• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of molcellbPermissionsJournals.ASM.orgJournalMCB ArticleJournal InfoAuthorsReviewers
Mol Cell Biol. Dec 2009; 29(23): 6220–6231.
Published online Sep 21, 2009. doi:  10.1128/MCB.01081-09
PMCID: PMC2786699

The Silencing Domain of GW182 Interacts with PABPC1 To Promote Translational Repression and Degradation of MicroRNA Targets and Is Required for Target Release[down-pointing small open triangle]

Abstract

GW182 family proteins are essential in animal cells for microRNA (miRNA)-mediated gene silencing, yet the molecular mechanism that allows GW182 to promote translational repression and mRNA decay remains largely unknown. Previous studies showed that while the GW182 N-terminal domain interacts with Argonaute proteins, translational repression and degradation of miRNA targets are promoted by a bipartite silencing domain comprising the GW182 middle and C-terminal regions. Here we show that the GW182 C-terminal region is required for GW182 to release silenced mRNPs; moreover, GW182 dissociates from miRNA targets at a step of silencing downstream of deadenylation, indicating that GW182 is required to initiate but not to maintain silencing. In addition, we show that the GW182 bipartite silencing domain competes with eukaryotic initiation factor 4G for binding to PABPC1. The GW182-PABPC1 interaction is also required for miRNA target degradation; accordingly, we observed that PABPC1 associates with components of the CCR4-NOT deadenylase complex. Finally, we show that PABPC1 overexpression suppresses the silencing of miRNA targets. We propose a model in which the GW182 silencing domain promotes translational repression, at least in part, by interfering with mRNA circularization and also recruits the deadenylase complex through the interaction with PABPC1.

In multicellular eukaryotes, the regulation of gene expression by microRNAs (miRNAs) is critical for biological processes as diverse as cell differentiation and proliferation, apoptosis, metabolism, and development (4). To exert a regulatory function, miRNAs associate with Argonaute proteins to form RNA-induced silencing complexes, which repress translation and trigger the degradation of target mRNAs (4, 10, 16). The extent to which translational repression and degradation contribute to silencing depends on the specific target-miRNA combination; some targets are regulated predominantly at the translational level, whereas others can be regulated mainly at the mRNA level (3). A large-scale proteomic analysis performed in parallel with measurements of mRNA levels showed that for the vast majority of miRNA targets, silencing correlates with changes at both the protein and mRNA levels (1, 27).

In animal cells, the degradation of miRNA targets is initiated by deadenylation and decapping, which are followed by the exonucleolytic decay of the mRNA body (2, 3, 9, 11, 12, 17, 19, 24, 30, 31). miRNA-dependent mRNA degradation requires a variety of proteins: an Argonaute and a GW182 protein, the CCR4-NOT deadenylase complex, the decapping enzyme DCP2, and several decapping activators including DCP1, Ge-1, HPat, EDC3, and Me31B (also known as RCK/p54) (3, 6, 9, 12, 19). Several studies previously demonstrated that miRNAs trigger deadenylation and decapping even when the mRNA target is not translated (9, 12, 19, 24, 30, 31), indicating that mRNA decay is not merely a consequence of a primary effect of miRNAs on translation but rather is an independent mechanism by which miRNAs silence gene expression.

Although how miRNAs trigger mRNA degradation is well established, the mechanisms driving the inhibition of translation are unclear. Multiple mechanisms have been proposed: the displacement of eukaryotic initiation factor 4E (eIF4E) from the mRNA cap structure, interference with the function of the eIF4F complex, a block of 60S ribosomal subunit joining, or an inhibition of translation elongation (4, 10, 16). Regardless of the precise mechanism, the translational repression of miRNA targets also requires GW182 family proteins (11, 13).

GW182 proteins are essential components of the miRNA pathway in animal cells, as their depletion suppresses miRNA-mediated gene silencing (reviewed in references 8 and 13). Recent studies have revealed that the silencing activity of these proteins resides predominantly in a bipartite silencing domain containing the middle and C-terminal regions (14, 22, 33). The precise molecular function of the GW182 silencing domain is not fully understood, yet it is known that the domain is not required for GW182 proteins to interact with Argonaute proteins or to localize to P bodies (3, 14, 22). Furthermore, when the silencing domains of GW182 proteins are artificially tethered to mRNAs, their expression is silenced; therefore, tethering bypasses the requirement for Argonaute proteins and miRNAs (5, 22, 33). These observations suggest that the silencing domains of GW182 proteins exhibit intrinsic silencing activity and therefore likely play a role at the effector step of silencing (13, 14, 22, 33).

Here we investigate what role the Drosophila melanogaster GW182 silencing domain plays in the miRNA pathway. Overall, our results reveal that the very C-terminal region of this domain is required for the release of GW182 from silenced mRNPs. Indeed, we unexpectedly found that we could detect D. melanogaster GW182 bound to miRNA targets only in cells depleted of components of the deadenylase complex. These results suggest that GW182 dissociates from Argonaute-1 (AGO1) and miRNA targets at a step of silencing downstream of deadenylation. In contrast, GW182 mutants lacking the C-terminal region remain stably bound to miRNA targets, even in wild-type cells, indicating that this region plays a role in the dissociation of GW182 from effector complexes. We further show that the bipartite silencing domain of GW182 interacts with PABPC1 and interferes with the binding of PABPC1 to eIF4G. The interaction of GW182 with PABPC1 is also required for the degradation of miRNA targets, most likely because the interaction facilitates the recruitment of the CCR4-NOT deadenylase complex. Accordingly, overexpressing PABPC1 suppresses the silencing of miRNA targets. Our findings uncover an unexpected role for PABPC1 in the miRNA pathway.

MATERIALS AND METHODS

DNA constructs.

