Opposing roles of microRNA Argonautes during Caenorhabditis elegans aging

Abstract

Argonaute (AGO) proteins partner with microRNAs (miRNAs) to target specific genes for post-transcriptional regulation. During larval development in Caenorhabditis elegans, Argonaute-Like Gene 1 (ALG-1) is the primary mediator of the miRNA pathway, while the related ALG-2 protein is largely dispensable. Here we show that in adult C. elegans these AGOs are differentially expressed and, surprisingly, work in opposition to each other; alg-1 promotes longevity, whereas alg-2 restricts lifespan. Transcriptional profiling of adult animals revealed that distinct miRNAs and largely non-overlapping sets of protein-coding genes are misregulated in alg-1 and alg-2 mutants. Interestingly, many of the differentially expressed genes are downstream targets of the Insulin/ IGF-1 Signaling (IIS) pathway, which controls lifespan by regulating the activity of the DAF-16/ FOXO transcription factor. Consistent with this observation, we show that daf-16 is required for the extended lifespan of alg-2 mutants. Furthermore, the long lifespan of daf-2 insulin receptor mutants, which depends on daf-16, is strongly reduced in animals lacking alg-1 activity. This work establishes an important role for AGO-mediated gene regulation in aging C. elegans and illustrates that the activity of homologous genes can switch from complementary to antagonistic, depending on the life stage.

Fig 1. Expression of ALG-1 and ALG-2 during adulthood.

(A) Western blot of FLAG::GFP-tagged ALG-1 in protein samples from L4 and adult stage transgenic animals. Tubulin levels serve as a loading control. (B) Western blot of FLAG::RFP-tagged ALG-2 in protein samples from L4 and adult stage transgenic animals. A sample from L4 stage non-transgenic (nt) animals demonstrates the specificity of the anti-FLAG antibody. Tubulin levels serve as a loading control. (C) Expression of endogenous ALG-1 tagged with GFP in L4, 2d and 5d adult animals visualized by fluorescence microscopy. The panels of non-transgenic animals show that auto-fluorescence contributes to the signal detected in the intestine of adults. Micrographs were captured at 40x magnification with 26 ms exposure. (D) Expression of endogenous ALG-2 tagged with RFP in L4, 2d and 5d adult animals visualized by fluorescence microscopy. The panels of non-transgenic animals show that auto-fluorescence contributes to the signal detected in the intestine of adults. Micrographs were captured at 40x magnification with 450 ms exposure. (E) Simultaneous expression of GFP::ALG-1 and RFP::ALG-2 in L4 and 5d adult animals visualized by fluorescence microscopy. Micrographs were captured at 40x magnification with equivalent exposures at L4 and 5d adult stages. (F) qRT-PCR analyses of alg-1 and alg-2 mRNA levels in adults relative to L4 stage in the spe-9(hc88) strain, averaged from three independent experiments. The sterile spe-9 background was used to avoid potential signal from progeny animals. The error bars represent SEMs. **P<0.01, ***P<0.001 (t-test).

Fig 2. Opposite effects of alg-1 and alg-2 on lifespan.

(A) Survival curves showing reduced lifespan in alg-1(gk214) (red) and increased lifespan in alg-2(ok304) (blue) compared to WT (black). ****P<0.0001 (log-rank). Statistical analyses of each replicate for all lifespan assays are shown in S1 Table. (B) Gene structure of alg-2 (a and b isoforms) depicting regions targeted by RNAi and locations of genetic mutations. Sequence identity with alg-1 in regions covered by the alg-2 RNAi constructs is indicated by percentages. Locations of ap426 and ok304 deletions are depicted in green and blue, respectively (see text for details). (C) Survival curves showing opposite effects of RNAi targeting the coding sequence (CDS) (red) versus the 3’UTR (blue) of alg-2 compared to vector control RNAi (black). ****P<0.0001 (log-rank). (D) Survival curves showing that, compared to WT (black), a new alg-2 loss of function strain (alg-2(ap426)) (green) has an extended lifespan similar to that of alg-2(ok304) (blue). ****P<0.0001 (log-rank).

Fig 3. Increased alg-1 is insufficient for lifespan extension.

