Fig 2: Effects of FBW2 overexpression or loss of function in silencing mutants restores high-molecular-weight AGO1 complexes(A) Representative pictures of 27-day-old plants of Col-0, hyl1-2, hen1-6, ago1-27, and their crosses with fbw2-4 or 35S:3HA-FBW2 (FBW2OE) as indicated (based on two biological replicates; see also Figure S3A). Right: quantitative analysis of the experiment represented on the left, with n > 12 plants per genotype. °p < 0.05 and °°°p < 0.001 (Student’s t test) compared with Col-0. *p < 0.05, **p < 0.01, and ***p < 0.001 (Student’s t test) compared with the corresponding single mutant.(B and C) Western blot of protein extracts from the same seedlings as indicated in Figure S3A (for biological replicates see Figures 5 and S3–S5). CB staining was used as loading control. AGO1 signal was quantified by ImageJ, normalized to the corresponding CB. Numbers below the panel indicate relative to the corresponding mutants (hyl1-2, hen1-6 and ago1-27, respectively) set at 1.0. @ indicates hybridization with the corresponding antibodies.(D and E) Gel filtration analysis of AGO1-based RISC complexes in Col-0, hyl1-2, hy1-2 fbw2-4, and fbw2-4 13-day-old seedlings (a biological replicate is shown in Figure S4). Proteins of known molecular weight are shown on top of the blot. CB staining was used as loading control and “@” indicates hybridization with the AGO1 antibody. (E) Shown are protein and sRNA analysis of the input fraction prior to gel filtration. Methylene blue (MB) staining of the membrane was used as loading control.(F) sRNA analysis from even fractions, spanning the same range (from the GF of [D]). For this analysis, 10µg of the RNA per lane was loaded. The “@” symbol indicates hybridization with the indicated oligonucleotide probes.

Fig 3: AGO1 associates with COOLAIR and influences COOLAIR processing. (A) Schematic diagram showing FLC gene structure and COOLAIR transcripts. Black and gray boxes represent FLC and COOLAIR exons, respectively. Black and gray lines represent FLC and COOLAIR introns, respectively. The arrow indicates the transcription start site (TSS). Short black lines with letters underneath indicate positions of amplicons in qPCR amplification. (B) RIP–qPCR analyzing HA-AGO1 enrichment on nascent COOLAIR transcript. Wild-type Col-0 was used as background control. The x axis corresponds to the fragments shown in A. Data are mean ± SD (n = 3). (C) The ratio of proximal-to-distal isoforms of COOLAIR transcripts (refer to the schematic in A) in various genotypes relative to Col-0. Data are mean ± SEM (n = 3). (D) The splicing efficiency of distal intron (spliced/unspliced) determined through chromatin-bound RNA analysis. Data are normalized to Col-0. Data are mean ± SD (n = 3).

Fig 4: AGO1 promotes COOLAIR R-loop removal and FLC chromatic silencing. (A) DRIP–qPCR determining R-loop level in Col-0 and ago1-25. The number on the x axis is the distance to FLC TSS. Data are percentage of input normalized to a region without COOLAIR transcript (between 6,000 and 7,000 base pairs to FLC TSS). Data are presented as mean ± SEM (n = 3). (B) ChIP analysis of H3K4me1 level at FLC in various genotypes. The number on the x axis is the distance to FLC TSS. Data are mean ± SD (n = 3).

Fig 5: Comparison of AGO1 and AGO5 activity and loaded miRNAs during pollen development. A, Analysis of the cellular localization of AGO1 and AGO5–mCherry–AGO fusion proteins during pollen development. Drawings accompanying the images represent the localization of the different nuclei on each picture. B, Analysis of miRNA enrichment for AGO1 and AGO5 immunoprecipitated sRNAs compared to their respective input controls. Significance was assessed by means of a Mann–Whitney U test (**** denotes a significance of P < 0.00001). C, Principal component analysis of the small RNA libraries from different AGO immunoprecipitations and their input controls based on miRNA levels. Venn diagram shows the overlap between the miRNA families loaded in AGO1 and AGO5. D, 5′-nt enrichment for AGO-enriched miRNAs. E, Size distribution of miRNAs enriched in the AGO1 and AGO5 IPs. F, Accumulation of enriched miRNA families in AGO1 and AGO5. In (C)–(F) miRNAs were considered enriched if their accumulation was higher than two-fold their accumulation in the input library or if it was only detected in the AGO immunoprecipitated library. G, Venn diagram showing the overlap of miRNA families enriched more than two fold in the mature pollen grain (blue circle) and miRNA families enriched more than two-fold in AGO5 IP compared to AGO1 IP. Significance (P = 0.032807) was calculated using the Chi-square statistic with Yates correction.