PUF Protein-mediated Deadenylation Is Catalyzed by Ccr4p*

PUF proteins control gene expression by binding to the 3′-untranslated regions of specific mRNAs and triggering mRNA decay or translational repression. Here we focus on the mechanism of PUF-mediated regulation. The yeast PUF protein, Mpt5p, regulates HO mRNA and stimulates removal of its poly(A) tail (i.e. deadenylation). Mpt5p repression in vivo is dependent on POP2, a component of the cytoplasmic Ccr4p-Pop2p-Not complex that deadenylates mRNAs. In this study, we elucidate the individual roles of the Ccr4p and Pop2p deadenylases in Mpt5p-regulated deadenylation. Both in vivo and in vitro, Pop2p and Ccr4p proteins are required for Mpt5p-regulated deadenylation of HO. However, the requirements for the two proteins differ dramatically: the enzymatic activity of Ccr4p is essential, whereas that of Pop2p is dispensable. We conclude that Pop2p is a bridge through which the PUF protein recruits the Ccr4p enzyme to the target mRNA, thereby stimulating deadenylation. Our data suggest that PUF proteins may enhance mRNA degradation and repress expression by both deadenylation-dependent and -independent mechanisms, using the same Pop2p bridge to recruit a multifunctional Pop2p complex to the mRNA.

Regulation of mRNA stability, translation, and localization ensure that a given mRNA produces the right amount of protein at the proper time and place. These events are often controlled by elements in the 3Ј-untranslated region (3Ј-UTR) 2 of the mRNA (1,2). mRNA stability and translational regulation are linked to cytoplasmic changes in poly(A) tail lengths (3,4). In particular, poly(A) shortening (deadenylation) is correlated with translational repression and mRNA decay (1,2). Specific regulatory proteins and micro-RNAs bind to 3Ј-UTR elements to promote poly(A) shortening and either repress translation or destroy the mRNA, or both (2,(5)(6)(7).
The Saccharomyces cerevisiae protein Mpt5p is a member of one such family of regulatory proteins, the so-called PUF proteins (8). These proteins promote deadenylation, decay, and translational repression. Mpt5p binds to the 3Ј-UTR of multi-ple target mRNAs (9 -11) and stimulates their deadenylation and decay. In particular, Mpt5p binds a regulatory element in the 3Ј-UTR of HO mRNA and causes rapid deadenylation and decay of that mRNA (12,13). The HO endonuclease is tightly controlled at multiple levels to prevent inappropriate matingtype switching and aberrant double-stranded DNA breaks (14). The Mpt5p repressor contributes to that regulation; in its absence, aberrant switching occurs at high frequency (12).
Recently, using a genetic assay, we showed that repression by Mpt5p requires the POP2 gene and that PUF proteins, including Mpt5p, bind directly to Pop2p (13). Pop2p is a subunit of the major cytoplasmic deadenylase complex, the Ccr4p-Pop2p-Not complex (15,16), thus providing a direct link between PUF proteins and the deadenylation machinery. Two of the Ccr4p-Pop2p-Not complex subunits, Pop2p and Ccr4p, bear sequence similarity to nucleases, and both proteins have been reported to possess deadenylase activity in vitro (17)(18)(19)(20). However, Ccr4p is thought to be the predominant deadenylase in yeast, at least under standard growth conditions (17,18,21). The finding that Pop2p was critical for PUF-mediated regulation of HO mRNA suggested that it might act as a deadenylase on that mRNA (13); it is controversial whether it contributes general deadenylase activity in vivo (15,17,19,20,22,23). The Pop2p deadenylase may be regulated (20,21,23), and Mpt5p might stimulate its enzymatic activity as well as target it to specific mRNAs.
The individual roles of the Pop2p and Ccr4p deadenylases in regulated mRNA decay are not understood. We sought to determine which deadenylase was responsible for PUF-stimulated deadenylation and to delineate the roles of Pop2p and Ccr4p. Our data suggest that Pop2p acts as a bridge through which the PUF protein recruits the Ccr4p enzyme to the mRNA. Additional data suggest that PUF proteins may repress expression by deadenylation-dependent and -independent mechanisms, using the same Pop2p bridge.

