P Granules Protect RNA Interference Genes from Silencing by piRNAs

Abstract

P granules are perinuclear condensates in C. elegans germ cells proposed to serve as hubs for self/non-self RNA discrimination by Argonautes. We report that a mutant (meg-3 meg-4) that does not assemble P granules in primordial germ cells loses competence for RNA-interference over several generations and accumulates silencing small RNAs against hundreds of endogenous genes, including the RNA-interference genes rde-11 and sid-1. In wild type, rde-11 and sid-1 transcripts are heavily targeted by piRNAs and accumulate in P granules but maintain expression. In the primordial germ cells of meg-3 meg-4 mutants, rde-11 and sid-1 transcripts disperse in the cytoplasm with the small RNA biogenesis machinery, become hyper-targeted by secondary sRNAs, and are eventually silenced. Silencing requires the PIWI-class Argonaute PRG-1 and the nuclear Argonaute HRDE-1 that maintains trans-generational silencing of piRNA targets. These observations support a "safe harbor" model for P granules in protecting germline transcripts from piRNA-initiated silencing.

Keywords: Argonautes; P granules; RNA-mediated interference; epigenetic silencing; piRNAs.

Fig. 1:. Segregation of P granules in wild-type and meg-3 meg-4 embryos

Schematics of C. elegans embryos at successive stages of development from the 1-cell zygote to the first larval stage post hatching. RNA polymerase II activity is repressed in the P lineage until gastrulation when P₄ divides to generate Z2 and Z3. In wild-type, P granules (green dots) are segregated preferentially with the germ plasm (lighter green color) to the P lineage that gives rise the primordial germ cells Z2 and Z3. In meg-3 meg-4 mutants, P granules are partitioned to all cells and are eventually dissolved/turned over. Germ plasm, however, segregates normally in meg-3 meg-4 mutants. Despite lacking maternal P granules, meg-3 meg-4 mutants assemble perinuclear P granules de novo during late embryogenesis and into the first larval stage (Wang et al., 2014).

Fig. 2:. meg-3 meg-4 mutants lose competency for RNA-interference and are defective in the production of secondary siRNAs.

A. Graph showing the percentage of viable embryos laid by hermaphrodites of the indicated genotypes upon treatment with pos-1 dsRNA. First two bars depict the embryonic viability from populations of ^~20 hermaphrodites fed starting at the L1 stage (each dot represents an experiment performed on a distinct population). On average, roughly 200 embryos were scored per RNAi experiment. The following two bars represent the percent viable progeny of motherŝ16 hours following injection with 200 ng/uL of pos-1 dsRNA (each dot represents the progeny of a single injected young adult hermaphrodite that laid more than 15 embryos). The last three bars represent viable progeny from M2Z2, M1Z0, and M0Z1 hermaphrodites fed starting at the L4 stage (each dot represents the progeny of a single hermaphrodite that laid more than 15 embryos). The “M” and “Z” designations refer to the number of wild-type meg-3 meg-4 alleles present in the mother (M) or hermaphrodite (Z) tested for RNAi. Bar height represents the mean; error bars represent the standard deviation. P-values were calculated using an unpaired t-test. B. Graph showing the percentage of viable embryos among broods (^~12 mothers) laid by newly generated meg-3 meg-4 hermaphrodites fed with bacteria expressing pos-1 dsRNA (from L4 stage). Three independently derived strains are shown. “Generation” refers to the number of generations since the meg-4 gene was deleted by genome editing in the starting strain carrying only a meg-3 deletion. See Fig. S1C for RNAi sensitivity of three sibling strains carrying only the original meg-3 deletion. See Fig. S1D for CRISPR breeding scheme. C. Genome browser view of sRNA reads mapping to the pos-1 locus in adult hermaphrodites of indicated genotypes fed with bacteria expressing a dsRNA trigger (red in figure) against a central region of the pos-1 locus. D. Graphs showing the abundance of sRNA reads mapping to the pos-1 locus in adult hermaphrodites of the indicated genotypes fed pos-1 RNAi. The upper panel shows primary sRNAs (directly derived from the ingested trigger), the bottom graph shows all sRNAs (both primary and secondary) from phosphatase treated library samples. Bar height represents the mean; error bars represent the standard deviation; p-values were calculated using an unpaired t-test.

Fig. 3:. meg-3 meg-4 mutants misregulate sRNAs that target hundreds of loci.

