AGO1 interacts in a mutually exclusive manner with DCR-1 and GW182. (A) AGO1 coimmunoprecipitates with an ∼175-kDa protein, GW182, from S2 cells. The immunoprecipitated proteins were stained with silver. (n.i.) Non-immune IgG used as negative control; (h.c. and l.c.) heavy and light chains of the antibodies, respectively. (B) Western blotting analyses show that anti-GW182 antibodies produced in our laboratory specifically recognize GW182 in S2 cells. (C) Anti-AGO1 immunoprecipitates from S2 cells contain both AGO1 and GW182. (D) AGO1, but not AGO2, DCR-1, and DCR-2, is detected in the immunoprecipitated complex with anti-GW182 from S2 cells. (E) The PIWI proteins, Aub and Piwi, are not detected in anti-GW182 immunoprecipitates obtained from fly ovaries.

Characterization of two AGO1-containing complexes immunoprecipitated from S2 cells. (A) GW182 is not detected in the AGO1-DCR-1 complex immunoprecipitated from S2 cells using anti-LOQS. It should be noted that 10-fold more S2 lysate was used for n.i. (non-immune IgG; negative control) and anti-LOQS immunoprecipitations compared to those for anti-AGO1 and anti-GW182 immunoprecipitations to equalize the amounts of AGO1 in the immunoprecipitates. (Lowest panel, *) Bands corresponding to antibodies (heavy chains). (B) RNAs were extracted from the immunoprecipitated complexes (shown in A) and subjected to Northern blotting with DNA oligos for bantam and the opposite strand, bantam*. Anti-AGO1 and anti-GW182 immunoprecipitates almost exclusively contain single-stranded mature bantam. On the other hand, anti-LOQS immunoprecipitates show very weak signals for bantam precursor and bantam/bantam* duplexes. (C) Immunoprecipitated complexes (shown in A) were subjected to in vitro target RNA cleavage assays using a target RNA (bantam-130) showing full complementarity to bantam. The AGO1-GW182 complex, but not the AGO1-DCR-1 complex, exhibits bantam-dependent Slicer activity.

The AGO1-DCR-1 complex exhibits pre-miRNA processing activity, but the AGO1-GW182 does not. (A) Immunoprecipitated complexes (Fig. 2A) were subjected to in vitro pre-miRNA processing assays using pre-let-7 labeled with ³²P. The AGO1-GW182 complex does not show the activity because the complex contains little or no DCR-1. (B) Immunoprecipitated complexes (Fig. 2A) were first subjected to in vitro pre-miRNA processing using pre-let-7 (cold), and then target RNA cleavage assays were performed. The AGO1-GW182 complex, which did not show pre-let-7 processing activity in A, does not cleave the target (let-7 target; L7Pv), whereas the AGO1-DCR-1 complex shows the activity, indicating that the AGO1 that was originally contained in the complex was loaded with mature let-7 processed by the immunoprecipitated complex. (C) pre-let-7 processing was performed using let-7-130 as a target. After the reaction, the supernatant and bead fractions were separated by centrifugation, and then target RNA cleavage assays were performed as in B. The fractions were also subjected to Western blotting using anti-DCR-1 and anti-AGO1. AGO1 is detected both in the supernatant (S) and bead (B) fractions of anti-LOQS immunoprecipitates, whereas DCR-1 is only detected in the bead fraction, indicating that AGO1 dissociates from DCR-1 upon miRNA loading. Such dissociation was confirmed by performing target RNA cleavage assays. The cleavage products are observed in both the supernatant and bead fractions of anti-LOQS immunoprecipitates. The supernatant fraction of anti-AGO1 immunoprecipitates shows no activity to cleave the target RNA, because AGO1 stayed on the beads during the assays. The bead fraction of anti-LOQS shows stronger cleavage activity than expected. This was because pre-miRNA processing continuously occurred even after the target RNA cleavage reaction was started. The lowest panel shows that even under conditions where pre-let-7 was not added, AGO1 was released to the supernatant, indicating that AGO1 displacement from DCR-1 occurs in an miRNA-loading-independent manner. (D) GW182 and DCR-1 were depleted from S2 cells by RNAi, and RNAs isolated from individual cells were subjected to Northern blotting using a DNA oligo for bantam. When DCR-1 was depleted, pre-bantam was aberrantly accumulated, whereas GW182 depletion did not cause such an effect. The lower panel (Western blot) shows the high efficiency of DCR-1 and GW182 depletion in each sample. (E) GW182 and DCR-1 are depleted from S2 cells as in D, and the resultant cell lysates were subjected to pre-miRNA processing assays. Only DCR-1 depletion causes a severe decrease in pre-miRNA processing activity.

Visualization of miRLC and miRISC on native agarose gels. (A) pre-let-7 was labeled with ³²P and incubated in S2 cell lysates in the presence of EDTA for 3 min and 30 min. After 30 min of incubation, Mg²⁺ was added to the reaction mixture to induce DCR-1 activity. Further incubation was carried out as indicated. The resultant samples at each time point were loaded onto an agarose gel, and complexes containing either pre-let-7 or mature let-7 were visualized after running the gel. Simultaneously, siRNA duplex labeled with ³²P was incubated in the S2 lysates in the presence of EDTA or Mg²⁺. After 30 min of incubation, both the lysates were loaded on the same gel system. The input lanes on the far left and right sides show pre-miRNA and siRNA duplexes by themselves, respectively. The bands with an asterisk and two asterisks correspond to miRLC and miRISC, respectively. The bands corresponding to siRLC and siRISC are as indicated on the gel. The band shown with a black triangle is an unknown product, which did not appear in previous equivalent experiments (Miyoshi et al. 2005) but appeared in this study. (B) The bands indicated as miRLC and miRISC in A were isolated from the gel, and RNAs isolated were subjected to a native acrylamide gel. The miRLC contained pre-let-7, whereas miRISC contained mostly single-stranded let-7. (Input lane) Contains pre-let-7 as the starting material. Pre-let-7 incubated in S2 lysates was also subjected to the same gel (S2 lysate + pre-let-7), which shows the migration patterns of pre-let-7, let-7/let-7*, and mature let-7.

The requirement for various domains of AGO1 in binding with miRNA, DCR-1, and GW182. (A) Two mutants of AGO1, the PAZ and F2V2 mutants, were produced by mutagenesis and employed for binding assays with miRNA, DCR-1 and GW182. Proteins associated with Flag-AGO1-WT and Flag-AGO1-PAZ and Flag-AGO1-F2V2 mutants were blotted with anti-DCR-1, anti-GW182, and anti-Flag antibodies. All interacted with DCR-1, but the F2V2 mutant failed to associate with GW182, unlike the wild type (WT) and the PAZ mutant. Northern blotting revealed that neither of the mutants associates with mature bantam. (B) The F2V2 mutant (red), which lacks binding capacity to GW182, fails to accumulate in P-bodies, whereas the PAZ mutant (red) co-localizes with DCP1 (green), a marker for P-bodies. (C) The D-D-H motif in the PIWI domain was mutated to A-D-H or D-A-H. Both mutants were expressed in S2 cells in a Flag-tagged form. Both mutants interacted with DCR-1 and GW182 and also mature single-stranded miRNA. In vitro target RNA cleavage assays show that both mutants lack Slicer activity, indicating that Slicer activity of AGO1 is dispensable for miRNA loading. The Δpiwi mutant interacts with DCR-1 but not with GW182, indicating that the domains required for binding with DCR-1 and GW182 are different.