Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Nat Struct Mol Biol. Author manuscript; available in PMC 2011 Feb 1.
Published in final edited form as:
PMCID: PMC2916640

Unique Functionality of 22 nt miRNAs in Triggering RDR6-Dependent siRNA Biogenesis from Target Transcripts in Arabidopsis


RNA interference pathways may involve amplification of secondary siRNAs by RNA-dependent RNA polymerases. In plants, RDR6-dependent secondary siRNAs arise from transcripts targeted by some microRNA (miRNA). Here, Arabidopsis thaliana secondary siRNA from mRNA, and trans-acting siRNA, are shown to be triggered through initial targeting by 22 nt miRNA that associate with AGO1. In contrast to canonical 21 nt miRNA, 22 nt miRNA primarily arise from foldback precursors containing asymmetric bulges. Using artificial miRNA constructs, conversion of asymmetric foldbacks to symmetric foldbacks resulted in production of 21 nt forms of miR173, miR472 and miR828. Both 21 and 22 nt forms associated with AGO1 and guided accurate slicer activity, but only 22 nt miRNA were competent to trigger RDR6-dependent siRNA from target RNA. These data suggest that AGO1 functions differentially with 21 and 22 nt miRNA to engage the RDR6-associated amplification apparatus.


During RNA interference (RNAi), double stranded RNA (dsRNA) is processed by Dicer, a dsRNA-specific RNase III class ribonuclease, into small ~20–30 nt RNA duplexes. Typically, one strand of the duplex preferentially associates with an Argonaute protein to form an effector complex (RNA-induced silencing complex, or RISC), that targets and silences transcripts based on sequence complementarity1. In plants, fungi, nematodes, and some other organisms, this process can be amplified through production of secondary short interfering RNAs (siRNAs) after transcription by RNA-dependent RNA polymerase (RdRp) on the primary target RNA2. As exemplified in Caenorhabditis elegans, primary siRNA, secondary siRNA and other small RNA may serve as guides in association with distinct Argonaute proteins35. Additionally, dsRNA precursors or siRNA have the potential to act non-cell-autonomously in plants and some animals, leading to spread of silencing signals, subsequent amplification, and transgenerational silencing in some cases2,6.

In amplification-competent organisms, the selective recruitment and activity of RdRp on primary targets is a key step that governs whether or not secondary siRNA and amplification occurs. The partially redundant RdRps RRF-1 and EGO-1 from C. elegans, QDE-1 from Neurospora crassa, Rdp1 from Schizosaccharomyces pombe, and RNA-DEPENDENT RNA POLYMERASE6 (RDR6) from Arabidopsis thaliana transcribe RNA in a primer-independent manner712. A. thaliana may encode six RdRps, at least three of which generate long dsRNA that serve as DICER-LIKE (DCL) substrates in several distinct pathways2,13. RDR6, which frequently functions in combination with DCL4, participates in posttranscriptional silencing of exogenous targets (transgene transcripts, some viral RNA), as well as several types of endogenous transcripts13,14. Indiscriminant entry of transcripts into the RDR6-dependent siRNA amplification pathway does not occur in plants. Amplification of secondary siRNA occurs, however, from a minority of small RNA-targeted transcripts, and the genetic requirements for biogenesis of these siRNA resemble those of trans-acting siRNA (tasiRNA)1520. Tightly controlled entry of transcripts into amplification pathways makes intuitive sense, as feed-forward amplification of silencing signals could conceivably result in runaway suppression of off-target transcripts from related gene family members. In Arabidopsis, the XRN family of exonucleases functions to degrade sliced transcript fragments21, and sequencing analysis using an xrn4 mutant suggests that XRN4 is antagonistic to secondary siRNA biogenesis22. How secondary siRNA-producing transcripts are routed differentially from most other target transcripts is understood poorly.

Plant tasiRNA represent a case in which an RdRp-dependent, secondary siRNA-producing pathway has evolved to produce specialized small RNA that function to repress co-evolved targets. The four known families of tasiRNA (TAS1-TAS4) in Arabidopsis are particularly informative as experimental systems to understand formation of secondary siRNA. tasiRNAs form from primary TAS transcripts that are initially targeted and sliced by AGO1-miR173 (TAS1, TAS2), AGO1-miR828 (TAS4) or AGO7-miR390 (TAS3) complexes18,2326. TAS1, TAS2 and TAS4 tasiRNA-generating transcripts are cleaved at a 5'-proximal site. The 3' RNA product is then transcribed by RDR6, leading to dsRNA that is sequentially processed by DCL4 to yield phased tasiRNA in register with the cleavage site18,23,2630.

Here, the basis for selective entry of miRNA-targeted transcripts through the RDR6-dependent siRNA-generating pathway was investigated. Combining genome-wide analyses of 21 nt siRNA and miRNA-targeting patterns with directed experiments, miRNA length was found to be a key determinant in triggering amplification in the context of AGO1-miRNA complexes. In addition, determinants within MIRNA foldbacks that govern miRNA size were identified.


