Identification of piRNA Binding Sites Reveals the Argonaute Regulatory Landscape of the C. elegans Germline

Abstract

piRNAs (Piwi-interacting small RNAs) engage Piwi Argonautes to silence transposons and promote fertility in animal germlines. Genetic and computational studies have suggested that C. elegans piRNAs tolerate mismatched pairing and in principle could target every transcript. Here we employ in vivo cross-linking to identify transcriptome-wide interactions between piRNAs and target RNAs. We show that piRNAs engage all germline mRNAs and that piRNA binding follows microRNA-like pairing rules. Targeting correlates better with binding energy than with piRNA abundance, suggesting that piRNA concentration does not limit targeting. In mRNAs silenced by piRNAs, secondary small RNAs accumulate at the center and ends of piRNA binding sites. In germline-expressed mRNAs, however, targeting by the CSR-1 Argonaute correlates with reduced piRNA binding density and suppression of piRNA-associated secondary small RNAs. Our findings reveal physiologically important and nuanced regulation of individual piRNA targets and provide evidence for a comprehensive post-transcriptional regulatory step in germline gene expression.

Figure 1. PRG-1 CLASH identifies piRNA-target chimeras in C. elegans

(A) CLASH workflow. PRG-1 (gray oval), piRNA (red line), and mRNA (black line). Linkers indicated as blue (3’) and purple (5’) rectangles. (B) Silver stain analysis of GFP IP complexes (red arrow) purified from wild type (WT) and gtf∷prg-1 worms with or without UV irradiation. (C-D) Autoradiography (left) and western blot (right) of FLAG IP complexes released by TEV cleavage after GFP IP. Panels C and D show independent samples. Red line shows the region excised for library preparation. (E) Polyacrylamide gel showing library products (red line) isolated for sequencing. (F) Summary of combined data from two CLASH replicates. (G) Normalized CLASH counts per gene for soma-specific genes, germline-specific genes, and both. CLASH reads normalized as described in STAR Methods. (H) Predicted binding energies (ΔG, kcal/mol) between piRNAs and target sites identified by CLASH (red) or between randomly matched pairs (blue). (I) Box plots of CLASH counts per target site with increasing binding energy or piRNA abundance (low, 0–33%; medium, 33–66%; high, 66–100%). Median, solid black line. Significant differences between groups indicated by p-values. See also Figure S1.

Figure 2. 22G-RNAs peak at the center and ends of piRNA binding sites

(A–H) 22G-RNA 5’ ends cloned from wild-type (red) or prg-1 (blue) worms mapped at single-nucleotide resolution to an extended 40-nt window around CLASH-defined piRNA target sites. Each plot is centered on a 21-nt piRNA, shown schematically. WAGO (A, C, E) and CSR-1 (B, D, F) targets analyzed separately. All hybrids (A and B), hybrids with ΔG < – 20 kcal/mol (C and D), hybrids from the high abundance piRNAs with ΔG < −20 kcal/mol (E and F), and control target regions >100-nt from the defined piRNA target sites (G and H). See also Figure S2.

Figure 3. piRNAs target with miRNA-like seed and 3’ supplementary base pairing

(A) Heat maps of Watson-Crick base-pairing (black pixels) at each piRNA position for piRNA-mRNA chimeras detected at least 5 times (left), and for negative control (random target) sequences. (B) Chart showing the percentage of CLASH reads (piRNA-target chimeras; black) with complementarity at the indicated positions. Approximately 70% of piRNA-target interactions possess the tested complementarities. Target sequences with shuffled dinucleotides served as control (gray). (C) A 4-mer sliding window search for perfect Watson-Crick base pairing between piRNAs and CLASH-defined targets. (D) Ratios of G:C (red) or A:T (blue) base pairing in piRNA-mRNA duplexes, after deducting ratios from random control. (E) Percent of seed-matched target sites (left) with the indicated nucleotide at target position 1 (t1; opposite piRNA position 1). Randomized target sequences with shuffled dinucleotides (center) and trinucleotides (right) serve as controls. Data expressed as mean ± 2 s.e.m. from two replicates. See also Figure S3.

