Nuclear RNAi maintains heritable gene silencing in Caenorhabditis elegans

Abstract

RNA interference (RNAi) is heritable in Caenorhabditis elegans; the progeny of C. elegans exposed to dsRNA inherit the ability to silence genes that were targeted by RNAi in the previous generation. Here we investigate the mechanism of RNAi inheritance in C. elegans. We show that exposure of animals to dsRNA results in the heritable expression of siRNAs and the heritable deposition of histone 3 lysine 9 methylation (H3K9me) marks in progeny. siRNAs are detectable before the appearance of H3K9me marks, suggesting that chromatin marks are not directly inherited but, rather, reestablished in inheriting progeny. Interestingly, H3K9me marks appear more prominently in inheriting progeny than in animals directly exposed to dsRNA, suggesting that germ-line transmission of silencing signals may enhance the efficiency of siRNA-directed H3K9me. Finally, we show that the nuclear RNAi (Nrde) pathway maintains heritable RNAi silencing in C. elegans. The Argonaute (Ago) NRDE-3 associates with heritable siRNAs and, acting in conjunction with the nuclear RNAi factors NRDE-1, NRDE-2, and NRDE-4, promotes siRNA expression in inheriting progeny. These results demonstrate that siRNA expression is heritable in C. elegans and define an RNAi pathway that promotes the maintenance of RNAi silencing and siRNA expression in the progeny of animals exposed to dsRNA.

Fig. 1.

Nuclear RNAi maintains heritable gene silencing. (A) Animals of the indicated genotypes expressing sur-5p::gfp were fed Escherichia coli expressing ± gfp dsRNA. A similar feeding RNAi approach is used for the remainder of the RNAi experiments in this article. Overlaid florescent and light images of representative adult animals not exposed to gfp dsRNA (column 1) or exposed to gfp dsRNA (column 2) are shown. Column 3 shows embryos from animals in column 2 that were isolated by hypochlorite treatment and grown in the absence of gfp dsRNA. Overlaid florescent and light images of larval-stage 3 (L3) progeny are shown. In all conditions, >98% of animals exhibited phenotypes similar to that shown in images. See Fig. S1 for nrde-2 gfp RNAi inheritance data. The following genotypes were used in A–D: nrde-1(gg088), nrde-2(gg091), nrde-3(gg066), and nrde-4(gg129). Unless otherwise indicated, these same nrde alleles are used for the remainder of the experiments in this article. (B) Florescent image of embryos obtained by hypochlorite treatment of animals treated ± gfp dsRNA. (C) Animals were exposed to dpy-11 dsRNA. F1 progeny were grown in the absence of dpy-11 dsRNA. Images show light microscopy of representative L3 progeny. In all conditions, >98% of animals exhibited phenotypes similar to that shown in images. (D) Animals of the indicated genotypes were grown on pos-1 dsRNA expressing bacteria or pos-1 dsRNA expressing bacteria diluted to the indicated concentration with E. coli not expressing pos-1 dsRNA. Animals were allowed to lay a brood overnight (in the absence of dsRNA), and, 24 h later, the percentage of these F1 embryos that hatched were scored (n = 4 ± SD). rde-4 is required for the initiation of RNAi silencing (22). The rde-4 allele used was ne301. See Fig. S2 for nrde-1 and nrde-4 exposed to pos-1 dsRNA data.

Fig. 2.