Secondary siRNAs originate from some targeted mRNAs

Six A. thaliana small RNA libraries were generated from aerial plant tissue and sequenced in a multiplexed format, yielding an average of 658,999 reads/library that mapped to at least one genomic position. Among 20–24 nt reads, 53.9% were 21 nts, the vast majority of which corresponded to miRNA (Supplementary Fig. 1). A significant proportion of 20–24 nt reads (74.4%) mapped to 979 annotated transcripts, 55 of which are known to be targeted by one or more miRNA or tasiRNA (Fig. 1a, Supplementary Table 1). Although targeted transcripts represented only 5.6% of the 979 small RNA-generating transcripts, reads from the targeted set represented 27.0% of reads from the entire transcript set. Among reads mapping to the targeted transcripts, 85.4% were 21 nt in length, whereas the majority (52.9%) of reads derived from non-targeted transcripts were 24 nts in length (Fig. 1a,b). Twenty-one-nucleotide siRNA originating from the targeted transcripts were shown previously to be largely RDR6- and DCL4-dependent15. Thus, among Arabidopsis siRNA originating from annotated transcripts, those from targeted transcripts are more likely to yield siRNA of 21 nts in length.

Figure 1
Small RNA from Arabidopsis annotated transcripts. (a) Proportion of small RNA-generating transcripts that are targeted (at single or multiple sites) by miRNA or tasiRNA, or non-targeted. Outer rings show proportion of small RNA size. (b) Mean 20–24 ...

A total of 280 miRNA-target or tasiRNA-target pairs were collated from published data (Supplementary Table 2). Given that multigene families with many members targeted by a specific miRNA or tasiRNA family tend to bias representation on this list, “non-redundant” sets of miRNA-target and tasiRNA-target pairs were generated. One hundred non-redundant set iterations (50 pairs in each) contained only one paralog/target family/small RNA, with the targeted family member chosen randomly for each family in each iteration. Most targeted transcripts from the non-redundant list yielded no 21 nt siRNA (Fig. 1c, Supplementary Table 2). However, on average 33.8% of targets yielded at least a few siRNAs, with several targets yielding highly abundant 21 nt siRNA (Fig. 1c). Levels of these abundant siRNAs were decreased in the dcl2-1 dcl3-1 dcl4-2 (termed dcl234) triple mutant (Fig. 1c).

In exploring the differences between small RNA-target pairs that led to secondary siRNA production and those that did not, it was noticed that secondary siRNA-generating transcripts were frequently targeted by miRNA or tasiRNA of 22 nts in length (Fig. 1d). Targets that interacted with miRNA or tasiRNA containing less than 22 nts yielded significantly (p=0.034, permutation test) less siRNA on average. Additionally, 47% of singly targeted transcripts and 43% of multiply targeted transcripts presented in Figure 1a were targeted by a small RNA of predominantly 22 nts in length. The biogenesis and activity of 22 nt miRNAs were therefore subjected to further analysis.

Asymmetric MIRNA foldbacks yield 22 nt miRNA

The vast majority of Arabidopsis miRNA reads were either 21 or 22 nts in length, with 22 nt families representing 22.5% of all annotated MIRNA families (Fig. 2a). The overall abundance of 22 nt miRNA reads, however, was relatively low (2.3%). Both the percentage of 22 nt-generating MIRNA families, and the numbers of 22 nt miRNA reads, were comparably low in rice (Fig. 2a; Supplementary Table 3,4). Size was also a distinct characteristic for most MIRNA families in both Arabidopsis and rice, with relatively few families expressing a mixed size distribution. Comparing the proportions of 21 or 22 nt miRNA from non-redundant loci in Arabidopsis and rice revealed a distinct bimodal distribution; over 90% of all MIRNA loci generated miRNA that were >80% either 21 nt or 22 nt in length (Fig. 2b; Supplementary Table 3,4). To understand the basis for 21 vs. 22 nt miRNA formation, foldback base-pairing patterns of Arabidopsis MIRNA were examined. Strikingly, those Arabidopsis foldbacks that contained asymmetry in the form of a non-paired nucleotide within the miRNA sequence were significantly enriched for 22 nt miRNA reads (p<2.6 10−9, Wilcoxon rank sum test) (Fig. 2c). In nearly all cases of predominantly 22 nt miRNA, including miR173, miR393 and miR472, an asymmetric non-paired base was detected; there were a few exceptions, such as miR828, which arose from a foldback containing only symmetric mispairs within the miRNA/miRNA* segment (Fig. 2d, Supplementary Table 3).

Figure 2
MIRNA foldback asymmetry leads to formation of 22 nt miRNAs. (a) Mean proportions of distinct miRNA size classes in read datasets (top), and of predominant size class for MIRNA families (bottom), from Arabidopsis and rice. The rice miRNA were from a filtered ...

To determine if the asymmetric positions within miRNA sequences of foldbacks led to formation of the 22 nt size miRNA, artificial miRNA (amiRNA) were expressed using foldbacks containing either asymmetric (wild-type configuration) or symmetric miRNA/miRNA* segments. These were produced using the MIR390a foldback, in which the miR390/miR390* sequences were replaced by miR173/miR173* sequences25. The symmetric foldback was engineered by the addition of a U residue to base-pair with the normally mispaired A residue within the MIR173 foldback (Fig. 3a). The authentic MIR173 foldback, and the asymmetric and symmetric amiRNA foldbacks, were coexpressed with the tasiRNA-generating construct, 35S:TAS1c, in transient assays using Nicotiana benthamiana23, and both miRNA and tasiRNA accumulation (see below) were measured in blot assays and by sequence analysis. Neither miR173 nor TAS1c tasiRNA are conserved in Nicotiana species, so accumulation of these small RNAs is entirely dependent on the transient expression assay. The asymmetric foldback, from either the authentic (35S:MIR173) or artificial (35S:amiR173) constructs, yielded predominantly 22 nt-size miR173 in blot assays (Fig. 3b). In contrast, the symmetric foldback (35S:amiR173-21) yielded a 21 nt-size miR173 (Fig. 3b). Sequence analysis of small RNAs from these assays confirmed that the asymmetric foldbacks yielded predominantly 22 nt miR173, although a substantial amount of off-size products of 19 and 20 nts were sequenced (Fig. 3c, Supplementary Fig. 2, Supplementary Table 5). The symmetric foldback, however, yielded miR173 that was 91% 21 nt in length (Fig. 3c, Supplementary Fig. 2, Supplementary Table 5). The 21 nt form lacked the 3'C residue present in authentic miR173.