EXPERIMENTAL PROCEDURES
Strains and Plasmids-The wild-type BY4742 yeast strain and isogenic strains with gene-specific deletions of POP2, CCR4, PAN2, CAF120, CAF130, NOT3, and NOT4 were obtained from Open Biosystems. These deletion strains were created by PCR-mediated gene modification using the kanamycin/G418 resistance marker. The MPT5-TAP strain (Open Biosystems) was created in the S288C strain background by integrating a C-terminal TAP tag onto the coding sequence of MPT5 using PCR-mediated gene modification with a HIS3 marker.
POP2 and CCR4 expression plasmids were created in the high copy vector pACG1-NT and contained N-terminal His 6 and T7 epitope tags that could be cleaved off using TEV prote-* This work was supported by National Institutes of Health Grant GM50942.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ase. The ADH1 promoter and 3Ј-UTR were used for expression, and the plasmids carried the Zeocin resistance marker. The active site mutant POP2 contains two missense mutations, S188A/E190A, described by Thore et al. (20). The active site mutant CCR4 contains missense mutation E556A described by Chen et al. (18). Both mutants were created by QuikChange (Stratagene) site-directed mutagenesis.
High Resolution Northern Blotting-RNA was extracted from samples using the hot acidic phenol method. Cell pellets were resuspended in 500 ul of TENS (10 mM Tris-HCl, pH 7.5, 1 mM EDTA, 0.3 M NaCl, 0.2% SDS) and extracted with 500 l of 1:1 phenol pH 4.6:chloroform by incubating at 70°C for 5 min followed by vigorous vortexing for 2 min. RNA was then extracted a second time with chloroform, precipitated, and resuspended in sterile water.
The decay of HO mRNA poly(A) tails was analyzed over the indicated time course following addition of the transcriptional inhibitor thiolutin (gift from Pfizer) at 20 g/ml as previously described (24). After RNA isolation, 10 g of each total RNA sample was cleaved with ribonuclease H (Promega) and an antisense HO oligonucleotide (5Ј-GGACAGCATCAAACTG-TAAGATTCCGCCAC-3Ј) (Integrated DNA Technologies). When indicated, oligo(dT) 15 was added to the ribonuclease H reactions. The RNA was then analyzed by 6% polyacrylamideurea-denaturing PAGE and transferred to nylon membranes using 1ϫ TBE (Tris borate-EDTA) in a Bio-Rad Trans-Blot apparatus. Blots were probed with an antisense HO 3Ј-UTR probe or an antisense SCR1 probe. The blots were imaged using a Typhoon PhosphorImager (GE Healthcare).
Co-immunoprecipitation Analysis-Co-immunoprecipitations were performed as described by Goldstrohm et al. (13), using the MPT5-TAP strain with plasmids expressing the indicated T7 epitope-tagged POP2 or CCR4 proteins.
PUF Repression Growth Assays-Mpt5p-mediated repression assays were performed using the reporter gene construct YCp33 HO promoter-HIS3-HO 3Ј-UTR and YEp181 MPT5, described by Goldstrohm et al. (13), in either wild-type BY4742 or pop2 strains. Wild-type or catalytically inactive mutant POP2 were expressed from vector pACG1-NT. Colonies from each test strain were isolated and grown to mid-log phase at 30°C, and the indicated number of cells was plated on minimal medium with or without histidine in the presence of 300 g/ml of Zeocin (Invivogen). The His3p competitive inhibitor 3-aminotriazole was added to medium lacking histidine at a concentration of 1 mM to increase stringency of the growth assay.