Luciferase reporters and plasmids for the expression of miRNAs and hemagglutinin (HA)- or green fluorescent protein (GFP)-tagged AGO1 and GW182 were described previously (3, 9, 11). Renilla luciferase cloned between the KpnI and XhoI sites of vector pAc5.1A (Invitrogen) served as a transfection control. The AGO1 and GW182 mutants used in this study were described previously (14). Plasmids for the expression of HA-tagged eIF4E and eIF4G were obtained by inserting the eIF4E (CG4035) and eIF4G (CG10811) open reading frames (ORFs) into EcoRI-NotI and NotI-XbaI sites, respectively, of vector pAc5.1-λN-HA. An expression plasmid encoding C-terminally V5-tagged PABPC1 was obtained by inserting the PABPC1 (CG5119) ORF into the KpnI-XhoI sites of vector pAc5.1-λN-V5 in frame with the C-terminal V5 epitope. Glutathione S-transferase (GST) cDNA was cloned into the EcoRI-NotI sites of vector pAc5.1-λN-HA-V5. Plasmids for the expression of deadenylase complex components were obtained by inserting the corresponding cDNAs into vector pAc5.1-EGFP using the restriction sites EcoV-NotI (POP2, which is related to CAF1), EcoRI-NotI (CCR4), HindIII-XbaI (NOT2), HindIII-NotI (NOT3/5), and XhoI-BstBI (NOT4).

RNA interference, transfections, and luciferase assays.

RNA interference and complementation assays were performed as described previously (3, 9, 14). S2 cells were transfected in six-well plates using Effectene transfection reagent (Qiagen). For coimmunoprecipitation assays, the transfection mixtures contained 0.5 μg of firefly luciferase (F-Luc) reporter plasmid, 0.5 μg of the Renilla transfection control, and either 0.5 μg of plasmid expressing miR-9b, 0.8 μg of plasmid expressing miR-279 primary transcripts, or the corresponding vector without an insert (unless indicated otherwise) (e.g., see Fig. Fig.1).1). In addition, 0.5 μg of plasmids expressing recombinant proteins was cotransfected. For the tethering assay, the following plasmids were cotransfected: 0.1 μg reporter plasmid (F-Luc-5BoxB or F-Luc), 0.4 μg pAc5.1-R-Luc as a transfection control, and 0.025 μg of plasmids expressing λN-HA-protein fusions. In the experiments shown in Fig. 8B to E, the transfection mixtures contained 0.1 μg of F-Luc reporter plasmid, 0.4 μg of the Renilla transfection control, and either 0.2 μg of plasmid expressing miRNA primary transcripts or the corresponding vector without an insert. For PABPC1 overexpression, 0.2 μg of plasmids expressing full-length PABPC1 or fragments (lacking the λN tag) were included in the transfection mixtures. In all experiments, F-Luc and Renilla luciferase activities were measured 3 days after transfection using a dual-luciferase reporter assay system (Promega).

FIG. 1.
miRNA-dependent AGO1 association with miRNA targets. (A to D) S2 cells were transfected with a mixture of four plasmids: one expressing the F-Luc-Nerfin-1 reporter, another expressing miR-279 (increasing amounts as indicated) or the corresponding empty ...
FIG. 8.
Overexpression of PABPC1 suppresses miRNA-mediated silencing. (A) S2 cells were transfected with a mixture of three plasmids: one expressing the F-Luc-5BoxB reporter; another expressing the λN-HA peptide (white bars), λN-HA-GW182 (dark ...

Coimmunoprecipitations and Western blots.

The interaction of GW182 with endogenous miRNAs and AGO1 was tested as described previously (11). Antibodies to AGO1 (dilution, 1:1,000) were purchased from Abcam (catalog number ab5070). For coimmunoprecipitation assays, S2 cells (10 × 106 to 12 × 106 cells) were collected 3 days after transfection, washed with phosphate-buffered saline, and lysed in 0.5 ml of NET buffer (50 mM Tris [pH 7.4], 150 mM NaCl, 1 mM EDTA, 0.1% NP-40) supplemented with protease inhibitors. Cells were lysed by three 30-s sonications, followed by a 15-min incubation on ice. Cells were spun at 16,000 × g for 15 min at 4°C. Aliquots (1/5) of the cleared lysate (input) were kept aside for both RNA extraction and Western blotting analysis. Anti-HA antibodies (Covance Research Products) were added to the cleared lysates (2.5 μl/2 × 106 cells). After 1 h at 4°C, 25 μl of protein G-agarose (Roche) that had been preincubated with 0.5 mg of yeast RNA for 1 h at 4°C was added, and the mixtures were rotated for 1 h at 4°C. Beads were washed four times with NET buffer and once with NET buffer without NP-40. Beads were resuspended in 100 μl of NET buffer without NP-40, and one-third of the bead suspension was used to examine the efficacy of the immunoprecipitation by Western blotting. The remaining bead fractions were used for RNA analysis. For the immunoprecipitations shown in Fig. Fig.6,6, anti-GFP antibodies were used, and beads were directly resuspended in protein sample buffer.

FIG. 6.
The silencing domain of GW182 interacts with PABPC1. (A and B) Lysates from S2 cells coexpressing GFP-tagged F-Luc, wild-type GW182, or GW182 mutants together with V5-tagged PABPC1 were immunoprecipitated using a polyclonal anti-GFP antibody. Inputs (4%) ...

RNA extraction, reverse transcription (RT), and real-time qPCR.

Samples corresponding to input and immunoprecipitates were resuspended in Trizol-LS reagent (Invitrogen). RNA was prepared according to the manufacturer's protocol. DNase treatment was performed by using the Turbo DNA-free kit (Ambion) for 30 min at 37°C. RNAs were detected via cDNA synthesis and real-time quantitative PCR (qPCR). cDNAs were synthesized with Moloney murine leukemia virus reverse transcriptase (Fermentas) and random primers (Sigma) according to the manufacturers' protocols. qPCR analysis was performed using gene-specific primer pairs (indicated below) and SYBR green PCR master mix (Applied Biosystems). Each sample was analyzed in triplicate. Results were evaluated by the comparative threshold cycle method (26) by using Renilla luciferase mRNA as the invariant control gene. Primer sequences for the F-Luc-Nerfin reporter targeting the 3′ untranslated region of Nerfin are forward primer 5′-CTCTAGATTTAGCTAGTTTAGGAACTTTT-3′ and reverse primer 5′-ATTAAGGGTAGAATATTAACTGGAAAAT-3′. Primer sequences for the F-Luc-Vha68-1 reporter targeting the ORF of F-Luc are forward primer 5′-ATGGCCGAAGACGCCAAAAACATAAAG-3′ and reverse primer 5′-AATAACGCGCCCAACACCGGCA-3′. Primer sequences for the Renilla luciferase reporter are forward primer 5′-CGGGGTACCAACATGACTTCGAAAGTTTATGA-3′ and reverse primer 5′-CTCGAGTTGTTCATTTTTGAGAACTC-3′.