Figure 3
Production and activities of 21 and 22 nt miR173 forms. (a) Foldbacks of from wild-type MIR173, amiR173 and amiR173-21. Artificial miRNAs were engineered within the MIR390a foldback. miRNA guide and miRNA* strands are represented with green and red, respectively. ...

Several previous reports collectively showed that miR173 formation requires DCL1, but not DCL23032. DCL2, however, was shown to produce 22 nt siRNA from endogenous Arabidopsis dsRNA27,29. To determine if DCL2, or siRNA-generating DCL3 and DCL4, are necessary for 22 nt miRNA, miRNA reads sequenced from Col-0 and the dcl234 triple mutant were compared. Accumulation and family representation levels of most miRNAs in the dcl234 triple mutant, regardless of size, were similar to those in wild-type Col-0 plants (Fig. 2a, Supplementary Fig. 3a,b). miRNA reads from Col-0 and dcl234 mutant plants for most families, except for two (miR822 and miR839, which are known to be DCL4-dependent18), were highly correlated (Pearson product-moment correlation coefficient [r] = 0.95) (Supplementary Fig. 3b). A high correlation (r = 0.94) was shown when only 22 nt miRNA reads were compared (Supplementary Fig. 3b). Therefore, all known 22 nt miRNA in Arabidopsis are likely DCL1-dependent.

22 nt forms of miR173, miR472, and miR828 trigger siRNAs

As shown previously23,30 and in Figure 3b, coexpression of 35S:MIR173 with 35S:TAS1c triggered tasiRNA biogenesis. Authentic-size miR173 generated from 35S:amiR173 also triggered TAS1c tasiRNA formation (Fig. 3b, lane 5). The tasiRNA pools triggered by both 22 nt miR173 sources were highly phased, with the register set by the miR173 cleavage site (Fig. 3c, Supplementary Fig. 2, Supplementary Table 6). The 21 nt form produced from the amiR173-21 foldback, however, possessed very low tasiRNA trigger activity, and the few TAS1c-derived siRNA that were produced exhibited no phasing pattern (Fig. 3b, c; Supplementary Fig. 2; Supplementary Table 6).

The lack of tasiRNA-trigger activity of miR173-21 could be due to one or more reasons, including failure to associate with AGO1, failure to interact with and guide cleavage of the TAS1c primary transcript, or loss of post-cleavage functions necessary to recruit RDR6 to the precursor transcript. Mi et al. showed that, among AGO1, AGO2, AGO4 and AGO5, miR173 associated most commonly with AGO1. To confirm that both 21 and 22 nt forms of miR173 associated with AGO1, TAS1c biogenesis assays in N. benthamiana were done with co-expression of HA-tagged AGO1, followed by coimmunoprecipitation analysis. Both 22 and 21 nt forms of miR173 co-immunoprecipitated with HA-AGO1 (Fig. 3d, lanes 4, 6 and 8). TAS1c tasiR255, produced in the presence of 22 nt miR173, also co-immunoprecipitated with HA-AGO1 (Fig. 3d, lanes 4 and 6). As HA-AGO1-nonassociated controls, accumulation TAS1c 3'D2(−) tasiRNA, which contains an AGO1-nonpreferred 5'A, and U6 RNA were analyzed. No or very low levels of TAS1c 3'D2(−) tasiRNA and U6 RNA were detected in the immunoprecipitated samples (Fig. 3d), indicating that HA-AGO1 was selective. To assess more broadly whether or not AGO1 associates preferentially with 21 or 22 nt miRNA, small RNAs from total RNA (input fraction) and immunoprecipitated HA-AGO1 complexes (IP fraction) were identified by high-throughput sequencing (two replicates each). Of the miRNAs that were enriched at least 2-fold in the IP fraction, miRNA families that were predominantly 20–24 nts were recovered, suggesting that AGO1 does not include or exclude miRNA based on length (Supplementary Fig. 4). All predominantly 22 nt miRNA families that met read thresholds for inclusion in the enrichment calculation were enriched in the IP fraction. Additionally, for two predominantly 21 nt miRNAs (miR397a and miR167a,b,d) that had a >30% subpopulation of 22 nt variants (Fig. 2b), both the 21 and 22 nt variants were enriched more than 2-fold in the IP fraction (Supplementary Fig. 4).

To assess guide function of 22 and 21 nt miR173 forms, 5' RACE assays were done for each co-expression assay using primers to detect the 3' product of miR173-guided cleavage. Products were detected in all three co-expression assays, and sequencing confirmed that cleavage guided by both 22 and 21 nt miR173 forms occurred at the canonical target site (Fig. 3e,f). Furthermore, both 22 and 21 nt miR173 forms destabilized TAS1c full-length transcripts to very low or undetectable levels (Supplementary Fig. 5).

