The AGO complexes were immunopurified by using peptide-specific antibodies and separated on 10% SDS-PAGE (upper panels). For AGO1, AGO2, and AGO5, preimmune antisera were used for the control purifications, and for AGO4, untransformed plant materials were used for the control purification. The proteins were visualized by silver staining. The positions of protein size markers, electrophoresed in parallel, are shown to the left of each gel. The AGO protein bands are indicated by solid arrowheads. Small RNAs were extracted from each AGO complex, analyzed by denaturing polyacrymide gels, and visualized by SYBR-gold staining (lower panels). The positions of RNA size markers are shown to the left.

(A) Size distribution of sequenced small RNAs in total RNA (Total) (Rajagopalan et al., 2006) and those bound by AGO1, AGO2, AGO4, and AGO5 complexes.
(B) Pie charts summarizing the annotation of small RNA populations in total RNA (Total) and those bound by the four AGO complexes. Some small RNAs can match more than one category of the sequences listed, so the sum of the numbers may be bigger than the input total number.
(C) Comparison between small RNA populations bound by the four AGO complexes. The numbers represent the unique small RNA reads that overlap between the AGO complexes.

(A) The relative frequency of each 5′ terminal nucleotide of small RNAs in total RNA (Total) (Rajagopalan et al., 2006) and those bound by AGO1, AGO2, AGO4, and AGO5 complexes.
(B) The relative nucleotide bias at each position of the small RNAs in total RNA (Total) and those bound by AGO1, AGO2, AGO4, and AGO5. The graphics were made by using WebLogo (Crooks et al., 2004). The sequence conservation at each position is indicated by the overall height of the stack of symbols (U, A, C, and G), while the relative frequency of each nucleotide is represented by the height of the corresponding symbol.
(C) The relative frequency of each 5′ terminal nucleotide of different categories of small RNAs in total RNA (Total) (Rajagopalan et al., 2006) and those bound by AGO1, AGO2, AGO4, and AGO5 complexes.

(A) Examples of miRNAs and miRNA*s that were specifically recruited by distinct AGO complexes. The 5′ terminal nucleotide of each small RNA is indicated in the parentheses. The relative abundance of each small RNA in total RNA (Total) and the AGO complexes is represented by its normalized reads per million sequences.
(B) tasiRNAs produced from TAS1a locus were recruited by distinct AGO complexes. The 5′ terminal nucleotide of each tasiRNA is indicated. The relative abundance of each tasiRNA in total RNA (Total) and the AGO complexes is represented by its normalized reads per million sequences. The solid arrowheads indicate the ends of each tasiRNA. The red line represents the miR173 target region. See Figure S6 for analyses on tasiRNAs from other loci.
(C) Detection of miRNAs and miRNA*s in total RNA, AGO1, AGO2, AGO5, and control immunoprecipitates. The northern blots were stripped and reprobed multiple times. A silver-stained gel is shown to indicate that comparable amounts of each AGO complex were used for RNA preparation. The solid arrowhead indicates the bands of AGO proteins. End-labeled 21 and 24 nt synthetic RNA oligos were electrophoresed in parallel and used as size markers.

(A) AGO1, AGO2, AGO5, and control immunoprecipitates, as indicated, were incubated with single-stranded ³²P-labeled 21 nt siRNAs bearing the indicated 5′ terminal nucleotides. Mixtures were irradiated with UV and resolved by 10% SDS-PAGE. The cross-linked products are indicated by solid arrowheads. Gels with shorter exposure (middle panels) are shown to indicate that equal amounts of siRNAs were added into each reaction. Silver-stained gels (bottom panels) are shown as controls for the proteins used in the crosslinking reactions.
(B) In the UV-crosslinking reactions containing AGO1 immunoprecipitates and labeled siRNAs with 5′terminal U (U21, upper panel) or AGO2 immunoprecipitates and labeled siRNAs with a 5′ terminal A (A21, lower panel), increasing amounts (0.02–20 pmols) of competitors, unlabeled U21 and A21, were added. The crosslinked products are indicated by solid arrowheads.
(C) The signals of crosslinked products from three repeats of the competition experiment described in (B) were quantified and the average (± standard deviation) values are shown.

(A) Schematic drawing of the structures of wild-type and mutant miRNA precursors (miR393b as an example). For miR391 and miR393b, the 5′ terminal nucleotides of both the miRNA and miRNA* were changed. For amiR-trichome, miR163.1 and miR166 g, the 5′ terminal U of each miRNA was changed to an A. For miR390a and miR163.2, the 5′ terminal A was replaced with a U in the mutant. In each mutant, when necessary, additional mutations were introduced at the position 19 of the miRNA* strands to maintain the stem-loop structures: for miR391, 163.1, 163.2, 166g, and amiR-trichome, the A’s were replaced with U’s; for miR390, the U was replaced with an A.
(B) Primer extension experiments were performed with total RNAs prepared from N. benthamiana transfected with the indicated constructs using 18 nt 5′ end-labeled DNA primers that are complementary to positions 4–18 of each miRNA. The extension products were 21 nt in length, indicating that wild-type and mutant miRNA precursors produced miRNAs with correct 5′ terminal nucleotides.
(C) Changing the 5′ terminal nucleotide of a miRNA or miRNA* could redirect its loading into a different AGO complex. Wild-type or mutant miRNAs were expressed together with TAP-tagged AGO1 or AGO2 in N. benthamiana by Agrobacterium-mediated transfection. AGO1 and AGO2 complexes were purified. Northern blot analyses were performed with total RNA and RNAs extracted from the AGO complexes using end-labeled DNA oligos complementary to the specified miRNAs and miRNA*s. The positions of RNA size markers are shown to the left.
(D) TAP-tagged chimeric proteins, AGO1^PAZ:AGO2^Mid+PIWI and AGO2^PAZ:AGO1^Mid+PIWI, were expressed together with wild-type and mutant miR390a. Northern blot analyses were performed with total RNA and RNAs extracted from the chimeric AGO complexes using end-labeled DNA oligos complementary to the miR390a. The positions of RNA size markers are shown to the left.

(A) Primer extension experiments were performed with total RNAs prepared from Arabidopsis plants transformed with the indicated constructs using 18 nt 5′ end-labeled DNA primers that are complementary to positions 4–18 of amiR-trichome. The extension products were 21 nt in length, indicating that wild-type and mutant miRNA precursors produced miRNAs with correct 5′ terminal nucleotides.
(B) Northern blot analyses were performed with total RNA and RNAs extracted from the AGO complexes using end-labeled DNA oligos complementary to amiR-trichome. The positions of RNA size markers are shown to the left.
(C–E) Transgenic Arabidopsis overexpressing empty pre-miR168 backbone.
(F–H) Transgenic Arabidopsis overexpressing amiR-trichome.
(I–K) Transgenic Arabidopsis overexpressing amiR-trichomeM.
(L) Relative expression levels of CPC in the transgenic Arabidopsis overexpressing amiR-trichome, amiR-trichomeM, or pre-miR168 empty backbone. The CPC gene-expression level was normalized using the signal from the actin gene. The average (± SD) values from three repeats of quantitative reverse-transcription PCR are shown.