Fig 2: Immunolocalization of AGO1 (a), AGO5 (b), AGO2 (c), AGO4 (d), and AGO9 (e) at different representative meiotic stages in Arabidopsis meiocytes. Representative images of different meiotic stages in meiocytes as indicated in the left column: Leptotene (a1, b1, c1, d1, e1); Zygotene (a2, b2, c2, d2, e2); Pachytene (a3, b3, c3, d3, e3); Diplotene (a4, b4, c4, d4, e4); Diakinesis (B5), Metaphase I (a5, c5, d5, e5) Prophase II (a6, b6, d6, e6); Metaphase II (c6); Tetrad (a7, b7, c7, d7, e7). Immunostaining with antibodies is shown in red and counterstaining of nucleus and chromosomes with DAPI is shown in gray and indicated at the top of each image series. Bar indicates 10 µm and is shown as a reference of cell size

Fig 3: Analysis of miRNA TE targets in the uninuclear, binuclear, and mature pollen grain identified by PARE sequencing. A, Number of miR845a- and miR845b-targeted TEs identified in uninuclear and mature pollen grain PARE libraries using high-confidence parameters. B, Representative examples of miR845a targeting positions in TEs. The location of the two different PBSs is highlighted in gray. C, Distribution of 5'-ends of PARE reads around the confirmed miR845-targeted TEs cleavage site (located at coordinate 0 in the X-axis) in a 30-nt window. Gray zone represents the physical position covered by the bound miRNA. D, Distribution of 5'-ends of PARE reads around the confirmed miRNA-targeted TEs cleavage site (located at coordinate 0 in the X-axis) in a 100-nt window. Gray zone represents the physical position covered by the bound miRNA. E, High-confidence identified miRNA-targeted TEs in PARE libraries. Target TE and its associated miRNA are indicated together with a heatmap of the presence of the target site in the different number of unicellular or mature PARE libraries in different shades of gray, and the level of accumulation of PARE reads in a 20-nt window surrounding the miRNA target site compared to the total accumulation of PARE reads for that gene for each individual bioreplicate studied with a blue-red gradient. F, Accumulation size profile of TE-derived siRNAs of PARE identified miRNA-targeted TEs. G, Number of targeted genes (left) and TEs (right) in uninuclear and mature pollen PARE libraries. Values for each bioreplicate are shown. H, Total accumulation values (in reads per million) of TE-targeting miRNA families (shown in (B) and including miR845) in uninuclear and mature pollen sRNA libraries. P-value was obtained using a paired t test. I, Number of identified TE-miRNA target sites in the mature pollen grain enriched more than two-fold (black) or not enriched (white) in AGO5 and AGO1.

Fig 4: Analysis of miR845 dynamics and its target TEs during pollen development. A, Accumulation percentages for miR845a and b during pollen development. B, Sequence comparison of miR845a and miR845b with the 5' nucleotide highlighted. C, Percentage of the impact of miR845a and miR845b on total AGO1 and AGO5 immunoprecipitated miRNAs. D, Accumulation size profile of TE-derived siRNAs of predicted miR845-targeted TEs. E, Accumulation of 21-, 22-, and 24-nt sRNAs from miR845-targeted TEs during pollen development shown as percentages of total 20–25-nt sRNAs derived from those TEs. F, Accumulation profile of all TE-derived sRNAs. G, Accumulation of 21-, 22-, and 24-nt sRNAs from TEs during pollen development shown as percentages of total 20–25-nt easiRNAs.

Fig 5: Proposed model. CMV infection leads to the accumulation of its genomic RNAs which are recognized by the RNA silencing machinery producing 21- (mainly) and 22-nt vsiRNAs. These vsiRNAs are then incorporated into the endogenous antiviral machinery mediated by AGO2, AGO1, AGO5, and AGO7 and inhibited by the 2b protein. As a side effect of this antiviral activity, vsiRNAs are also incorporated into the endogenous RNA silencing pathways mediating the silencing of endogenous genes. Together with this, 2b binds to endogenous mRNA-derived sRNAs and is involved in the transcriptional regulation of these genes