Because the 21 nt form derived from the 35S:amiR173-21 foldback contained a 3'A rather than a 3'C residue, as is present in 22 nt authentic miR173, the effect of 3' nucleotide identity on both 21 and 22 nt forms of miR173 was examined. First, a construct (35S:amiR173-21-3'C) that generated a symmetric foldback yielding a 21 nt form containing a 3'C was tested (Supplementary Fig. 5a). This 21 nt-size miR173 failed to trigger tasiRNA accumulation despite guiding accurate cleavage of the TAS1c transcript (Supplementary Fig. 5b,c lane 2). Second, to assess the possibility of a preference for a specific 3' nucleotide in a 22 nt context, 22 nt miR173 variants containing 3' G, A or U were tested. Each of the variants accumulated to within 36% of the level of 22 nt miR173 containing a natural 3'C and guided accurate cleavage of the TAS1c transcript (Supplementary Fig. 5b,c). Interestingly, although each of the 22 nt 3'G, 3'A and 3'U variants triggered tasiRNA accumulation, the levels of tasiRNA generated were relatively low (Supplementary Fig. 5b).

To determine if the 22 nt requirement to trigger tasiRNA is specific to miR173, 22 and 21 nt artificial miRNAs for both miR472 and miR828 were generated using the MIR390a-based foldback, and tested using the TAS1c tasiRNA transient assay system. Both asymmetric and symmetric foldbacks were engineered following the same strategy as for the miR173 variants (Fig. 4a) and co-expressed with modified TAS1c constructs (35S:TAS1c-472 and 35S:TAS1c-828) in which the normal miR173 target site was substituted with sites recognized by miR472 or miR828. Based on mobility of miR472 and miR828 forms in blot assays, 22 and 21 nt variants were generated as predicted from asymmetric and symmetric foldbacks, respectively, although the asymmetric miR828 foldback yielded a mixture of both size classes (Fig. 4b). Importantly, tasiRNA were detected only in those samples containing 22 nt variants of miR472 and miR828 (Fig. 4b). Like the miR173 size variants, both the 22 and 21 nt forms of miR472 and miR828 guided accurate cleavage of the engineered tasiRNA transcripts (Fig. 4b,c) and co-immunoprecipitated with HA-AGO1 (Fig. 4d). These data clearly show that 22 nt forms of three Arabidopsis miRNAs possess unique functionality to direct tasiRNA or secondary siRNA biogenesis.

Figure 4
Production and activities of 21 and 22 nt miR472 and miR828 forms. (a) Foldbacks of amiR828, amiR828-21, amiR472, and amiR472-21. (b) Accumulation of miR472, miR828 and modified TAS1c tasiRNA (siR255), and 5'RACE to detect miRNA-guided cleavage products ...


This study associated a unique class of Arabidopsis miRNA - those that are 22 nts in length - with triggering formation of secondary, RDR6/DCL4-dependent 21 nt siRNAs from primary miRNA targets. While both 22 nt and 21 nt forms of miR173, miR472 and miR828 associated with AGO1 and guided accurate cleavage of target transcripts, advancement of target RNA fragments through the 21 nt siRNA biogenesis pathway was a unique property of the 22 nt miRNA-target RNA pairs. Previous studies provided hints that 22 nt miRNAs might be associated with secondary siRNA. Arabidopsis MIR168b was shown to generate both 21 and 22 nt forms of miR16833, which targets the transcript encoding AGO1 and leads to secondary siRNAs in a homeostatic regulatory loop33,34. In rice, 22 nt miRNA target sites were identified in several abundant siRNA-generating transcripts35, suggesting that the 22 nt property may be conserved in angiosperms.

The fact that most plant miRNAs, particularly those from the most highly expressed families, are 21 nt in length may explain why secondary siRNA from most targeted transcripts are absent or in low abundance. While the tasiRNA represent highly refined examples of functional, discrete secondary siRNAs from a non-coding RNA, it is interesting to consider potential functions for the cases in which abundant siRNAs originate from protein-coding transcripts. In at least some examples, secondary siRNA might specifically target transcripts from related family members. For gene families that are evolving or expanding rapidly, secondary targeting may suppress dosage effects. Evidence for the functionality of secondary siRNA from large multigene family transcripts has been presented15,36,37.

It is proposed that 22 nt miRNA-AGO1 complexes, but not 21 nt miRNA-AGO1 complexes, mark transcripts for RDR6-dependent siRNA formation. This could involve AGO1 adopting one of two states - an amplification-trigger state and a non-trigger state. Both states would be competent to interact with and slice targets containing a suitable target site, but only the trigger state would recruit RDR6 to the de-capped (sliced) target. Previous results from experiments with a non-cleavable target of miR173 suggested that secondary siRNA formation depends on a functional cleavage site30. Several studies have also shown that all known tasiRNA, and many other RDR6-dependent siRNA in Arabidopsis15,18,19,23, originate preferentially adjacent to target cleavage sites. Recruitment of RDR6, therefore, may depend on target cleavage.

Conformational changes of well-characterized Argonaute proteins are known to occur upon interaction with target RNA and release of the guide 3' end from the PAZ domain38,39. How a 22 nt RNA would effect an AGO1 conformation distinct from a 21 nt RNA is not immediately obvious. The proposed trigger state, however, could involve a direct interaction with RDR6 or associated factors, such as SGS316,26. To date, there are no data suggesting direct interaction between AGO1 and RDR6 or known RDR6-associated factors. Alternatively, conformationally distinct AGO1-small RNA complexes may interact differentially with GW domain proteins, which interact directly with AGO proteins as scaffolds to mediate interactions with other factors to suppress translation and promote de-adenylation40. Arabidopsis AGO1-specific GW proteins are not yet characterized, although AGO4 likely interacts with GW domains in an RNA PolV subunit (NRPDE1) and the transcription elongation factor SPT5L4143.

