A Family of Argonaute-Interacting Proteins Gates Nuclear RNAi

Abstract

Nuclear RNA interference (RNAi) pathways work together with histone modifications to regulate gene expression and enact an adaptive response to transposable RNA elements. In the germline, nuclear RNAi can lead to trans-generational epigenetic inheritance (TEI) of gene silencing. We identified and characterized a family of nuclear Argonaute-interacting proteins (ENRIs) that control the strength and target specificity of nuclear RNAi in C. elegans, ensuring faithful inheritance of epigenetic memories. ENRI-1/2 prevent misloading of the nuclear Argonaute NRDE-3 with small RNAs that normally effect maternal piRNAs, which prevents precocious nuclear translocation of NRDE-3 in the early embryo. Additionally, they are negative regulators of nuclear RNAi triggered from exogenous sources. Loss of ENRI-3, an unstable protein expressed mostly in the male germline, misdirects the RNAi response to transposable elements and impairs TEI. The ENRIs determine the potency and specificity of nuclear RNAi responses by gating small RNAs into specific nuclear Argonautes.

Keywords: 22G-RNA; Argonaute; NRDE-3; RNA interference; RNAi; piRNA; trans-generational epigenetic inheritance.

Graphic abstract

Figure 1. ENRI-1 and ENRI-2 are uncharacterized NRDE-3 interacting proteins

(A) Table indicating the proteins that were IPed in the indicated genetic backgrounds. Peptide coverage of the listed interactors is indicated in percentage, as well as the number of replicates in which the interactor was found. (B) Percent identity between each protein is listed, and predicted disordered regions marked with a black wavy line above each protein. Protein schematics are to scale. (C) Input and FLAG IPs from wild-type or lap::nrde-3 embryos were probed with an antibody to ENRI-1 to confirm the interaction observed in MudPIT. LAP::NRDE-3 was detected with a GFP antibody. (D) Wild-type and enri-2::3xflag embryo lysates were subjected to NRDE-3 IP using pre-immune (PI) serum as a negative control. Input and IPs were probed with antibodies to NRDE-3, ENRI-1, and FLAG. (E) Table depicting ubiquitin-proteasome components found to interact with ENRI-1 and ENRI-2 and those shared with LAP::NRDE-3 (bolded). See also Figure S1, Table S1 and Table S2.

Figure 2. ENRI-1 and ENRI-2 suppress nuclear RNAi

(A) Western blots of lysates from wild-type and enri-1 knock-out embryos probed with antibodies to ENRI-1 and tubulin. (B) Schematic detailing the contributions of cytoplasmic and nuclear RNAi and expected outcomes. (C) Quantification of animals that survive lir-1 RNAi (n >15 biological replicates). The enri-1;enri-2 double mutant comprised alleles enri-1(tag1601) and enri-2(tag1609). (D) Blistering of animals was scored following bli-1 RNAi (n > 3 biological replicates). (E) Complete paralysis of animals exposed to unc-22 RNAi was quantified (n > 6 biological replicates). (F) Animals with a coiler phenotype were quantified (n > 3 biological replicates). For all RNAi assays: Tukey’s multiple comparisons test was conducted with p-value: 0.05 – 0.0332 (*), 0.0333 – 0.0021 (**), 0.0022 – 0.0002 (***), < 0.0001 (****), and data is depicted as mean ± SD. See also Figure S2 and S3.

Figure 3. ENRI-1 and ENRI-2 interact with unloaded NRDE-3.

(A) Total lysates and NRDE-3 IPs from enri-2::3xflag, eri-1(mg366), and enri-2::3xflag;eri-1(mg366) embryos were probed with NRDE-3, FLAG and ENRI-1 antibodies. (B) Top: Schematic of the 2’O-methyl pull-down. Bottom: Input and pull-down fractions were probed with FLAG and ENRI-1 antibodies. (C) RNA extracted from FLAG IPs performed in transgenic lap::nrde-3 and enri-1::3xflag embryos was probed for 22G-RNAs mapping to siR-26-1 and X-cluster. FLAG antibody was used to detect both LAP::NRDE-3 and ENRI-1::3xFLAG. Dashed line indicates an unrelated lane was removed from the siR-26-1 northern blot. (D) RNA extracted from NRDE-3 and ENRI-2::3xFLAG IPs was probed for the 22G-RNAs mapping to the X-cluster. PI = pre-immune serum. NRDE-3 antibody and FLAG antibody were used to detect NRDE-3 and ENRI-2::3xFLAG respectively. (E) Top: schematic outlining the fragments of NRDE-3 fused to GST. Bottom: 50% of the pull-down was loaded onto a 10% gel for western blot analysis using an anti-6xHIS antibody. The other 50% was loaded onto a gel for Coomassie staining (Figure S3E, F). Each pull-down was performed at least 4 times and a representative blot for each is shown. See also Figure S3.