RESULTS

GW182 is not required for target recognition by AGO1.

Previous studies showed that a GW182 mutant that lacks the silencing domain (GW182-ΔSD; i.e., lacking the middle and C-terminal regions but not the RNA recognition motif [RRM]) is inactive in silencing (14) despite the finding that this mutant interacts with AGO1 and miRNAs. Indeed, in cells expressing GW182-ΔSD but lacking endogenous GW182, miRNA targets are neither translationally repressed nor degraded (14). These results might suggest that GW182-ΔSD cannot recruit deadenylase complex components or decapping activators, nor can it interfere with translation (14), but an alternative explanation could be that GW182-ΔSD does not associate with miRNA targets. To distinguish between these two possibilities, we compared the association of wild-type GW182 and GW182-ΔSD with miRNA targets in the absence or presence of cognate miRNAs. To do this, we immunoprecipitated GW182 and used real-time RT-qPCR to analyze the levels of miRNA reporters coimmunoprecipitating with GW182.

To validate our approach, we first analyzed the association of HA-tagged AGO1 with miRNA reporters. The reporters consisted of the F-Luc ORF followed by the 3′ untranslated regions of the D. melanogaster genes nerfin-1 (silenced by miR-9b and miR-279) and vha68-1 (silenced by miR-9b) (3). These reporters are silenced predominantly at the translational level (ca. fivefold repression), although in our experiments, mRNA levels were also reduced 1.5- and 2-fold, respectively (data not shown; 3, 9). Plasmids expressing Renilla luciferase and HA-GST served as a transfection control and as a negative control for the immunoprecipitations, respectively.

When increasing amounts of a plasmid encoding miR-279 were cotransfected with the F-Luc-Nerfin-1 reporter, the level of F-Luc expression decreased (Fig. (Fig.1A).1A). Increasing the amount of miR-279 increased silencing efficacy, which correlated with the amount of reporter mRNA that coimmunoprecipitated with HA-AGO1 (Fig. (Fig.1B).1B). In contrast, in the absence of miR-279, AGO1 did not coimmunoprecipitate more target mRNA than the negative control, HA-GST (Fig. 1B and D). The efficacies of AGO1 or GST immunoprecipitations were comparable irrespective of the amounts of miR-279 (Fig. (Fig.1E).1E). Furthermore, AGO1 preferentially associated with the F-Luc-Nerfin-1 reporter only in the presence of miR-279 or miR-9b but not in the presence of a noncognate miRNA such as miR-12, which does not silence the F-Luc-Nerfin-1 reporter (Fig. 1F and G). It should be noted that silencing was more efficient in cells overexpressing AGO1 than in cells overexpressing GST (Fig. (Fig.1A1A versus C), suggesting that endogenous AGO1 is limiting under our experimental conditions.

To further validate data from the coimmunoprecipitation assays, we tested whether reporter mRNAs coimmunoprecipitate with AGO1 mutants that do not interact with miRNAs and/or GW182. In particular, we tested an AGO1 mutant carrying six mutations in the PAZ domain (PAZ6) (14). This mutant does not interact with miRNAs and therefore is inactive in silencing. We also tested an AGO1 mutant carrying valine substitutions of phenylalanine 594 (F594V) and F629V (referred to as F2V2), which is inactive in silencing because it interacts with neither miRNAs nor GW182 (11). The PAZ6 and F2V2 AGO1 mutants did not coimmunoprecipitate the F-Luc-Nerfin-1 reporter above background levels both in the presence and in the absence of miRNAs (Fig. 2A and B). Collectively, these results clearly demonstrate that the association between AGO1 and miRNA reporters detected in our assay reflects the formation of silenced mRNP complexes.

FIG. 2.
AGO1 requires miRNAs but not GW182 for target recognition. (A and B) Coimmunoprecipitation of the F-Luc-Nerfin-1 reporter with wild-type AGO1 or mutants in the absence or presence of miR-279, analyzed as described in the legend of Fig. Fig.1. ...

Finally, we tested whether AGO1 mutants that are impaired in binding to GW182 but that still interact with miRNAs can coimmunoprecipitate mRNA targets. In particular, we tested AGO1 mutants carrying an alanine substitution at residue F777, either alone or in combination with R771. In previous complementation assays, these mutations impaired AGO1 silencing activity (14). In our experiments, the AGO1 mutants coimmunoprecipitated the F-Luc-Nerfin-1 reporter in an miR-279-dependent manner and as efficiently as wild-type AGO1 (Fig. 2A and B). Wild-type AGO1 and mutants were expressed and immunoprecipitated at comparable levels (Fig. (Fig.2C).2C). Note that under our experimental conditions (i.e., 0.5 μg of transfected plasmid), AGO1 mutants did not inhibit silencing in a dominant-negative manner. We conclude that GW182 is not required for miRNA target recognition.

GW182 is not detectably bound to silenced mRNAs.

Next, we examined the association of GW182 with miRNA targets. Unexpectedly, in the presence of cognate miRNAs, GW182 did not coimmunoprecipitate mRNA reporters above background levels (Fig. 3A to F), although the protein was detected in the immunoprecipitates (Fig. (Fig.3G).3G). One possible explanation for this observation is that the levels of HA-GW182 are too low relative to those of endogenous GW182, preventing the tagged protein from being incorporated into functional complexes at detectable levels. However, even in cells depleted of endogenous GW182, HA-GW182 did not coimmunoprecipitate the reporters (data not shown), despite the finding that in these cells, HA-GW182 rescues silencing (14). Furthermore, HA-GW182 coimmunoprecipitated endogenous AGO1 and Bantam miRNA, confirming that this protein is functional (Fig. (Fig.3H,3H, lane 5). Also as expected, a negative control GW182 mutant (GW-AA) that carries alanine substitutions of all N-terminal GW repeats associated with neither AGO1 nor Bantam (Fig. (Fig.3H,3H, lane 6).