While it is clear that AGO1 exhibits a preference for small RNAs with a 5'U, which docks within a MID domain binding pocket, there are no clear structural or functional data supporting a 3' nucleotide preference for interaction with the PAZ domain. Among 22 nt miRNAs in Arabidopsis, a 3'C is most common, although this implies neither a binding preference nor particular functionality. Within a 22 nt context, a 3'C promoted siRNA-trigger function of miR173 most effectively. However, there was no strict requirement for any specific 3' nucleotide, and the 22 nt siRNA-triggers miR828 and miR393 both contain 3'A. The studies presented here do not allow quantitative comparison of the affinity of AGO1 for 21 vs. 22 nt miRNAs, or for 22 nt miRNAs with different 3' nucleotides.

The finding that 22 nt miRNA trigger RDR6-dependent secondary siRNAs leads to the question of whether or not other siRNA amplification pathways have guide RNA-specific activities to recruit RDR1 or RDR2. Signals for amplification mediated by RDR1 and RDR6 during antiviral RNAi are not known, although results presented here suggest 22 nt siRNA, working through AGO1 or other AGO proteins, might be plausible triggers. As DCL2 is known to both promote a full silencing response against several viruses32,4447 and to generate 22 nt siRNAs27,29, this idea may have support. DCL2 was also shown to promote DCL4-dependent posttranscriptional silencing of several transgenes48. Furthermore, overproduction of DCL2-dependent 22 nt siRNA may account for the severe phenotypes of dcl134 and dcl14 mutant Arabidopsis plants, in which DCL2 might generate 22 nt siRNA from dsRNA that would normally be processed to 21 nt siRNA by DCL432.

Finally, the basis for production of either a 21 nt or a 22 nt miRNA depends on the nature of the foldback in plants. Most Arabidopsis miRNAs are formed by DCL1-mediated processing first at the loop-distal end of the foldback, then at the loop proximal end4951. DCL1, like other Dicers, functions as a molecular ruler. For the loop-proximal cuts, DCL1 “measures” the length of helical RNA from the PAZ domain-bound end to the positions juxtaposed to the active centers5254. For foldbacks that are symmetrically paired within the miRNA/miRNA* region, it is inferred that DCL1 measures a length equivalent to 21 base pairs in most cases. However, it is also inferred that length of an A-form helix containing a single asymmetric bulge is frequently equivalent to that of a symmetrically paired helix. Three-dimensional modeling of 21 nt- and 22 nt-generating MIR173 foldbacks using MC-Fold and MC-Sym55 supports this idea. The most likely models show that an unpaired nucleotide does not substantially increase the length of the miRNA/miRNA* duplex region (Supplementary Fig. 6). This explains why, when the asymmetric nucleotide occurs within the miRNA sequence, the resulting miRNA contains a non-canonical 22 nt length. Given the unique functionality of 22 nt miRNA to trigger siRNA amplification, it seems reasonable to propose that foldback structures with asymmetric positions, and that would result in 22 nt miRNA, should be under evolutionary constraints.


Plant materials and growth conditions

N. benthamiana and A. thaliana plants, including the previously described dcl2-1 dcl3-1 dcl4-2 triple mutant47, were grown in normal greenhouse conditions with supplemental light on a 16hr light/ 8 hr dark cycle.

Transgene constructs

35S:TAS1c and 35S:MIR173a derived constructs were generated in pMDC32 and described previously23,25,30,57. MIR390a-derived artificial miRNA constructs were designed by ligating overlapping oligonucleotides into a pMDC32-derived vector containing ~200bp upstream and downstream of the MIR390a foldback.

Transient expression assays

Transient expression assays in N. benthamiana leaves were carried out as previously described58 with Agrobacterium tumefaciens GV3101. Bacterial cultures were resuspended at OD600=1.0. When multiple constructs were co-expressed, equal amounts of each culture were used. Plants were grown in the greenhouse after infiltration, and zones of infiltration were harvested 48 hours post-infiltration for RNA isolation.

RNA Blot assays

Total RNA from Arabidopsis and N. benthamiana was isolated using TRIzol reagent (Invitrogen). Two chloroform extractions were done and RNA was precipitated in an equal volume of isopropanol for 20 min. Triplicate samples from pools of N. benthamiana infiltrated leaves were analyzed. RNA blot assays were done as described30. Briefly 5 μg, 10 μg or 20 μg of total normalized RNA was resolved by denaturing polyacrylamide-gel electrophoresis for small RNA analysis, or 5 μg of total RNA was resolved by denaturing 1.5% agarose-gel electrophoresis. RNA was transferred to positively charged nitrocellulose membrane. DNA or LNA probes were end-labeled using [32]ATP and Optikinase (USB). Probes were hybridized to RNA on membranes in Sigma Perfect-Hyb at 68°C (HMW blots) or 38–42°C (small RNA blots). An Instant Imager (Packard Bioscience) was used to measure blot hybridization signals.

Small RNA sequencing

Small RNA amplicons from transient assays in N. benthamiana were prepared as described57,59. Four synthetic oligoribonucleotides were added in different amounts (Std11, 0.0001 pmol; Std6, 0.001 pmol; Std3, 0.01 pmol; Std2, 0.1 pmol) to 90 μg total RNA per sample before amplicon preparation. Sequencing-by-synthesis was done with five pmol of each amplicon using an Illumina Genome Analyzer II (GAII, www.illumina.com). Reads were parsed using the first six nucleotides of the 3' adapter (CTGTAG). Reads were normalized per million as described59. Read proportions were based on total reads (18–24 nts) that matched perfectly within a 29 base window surrounding the miR173 or miR173* sequences from samples containing 35S:MIR173a, 35S:amiR173 and 35S:amiR173-21. Radar plots to display phasing from 35S:TAS1c were generated as previously described19,25. Small RNA amplicons were generated and analyzed from HA-AGO1 IP samples as described25.

