The Caenorhabditis elegans HEN1 ortholog, HENN-1, methylates and stabilizes select subclasses of germline small RNAs

Abstract

Small RNAs regulate diverse biological processes by directing effector proteins called Argonautes to silence complementary mRNAs. Maturation of some classes of small RNAs involves terminal 2'-O-methylation to prevent degradation. This modification is catalyzed by members of the conserved HEN1 RNA methyltransferase family. In animals, Piwi-interacting RNAs (piRNAs) and some endogenous and exogenous small interfering RNAs (siRNAs) are methylated, whereas microRNAs are not. However, the mechanisms that determine animal HEN1 substrate specificity have yet to be fully resolved. In Caenorhabditis elegans, a HEN1 ortholog has not been studied, but there is evidence for methylation of piRNAs and some endogenous siRNAs. Here, we report that the worm HEN1 ortholog, HENN-1 (HEN of Nematode), is required for methylation of C. elegans small RNAs. Our results indicate that piRNAs are universally methylated by HENN-1. In contrast, 26G RNAs, a class of primary endogenous siRNAs, are methylated in female germline and embryo, but not in male germline. Intriguingly, the methylation pattern of 26G RNAs correlates with the expression of distinct male and female germline Argonautes. Moreover, loss of the female germline Argonaute results in loss of 26G RNA methylation altogether. These findings support a model wherein methylation status of a metazoan small RNA is dictated by the Argonaute to which it binds. Loss of henn-1 results in phenotypes that reflect destabilization of substrate small RNAs: dysregulation of target mRNAs, impaired fertility, and enhanced somatic RNAi. Additionally, the henn-1 mutant shows a weakened response to RNAi knockdown of germline genes, suggesting that HENN-1 may also function in canonical RNAi. Together, our results indicate a broad role for HENN-1 in both endogenous and exogenous gene silencing pathways and provide further insight into the mechanisms of HEN1 substrate discrimination and the diversity within the Argonaute family.

Figure 1. Methylation of 21U RNAs Requires C. elegans HEN1 Ortholog HENN-1.

A) HENN-1 is required for 21U RNA methylation. Endogenous (xkSi1) and germline-specific (xkSi2) expression of henn-1::gfp rescue 21U RNA methylation in henn-1(tm4477) mutant embryo. Total embryo RNA of the indicated genotypes was β-eliminated (βe +) or control treated (βe −) and probed for piRNA 21UR-4292. prg-1(tm872) lacks 21U RNAs and is included as a negative control. Below, ethidium bromide staining of 5.8S rRNA is shown. Additional 21U RNA northern blots are shown in Figure S3A. B) C. elegans miRNAs are unmethylated. Total embryo RNA was probed for miR-1. Variable intensity of 5.8S rRNA bands in embryo indicates unequal loading.

Figure 2. HENN-1 Stabilizes 21U RNAs.

A) Loss of henn-1 impairs 21U RNA accumulation in adult, embryo, and early larva. Levels of 21UR-1848 were assayed by Taqman qPCR in embryo and every four hours across development of wild-type and henn-1(tm4477) mutant animals at 25°C. Standard deviation is shown for biological triplicates. Taqman qPCR data for eight additional 21U RNAs are shown in Figure S4. B) Effects of loss of henn-1 are restricted to its small RNA substrates. Levels of miR-1 across development were assayed by Taqman qPCR. Standard deviation is shown for biological triplicates. Additional Taqman qPCR data for miRNAs are shown in Figure S5. C) Loss of henn-1 impairs Tc3 transposase silencing primarily in early L1 larva. Tc3 transposase mRNA levels were assayed by qPCR across development and normalized to mRNA levels of eft-2, an abundantly expressed housekeeping gene. prg-1(tm872) lacks 21U RNAs and is included as a positive control for Tc3 upregulation. Significant zero and four hour time points are expanded at right (*: P = 0.0251; **: P = 0.0250, two-tailed t-test). Standard deviation is shown for biological triplicates. E, embryo; hr, hour.

Figure 3. HENN-1 Selectively Methylates ERGO-1 Class 26G RNAs in an ERGO-1–Dependent Manner.

A) HENN-1 is required for ERGO-1 class 26G RNA methylation and stability. Total β-eliminated (βe +) or control treated (βe −) embryo RNA of the indicated genotypes was probed for ERGO-1 class 26G RNA 26G-O7. eri-1(mg366) lacks 26G RNAs and is included as a negative control. Asterisk indicates signal corresponding to cross-hybridization with unmethylated 22G RNAs. Below, ethidium bromide staining of 5.8S rRNA. Additional ERGO-1 class 26G RNA northern blots are shown in Figure S3B. B) ALG-3/ALG-4 class 26G RNAs are unmethylated. Total him-8(e1489) male RNA was probed for ALG-3/ALG-4 class 26G RNA 26G-S5. An additional ALG-3/ALG-4 class 26G RNA northern blot is shown in Figure S3C. C) 21U RNAs are methylated in a HENN-1-dependent manner in both female and male germlines. Total RNA of the indicated genotypes from fem-1(hc17) female or him-8(e1489) male was probed for female germline-enriched piRNA 21UR-4292 or male germline-enriched piRNA 21UR-5941, respectively.

Figure 4. ERGO-1 Is Required for Methylation of 26G RNAs.

