for Resistance to Starvation Stress, which Diminishes ifg-1 Expression

(A and B) Individual stress response genes upregulated upon ifg-1 inhibition and important for life span were attenuated via RNAi feeding for 48 hr prior to starvation in the ifg-1 mutant (A) and wild-type N2 backgrounds (B). (C) The P170 form of IFG-1 required for cap-mediated mRNA circularization was resolved by western blotting after 48 hr of feeding or fasting. (D) Densitometry results for IFG-1 levels from three biological replicates after 48 hr of fasting or feeding are shown. All experiments were started on the first day of adulthood. Quantitation of IFG-1 was normalized to β-actin. Results are expressed as mean ± SEM.

Figure 5. UTR and ORF Length Is Associated with Altered Ribosomal Loading upon Suppression of ifg-1

(A) Comparison of average 3′UTR length between translationally upregulated and downregulated genes. Results are expressed as mean ± SEM. “Array” for the blue bar indicates all genes on the microarray. (B) Comparison of average ORF length between translationally upregulated and downregulated genes. Results are expressed as mean ± SEM. (C) Histogram indicating the number of genes with the ORF length indicated. The bracketed area shows that, starting from average ORF length for all genes on the microarray, 34% of genes are subject to differential posttranscriptional regulation based on ORF length. (D) Lowess smooth plot of the ratio of TI (ifg-1 RNAi: control RNAi) and p value versus ORF length for the entire array.

Figure 6. Biochemical and Bioinformatics Analyses Point to Differential Changes in Protein Expression According to Message Length When ifg-1 Is Attenuated

(A) Results of quantitative mass spectrometry of protein lysates comparing adult nematodes fed on control RNAi with ifg-1 RNAi for 2 days. Results are expressed as mean ± SEM of changes in aa (amino acid) length of genes that were increased or decreased upon ifg-1 inhibition. (B) Bioinformatic analysis of the relationship between ORF length and GO categories. Genes were separated on the basis of ORF length (bases) into five bins containing an equal number of genes. The ORF lengths for bins 1–5 are 0–522, 523–873, 874–1101, 1102–1587 and 1588–39300, respectively. From each bin, the total number of those corresponding to “Biological process” GO categories of stress response, translation, and ribonucleotide biosynthesis were compared with the average representation for all GO categories containing more than 100 genes (220 categories total).