Figure 4. ENRI-1 and ENRI-2 localize to embryos and prevent premature nuclear translocation of NRDE-3.

(A) ENRI-1::GFP localization in a dissected young adult hermaphrodite germline. (B) ENRI-2::GFP expression in various embryo stages. (C) ENRI-2::GFP expression in the young adult hermaphrodite. Germline outline is marked by dashed lines. (D) Localization of LAP::NRDE-3 in developing embryos in indicated genetic backgrounds. Scale bars for reference.

Figure 5. Loss of the enri genes leads to deregulation of NRDE-3-associated small RNAs

(A) (B) MA plot (where M is the difference between log intensities and A is the average log intensity for a dot in the plot) of (A) input or (B) FLAG-IP 22G-RNA reads mapping antisense to individual genes from LAP::NRDE-3 (enri-1(tag1609);enri-2(tag1609)) embryos compared to 22G-RNA reads from LAP::NRDE-3 (WT). (C) (D) MA plot of (C) input or (D) FLAG-IP 22G-RNA reads mapping antisense to annotated transposable elements from LAP::NRDE-3 (enri-1;enri-2) embryos compared to 22G-RNA reads from LAP::NRDE-3 (WT). Shown in red are significantly upregulated 22G-RNAs and shown in blue are significantly downregulated 22G-RNAs. Only 22G-RNAs that displayed log2(fold-change) > 1, log2(average reads) > 4 and padj < 0.05 were considered significantly changed from wild-type. (E) Heat map of all 22G-RNAs mapping to individual genes recovered in LAP::NRDE-3 IPs. (F) Heat map of 22G-RNAs mapping to all annotated germline-specific RNAi targets recovered in LAP::NRDE-3 IPs. (G) Heat map of 22G-RNAs mapping to all annotated soma-specific RNAi targets recovered in LAP::NRDE-3 IPs. See also Figures S4, S5, and S6.

Figure 6. ENRI-3: Small RNA sorting in the male germline

(A) Left: ENRI-3::GFP localization in the developing gonad of an L4 hermaphrodite. Region of spermatogenesis indicated as “male germline”. Right: ENRI-3::GFP localization in young adult hermaphrodites. White arrowhead indicates sperm. Bright signal to the right is likely residual bodies (anucleate cell bodies that detach from maturing spermatids). (B) ENRI-3::GFP, HIS-58::mCherry, and HIS-72::mCherry localization in adult male germline. White arrowhead indicates residual body. (C) and (D) MA plots of total 22G-RNAs mapping to individual genes in a (C) enri-2(tag1609);enri-3(qe19) double mutant or a (D) enri-3(qe17) mutant. (E) (F) MA plots of total 22G-RNAs mapping to annotated transposable elements in a (E) enri-2(tag1609);enri-3(qe19) double mutant or (F) enri-3(qe17) mutant. A filter was set such that only 22G-RNAs that displayed log2(fold-change) > 1, log2(average reads) > 4 and padj < 0.05 were considered significantly changed from wild-type. (G) A table indicating the number of worms displaying the indicates phenotypes. Worms were grown at 20°C for two generations and scored at the gravid adult stage. Dpy = dumpy, unc = uncoordinated, p.v. = protruding vulva (H) The brood size of enri mutants in the oma-1(zu405) background was scored for 8 generations following injection of P0s with oma-1 dsRNA (50ng/ul). See also Figure S6 and S7.

Figure 7. Model of ENRI function.

(A) In the soma, ENRI-1/2 gate NRDE-3 by preventing precocious and inappropriate loading of NRDE-3 with 22G-RNAs produced downstream of the 21U-RNA and/or 26G-RNA pathways. ENRI-1/2 association with NRDE-3 could be regulated by PTMs. (B) We propose that ENRI-1 and ENRI-2 sequester NRDE-3 in the early embryo to prevent loading of 22G-RNAs downstream of germline RNAi pathways. (C) We propose that ENRI-3 gates one or many of the Argonautes acting in the germline and loss of ENRI-3 impacts multiple pathways leading to impaired inheritance of RNAi and activation of transposons.