Supplementary Figure 2: Signal and background constraints for mime-seq.

(a) Schematic of the domain structure of plant and animal HEN1. The Arabidopsis enzyme has two dsRNA binding domains that define its specificity for short duplexes. The animal methyltransferases, despite similar enzymatic activity, do not bind or act on dsRNA and in fact do not seem to share a common ancestor with Ath-HEN1. (b) Pie charts showing the proportion of miRNAs and other small RNAs sequenced before and after oxidation of RNA obtained from wild-type, non-trangenic C. elegans L1 larvae. While miRNAs are strongly depleted, the libraries generated after oxidation still contained 5-10% of reads corresponding to miRNAs, mostly representing 26 miRNAs (Suppl. Table 6). Given their precursor structure and association with other Argonautes, and the fact that they are depleted upon oxidation in henn-1(0) background (Suppl. Tables 3, 4, 5), these are endogenous targets of Cel-HENN-1 (see discussion and Suppl. Table 6). (c) Duplex structures of two of the miRNAs that are endogenously methylated by Cel-HENN-1. (d) Comparison of the relative abundances of all expressed mature miRNAs in Drosophila S2 cells depleted of endogenous Dme-Hen1 by CRISPR-Cas9 genome editing (S2-hen1^KO) and expressing Ath-HEN1, before and after oxidation treatment. Spearman correlation coefficient and associated p-Value are shown. Independent repeats = 1. (e) Depletion of endogenous RNA methyltransferase HENN-1 does not affect miRNA relative abundance in worms. Comparison of the relative abundances of all confidently detected (>10 cpm in the average of all untreated samples) mature miR strands in wt, non-transgenic L1 worms versus animals deficient for endogenous HENN-1. Spearman correlation coefficient and associated p-Value are shown. Independent repeats = 1. (f) Comparison of the relative abundances of all confidently detected (>10 cpm in the average of all untreated samples) mature miR strands in henn-1(0) worm L1s expressing rps-5^prom::Ath-HEN1, before and after oxidation treatment. Spearman correlation coefficient and associated p-Value are shown. In blue are miRNAs that are significantly depleted after oxidation (p-Value <0.001) and were not detected as enriched in any other profiling experiment in this study (see text and Supplementary Table 2). Independent repeats = 2. (g) Volcano plot of the data in (f) showing the respective p-Values for all fold-changes. Independent repeats = 2.