Fig. 7: G:U hpRNAs are differently processed from traditional hpRNAs.

a Northern blot hybridization to detect antisense sRNAs from T2 hpEIN2[WT] and hpEIN2[G:U] Arabidopsis plants using sense RNA transcripts of the 200 bp EIN2 target sequence as probe (see Fig. 2 and Supplementary Fig. 4a for RNAi phenotypes). Asterisks indicate samples for sRNA deep sequencing. b Detection of antisense sRNAs (upper panel) and long dsRNA (lower panel) using the same EIN2 probe as in (a). Ethidium bromide-stained rRNA is used as the loading control. c β-elimination (NaIO4 treatment) assay confirming similar 3′-O-methylation between hpEIN2[WT] and hpEIN2[G:U]-derived siRNAs (If hydroxyls on the 3′-terminal ribose are unmethylated, NaIO4 oxidizes them to form an unstable dialdehyde that leads to β-elimination of the terminal nucleoside and an approximately 2-nt downward mobility shift). d Alkaline phosphatase (CIP) treatment of sRNAs. The same gel blot was sequentially hybridized with the sense EIN2, trans-acting siRNA255 (tasiR255), and miR168 probes. Note that CIP treatment resulted in slowed but more similar siRNA gel migration between the two hpRNA designs. Also, note that tasiR255 showed a greater gel mobility shift than miR168 (see the different gaps between the two short red lines that indicate the average position of the CIP-treated and untreated tasiR255 and miR168 bands. It was unclear why the intensity of tasiR255 band was markedly reduced upon CIP treatment). e Detection of hpEIN2[WT] and hpEIN2[G:U]-derived siRNAs together with 5′ labeled 21 and 24 nt (5′p-GUS21Me, 5′p-GUS24Me) or un-labeled 24 nt (5′OH-GUS24Me) synthetic GUS sRNA markers (Supplementary Data 1), and in vitro transcribed mono- (5′P), di- (5′PP) and tri- (5′PPP) phosphorylated EIN2 sRNA markers (see Methods). All marker samples were mixed with 2 µg total RNA of WT Col-0 before loading. The un-labeled GUS24 and EIN2 sRNA markers were visualized by hybridization with respective antisense oligonucleotide probes. Note that the gel migration of 24 nt sRNA markers was slowest for the non-phosphorylated 5′OH-GUS24Me and the fastest for the di- and tri-phosphorylated EIN2 markers. The Col/hp[WT] sample looked degraded on this gel but another gel was run to verify the sRNA pattern (Supplementary Fig. 11b). Source data are provided as a Source Data file.