PIWI associated siRNAs and piRNAs specifically require the Caenorhabditis elegans HEN1 ortholog henn-1

Abstract

Small RNAs--including piRNAs, miRNAs, and endogenous siRNAs--bind Argonaute proteins to form RNA silencing complexes that target coding genes, transposons, and aberrant RNAs. To assess the requirements for endogenous siRNA formation and activity in Caenorhabditis elegans, we developed a GFP-based sensor for the endogenous siRNA 22G siR-1, one of a set of abundant siRNAs processed from a precursor RNA mapping to the X chromosome, the X-cluster. Silencing of the sensor is also dependent on the partially complementary, unlinked 26G siR-O7 siRNA. We show that 26G siR-O7 acts in trans to initiate 22G siRNA formation from the X-cluster. The presence of several mispairs between 26G siR-O7 and the X-cluster mRNA, as well as mutagenesis of the siRNA sensor, indicates that siRNA target recognition is permissive to a degree of mispairing. From a candidate reverse genetic screen, we identified several factors required for 22G siR-1 activity, including the chromatin factors mes-4 and gfl-1, the Argonaute ergo-1, and the 3' methyltransferase henn-1. Quantitative RT-PCR of small RNAs in a henn-1 mutant and deep sequencing of methylated small RNAs indicate that siRNAs and piRNAs that associate with PIWI clade Argonautes are methylated by HENN-1, while siRNAs and miRNAs that associate with non-PIWI clade Argonautes are not. Thus, PIWI-class Argonaute proteins are specifically adapted to associate with methylated small RNAs in C. elegans.

Figure 1. Endogenous siRNA sensor design and validation.

(A) Diagram of GFP control (ubl-1::GFP) and siRNA sensor (ubl-1::GFP-siR-1-sensor) transgenes. Grey rectangles are exons. (B) Images show GFP expression in C. elegans containing control or siRNA sensor transgenes. Upper panel, GFP fluorescence in whole worms and embryos. Lower panels, antibody stained GFP in dissected germlines. (C) Images show GFP fluorescence in control- and siRNA sensor-transgenic C. elegans treated with vector or ergo-1 RNAi. (D) RNA blot assay of GFP mRNA levels for three biological replicates of C. elegans containing control or siRNA sensor transgenes. EtBr stained rRNAs are shown as a loading control. (E) RNA and protein blot assays for GFP from three biological replicates of siRNA sensor transgenic C. elegans treated with vector or ergo-1 RNAi. EtBr stained rRNAs and antibody stained Actin protein are shown as loading controls. (F) Relative GFP protein and mRNA levels from the ubl-1::GFP-siR-1-sensor transgene following ergo-1 or vector RNAi. Protein levels were quantified from Western blots and mRNA levels were measured using qRT-PCR. (G) Images show GFP fluorescence in control- and siRNA sensor-transgenic wild type and nrde-3 mutant C. elegans.

Figure 2. Small RNA formation from control and siRNA sensor transgenes.

(A) Size and 5′ nt distributions of GFP-derived small RNAs deep sequenced from C. elegans containing control or siRNA sensor transgenes. (B) Normalized reads (reads per million total reads) mapping to GFP mRNA from control- and siRNA sensor-transgenic C. elegans deep sequencing libraries. (C) Small RNA distribution across the control GFP transgene. (D) Small RNA distribution across the siRNA sensor transgene. Inset, RNA blot assay for 22G siR-1 from control- and siRNA sensor-transgenic C. elegans. EtBr stained tRNAs are shown as a loading control.

Figure 3. Sequence complementarity requirements for 22G siR-1-target recognition.

(A) Diagram of the wild type siRNA sensor (ubl-1::GFP-siR-1-sensor) and each of the target site mutants. Grey rectangles are exons. Images show GFP fluorescence in C. elegans containing wild type and mutant siRNA sensor transgenes. (B) Protein blot assay for GFP from wild type and target site mutant siRNA sensor transgenic C. elegans. Actin protein is shown as a loading control.

Figure 4. 26G siR-O7 acts in trans to trigger 22G siRNA formation from the X-cluster.