Purification of Deadenylase Complexes-Wild-type or catalytically inactive mutant Pop2p or Ccr4p complexes were purified from pop2 or ccr4 deletion strains as indicated in Fig. 4 and supplemental Fig. S1. Conditions identical to the TAP-POP2 purification (13) were used except that the target proteins had a T7 affinity tag at the N terminus with a TEV protease site to allow elution of the isolated complexes. All steps were done on ice or at 4°C. Cell extracts were prepared from 1-liter cultures grown to OD 660 nm of 1.5 in YPAD with 300 g/ml Zeocin (Invivogen). Cells were washed and lysed by glass bead bashing for 15 min in 1 packed cell volume of TNEMN150 (50 mM Tris-HCl, pH 8.0, 0.5% Nonidet P40, 1 mM EDTA, 2 mM MgCl 2 , and 150 mM NaCl) containing Complete protease inhibitors mixture (Roche Applied Science). The lysate was then centrifuged 10 min at 16,000 ϫ g, and the supernatant was removed and incubated with T7 monoclonal antibodies covalently linked to agarose beads (Novagen) for 4 h. Beads were washed four times with 150 bed volumes of TNEMN150 and then one time with 150 bed volumes of deadenylation buffer (50 mM Tris-HCl, pH 8.0, 20 mM NaCl, 0.1 mM MgCl 2 , 10% glycerol) (25). Bound proteins were eluted from beads in 1 bed volume of deadenylation buffer using 8 units of AcTEV protease (Invitrogen) for 12 h at 4°C. Purified complexes were evaluated by silver staining and by deadenylation assays.
In Vitro Deadenylation Assays-Deadenylation reactions were carried out in a 20-l volume in deadenylation buffer (50 mM Tris-HCl, pH 8, 20 mM NaCl, 0.1 mM MgCl 2 , 10% (v/v) glycerol) and 10 M nonspecific competitor RNA (synthetic oligoribonucleotide with sequence 5Ј-UCUAAUCGGGG-UACAAUUAUAUAAUAUA-3Ј). HO substrate RNA (Integrated DNA Technologies) with sequence 5Ј-AGUUUAAA-AAGUUGUAUGUAAUAAAAGUA 14 -3Ј was radioactively labeled with T4 polynucleotide kinase (Promega) at the 5Ј-end and added to the reactions at a final concentration of 10 nM. Recombinant purified GST-Mpt5p RBD (250 nM) was added to reactions where indicated. Purified Pop2p or Ccr4p complexes (10 ng each) were added to their respective reactions. In initial experiments, we carefully titrated these complexes to measure their deadenylase activity (supplemental Fig. S1), and then balanced the amount and activity used in the final experiments shown in Fig. 4.

RESULTS
CCR4 and POP2 Are Required for Deadenylation of HO mRNA in Vivo-Mpt5p stimulates deadenylation of HO mRNA in vivo and in vitro (13). We sought to determine the contributions of Pop2p and Ccr4p to PUF-stimulated deadenylation. We first analyzed the length of poly(A) on endogenous HO mRNA in wild-type cells and in strains with deletions of genes encoding subunits of Ccr4p-Pop2p-Not complex. We focused initially on a comparison of wild-type, pop2, and ccr4 deletion strains. We used thiolutin to inhibit transcription and then collected samples over a time course to observe the degradation of the poly(A) tails. Total RNA was extracted from each sample and cleaved with ribonuclease H and a specific antisense oligonucleotide to produce a short HO 3Ј-end fragment. To measure the poly(A) length of this 3Ј-fragment, we performed high resolution polyacrylamide electrophoresis and Northern blotting. The noncoding SCR1 RNA served as a loading control.
In pop2 cells, the distribution of poly(A) ranged form 20 to 60 nucleotides; poly(A) lengths below 20 nucleotides were not detected (Fig. 1A, lanes 8 -14). Over the time course, the longer tails were shortened to 20 -40 adenosines (lane 14) and the mRNA disappeared slowly, without further deadenylation (see below). In addition to this effect on steady-state mRNA distri-bution, the apparent rate of removal of the poly(A) tail was slower in pop2 cells (compare lanes 2-4 with lanes 9 -11).