Figure 4. Seed and 3’ supplementary pairing are required for silencing

(A)anti-gfp piRNA (red) and single-nucleotide mismatches (blue) from positions 2 to 21 on the piRNA target site in cdk-1∷gfp (black). (B and C) Graphs of the fraction of GFP-positive worms in the presence of anti-gfp piRNA with single-nucleotide mismatches at the indicated piRNA positions (numbers). Ten worms were randomly picked for fluorescence microscopy at the F2 generation (B) and later generations F4, F6, and F8 (C). Data expressed as mean ± 2 s.e.m. of three experiments. (D) Western blots of CDK-1∷GFP in F4 and F8 worms with single-nucleotide mismatches at each piRNA position (from C). Negative control worms without anti-gfp piRNA. Positive control worms with fully match anti-gfp piRNA. (E) Schematic of 22G-RNAs targeting gfp in F4 cdk-1∷gfp worms with the indicated single-nucleotide mismatches (m2 = position 2 mismatch, etc.). Positions from 5’, central, and 3’ regions of the piRNA were randomly chosen for analysis. Scale bar, 5 reads per million. (F) Schematic of 22G-RNAs targeting gfp in cdk-1∷gfp worms with the with the indicated single-nucleotide mismatches (m3, m8, or m18) at the F2 and F8 generations. Scale bar, 5 reads per million. See also Figure S4.

Figure 5. 21ur-4863 and 21ux-1 suppress xol-1 function in sex determination

(A) Diagram of silent mutations at positions t2, t11, and t14 of the 21ux-1 target site in xol-1 (green) to create single-nucleotide mismatches with 21ux-1. (B) Western blot (anti-FLAG) of GFP∷FLAG∷XOL-1 levels in gfp∷flag∷xol-1 transgenic worms (+) with intact 21ux-1 (WT), 21ux-1 deletion, or xol-1 silent mutations at position t2, t11, or t14. N2 worms serve as negative control (−). (C) Bar graphs of percent viable progeny of WT, prg-1 loss of function, 21ux-1 deletion, or xol-1 single-nucleotide mismatch (t2, t11, t14) worms treated with sex-1(RNAi). n > 150 progeny per experimental group. Data expressed as mean ± 2 s.e.m. of three experiments. (D) DIC images (upper panel) of typical hermaphrodite, male, and pseudomale worms. Bar graphs (lower panel) show the percent pseudomale in WT, prg-1, 21ux-1 deletion, and xol-1 single-nucleotide mismatch (t2, t11, t14) worms treated with sex-1(RNAi). n > 100 progeny per experimental group. Data expressed as mean ± 2 s.e.m. of three experiments. (E) Distribution of chimeric xol-1 reads (red) identified by CLASH, and the distribution of xol-1 22G-RNAs (blue) in prg-1 mutant and WT worms. Locations of 21ur-4863 (upper) and 21ux-1 target sites in xol-1 gene indicated by inverted black triangles. Sequences and base pairing (right) of 21ur-4863:xol-1 (upper) and 21ux-1:xol-1 (lower) chimeras. piRNA expression level, number of chimeric reads, and binding energy (ΔG, kcal/mol) indicated above each chimera. Distribution of 22G-RNAs at single-nucleotide resolution shown below each chimera. (F) Bar graph of xol-1 mRNA levels in WT, prg-1,21ur-4863 deletion, and 21ux-1 deletion worms measured by RT-qPCR. actin mRNA served as the internal control. Data expressed as mean ± s.d. of three experiments. (G) Western blot (anti-FLAG) of GFP∷FLAG∷XOL-1 (top) levels in WT, 21ur-4863 deletion, and 21ux-1 deletion worms. Alpha-tubulin (bottom) was probed as a loading control. (H and I) Bar graphs of percent viable (H) and pseudomale (I) progeny of WT, prg-1, and 21ur-4863 deletion worms treated with sex-1(RNAi). n > 500 per experimental group. Data expressed as mean ± 2 s.e.m. of three experiments. See also Figures S5.

Figure 6. CSR-1 prevents piRNA binding to its targets

(A) Box-and-whisker plots of piRNA binding site density in 3,820 CSR-1 targeted mRNAs identified by CLASH in both WT and CSR-1^depleted worms. piRNA binding density expressed as the number of normalized CLASH reads per kilobase (target mRNA length) per million mapped reads (RPKM) in WT or CSR-1^depleted. Outliers were removed. Mean, median, and p-values are indicated. CLASH reads normalized as described in STAR Methods. (B) Scatterplot of unique piRNA binding site counts per gene for each CSR-1 target in CSR-1^depleted versus WT. For 1466 CSR-1 targets, the number of piRNA binding sites increased >2-fold after depleting CSR-1. A representative target, dhc-1, shown in panel (C). (C) Distribution of dhc-1 chimeric reads from WT and CSR-1^depleted worms. (D) Scatter plot of the change in mRNA abundance between WT and CSR-1^depleted worms versus the change of piRNA binding density for 3,820 CSR-1 targets. (E) Box plots of the change in mRNA expression levels between WT and CSR-1^depleted worms for 5 sets of genes with different levels of piRNA binding site changes. Median and mean indicated. Each set is significantly different from the prior one, as indicated by the p-values. See also Figures S6.

Figure 7. Model for a regulatory landscape of piRNAs in the C. elegans germline