FIG. 3.
GW182 does not coimmunoprecipitate silenced mRNPs. (A to F) Coimmunoprecipitation of the F-Luc-Nerfin-1 or the F-Luc-Vha68-1 reporter with HA-GW182 in the absence or presence of cognate miRNAs, analyzed as described in the legend of Fig. Fig. ...

One potential reason for the inability of HA-GW182 to coimmunoprecipitate F-Luc reporters may be that the HA tag is not accessible in silenced mRNPs. To examine this possibility, we performed coimmunoprecipitations using N-terminally GFP-tagged or C-terminally V5-tagged GW182. However, again, GFP-GW182 and GW182-V5 did not coimmunoprecipitate mRNA reporters. Under the same conditions, GFP-AGO1 associates with the reporters, as observed for HA-AGO1 (see Fig. S1A to S1D in the supplemental material). GFP- and V5-tagged GW182 complemented silencing in cells depleted of endogenous GW182, indicating that the tags do not interfere with the activity of the protein (see Fig. S1E and S1F in the supplemental material).

Our finding that GW182 coimmunoprecipitates AGO1 and miRNAs but not miRNA targets suggests that GW182 may dissociate before or immediately after target recognition. Alternatively, the GW182 interaction with AGO1 may be loosened after target binding so that the complex dissociates under our coimmunoprecipitation conditions. However, the results described below and the observation that GW182 did not coimmunoprecipitate miRNA reporters, even when coimmunoprecipitations were performed under conditions of low salt (50 mM NaCl instead of 150 mM) (data not shown), suggest that GW182 is not stably bound to silenced mRNAs.

GW182 coimmunoprecipitates miRNA targets in cells depleted of NOT1.

GW182 is required for the degradation of miRNA targets and can trigger the deadenylation and decapping of bound mRNAs independently of Argonaute proteins (i.e., in tethering assays [3, 5, 11, 12]), so we predicted that GW182 should be bound to the reporter prior to deadenylation. To investigate whether deadenylation triggers GW182 to dissociate from reporters, we performed coimmunoprecipitation assays with cells depleted of NOT1, a component of the deadenylase complex whose depletion inhibits deadenylation in S2 cells (3, 29). Consistent with our expectation, in the presence of miRNAs, GW182 preferentially coimmunoprecipitated the reporters in NOT1-depleted cells but not in control cells (Fig. 4A and B). In NOT1-depleted cells, the F-Luc-Nerfin-1 reporter remains silenced, as reported previously (3) (Fig. 4A and B), although NOT1 levels were reduced (Fig. (Fig.4C).4C). We therefore conclude that GW182 dissociates from miRNA targets at a step of silencing downstream of deadenylation.

FIG. 4.
GW182 dissociates from silenced mRNAs after deadenylation. (A and B) Control S2 cells (treated with GFP double-stranded RNA) or cells depleted of NOT1 were cotransfected with a mixture of four plasmids: one expressing the F-Luc-Nerfin-1 reporter; another ...

The C-terminal domain of GW182 is required for target release.

Because a GW182 mutant lacking the silencing domain (GW182-ΔSD; i.e., lacking the middle and C-terminal regions but not the RRM) fails to trigger the deadenylation and decay of miRNA targets, we predicted that this mutant should remain bound to the mRNA target. In agreement with our expectation, GW182-ΔSD coimmunoprecipitated the F-Luc-Nerfin-1 reporter in the presence of miR-9b (Fig. (Fig.5B5B).

FIG. 5.
The C-terminal domain is required for GW182 to dissociate from silenced mRNAs. (A) Schematic diagram showing GW182 domain organization. The N-terminal AGO-binding domain is indicated. UBA, ubiquitin associated-like domain; Q-rich, region rich in glutamine ...

The silencing domain is bipartite and comprises the GW182 middle and C-terminal regions (Fig. (Fig.5A)5A) (13, 14). We therefore tested whether GW182 mutants lacking either the middle region (GW182-ΔM) or the very C-terminal region (GW182-ΔC-term) associated with the reporter. These GW182 mutants partially rescued silencing in cells depleted of endogenous GW182 (14). In the presence of miR-9b, GW182-ΔC-term preferentially coimmunoprecipitated the F-Luc-Nerfin-1 reporter, similar to GW182-ΔSD (Fig. (Fig.5B).5B). In contrast, the GW182-ΔM mutant failed to coimmunoprecipitate the reporter (Fig. (Fig.5B).5B). Wild-type GW182 and mutants were expressed and immunoprecipitated at comparable levels (Fig. (Fig.5C).5C). These results show that the C-terminal region is required for GW182 to dissociate from AGO1 and miRNA targets. Thus, although both the middle and C-terminal regions of GW182 are required for silencing, these regions have some distinct functions.

The GW182-AGO1 interaction is enhanced in NOT1-depleted cells.

We previously showed that GW182-ΔSD and GW182-ΔC-term interact with AGO1 more efficiently than wild-type GW182 (14) (Fig. (Fig.5D,5D, lanes 10 and 11 versus lane 8). Our observations suggest that the mutants bind more tightly to AGO1 because they remain bound to the target mRNA. If this was true, then the interaction between wild-type GW182 and AGO1 should be enhanced in NOT1-depleted cells, as GW182 remains bound to the target under these conditions. Consistent with our expectation, in NOT1-depleted cells, GW182 coimmunoprecipitated endogenous AGO1 as efficiently as the mutants lacking the C-terminal domain (Fig. (Fig.5D,5D, lane 20 versus lane 22, and see Fig. S2 in the supplemental material). In control cells, in contrast, GW182 coimmunoprecipitated threefold less AGO1 than mutants lacking the C-terminal domain (Fig. (Fig.5D,5D, lane 8 versus lane 10, and see Fig. S2 in the supplemental material). As a negative control, a GW182 mutant that carries alanine substitutions of all N-terminal GW repeats (GW-AA) did not coimmunoprecipitate AGO1 in control or depleted cells (Fig. (Fig.5D,5D, lanes 12 and 24). These results provide further support for the conclusion that GW182 dissociates from AGO1 and miRNA targets at a step of silencing downstream of deadenylation.