(A) Survival curves of WT (black) and alg-2(ok304) (blue) strains transferred to vector control (solid) or alg-1 (dashed) RNAi at the L4 stage. ****P<0.0001 (log-rank). (B) Top: Western blot of ALG-1 protein levels in spe-9(hc88) and spe-9(hc88); alg-2(ok304) animals during L4 and adulthood. Bottom: Quantification of protein levels from four independent experiments. The samples were normalized to L4, and the error bars represent SEMs. **P<0.01, *P<0.05 (t-test). (C) Expression of endogenous ALG-1 tagged with GFP in WT versus alg-2(ok304) backgrounds visualized by fluorescence microscopy. Micrographs of heads & tails were captured at 40x magnification with 76 ms exposure. (D) Depiction of ALG-1 and miRNA binding sites in the alg-1 3’UTR detected by iCLIP and miRNA::target chimera analyses reported in Broughton et al., 2016 [32]. (E) Top: Western blot of ALG-1 protein levels expressed from the gene containing its native or a swapped 3’UTR during L4 and adulthood. Bottom: Quantification of protein levels from three independent experiments. The samples were normalized to WT L4, and the error bars represent SEMs. **P<0.01 (t-test). (F) Survival curves of the indicated strains, showing that replacement of the native alg-1 3’UTR with the Y45F10D.4 (swap) 3’UTR is not sufficient to produce a lifespan phenotype. Not significant (n.s.); ****P<0.0001 (log-rank).

Fig 4. Altered miRNA and mRNA expression in alg-1 and alg-2 mutants.

(A) TaqMan analyses of the indicated miRNAs in alg-1(gk214) (red) and alg-2(ok304) (blue) compared to WT animals at day 5 of adulthood, averaged from five independent experiments. The error bars represent SEMs. *P<0.05, **P<0.01, ***P<0.001 (t-test). The published lifespan phenotypes observed with reduced expression (Δ), overexpression (OEX), and corresponding references (ref) are indicated. ↑ (increased lifespan), ↓ (decreased lifespan), NE (no effect), ND (not determined). (B) Enrichment of specific miRNAs with ALG-1 (red) or ALG-2 (blue) detected by sequencing of small RNAs that co-immunoprecipitated (co-IP) with each AGO protein at day 5 of adulthood averaged from 2 independent experiments. See S2 Table for a complete list of miRNAs reproducibly enriched for association with ALG-1 or ALG-2. (C) Overlap of all genes significantly (P<0.05) up- or down-regulated at day 5 of adulthood in alg-1(gk214) (red) and alg-2(ok304) (blue) compared to WT animals. Overlap of down-regulated genes in alg-1(gk214) and alg-2(ok304) is more than expected by chance (P<0.0001, hypergeometric test). The top ranked gene ontology terms identified by DAVID analysis are listed. See S3 Table and S4 Table for complete RNA-seq and DAVID results. (D) Enrichment and depletion of seed pairing (nucleotides 2–7) for the indicated miRNAs with genes differentially expressed in alg-1(gk214) versus WT animals. The fold difference shown is in comparison to the fraction of seed-pairing sites detected in genes with unchanged expression patterns in the alg-1 mutants. ***P<0.0001 (Chi Squared with Yates Correction). (E) Percent of conserved lin-4 and miR-71 targets predicted by TargetScan [36, 37] that are differentially regulated in alg-1(gk214).

Fig 5. Differential regulation of the IIS pathway by alg-1 and alg-2.

(A) Left: Average DAF-16::GFP intestinal nuclear localization score for WT, daf-2(e1370), alg-1(gk214), alg-2(ok304) from four blinded scorers. The error bars represent SEMs. ****P<0.0001, *P<0.05 (t-test). Right: representative images of scoring key showing (0) no intestinal nuclear localization, (1) nuclear localization restricted to the most anterior pair of intestinal nuclei, denoted by an arrow, (2) moderate nuclear localization in the anterior pair and additional intestinal nuclei, indicated by arrowheads, and (3) strong nuclear localization in intestinal nuclei. (B) The two upper bars show the overlap of genes up- or down-regulated at least 4-fold in alg-1(gk214) with Class I (purple) or Class II (orange) genes [28]. The two lower bars show the overlap of all genes up- or down-regulated in alg-2(ok304) with Class I or II genes [28]. Enrichment: Class I genes and down in alg-1 (****P<0.0001), Class II genes and down in alg-1 (****P<0.0001), Class I genes and up in alg-2 (****P<0.0001), Class II genes and down in alg-2 (****P<0.0001) (Hypergeometric test using fractions of Class I and II genes across genome). (C) Survival curves of single and double mutants of the alg-1(gk214), daf-2(e1370), daf-16(mu86) strains compared to WT. ****P<0.0001 (log-rank). (D) Survival curves of single and double mutants of the alg-2(ok304), daf-2(e1370), daf-16(mu86) strains compared to WT. Not significant (n.s.); ****P<0.0001 (log-rank). (E) Survival curves of WT and alg-2(ok304) strains transferred to vector control or asah-1 (RNAi) at the L4 stage. Not significant (n.s.); ****P<0.0001 (log-rank). (F) Survival curves of WT and alg-2(ok304) strains transferred to vector control or cest-1 (RNAi) at the L4 stage. Not significant (n.s.); ****P<0.0001 (log-rank).