CCR4-NOT deadenylates mRNA associated with RNA-induced silencing complexes in human cells

Abstract

MicroRNAs (miRNAs) repress gene expression posttranscriptionally by inhibiting translation and by expediting deadenylation so as to trigger rapid mRNA decay. Their regulatory influence is mediated by the protein components of the RNA-induced silencing complex (RISC), which deliver miRNAs and siRNAs to their mRNA targets. Here, we present evidence that CCR4-NOT is the deadenylase that removes poly(A) from messages destabilized by miRNAs in human cells. Overproducing a mutationally inactivated form of either of the catalytic subunits of this deadenylase (CCR4 or CAF1/POP2) significantly impedes the deadenylation and decay of mRNA targeted by a partially complementary miRNA. The same deadenylase initiates the degradation of "off-target" mRNAs that are bound by an imperfectly complementary siRNA introduced by transfection. The greater inhibitory effect of inactive CAF1 or POP2 (versus inactive CCR4) suggests a predominant role for this catalytic subunit of CCR4-NOT in miRNA- or small interfering RNA (siRNA)-mediated deadenylation. These effects of mi/siRNAs and CCR4-NOT can be fully reproduced by directly tethering RISC to mRNA without the guidance of a small RNA, indicating that the ability of RISC to accelerate deadenylation is independent of RNA base pairing. Despite its importance for mi/siRNA-mediated deadenylation, CCR4-NOT appears not to associate significantly with RISC, as judged by the failure of CAF1 and POP2 to coimmunoprecipitate detectably with either the Ago or TNRC6 subunit of RISC, a finding at odds with deadenylase recruitment as the mechanism by which RISC accelerates poly(A) removal.

FIG. 1.

Dominant-negative effect of catalytically inactive CCR4-NOT subunits. The deadenylation and decay of BG+L7 mRNA were monitored after transiently inducing its synthesis in HeLa cells transfected with a gene encoding a wild-type or mutationally inactivated (mut) V5-tagged deadenylase subunit (CAF1, CAF1-D40A/E42A; POP2, POP2-D40A/E42A; CCR4a, CCR4a-E240A; PAN2, PAN2-D977A; or PARN, PARN-D28A/D382A) and an AG-GAPDH gene. (A) Decay of BG+L7 mRNA. Equal amounts of cytoplasmic RNA isolated at time intervals after transient induction of transcription were analyzed by Northern blotting to detect BG+L7 mRNA. For comparison, the slow decay of BG mRNA lacking a let-7-responsive L7 element was also examined in the absence of deadenylase subunit overproduction (BG, top left). (B) Graphs of the concentration of BG+L7 mRNA as a function of time. All values were normalized to AG-GAPDH mRNA, a constitutively transcribed internal standard. (C) Deadenylation of BG+L7 mRNA. The cytoplasmic RNA samples analyzed in panel A were subjected to site-specific cleavage by RNase H in the presence of an oligodeoxynucleotide complementary to codons 74 to 81 within the coding region. The 3′ BG+L7 RNA fragments thereby produced were analyzed by electrophoresis and blotting, using markers (M) that corresponded in size to reporter mRNA 3′ fragments bearing no poly(A) or a 160-nucleotide poly(A) tail. (D) Immunoblot analysis of ectopically expressed V5-tagged deadenylase subunits. β-Actin served as an internal standard.

FIG. 2.

Effect of CAF1 or POP2 depletion on miRNA-mediated deadenylation. The deadenylation of BG+L7 mRNA was monitored after transiently inducing its synthesis in HeLa cells transfected with an siRNA complementary to CAF1 or POP2 mRNA or with a noncomplementary control siRNA (siNC). The CAF1- or POP2-specific siRNAs reduced the concentration of CAF1 mRNA by 70% and the concentration of POP2 mRNA by 67%, as determined by real-time RT-PCR.

FIG. 3.

Off-target deadenylation mediated by an siRNA. The deadenylation and decay of BG+G_M mRNA were monitored after transiently inducing its synthesis in HeLa cells transfected with a partially complementary (siEGFP) or noncomplementary (siNC) siRNA. (A) RNA duplex expected for siEGFP base paired with element G_M. (B) Decay of BG+G_M mRNA. Equal amounts of cytoplasmic RNA isolated at time intervals after transient induction of transcription were analyzed by Northern blotting. AG-GAPDH mRNA served as an internal standard. (C) Deadenylation of BG+G_M mRNA. The cytoplasmic RNA samples analyzed in panel B were subjected to site-specific cleavage by RNase H and Northern blot analysis. (D) Effect of CAF1 or POP2 depletion on siEGFP-mediated deadenylation of BG+G_M mRNA.

FIG. 4.

Deadenylation mediated by tethered RISC. The deadenylation and decay of BG+boxB mRNA were monitored after transiently inducing its synthesis in HeLa cells transfected with a gene encoding HA-tagged human Ago1, Ago2, Ago3, Ago4, or catalytically inactive Ago2mut (Ago2-D597A) fused (N-HA-Ago) or not fused (HA-Ago) to the boxB-binding domain of the bacteriophage lambda N protein. (A) Decay of BG+boxB mRNA. Equal amounts of cytoplasmic RNA isolated at time intervals after transient induction of transcription were analyzed by Northern blotting. AG-GAPDH mRNA served as an internal standard. (B) Deadenylation of BG+boxB mRNA. The cytoplasmic RNA samples analyzed in panel A were subjected to site-specific cleavage by RNase H and Northern blot analysis. (C) Dominant negative effect of catalytically inactive CAF1 or POP2 on deadenylation mediated by tethered N-HA-Ago2. (D) Immunoblot analysis of ectopically expressed HA-tagged Ago proteins. β-Actin served as an internal standard.

FIG. 5.

Failure of deadenylase subunits to coimmunoprecipitate with RISC. HEK 293T cells were cotransfected with genes encoding a FLAG- or HA-tagged RISC subunit and either a V5-tagged deadenylase subunit (POP2, CAF1, CCR4a, CCR4b, PAN2, or PARN), another V5-tagged RISC subunit (positive control), or V5-tagged firefly luciferase (negative control). The FLAG- or HA-tagged RISC subunit was immunoprecipitated, and copurifying V5-tagged proteins were detected by immunoblotting. Alternatively, HEK 293T cells were cotransfected with a gene encoding CAF1 with a V5 epitope tag at either the N terminus (V5-CAF1) or C terminus (CAF1-V5) and a second gene encoding an HA-tagged RISC subunit (Ago2, GW182, or TNRC6B), CCR4-NOT subunit (CCR4a or CCR4b), or β-galactosidase (negative control). The HA-tagged protein was immunoprecipitated, and immunoblotting was performed to detect any copurifying V5-tagged CAF1. (A) V5-tagged proteins that copurify with FLAG-Ago2. (B) V5-tagged proteins that copurify with N-HA-GW182 (TNRC6A). (C) HA-tagged proteins with which N-terminally tagged V5-CAF1 copurifies. (D) HA-tagged proteins with which C-terminally tagged CAF1-V5 copurifies. HA- or V5-tagged proteins were also detected in cell extracts to verify their production. The ubiquitous band marked with an asterisk is the cross-reacting monoclonal anti-FLAG or anti-HA antibody used for immunoprecipitation.