The GW182 silencing domain interacts with PABPC1.

The results described above show that the GW182-ΔSD mutant binds miRNA targets; consequently, they suggest that this mutant lacks silencing activity because it cannot recruit the deadenylase and decapping complex and/or interfere with translation. To test this hypothesis, we examined whether wild-type GW182 and mutant GW182 interact with components of the deadenylase complex (POP2 and NOT3/5) or with decapping activators (DCP1 and Me31B). We also tested for interactions with PABPC1 and components of the eIF4F complex (eIF4E and eIF4G), because it was previously proposed that miRNAs interfere with the function of these proteins (23, 30). We observed that GW182 coimmunoprecipitated PABPC1 but not the other factors tested (Fig. (Fig.6A6A and see Fig. S3 in the supplemental material).

The GW182-PABPC1 interaction was observed in the presence of RNase A, suggesting that it is not RNA mediated (Fig. (Fig.6A,6A, lane 16). This is consistent with the observation that PABPC1 still associates with the GW182 mutant (GW-AA), which does not interact with AGO1 (Fig. (Fig.6A,6A, lanes 10 and 17). Thus, GW182 interacts with PABPC1 independently of AGO1 and miRNA targets. Interestingly, the GW182 mutant lacking the silencing domain (ΔSD) (lacking the middle and C-terminal regions) failed to interact with PABPC1, but mutants lacking either the middle or C-terminal regions did interact with PABPC1 although slightly less efficiently than did wild-type GW182 (Fig. (Fig.6A,6A, lanes 12 to 14 and 19 to 21, and B, lanes 13 to 15). Deleting the GW182 RRM did not affect the interaction with PABPC1 (Fig. (Fig.6A,6A, lane 11). Therefore, both the middle and C-terminal regions (the silencing domain), but not the RRM, are required for GW182 to interact with PABPC1.

The silencing domain is also sufficient for PABPC1 binding. Indeed, protein fragments containing the middle and C-terminal regions (including or excluding the RRM) interacted with PABPC1 equally well (Fig. (Fig.6D,6D, lanes 8 and 9). We conclude that the GW182 middle and C-terminal regions define a single functional bipartite domain required for both binding to PABPC1 and the silencing of miRNA targets.

Conserved motif III together with the M2 and C-terminal regions define the minimal PABPC1-binding domain in GW182.

The middle region of GW182 is not highly conserved, with the exception of a motif of about 40 residues (conserved motif III, also known as DUF [33]) (Fig. (Fig.6C).6C). Moreover, Zipprich et al. (33) previously showed that sequences upstream of motif III are not required for silencing. Based on these observations, the middle region can be divided into three segments: M1, motif III, and M2 (Fig. (Fig.6C).6C). To define which sequences within the middle region contribute to PABPC1 binding, we tested whether GW182 mutants lacking either the M2 region or motif III interacted with PABPC1. We observed that these mutants interacted with PABPC1 more efficiently than did a GW182 mutant lacking the C-terminal region (Fig. (Fig.6B,6B, lanes 16 and 17 versus lane 14). These results suggest that the C-terminal region provides a major contribution to the interaction with PABPC1. PABPC1 binding was abolished when, in addition to the C-terminal region, motif III and the M2 region were deleted (Fig. (Fig.6D,6D, lane 18).

To examine whether motif III and the M2 and C-terminal regions were also sufficient for PABPC1 binding, we tested fragments containing the entire silencing domain or the different regions in various combinations. We observed that a GW182 fragment containing motif III and the M2 and C-terminal regions was sufficient to interact with PABPC1 (Fig. (Fig.6D,6D, lane 12 versus lane 9). Fragments lacking motif III interacted with PABPC1 although less efficiently, indicating that motif III contributes but is not absolutely required (Fig. (Fig.6D,6D, lane 11). We conclude that motif III together with the M2 and C-terminal regions defines the minimal PABPC1-binding domain in GW182.

The GW182 silencing domain competes with eIF4G for binding to PABPC1.

To address the functional significance of the GW182-PABPC1 interaction for silencing, we first determined which domain of PABPC1 interacted with GW182. PABPC1 contains four conserved RRMs (RRM1 to RRM4) followed by a linker region and a C-terminal domain (termed the PABC domain [7]). We observed that GW182 interacts with the PABPC1 N-terminal domain containing the four RRMs (Fig. (Fig.7A).7A). PABPC1 fragments comprising RRM1 and RRM2, RRM2 and RRM3, RRM3 and RRM4, or the linker and C-terminal domain (L+C) did not interact with GW182 above background levels (Fig. (Fig.7A).7A). These results indicate that the four RRMs are required for PABPC1 to interact with GW182.

FIG. 7.
PABPC1 acts as a bridging factor between the silencing machinery and translation and decay factors. (A) Lysates from S2 cells coexpressing F-Luc-GFP or GFP-GW182 together with V5-tagged wild-type PABPC1 or mutants were immunoprecipitated using a polyclonal ...

The interaction of GW182 with the RRMs of PABPC1 prompted us to examine whether GW182 may compete with eIF4G for PABPC1 binding. Support for this idea comes from the previously reported observation that the N-terminal domain of eIF4G interacts with RRM1 and RRM2 of PABPC1 (18, 20). Accordingly, we observed that D. melanogaster PABPC1 coimmunoprecipitated eIF4G (Fig. (Fig.7B,7B, lanes 1 to 5). Moreover, we found that if the GW182 silencing domain was added to preformed eIF4G/PABPC1 complexes, then it could compete away eIF4G binding to PABPC1 (Fig. (Fig.7B,7B, lanes 6 and 7). Together, these results indicate that GW182 and eIF4G interact with PABPC1 in a mutually exclusive manner.