(A) X-cluster region small RNA and mRNA sequencing reads are displayed above gene models. Reads corresponding to potential 22G siR-1 loci are shown in red. (B) 26G siR-O7 region small RNA and mRNA sequencing reads are displayed above gene models. Reads corresponding to 26G siR-O7 are shown in yellow. (C) Alignment of 26G siR-O7 with the shaded region of the X-cluster shown in A. (D) RNA blot assays of small RNAs in wild type and K02E2.11 mutant C. elegans. EtBr stained tRNAs are shown as a loading control. (E) qRT-PCR assay of small RNA levels in wild type and K02E2.11 mutant C. elegans. Wild type = 1.0. Error bars display standard deviation from the mean for two biological replicates. (F) Images show GFP fluorescence in control- and siRNA sensor-transgenic wild type and K02E2.11 mutant C. elegans. (G) Alignment of 26G siR-O7 with the siRNA sensor transgene 22G siR-1 target site region. (H) Images show GFP fluorescence in control- and siRNA sensor-transgenic wild type and rde-2 and rrf-1 mutant C. elegans.

Figure 5. henn-1 is required for 22G siR-1 activity.

(A) Images show GFP fluorescence in control- and siRNA sensor-transgenic C. elegans treated with vector or henn-1 RNAi. (B) Protein blot assays for GFP from C. elegans containing control or siRNA sensor transgenes and treated with vector, ergo-1, or henn-1 RNAi. Actin protein is shown as a loading control. One of three biological replicates is shown. (C) Images show GFP expression in control- and siRNA sensor-transgenic wild type and henn-1 mutant C. elegans. Upper panel, GFP fluorescence in whole worms. Lower panel, antibody stained GFP in dissected germlines. (D) Protein blot assay for GFP in wild type or henn-1 mutants containing control or siRNA sensor transgenes. Actin protein is shown as a loading control. One of three biological replicates is shown.

Figure 6. RNA silencing defects in henn-1 mutants.

(A) RNA blot assays of small RNAs in wild type and henn-1 mutant embryos and adults. For embryos, one of three biological replicates is shown. EtBr stained tRNAs are shown as a loading control. (B) qRT-PCR assay of small RNA levels in wild type and henn-1 mutant embryos. Wild type = 1.0. Error bars display standard deviation from the mean for three biological replicates. P values are for comparisons to wild type. (C) qRT-PCR assay of small RNA target mRNA levels in wild type and henn-1 mutant embryos. Wild type = 1.0. Error bars display standard deviation from the mean for three biological replicates. P values are for comparisons to wild type. (D) RNA blot assays for miRNAs in wild type and henn-1 mutant L4 larvae. EtBr stained tRNAs are shown as a loading control. (E) qRT-PCR assay of small RNA levels in wild type and henn-1 mutant L4 larvae. Wild type = 1.0. (F) qRT-PCR assay of ALG-3/4 target mRNA levels in wild type and henn-1 mutant embryos. Wild type = 1.0. Data shown for two independent experiments. (G) Box plots display brood size per individual wild type or henn-1 mutant grown at either 20°C or 25°C. n = 20 (20°C) or n = 30 (25°C) individuals per strain. P values are for comparisons to wild type.

Figure 7. High-throughput sequencing of methylated small RNAs.

(A) The Log₂ ratio of normalized reads (reads per million total reads) for piRNAs and miRNAs (left plot), or ERGO-1 and ALG-3/4 class 26G siRNA loci (right plot) in small RNA high-throughput sequencing libraries from β-eliminated and untreated RNA isolated from L4 larvae. Data points within the shaded region correspond to small RNAs that are depleted in the β-eliminated library. Data points outside the shaded region correspond to small RNAs that are enriched in the β-eliminated library. (B) Ratio of normalized small RNA reads in β-eliminated to untreated RNA high-throughput sequencing libraries. (C) Phylogenetic tree of D. melanogaster, H. sapiens and C. elegans Argonautes. The predominant small RNA type each Argonaute binds is indicated. (D) Trimming and tailing of small RNAs is displayed as the proportion of small RNA deep sequencing reads that contain 3′ untemplated nucleotides relative to the combined number of reads lacking untemplated nucleotides and those containing 3′ untemplated nucleotides.