A) ERGO-1 class 26G RNA 26G-O1 is unmethylated in the absence of ERGO-1. Total embryo wild-type (5 µg) or ergo-1(tm1860) (10 µg) β-eliminated (βe +) or control treated (βe −) RNA was probed for 26G-O1. B) Anti-ERGO-1 rabbit polyclonal antibody immunoprecipitates ERGO-1 complexes. ERGO-1 complexes were immunopurified from lysates of equalized protein concentration extracted from wild-type, henn-1(tm4477) mutant, or eri-1(mg366) mutant embryo. Aliquots of lysates and immunoprecipitates (RNA IP) were probed with anti-ERGO-1 antibody. ergo-1(tm1860) mutant lysate was run in parallel to ensure specificity of ERGO-1 detection (data not shown). C) ERGO-1 binds methylated and unmethylated 26G RNAs. Taqman RT-qPCR for the indicated ERGO-1 class 26G RNAs was performed on samples described in B. The eri-1(mg366) mutant lacks 26G RNAs and serves as a negative control to demonstrate specificity of 26G RNA detection by Taqman assay. Standard deviation is shown for technical duplicates. Results are representative of two independent RNA immunoprecipitation experiments.

Figure 5. HEN1 Stabilizes ERGO-1 Class, but Not ALG-3/ALG-4 Class, 26G RNAs.

A) Loss of henn-1 impairs ERGO-1 class 26G RNA accumulation at all stages. Levels of ERGO-1 class 26G RNA 26G-O3 were assayed by Taqman qPCR across development of wild-type and henn-1(tm4477) mutant animals at 25°C. Standard deviation is shown for biological triplicates. Taqman qPCR data for seven additional ERGO-1 class 26G RNAs are shown in Figure S8. B) ALG-3/ALG-4 class 26G RNAs are henn-1-independent. Levels of ALG-3/ALG-4 class 26G RNA 26G-S5 were assayed across the period of development in which ALG-3/ALG-4 class 26G RNAs are readily detectable. Standard deviation is shown for biological triplicates. Taqman qPCR data for two additional ALG-3/ALG-4 class 26G RNAs are shown in Figure S9. C) Loss of henn-1 may result in modest, sporadic defects in ERGO-1 class 26G RNA target silencing. Levels of eight target and two non-target mRNAs were assayed across development of wild-type and henn-1(tm4477) mutant animals at 25°C and normalized to eft-2. Expression in the henn-1(tm4477) mutant relative to wild-type is represented according to the red-green color scheme indicated in the right panel. Raw data is shown in Figure S10. E, embryo.

Figure 6. The henn-1 Mutant Exhibits Opposite RNAi Sensitivity Phenotypes in Soma and Germline.

A) henn-1(tm4477) mutant animals exhibit mildly enhanced somatic RNAi. Animals of the indicated genotype were plated as L1 larvae on lir-1 feeding RNAi diluted 1∶1 with empty vector (1/2 strength) and grown for 70 hours at 20°C. Data is quantified in part B. RNAi sensitivity data for knockdown of two additional somatic transcripts are shown in Figure S11. B) Endogenous expression of henn-1::gfp from xkSi1 rescues somatic RNAi sensitivity. Percent of animals reaching full size on lir-1 feeding RNAi of the indicated strength at 70 hours is plotted. N = 8 plates of >50 animals per strain. Standard deviation is shown. C) henn-1(tm4477) mutant animals exhibit defective germline RNAi. Brood size of animals plated at 20°C as L1 larvae on pos-1 feeding RNAi diluted 1∶2 with empty vector is plotted. N≥13 animals per strain. Mean and standard deviation are shown. RNAi sensitivity data for knockdown of four additional germline transcripts are shown in Figure S12. Alleles used in this figure: eri-1(mg366), prg-1(tm872), rde-4(ne301).

Figure 7. HENN-1 Is Broadly Expressed in C. elegans Germline.

A) The henn-1 mRNA expression profile is consistent with germline enrichment. Levels of henn-1 mRNA were assayed throughout development and normalized to eft-2 mRNA. Standard deviation is shown for biological triplicates. Non-normalized levels are shown in Figure S13A. B) HENN-1 is detected at all stages of development and in male. Lysates from animals of the indicated stages were probed with anti-HENN-1 rabbit polyclonal antibody. C) HENN-1 is abundant in hermaphrodite proximal germline and enriched in proximal oocyte nucleoplasm (inset). Extruded gonads of xkSi1; henn-1(tm4477) adult hermaphrodites were stained with anti-GFP mouse monoclonal and anti-HENN-1 rabbit polyclonal antibodies. D) HENN-1 is detectable in male proximal and distal gonad, with enrichment in residual bodies during spermatid maturation (inset). Extruded gonads of xkSi1; henn-1(tm4477) adult males were stained with anti-GFP and anti-HENN-1 antibodies. E) Expression of endogenous HENN-1 mirrors expression of HENN-1::GFP from transgene xkSi1. Extruded gonads of wild-type animals were stained with anti-HENN-1 antibody. F) Detection of HENN-1 proteins by immunostaining is specific. Extruded gonads of henn-1(tm4477) mutant animals were stained with anti-GFP and anti-HENN-1 antibodies. E, embryo.