PABPC1 interacts with components of the CCR4-NOT deadenylase complex.

The experiments described above suggest that GW182 binding to PABPC1 may contribute to silencing by interfering with the formation of the mRNA closed-loop conformation. This mechanism may facilitate translational repression but cannot fully account for the degradation of miRNA targets, which occurs even when the target is not translated (9, 12, 19, 24, 30, 31). Furthermore, the GW182 silencing domain triggers the deadenylation of bound mRNAs in tethering assays but does not interact with components of the deadenylase complex. We therefore asked whether these interactions could be mediated by PABPC1, which also plays a role in mRNA degradation (28). In agreement with this possibility, we observed that PABPC1 coimmunoprecipitated components of the CCR4-NOT complex (except POP2) in an RNA-independent manner (Fig. (Fig.7C).7C). These results suggest that PABPC1 may act as an adaptor molecule that bridges the interaction between the silencing machinery and general mRNA decay enzymes.

The GW182 interaction with PABPC1 is required for silencing of bound mRNAs.

Previous studies showed that GW182 promotes the translational repression and degradation of bound mRNAs in tethering assays (5, 15, 22, 33). To investigate whether the activity of GW182 is mediated by PABPC1, we performed tethering assays with cells overexpressing PABPC1 (Fig. (Fig.8A8A).

The overexpression of full-length PABPC1 relieved silencing mediated by GW182 (Fig. (Fig.8A).8A). Similar results were obtained with a PABPC1 fragment containing the four RRM domains (RRM1 to RRM4), which interacts with GW182 (Fig. (Fig.8A).8A). In contrast, a PABPC1 fragment that does not interact with GW182 (the RRM1+2 fragment) had no effect on silencing (Fig. (Fig.8A).8A). Importantly, in cells expressing the λN-HA peptide or λN-HA-AGO1-F2V2 (a negative control), PABPC1 overexpression did not interfere with the translation of the F-Luc-5BoxB reporter (Fig. (Fig.8A8A and data not shown). These results indicate that the suppression of GW182-mediated silencing observed in cells overexpressing PABPC1 is not due to a nonspecific stimulation of translation. These results also suggest that GW182 interacts with PABPC1 to repress the expression of bound mRNAs.

The GW182-PABPC1 interaction is required for silencing miRNA targets.

To further investigate the role of PABPC1 in silencing, we monitored miRNA activity in cells overexpressing PABPC1. The overexpression of full-length PABPC1 relieved the silencing of the F-Luc-Nerfin-1 or the F-Luc-Vha68-1 reporter (Fig. 8B and C). Similar results were obtained with a PABPC1 fragment containing the four RRM domains (RRM1 to RRM4), which interacts with GW182 (Fig. 8B and C). In contrast, PABPC1 fragments that do not interact with GW182 (RRM1+2 or L+C) had no effect on silencing (Fig. 8B and C). Thus, the overexpression of PABPC1 relieves miRNA-mediated silencing specifically, most likely by sequestering GW182 into inactive complexes.

The results obtained with the F-Luc-Nerfin-1 and F-Luc-Vha68-1 reporters were confirmed with the F-Luc-par-6 reporter, which is extensively degraded in the presence of miR-1 (9). Again, the overexpression of PABPC1 partially restored both F-Luc expression and mRNA levels (Fig. (Fig.8D8D and data not shown). In contrast, no effect was seen when PABPC1 fragments that do not interact with GW182 (e.g., RRM1+2 or L+C) were expressed (Fig. (Fig.8D).8D). In the absence of miRNAs, firefly and Renilla luciferase activities in control cells and in cells overexpressing PABPC1 were comparable (Fig. (Fig.8E8E and data not shown), indicating that the overexpression of PABPC1 does not stimulate translation. We conclude that irrespective of whether target mRNAs are repressed by blocking translation or triggering mRNA decay, the overexpression of PABPC1 inhibits miRNA-mediated silencing.

DISCUSSION

The data presented in this study provide insight into the role of GW182 in the miRNA pathway. First, GW182 dissociates from silenced mRNAs after deadenylation. Second, the release of GW182 from silenced mRNAs requires the very C-terminal domain of the protein. Third, GW182 interacts with PABPC1 through its silencing domain; this interaction is required for both degradation and translational repression of miRNA targets. Fourth, GW182 competes with eIF4G for binding to PABPC1. Together, these findings depict a scenario in which the GW182 silencing domain interferes with the eIF4G-PABPC1 interaction, thereby repressing translation and rendering the mRNA 5′ cap and poly(A) tail accessible to decay enzymes. Finally, we also show that PABPC1 may act as a protein adaptor, bridging GW182 and the general mRNA decay machinery. The implications of these findings are discussed below.

GW182 dissociates from silenced mRNAs.

Together with data from previous studies, our results indicate that GW182 and AGO1 form a complex that binds miRNA targets; this binding triggers translational repression and mRNA degradation. In D. melanogaster cells, GW182 dissociates from miRNA targets at a step of silencing downstream of deadenylation. Although the temporal order of events (translational repression versus deadenylation) remains to be established, these findings suggest that in D. melanogaster cells, GW182 is required to initiate but not to maintain silencing.

The observation that D. melanogaster GW182 dissociates from silenced mRNAs is surprising given that previous studies of Caenorhabditis elegans and human cells showed that GW182 orthologs coimmunoprecipitate miRNA targets (21, 32). The related C. elegans proteins AIN-1 and AIN-2 lack the C-terminal domain and therefore may associate stably with miRNA targets. Alternatively, species-specific variations in the affinity of GW182 for Argonaute proteins may account for these differences.

Role for PABPC1 in silencing.

Previously, we showed that a GW182 mutant that lacks the middle and C-terminal regions (i.e., the silencing domain) is inactive in silencing (14). Furthermore, the GW182 silencing domain alone can trigger both translational repression and the degradation of bound mRNAs (i.e., in tethering assays), indicating that this domain actively recruits decay factors and interferes with translation (5, 14, 22, 33). In this study, we show that the silencing domain interacts with PABPC1.