Small RNA analysis

Small RNA from A. thaliana were mapped to the genome and annotated transcriptome (TAIR9; http://www.arabidopsis.org), including annotated MIRNA precursors (miRBase v14; http://www.miRBase.org), using the CASHX pipeline59. Those transcripts to which four or greater small RNA reads/million (RPM) were mapped in at least four of the six libraries were defined as small RNA-generating transcripts. For analysis of small RNA-generating transcripts, annotated structural RNA, pseudogenes or transposable elements were removed.

Statistical analyses

All statistical analyses were done using R v2.9.2. Permutation tests were done using the R “twotPermutation” function, with 1 million simulations, from the DAAG package60. A Wilcoxon rank sum test was used to compare empirical logits for the ratio of 22 to 21 nt miRNA reads. The rank sum test was done using the R “wilcox.test” function from the stats package.


Modified RNA ligase-mediated 5' RACE was done as described30 with the Generacer kit (Invitrogen) using the following gene specific primers: TAS1c_707_5'_RACE [GATGATGCTTCTTCGCTACACCTCGGAG] and TAS1c_573_5'_RACE [AGCAACTGTTCTTTAGACGACTTGAAAATCTCAT]. 5' RACE products were gel purified using Invitrogen PureLink gel extraction kit, cloned in TOPO TA (Invitrogen), introduced into E. coli, screened for inserts, and sequenced.

Supplementary Material


A.C. was supported by a postdoctoral fellowship from the Ministerio de Ciencia e Innovacion (BMC-2008-0188). This work was supported by grants from the National Science Foundation (MCB-0618433 and MCB-0956526), National Institutes of Health (AI43288) and Monsanto Corporation.


Note: Supplementary information is available on the Nature Structural & Molecular Biology website.

Accession codes. High-throughput sequencing datasets for HA-AGO1-associated small RNA and input controls were deposited in Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo) under the series accession GSE22252. Datasets for Col-0 and dcl2-1 dcl3-1 dcl4-2 small RNA libraries were published previously47 and are available under the series accession GSE20197. Twenty-eight O. sativa Nipponbare libraries, as described in Supplementary Information, were also used.