We conclude that both Pop2p and Ccr4p are required for deadenylation of HO mRNA in vivo but that their roles differ. Ccr4p is required for initial deadenylation. Ccr4p is unable to remove HO poly(A) tails in the range of 5 to 20 nucleotides in the absence of Pop2p. Pop2p facilitates deadenylation from 20 to 60 nucleotides and is most critical for removal of shorter poly(A) 5-20 tails.
Deletion of Other Deadenylase Complex Components Does Not Affect HO Poly(A) Tails-To determine whether other mRNA decay factors were required for deadenylation of HO mRNA, we analyzed steady-state mRNA from a series of mutant strains, each lacking various turnover-related components. The same high resolution Northern blotting method as above was used to determine HO poly(A) tail lengths accurately (Fig. 1B). Deletion of CCR4 again caused a complete block in deadenylation (Fig. 1B, lane 2 versus 3). Mutants lacking Pan2p, the catalytic subunit of the Pan2p-Pan3p deadenylase complex, showed little effect on HO poly(A) tail lengths; a slight increase in the length of the longest poly(A) tails was observed (Fig. 1B,  lane 4) and is consistent with a role for the Pan2p-Pan3p complex in initial poly(A) tail trimming (26). Deletions of components of the Ccr4p-Pop2p-Not complex, including Caf120p, Caf130p, Not3p, and Not4p, had no apparent effect on HO poly(A) tail length (lanes [5][6][7][8]. Defects in other Ccr4p-Pop2p-Not complex components were not tested because those mutants were inviable. We conclude that the Pan2p deadenylase has very little effect on removal of HO poly(A) tails and that at least four proteins in the described Ccr4p-Pop2p-Not complex are not necessary for HO mRNA deadenylation. Deletions of these same genes also did not disrupt repression by Mpt5p in vivo (13).
POP2 Protein, but Not Its Enzymatic Activity, Is Required for HO Deadenylation in Vivo-The dependence of deadenylation and repression on Pop2p suggested that its deadenylase activity might be essential for PUF action. To test this idea, we determined whether Pop2p enzymatic activity was required for PUFstimulated deadenylation in vivo. We analyzed HO poly(A) tails in pop2 cells carrying a wild-type copy (POP2) or mutant form of POP2 (POP2 mt) on an episome ( Fig. 2A). The mutation in POP2 comprised two missense substitutions (S188A/E190A) that inactivate its deadenylase activity in vitro (20). HO mRNA in these strains was compared with the wild-type strain (Fig. 2A, . Both wild-type and mutant Pop2p were expressed at the same level (data not shown). To determine whether the mutant form of Pop2p still associated with Mpt5p, we determined whether Pop2p co-immunoprecipitated with Mpt5p (Fig. 2B). Both mutant and wild-type forms of Pop2p interact with Mpt5p in yeast extracts (Fig. 2B). We conclude that the POP2 protein is important for normal deadenylation of HO mRNA in vivo but its enzymatic activity is not.
Ccr4p Deadenylase Activity Is Crucial for HO Deadenylation in Vivo-To determine whether Ccr4p nuclease activity is required for deadenylation of HO mRNA, we introduced a wildtype or mutant CCR4 gene on an episome into ccr4 mutant cells and again assayed the length of poly(A) on HO mRNA (Fig. 2C). CCR4 deletion mutants exhibited a severe defect in deadenylation that was not rescued by empty vector (Fig. 2C, lane 2 versus  3). The wild-type distribution of poly(A) lengths was restored by episomal CCR4 (lane 4). A missense mutation in CCR4 gene (CCR4 mt), bearing a single amino acid substitution (E556A) that inactivates Ccr4p nuclease activity (18), completely abrogated the rescue of activity (lane 5). Both wild-type and mutant Ccr4p were expressed at nearly the same levels (data not shown) and co-immunoprecipitated with Mpt5p (Fig. 2D). We conclude that Ccr4p is the main deadenylase responsible for HO mRNA deadenylation in vivo.  figure, were determined based on migration of bands compared with an RNA molecular weight marker (Century Plus, Ambion). Where indicated by a "ϩ" at the top, oligo(dT) was added to the RNA and digested using ribonuclease H to remove the poly(A) tail, to provide a marker for deadenylated mRNA. Endogenous SCR1 RNA, detected by hybridization to the same blot, served as a loading control. B, HO mRNA poly(A) tail lengths in yeast strains with gene-specific deletions of mRNA decay factors. Northern blots of HO mRNA from wild-type (WT) or deletion strains indicated at the top. As described above, a 3Ј-UTR fragment of HO mRNA was generated by ribonuclease H cleavage. HO poly(A) tail lengths, on the left of the panel, were determined based on RNA molecular weight markers. Oligo(dT) treatment is indicated on the top by the "ϩ". SCR1 served as a loading control.