How does the interaction of GW182 with PABPC1 trigger silencing? A hint is provided by the observation that GW182 competes with eIF4G for binding to PABPC1. PABPC1 is known to bind to the mRNA poly(A) tail and interact with eIF4G, which binds to the 5′ cap structure; this interaction causes the mRNA to form a closed loop (7). A closed-loop conformation is thought to stimulate mRNA translation. Our results suggest that GW182 binding to PABPC1 can interfere with the formation of an mRNA closed loop. By interfering with mRNA circularization, GW182 could repress translation. Moreover, when the mRNA is in the open conformation, the 5′ cap and poly(A) tail are more accessible to mRNA decay enzymes. The possibility that miRNAs interfere with the formation of the mRNA closed-loop conformation was proposed previously but in that case only as a consequence of deadenylation (30).

The inhibition of mRNA circularization is likely to contribute to silencing but does not fully account for it because previous studies of D. melanogaster cells showed that miRNAs also silence transcripts that cannot circularize (i.e., mRNAs whose 3′ end is generated by ribozyme cleavage) (12). Additionally, in human and D. melanogaster cells, the poly(A) tail is not absolutely required for silencing (12, 25, 31). Thus, it is possible that GW182 interferes with additional functions of PABPC1 in translation (20). Finally, our results suggest that the GW182 interaction with PABPC1 could directly facilitate the recruitment of decay enzymes and contribute to miRNA target degradation independently of the translation status. Accordingly, we observed that PABPC1 associates with components of the CCR4-NOT1 complex, and a role for PABPC1 in recruiting components of the deadenylase complex was reported previously (28).

Another question that remains open is whether GW182 interacts with free PABPC1 molecules or with those bound in cis to the poly(A) tail of the mRNA target. The GW182 interaction with PABPC1 persists in the presence of RNase A, suggesting that the interaction is not mediated by RNA. Additionally, a GW182 mutant that does not interact with AGO1 still interacts with PABPC1. Finally, as mentioned above, unadenylated mRNAs can be silenced, suggesting that GW182 may also interact with free PABPC1.

In summary, the GW182 silencing domain functions at multiple stages in the silencing process; it interacts with PABPC1 to elicit translational repression and mRNA degradation, and (at least in D. melanogaster) it is involved in the release of GW182 from silenced mRNPs. An important area of future work will be to identify the molecular interactions that this domain in complex with PABPC1 establishes to bring about silencing.

Supplementary Material

[Supplemental material]

Acknowledgments

We are grateful to Elmar Wahle for providing antibodies to D. melanogaster NOT1.

This study was supported by the Max Planck Society, by grants from the Deutsche Forschungsgemeinschaft (FOR855 and the Gottfried Wilhelm Leibniz Program awarded to E.I.), and by the Sixth Framework Programme of the European Commission through SIROCCO Integrated Project LSHG-CT-2006-037900.

Footnotes

[down-pointing small open triangle]Published ahead of print on 21 September 2009.

Supplemental material for this article may be found at http://mcb.asm.org/.

The authors have paid a fee to allow immediate free access to this article.