1. Kim VN, Han J, Siomi MC. Biogenesis of small RNAs in animals. Nat Rev Mol Cell Biol. 2009;10:126–139. [PubMed]
2. Ghildiyal M, Zamore PD. Small silencing RNAs: an expanding universe. Nat Rev Genet. 2009;10:94–108. [PMC free article] [PubMed]
3. Guang S, et al. An Argonaute transports siRNAs from the cytoplasm to the nucleus. Science. 2008;321:537–541. [PMC free article] [PubMed]
4. Gu W, et al. Distinct argonaute-mediated 22G-RNA pathways direct genome surveillance in the C. elegans germline. Mol Cell. 2009;36:231–244. [PMC free article] [PubMed]
5. Claycomb JM, et al. The Argonaute CSR-1 and its 22G-RNA cofactors are required for holocentric chromosome segregation. Cell. 2009;139:123–134. [PMC free article] [PubMed]
6. Carthew RW, Sontheimer EJ. Origins and Mechanisms of miRNAs and siRNAs. Cell. 2009;136:642–655. [PMC free article] [PubMed]
7. Meister G, Tuschl T. Mechanisms of gene silencing by double-stranded RNA. Nature. 2004;431:343–349. [PubMed]
8. Makeyev EV, Bamford DH. Cellular RNA-dependent RNA polymerase involved in posttranscriptional gene silencing has two distinct activity modes. Mol Cell. 2002;10:1417–1427. [PubMed]
9. Sijen T, Steiner FA, Thijssen KL, Plasterk RH. Secondary siRNAs result from unprimed RNA synthesis and form a distinct class. Science. 2007;315:244–247. [PubMed]
10. Pak J, Fire A. Distinct populations of primary and secondary effectors during RNAi in C. elegans. Science. 2007;315:241–244. [PubMed]
11. Motamedi MR, et al. Two RNAi complexes, RITS and RDRC, physically interact and localize to noncoding centromeric RNAs. Cell. 2004;119:789–802. [PubMed]
12. Correa RL, Steiner FA, Berezikov E, Ketting RF. MicroRNA-directed siRNA biogenesis in Caenorhabditis elegans. PLoS Genet. 2010;6:e1000903. [PMC free article] [PubMed]
13. Voinnet O. Use, tolerance and avoidance of amplified RNA silencing by plants. Trends Plant Sci. 2008;13:317–328. [PubMed]
14. Mallory AC, Elmayan T, Vaucheret H. MicroRNA maturation and action--the expanding roles of ARGONAUTEs. Curr Opin Plant Biol. 2008;11:560–566. [PubMed]
15. Howell MD, et al. Genome-wide analysis of the RNA-DEPENDENT RNA POLYMERASE6/DICER-LIKE4 pathway in Arabidopsis reveals dependency on miRNA- and tasiRNA-directed targeting. Plant Cell. 2007;19:926–942. [PMC free article] [PubMed]
16. Peragine A, Yoshikawa M, Wu G, Albrecht HL, Poethig RS. SGS3 and SGS2/SDE1/RDR6 are required for juvenile development and the production of trans-acting siRNAs in Arabidopsis. Genes Dev. 2004;18:2368–2379. [PMC free article] [PubMed]
17. Vazquez F, et al. Endogenous trans-acting siRNAs regulate the accumulation of Arabidopsis mRNAs. Mol Cell. 2004;16:69–79. [PubMed]
18. Rajagopalan R, Vaucheret H, Trejo J, Bartel DP. A diverse and evolutionarily fluid set of microRNAs in Arabidopsis thaliana. Genes Dev. 2006;20:3407–3425. [PMC free article] [PubMed]
19. Axtell MJ, Jan C, Rajagopalan R, Bartel DP. A Two-Hit Trigger for siRNA Biogenesis in Plants. Cell. 2006;127:565–577. [PubMed]
20. Ronemus M, Vaughn MW, Martienssen RA. MicroRNA-targeted and small interfering RNA-mediated mRNA degradation is regulated by argonaute, dicer, and RNA-dependent RNA polymerase in Arabidopsis. Plant Cell. 2006;18:1559–1574. [PMC free article] [PubMed]
21. Chiba Y, Green PJ. mRNA degredation machinery in plants. J. Plant Biol. 2009;52:114–124.
22. Gregory BD, et al. A link between RNA metabolism and silencing affecting Arabidopsis development. Dev Cell. 2008;14:854–866. [PubMed]
23. Allen E, Xie Z, Gustafson AM, Carrington JC. microRNA-directed phasing during trans-acting siRNA biogenesis in plants. Cell. 2005;121:207–221. [PubMed]
24. Baumberger N, Baulcombe DC. Arabidopsis ARGONAUTE1 is an RNA slicer that selectively recruits microRNAs and short interfering RNAs. Proc Natl Acad Sci USA. 2005;102:11928–11933. [PMC free article] [PubMed]
25. Montgomery TA, et al. Specificity of ARGONAUTE7-miR390 interaction and dual functionality in TAS3 trans-acting siRNA formation. Cell. 2008;133:128–141. [PubMed]
26. Yoshikawa M, Peragine A, Park MY, Poethig RS. A pathway for the biogenesis of trans-acting siRNAs in Arabidopsis. Genes Dev. 2005;19:2164–2175. [PMC free article] [PubMed]
27. Xie Z, Allen E, Wilken A, Carrington JC. DICER-LIKE 4 functions in trans-acting small interfering RNA biogenesis and vegetative phase change in Arabidopsis thaliana. Proc Natl Acad Sci USA. 2005;102:12984–12899. [PMC free article] [PubMed]
28. Dunoyer P, Himber C, Voinnet O. DICER-LIKE 4 is required for RNA interference and produces the 21-nucleotide small interfering RNA component of the plant cell-to-cell silencing signal. Nat Genet. 2005;37:1356–1360. [PubMed]
29. Gasciolli V, Mallory AC, Bartel DP, Vaucheret H. Partially redundant functions of Arabidopsis DICER-like enzymes and a role for DCL4 in producing trans-acting siRNAs. Curr Biol. 2005;15:1494–1500. [PubMed]
30. Montgomery TA, et al. AGO1-miR173 complex initiates phased siRNA formation in plants. Proc Natl Acad Sci U S A. 2008;105:20055–20062. [PMC free article] [PubMed]
31. Park W, Li J, Song R, Messing J, Chen X. CARPEL FACTORY, a Dicer homolog, and HEN1, a novel protein, act in microRNA metabolism in Arabidopsis thaliana. Curr Biol. 2002;12:1484–1495. [PubMed]
32. Bouche N, Lauressergues D, Gasciolli V, Vaucheret H. An antagonistic function for Arabidopsis DCL2 in development and a new function for DCL4 in generating viral siRNAs. EMBO J. 2006;25:3347–3356. [PMC free article] [PubMed]
33. Vaucheret H. AGO1 homeostasis involves differential production of 21-nt and 22-nt miR168 species by MIR168a and MIR168b. PLoS One. 2009;4:e6442. [PMC free article] [PubMed]
34. Mallory AC, Vaucheret H. ARGONAUTE 1 homeostasis invokes the coordinate action of the microRNA and siRNA pathways. EMBO Rep. 2009;10:521–526. [PMC free article] [PubMed]
35. Johnson C, et al. Clusters and superclusters of phased small RNAs in the developing inflorescence of rice. Genome Res. 2009;19:1429–1440. [PMC free article] [PubMed]