The Ago NRDE-3 promotes RNAi inheritance by acting in the nuclei of inheriting progeny. (A) unc-1 is 1.2 cM from nrde-3. nrde-3(gg066) males were crossed with unc-1(e538) hermaphrodites. Non-Unc (cross) progeny (defined as P0) were exposed to dpy-11 dsRNA. F1 embryos were isolated and grown on E. coli not expressing dpy-11 dsRNA. F1 progeny were scored for Dpy/non-Dpy phenotypes and Unc/non-Unc phenotypes. The nrde-3(gg066) genotype of the predicted nrde-3(+/+) and nrde-3(−/−) animals was verified by PCR analysis. eri-1(mg366) animals exhibit an enhanced RNAi phenotype (44). An eri-1(mg366) background was used to facilitate phenotypic analysis. Note that 41/46 unc-1/+ (nrde-3/+) F1 animals were not Dpy, indicating that nrde-3(gg066) is semidominant in this assay. (B) Animals of the indicated genotypes were exposed to dpy-11 dsRNA. Representative light images of F1 L3 progeny are shown. In all conditions, >98% of animals exhibited phenotypes similar to that shown in images.

Fig. 3.

RNAi induces the heritable expression of siRNAs, which bind NRDE-3. (A) NRDE-3::FLAG::GFP–expressing animals were exposed to dsRNA. NRDE-3::FLAG::GFP was immunoprecipitated with anti-FLAG antibody (Sigma) at the indicated times and from the indicated generations. NRDE-3 coprecipitating RNA was isolated, converted to cDNA, and quantified by quantitative real-time PCR coupled with a TaqMan probe set (Materials and Methods). From each sample, the amount of immunoprecipitated NRDE-3 was determined by quantitative Western blotting, and data were normalized to NRDE-3 levels and expressed as fold enrichment [(dpy-11 RNAi)/(no RNAi)]. Signal detected in F4 embryos was arbitrarily defined as one (n = 3 ± SD). Similar results were observed when P0 embryos not exposed to dpy-11 dsRNA were defined as one (Fig. S6). emb, embryo. (B) Animals of the indicated genetic backgrounds were exposed to dpy-11 dsRNA, and NRDE-3 or NRDE-3(*PAZ) was immunoprecipitated from the L4 progeny of these animals. dpy-11 siRNAs were quantified as described in A. The level of dpy-11 siRNA bound to WT NRDE-3 in WT animals was defined as one. Data were normalized to the amount of NRDE or 3/NRDE-3(*PAZ) immunoprecipitated from each sample (n = 3 ± SEM). The rde-1 allele used was ne219.

Fig. 4.

Nuclear RNAi promotes siRNA perdurance in the inheriting generation. (A) Animals of the indicated genotypes expressing NRDE-3::FLAG::GFP were exposed to dpy-11 dsRNA. NRDE-3 was immunoprecipitated at the indicated times and indicated generation, and dpy-11 siRNAs were quantified as described in Fig. 3A. Data were normalized to the amount of NRDE-3 immunoprecipitated from each sample (n = 3 ± SEM). siRNA signal bound to WT NRDE-3 in WT worms was defined as one.

Fig. 5.

Nuclear RNAi directs heritable H3K9me. (A) Animals were exposed to ± dpy-11 dsRNA and subjected at the indicated times to ChIP of H3K9me. F1 embryos/animals were obtained from P0 animals exposed to dpy-11 dsRNA for 70 h. Extracts were cross-linked with formaldehyde, and H3K9me was immunoprecipitated with anti-H3K9me antibody (Millipore). Coprecipitating DNA was quantified by quantitative PCR. Data were normalized to coprecipitating eft-3 DNA. H3K9me levels in P0 animals (−RNAi) at the 0 time point was defined as one. Data are expressed as a ratio of H3K9me signal ± RNAi (n = 3 for 0–64 h; n = 2 for 72–120 h; data points ± SE). Animals from the 72-h, 96-h, and 120-h time points were grown on 5-fluoro-2′-deoxyuridine (F0503; Sigma) to prevent F1 (or F2) animals from interfering with the experiment. (B) Animals of the indicated genotypes were exposed to dpy-11 dsRNA. L4 progeny (equivalent to 48-h time point) were subjected to H3K9me ChIP. Data were normalized to eft-3 coprecipitating DNA, and dpy-11 H3K9me signal in WT animals not exposed to RNAi was defined as one. Data are expressed as a ratio of H3K9me signal ± RNAi (n = 3 ± SEM).