REFERENCES

1. Baek, D., J. Villén, C. Shin, F. D. Camargo, S. P. Gygi, and D. P. Bartel. 2008. The impact of microRNAs on protein output. Nature 455:64-71. [PMC free article] [PubMed]
2. Bagga, S., J. Bracht, S. Hunter, K. Massirer, J. Holtz, R. Eachus, and A. E. Pasquinelli. 2005. Regulation by let-7 and lin-4 miRNAs results in target mRNA degradation. Cell 122:553-563. [PubMed]
3. Behm-Ansmant, I., J. Rehwinkel, T. Doerks, A. Stark, P. Bork, and E. Izaurralde. 2006. mRNA degradation by miRNAs and GW182 requires both CCR4:NOT deadenylase and DCP1:DCP2 decapping complexes. Genes Dev. 20:1885-1898. [PMC free article] [PubMed]
4. Carthew, R. W., and E. J. Sontheimer. 2009. Origins and mechanisms of miRNAs and siRNAs. Cell 136:642-655. [PMC free article] [PubMed]
5. Chekulaeva, M., W. Filipowicz, and R. Parker. 2009. Multiple independent domains of dGW182 function in miRNA-mediated repression in Drosophila. RNA 15:794-803. [PMC free article] [PubMed]
6. Chu, C. Y., and T. M. Rana. 2006. Translation repression in human cells by microRNA-induced gene silencing requires RCK/p54. PLoS Biol. 4:e210. [PMC free article] [PubMed]
7. Derry, M. C., A. Yanagiya, Y. Martineau, and N. Sonenberg. 2006. Regulation of poly(A)-binding protein through PABP-interacting proteins. Cold Spring Harb. Symp. Quant. Biol. 71:537-543. [PubMed]
8. Ding, L., and M. Han. 2007. GW182 family proteins are crucial for microRNA-mediated gene silencing. Trends Cell Biol. 17:411-416. [PubMed]
9. Eulalio, A., J. Rehwinkel, M. Stricker, E. Huntzinger, S. F. Yang, T. Doerks, S. Dorner, P. Bork, M. Boutros, and E. Izaurralde. 2007. Target-specific requirements for enhancers of decapping in miRNA-mediated gene silencing. Genes Dev. 21:2558-2570. [PMC free article] [PubMed]
10. Eulalio, A., E. Huntzinger, and E. Izaurralde. 2008. Getting to the root of miRNA-mediated gene silencing. Cell 132:9-14. [PubMed]
11. Eulalio, A., E. Huntzinger, and E. Izaurralde. 2008. GW182 interaction with Argonaute is essential for miRNA-mediated translational repression and mRNA decay. Nat. Struct. Mol. Biol. 15:346-353. [PubMed]
12. Eulalio, A., E. Huntzinger, T. Nishihara, J. Rehwinkel, M. Fauser, and E. Izaurralde. 2009. Deadenylation is a widespread effect of miRNA regulation. RNA 15:21-32. [PMC free article] [PubMed]
13. Eulalio, A., F. Tritschler, and E. Izaurralde. 2009. The GW182 protein family in animal cells: new insights into domains required for miRNA mediated gene silencing. RNA 15:1433-1442. [PMC free article] [PubMed]
14. Eulalio, A., S. Helms, C. Fritzsch, M. Fauser, and E. Izaurralde. 2009. A C-terminal silencing domain in GW182 is essential for miRNA function. RNA 15:1067-1077. [PMC free article] [PubMed]
15. Eulalio, A., F. Tritschler, R. Büttner, O. Weichenrieder, E. Izaurralde, and V. Truffault. 2009. The RRM domain in GW182 proteins contributes to miRNA-mediated gene silencing. Nucleic Acids Res. 37:2974-2983. [PMC free article] [PubMed]
16. Filipowicz, W., S. N. Bhattacharyya, and N. Sonenberg. 2008. Mechanisms of post-transcriptional regulation by microRNAs: are the answers in sight? Nat. Rev. Genet. 9:102-114. [PubMed]
17. Giraldez, A. J., Y. Mishima, J. Rihel, R. J. Grocock, S. Van Dongen, K. Inoue, A. J. Enright, and A. F. Schier. 2006. Zebrafish MiR-430 promotes deadenylation and clearance of maternal mRNAs. Science 312:75-79. [PubMed]
18. Imataka, H., A. Gradi, and N. Sonenberg. 1998. A newly identified N-terminal amino acid sequence of human eIF4G binds poly(A)-binding protein and functions in poly(A)-dependent translation. EMBO J. 17:7480-7489. [PMC free article] [PubMed]
19. Iwasaki, S., T. Kawamata, and Y. Tomari. 2009. Drosophila Argonaute1 and Argonaute2 employ distinct mechanisms for translational repression. Mol. Cell 34:58-67. [PubMed]
20. Kahvejian, A., Y. V. Svitkin, R. Sukarieh, M. N. M'Boutchou, and N. Sonenberg. 2005. Mammalian poly(A)-binding protein is a eukaryotic translation initiation factor, which acts via multiple mechanisms. Genes Dev. 19:104-1113. [PMC free article] [PubMed]
21. Landthaler, M., D. Gaidatzis, A. Rothballer, P. Y. Chen, S. J. Soll, L. Dinic, T. Ojo, M. Hafner, M. Zavolan, and T. Tuschl. 2008. Molecular characterization of human Argonaute-containing ribonucleoprotein complexes and their bound target mRNAs. RNA 14:2580-2596. [PMC free article] [PubMed]
22. Lazzaretti, D., I. Tournier, and E. Izaurralde. 2009. The C-terminal domains of human TNRC6A, B and C silence bound transcripts independently of the Argonaute proteins. RNA 15:1059-1066. [PMC free article] [PubMed]
23. Mathonnet, G., M. R. Fabian, Y. V. Svitkin, A. Parsyan, L. Huck, T. Murata, S. Biffo, W. C. Merrick, E. Darzynkiewicz, R. S. Pillai, et al. 2007. MicroRNA inhibition of translation initiation in vitro by targeting the cap-binding complex eIF4F. Science 317:1764-1767. [PubMed]
24. Mishima, Y., A. J. Giraldez, Y. Takeda, T. Fujiwara, H. Sakamoto, A. F. Schier, and K. Inoue. 2006. Differential regulation of germline mRNAs in soma and germ cells by zebrafish miR-430. Curr. Biol. 16:2135-2142. [PMC free article] [PubMed]
25. Pillai, R. S., S. N. Bhattacharyya, C. G. Artus, T. Zoller, N. Cougot, E. Basyuk, E. Bertrand, and W. Filipowicz. 2005. Inhibition of translational initiation by Let-7 microRNA in human cells. Science 309:1573-1576. [PubMed]
26. Schmittgen, T. D., and K. J. Livak. 2008. Analyzing real-time PCR data by the comparative C(T) method. Nat. Protoc. 3:1101-1108. [PubMed]
27. Selbach, M., B. Schwanhäusser, N. Thierfelder, Z. Fang, R. Khanin, and N. Rajewsky. 2008. Widespread changes in protein synthesis induced by microRNAs. Nature 455:58-63. [PubMed]
28. Simón, E., and B. Séraphin. 2007. A specific role for the C-terminal region of the poly(A)-binding protein in mRNA decay. Nucleic Acids Res. 35:6017-6028. [PMC free article] [PubMed]
29. Temme, C., S. Zaessinger, S. Meyer, M. Simonelig, and E. Wahle. 2004. A complex containing the CCR4 and CAF1 proteins is involved in mRNA deadenylation in Drosophila. EMBO J. 23:2862-2871. [PMC free article] [PubMed]
30. Wakiyama, M., K. Takimoto, O. Ohara, and S. Yokoyama. 2007. Let-7 microRNA-mediated mRNA deadenylation and translational repression in a mammalian cell-free system. Genes Dev. 21:1857-1862. [PMC free article] [PubMed]
31. Wu, L., J. Fan, and J. G. Belasco. 2006. MicroRNAs direct rapid deadenylation of mRNA. Proc. Natl. Acad. Sci. USA 103:4034-4039. [PMC free article] [PubMed]
32. Zhang, L., L. Ding, T. H. Cheung, M.-Q. Dong, J. Chen, A. K. Sewell, X. Liu, J. R. Yates, and M. Han. 2007. Systematic identification of C. elegans miRISC proteins, miRNAs, and mRNA targets by their interactions with GW182 proteins AIN-1 and AIN-2. Mol. Cell 28:598-613. [PMC free article] [PubMed]
33. Zipprich, J. T., S. Bhattacharyya, H. Mathys, and W. Filipowicz. 2009. Importance of the C-terminal domain of the human GW182 protein TNRC6C for translational repression. RNA 15:781-793. [PMC free article] [PubMed]

Articles from Molecular and Cellular Biology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...