36. Chen HM, Li YH, Wu SH. Bioinformatic prediction and experimental validation of a microRNA-directed tandem trans-acting siRNA cascade in Arabidopsis. Proc Natl Acad Sci USA. 2007;104:3318–3323. [PMC free article] [PubMed]
37. Addo-Quaye C, Eshoo TW, Bartel DP, Axtell MJ. Endogenous siRNA and miRNA targets identified by sequencing of the Arabidopsis degradome. Curr Biol. 2008;18:758–762. [PMC free article] [PubMed]
38. Wang Y, et al. Structure of an argonaute silencing complex with a seed-containing guide DNA and target RNA duplex. Nature. 2008;456:921–926. [PMC free article] [PubMed]
39. Wang Y, et al. Nucleation, propagation and cleavage of target RNAs in Ago silencing complexes. Nature. 2009;461:754–761. [PMC free article] [PubMed]
40. Eulalio A, Tritschler F, Izaurralde E. The GW182 protein family in animal cells: new insights into domains required for miRNA-mediated gene silencing. RNA. 2009;15:1433–1442. [PMC free article] [PubMed]
41. El-Shami M, et al. Reiterated WG/GW motifs form functionally and evolutionarily conserved ARGONAUTE-binding platforms in RNAi-related components. Genes Dev. 2007;21:2539–2544. [PMC free article] [PubMed]
42. Bies-Etheve N, et al. RNA-directed DNA methylation requires an AGO4-interacting member of the SPT5 elongation factor family. EMBO Rep. 2009;10:649–654. [PMC free article] [PubMed]
43. He XJ, et al. An effector of RNA-directed DNA methylation in arabidopsis is an ARGONAUTE 4- and RNA-binding protein. Cell. 2009;137:498–508. [PMC free article] [PubMed]
44. Blevins T, et al. Four plant Dicers mediate viral small RNA biogenesis and DNA virus induced silencing. Nucleic Acids Res. 2006;34:6233–6246. [PMC free article] [PubMed]
45. Deleris A, et al. Hierarchical action and inhibition of plant Dicer-like proteins in antiviral defense. Science. 2006;313:68–71. [PubMed]
46. Diaz-Pendon JA, Li F, Li WX, Ding SW. Suppression of antiviral silencing by cucumber mosaic virus 2b protein in Arabidopsis is associated with drastically reduced accumulation of three classes of viral small interfering RNAs. Plant Cell. 2007;19:2053–2063. [PMC free article] [PubMed]
47. Garcia-Ruiz H, et al. Arabidopsis RNA-Dependent RNA Polymerases and Dicer-Like Proteins in Antiviral Defense and Small Interfering RNA Biogenesis during Turnip Mosaic Virus Infection. Plant Cell. 2010;22:481–496. [PMC free article] [PubMed]
48. Mlotshwa S, et al. DICER-LIKE2 plays a primary role in transitive silencing of transgenes in Arabidopsis. PLoS One. 2008;3:e1755. [PMC free article] [PubMed]
49. Mateos JL, Bologna NG, Chorostecki U, Palatnik JF. Identification of MicroRNA processing determinants by random mutagenesis of Arabidopsis MIR172a precursor. Curr Biol. 2010;20:49–54. [PubMed]
50. Song L, Axtell MJ, Fedoroff NV. RNA secondary structural determinants of miRNA precursor processing in Arabidopsis. Curr Biol. 2010;20:37–41. [PubMed]
51. Werner S, Wollmann H, Schneeberger K, Weigel D. Structure determinants for accurate processing of miR172a in Arabidopsis thaliana. Curr Biol. 2010;20:42–48. [PubMed]
52. Macrae IJ, et al. Structural basis for double-stranded RNA processing by Dicer. Science. 2006;311:195–198. [PubMed]
53. MacRae IJ, Zhou K, Doudna JA. Structural determinants of RNA recognition and cleavage by Dicer. Nat Struct Mol Biol. 2007;14:934–940. [PubMed]
54. Qin H, et al. Structure of the Arabidopsis thaliana DCL4 DUF283 domain reveals a noncanonical double-stranded RNA-binding fold for protein-protein interaction. RNA. 2010;16:474–481. [PMC free article] [PubMed]
55. Parisien M, Major F. The MC-Fold and MC-Sym pipeline infers RNA structure from sequence data. Nature. 2008;452:51–55. [PubMed]
56. Meyers BC, et al. Criteria for Annotation of Plant MicroRNAs. Plant Cell. 2008;20:3186–3190. [PMC free article] [PubMed]
57. Cuperus JT, et al. Identification of MIR390a precursor processing-defective mutants in Arabidopsis by direct genome sequencing. Proc Natl Acad Sci U S A. 2010;107:466–471. [PMC free article] [PubMed]
58. Llave C, Xie Z, Kasschau KD, Carrington JC. Cleavage of Scarecrow-like mRNA targets directed by a class of Arabidopsis miRNA. Science. 2002;297:2053–2056. [PubMed]
59. Fahlgren N, et al. Computational and analytical framework for small RNA profiling by high-throughput sequencing. RNA. 2009;15:992–1002. [PMC free article] [PubMed]
60. Maindonald JH, Braun J. Data analysis and graphics using R: an example-based approach. Cambridge University Press; Cambridge; New York: 2007. p. 502.
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • BioProject
    BioProject links
  • Gene
    Gene records that cite the current articles. Citations in Gene are added manually by NCBI or imported from outside public resources.
  • GEO DataSets
    GEO DataSets
    Gene expression and molecular abundance data reported in the current articles that are also included in the curated Gene Expression Omnibus (GEO) DataSets.
  • GEO Profiles
    GEO Profiles
    Gene Expression Omnibus (GEO) Profiles of molecular abundance data. The current articles are references on the Gene record associated with the GEO profile.
  • HomoloGene
    HomoloGene clusters of homologous genes and sequences that cite the current articles. These are references on the Gene and sequence records in the HomoloGene entry.
  • MedGen
    Related information in MedGen
  • Protein
    Protein translation features of primary database (GenBank) nucleotide records reported in the current articles as well as Reference Sequences (RefSeqs) that include the articles as references.
  • PubMed
    PubMed citations for these articles
  • SRA
    Massively-parallel sequencing project data in the Short Read Archive (SRA) that are reported in the current articles.
  • Substance
    PubChem chemical substance records that cite the current articles. These references are taken from those provided on submitted PubChem chemical substance records.
  • Taxonomy
    Taxonomy records associated with the current articles through taxonomic information on related molecular database records (Nucleotide, Protein, Gene, SNP, Structure).
  • Taxonomy Tree
    Taxonomy Tree

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...