Drosophila microRNAs are sorted into functionally distinct argonaute complexes after production by dicer-1

Abstract

Small interfering RNAs (siRNAs) and microRNAs (miRNAs) guide distinct classes of RNA-induced silencing complexes (RISCs) to repress mRNA expression in biological processes ranging from development to antiviral defense. In Drosophila, separate but conceptually similar endonucleolytic pathways produce siRNAs and miRNAs. Here, we show that despite their distinct biogenesis, double-stranded miRNAs and siRNAs participate in a common sorting step that partitions them into Ago1- or Ago2-containing effector complexes. These distinct complexes silence their target RNAs by different mechanisms. miRNA-loaded Ago2-RISC mediates RNAi, but only Ago1 is able to repress an mRNA with central mismatches in its miRNA-binding sites. Conversely, Ago1 cannot mediate RNAi, because it is an inefficient nuclease whose catalytic rate is limited by the dissociation of its reaction products. Thus, the two members of the Drosophila Ago subclade of Argonaute proteins are functionally specialized, but specific small RNA classes are not restricted to associate with Ago1 or Ago2.

Figure 1. Two models for the miRNA and siRNA pathways in Drosophila

(A) Small RNA biogenesis and RISC assembly are tightly coupled. miRNAs are exclusively loaded into Ago1 and siRNAs into Ago2 . (B) Small RNA biogenesis and RISC assembly are independent. After their production, small RNA duplexes are proposed to be actively sorted into distinct Ago proteins according to their structures, not their precursor origins.

Figure 2. Regulation of GFP reporter expression in cultured Drosophila S2 cells by endogenous miR-277

Clonally derived stable cell lines were generated that expressed control GFP unregulated by miR-277, GFP bearing two miR-277-complementary sites in its 3′ untranslated region (UTR), and GFP bearing in its 3′ UTR four miR-277-complementary sites, each containing three central mismatches to miR-277, producing a ‘bulge.’ Each cell line was transfected with a cholesterol-conjugated, 2′-O-methyl modified, antisense oligonucleotide (ASO) complementary to miR-277 or to an unrelated luciferase sequence. As a control, the unregulated GFP reporter cell line was transfected with GFP dsRNA. (A) Representative FACS profiles from a single experiment. (B) The average of mean fluorescence recorded in three trials. Error bars indicate ± one standard deviation. P-values were calculated using a two-sample T-test assuming equal variances.

Figure 3. Silencing of the reporter bearing two perfectly matched miR-277 complementary sites required components of both the miRNA biogenesis and the RNAi pathways, including Ago2, but is potentiated by depletion of Ago1

(A) Mean GFP fluorescence (average ± standard deviation for three or four trials). DsRNA-triggered RNAi was used to deplete the cells of the indicated protein. (B) Western blotting confirmed the success and specificity of the RNAi-mediated depletion for each protein. dcr-2(RNAi) reduced the abundance of both Dcr-2 and R2D2, as previously reported (Liu et al., 2003), but r2d2(RNAi) had no detectable effect on Dcr-2 abundance. The three isoforms of Loqs are indicated at the right of the Loqs panel. The bands above and below Ago2 correspond to cross-reacting proteins characteristically detected with this antibody.

Figure 4. Silencing of the reporter bearing four imperfectly matched miR-277 complementary sites required components of both the miRNA biogenesis pathway and Ago1, but not Dcr-2, R2D2, or Ago2, components of the RNAi pathway

(A) Over-expression of miR-277 from a mini-pri-miRNA transgene increased repression of the miR-277-regulated perfectly matched and bulged reporters. (B) Mean GFP fluorescence (average ± standard deviation for three or four trials). DsRNA-triggered RNAi was used to deplete the cells of the indicated protein.

Figure 5. Most endogenous miR-277 is not associated with Ago1 in S2 cells

(A) Northern analysis reveals that ago2(RNAi) reduced the steady-state abundance of miR-277, but not bantam, whereas ago1(RNAi) decreased the abundance of bantam, but not pre-bantam or miR-277. (B) Western blotting shows that immunoprecipitation of Ago1 depleted nearly all Ago1, but little or no Ago2, from S2 cell cytoplasmic extract. (C) Northern analysis shows that the majority of bantam, but not pre-bantam co-immunoprecipitated with Ago1. In contrast, the majority of endogenous and of over-expressed miR-277 remained, in the supernatant, unbound by Ago1. The asterisk marks non-specific hybridization of the miR-277 probe with 5S RNA.

Figure 6. In vivo in adult flies, miR-277 repression of the GFP reporter via perfectly complementary sites requires the loading activity of Dcr-2 and R2D2, but repression via bulged sites does not

(A) Representative Western blotting data of α-tubulin and GFP reporter in total lysates from adult flies of the indicated heterozygous (+/−) and homozygous (−/−) mutant genotypes. (B) The average (± standard deviation) GFP expression in homozygous mutant flies, relative to heterozygotes, for three (r2d2) or four trials of the experiment in (A). The dcr-2^L811fsX mutant lacks detectable Dcr-2 protein, whereas the dcr-2^G31R point mutant produces a Dcr-2 protein that cannot dice long dsRNA, but can nonetheless load small RNA duplexes, such as siRNA and miRNA/miRNA* duplexes, into Ago2.

Figure 7. Ago1 is a poor endonuclease

(A) Distinct cleavage kinetics distinguish Ago1- and Ago2-RISC. At approximately equal enzyme concentrations, the initial velocity for Ago2-RISC was ∼12-fold greater than that of Ago1-RISC. Cleavage by Ago2-RISC was linear throughout the reaction, as long as the substrate remained in vast excess, whereas cleavage by Ago1-RISC showed biphasic behavior, suggesting the product release step is rate-limiting. The RISC concentration estimated by burst analysis (∼3.2 nM; red arrow) correlated well with that measured by 2′-O-methyl ASO affinity capture (∼4.6 nM). (B) Pseudo-Michaelis-Menten and (C) Michaelis-Menten analyses of Ago1- and Ago2-RISC, respectively. Michaelis-Menten parameters are summarized in Table 1.