A. Scatter plot comparing sRNA abundance in wild-type (X-axis) and meg-3 meg-4 ^#1(Y-axis) hermaphrodites. Each dot represents an annotated locus in the C. elegans genome. Red dots represent loci with significantly upregulated or downregulated sRNAs comparing two biological replicates each for wild-type and meg-3 meg-4 ^#1. B. Pie chart showing the 619 genes with misregulated sRNAs in meg-3 meg-4 strains categorized according to the type of sRNAs that target these genes in wild-type. Note that 49.4% of these sRNAs are classified as HRDE-1-associated (Buckley et al., 2012). C. Venn diagrams showing the overlap between loci with upregulated or downregulated sRNAs in meg-3 meg-4 mutants and loci targeted by sRNAs that co-immunoprecipitate with HRDE-1 and CSR-1 (Buckley et al., 2012; Claycomb et al., 2009). D. Bar graph showing the average log2 fold difference in sRNA abundance for the indicated loci in the four meg-3 meg-4 strains compared to wild-type. The log2 fold change represents the average of two biological replicates for each genotype. Last grouping shows the mRNA abundance for each gene in the meg-3 meg-4 ^#1 adults as determined by RNAseq from two biological replicates. E. Venn diagrams showing the overlap between loci with downregulated sRNAs in meg-3 meg-4 mutants and loci with downregulated sRNAs in rde-11 mutants.

Fig. 4:. meg-3 meg-4 phenotypes are suppressed by loss-of-function mutations in hrde-1 and prg-1

A. Browser view of the rde-11/B0564.2 locus showing normalized sRNA reads in hermaphrodites of the indicated genotypes. B. Graph showing the percentage of viable embryos among broods laid by hermaphrodites of the indicated genotypes and fed bacteria expressing pos-1 dsRNA from the L1 stage. Each dot represents an independent RNAi experiment performed with a cohort of 15–20 hermaphrodites allowed to lay eggs for 1–2 hours. On average, over 200 embryos were scored per RNAi experiment. Note for prg-1; meg-3 meg-4, values were normalized to the levels of embryonic lethality the strain exhibits under non-RNAi conditions. Bar height and error bars represent the mean and standard deviation respectively; p-values were obtained using an unpaired t-test. C. Quantification of smFISH signal normalized to the average wild-type value. Each dot represents a single gonad. Center bar represents the mean and error bars indicate the standard deviation. P values were obtained through an unpaired t-test. See Fig S3G for regions quantified.

Fig. 5:. Localization of epigenetic factors during embryonic development in wild-type and meg-3 meg-4 mutants.

Photomicrographs of (A) wild-type and (B) meg-3 meg-4 embryos at the indicated developmental stages expressing fluorescently-tagged nuage proteins and HRDE-1. All tags were introduced at the endogenous locus by genome editing. Last column shows close-ups of a single primordial germ cell (PGC) at comma-stage. Image acquisition and display values were adjusted for each protein and therefore levels cannot be compared between proteins. Wild-type and meg-3 meg-4 panels for each fusion are comparable, except for panels with asterisks which were adjusted to visualize the much lower levels of fluorescence in meg-3 meg-4 mutants. See Fig. S4F for non-adjusted panels. Scale bars are 4 μm (embryo panels) and 2 μm (PGC panels).

Fig. 6:. Localization of rde-11 and sid-1 transcripts in wild-type and meg-3 meg-4 primordial germ cells.

A. Photomicrographs of primordial germ cells in comma-stage embryos hybridized to fluorescent probes to visualize rde-11 and sid-1 transcripts (yellow). Embryos also express GFP::PRG-1 fusion (green). Arrows point to nuclear transcripts. Stippled lines indicate cell outline. Scale bar is 2 μm. B. Graph showing the % of rde-11 and sid-1 transcripts in GFP::PRG-1 granules in wild-type vs meg-3 meg-4 primordial germ cells. Each dot represents one embryo. Error bars represent the standard deviation. P-values were obtained through an unpaired t-test. C. Graph showing the number of rde-11 and sid-1 transcripts in wild-type and meg-3 meg-4 primordial germ cells. Each dot represents one embryo. Mid bar represents the mean while error bars indicate the standard deviation. P-values were obtained through an unpaired t-test. A significant p-value was obtained between mRNA number in wild-type and meg-3 meg-4 for rde-11 mRNA but was not for sid-1 mRNA due to a single outlier.

Fig. 7:. Model illustrating the fate of rde-11 transcripts in wild-type and meg-3 meg-4 primordial germ cells.

In wild-type primordial germ cells, rde-11 transcripts (black) are transcribed by RNA polymerase II (blue), and accumulate in P granules (green) upon exit from the nucleus. In P granules, rde-11 transcripts are targeted by PRG-1/piRNA complexes (green) which slows their release into the cytoplasm. Few transcripts reach the cytoplasm (yellow) where mutator activity triggers production of secondary sRNAs (red) that load on HRDE-1 (pink). In meg-3 meg-4 primordial germ cells, rde-11 transcripts immediately disperse in the cytoplasm upon exit from the nucleus. In the cytoplasm, rde-11 transcripts are targeted by PRG-1/piRNA complexes and by mutator activity which triggers the production of secondary sRNAs. The secondary sRNAs are loaded on HRDE-1 stimulating its nuclear accumulation leading to silencing of the rde-11 locus.