Mpt5p Repression Is Not Affected by Pop2p Active Site
Mutations-To determine whether the enzymatic activity of Pop2p was required for PUF repression, we used a HIS3 reporter gene containing the HO 3Ј-UTR (13). Mpt5p binds to the HO 3Ј-UTR and specifically represses the HIS3-HO mRNA, which can be assayed on medium lacking histidine and containing 3-aminotriazole, a competitive inhibitor of the HIS3-encoded protein (Fig. 3). When Mpt5p is overexpressed in wildtype cells, the reporter mRNA is repressed and the cells no longer grow on medium lacking histidine (Fig. 3, WT strain). In pop2 mutant cells, Mpt5p-mediated repression was restored by either the wild-type Pop2p or the catalytically inactive Pop2p mutant (Fig. 3, pop2 strain). Thus, we conclude that whereas POP2 protein is necessary, its deadenylase activity is not. Because deletion of CCR4 has only a modest effect on Mpt5p repression (13), these findings suggest that deadenylation may not be the only mechanism involved in Mpt5p-mediated repression of the reporter mRNA (see "Discussion").
Pop2p Deadenylase Activity Is Not Required for PUF-stimulated Deadenylation in Vitro-To further examine the requirements for POP2 and CCR4 proteins, we exploited the in vitro system we recently developed that recapitulates PUF-mediated deadenylation (13). In this assay, deadenylation of synthetic RNAs containing a PUF binding site is catalyzed by Pop2p complexes purified from yeast, mixed with purified recombinant Mpt5p. We used this assay to determine whether the enzymatic activity of Pop2p was required for PUF-stimulated deadenylation. Either wild-type or mutant POP2 genes, with T7 epitope tags, were introduced into pop2 cells. The mutant Pop2p carried the two missense substitutions (S188A/E190A) that disrupt its enzyme activity (20). The POP2 proteins, with associated factors, were purified from yeast and assayed for deadenylation using HO 3Ј-UTR RNA substrates with 14 adenosines at their 3Ј-end. We titrated the amount and activity of these Pop2p complexes so that deadenylase activity was minimal under the conditions used (Fig. 4A, lanes 3-5 and 9 -11; supplemental Fig. 1A). Recombinant Mpt5p stimulated deadenylation by both the wild-type Pop2p and mutant Pop2p complexes and yielded a fully deadenylated product (Fig. 4A, lanes  6 -8 and 12-14). Therefore, Pop2p deadenylase activity is not essential for PUF-stimulated deadenylation in vitro.
We also purified Pop2p complexes from a ccr4 strain to determine whether Pop2p had any activity in the absence of FIGURE 2. Deadenylation of HO mRNA in vivo is dependent on the deadenylase activity of Ccr4p, but not that of Pop2p. A, Pop2 protein, but not its enzymatic activity, is required for HO deadenylation in vivo. Poly(A) tail distribution of HO mRNA was measured by denaturing polyacrylamide gels and Northern blotting of RNA from wild-type (WT) and pop2 deletion cells as indicated at the top. Empty plasmid vector, wild-type POP2, or catalytically inactive POP2 mutant (POP2 mt) were expressed in pop2 yeast cells. Poly(A) tail lengths are indicated on the left. Northern blot of SCR1 RNA served as a loading control. B, wild-type (POP2-T7) and catalytically inactive POP2 mutant (POP2-T7 mt) proteins co-immunoprecipitate with Mpt5p. Mpt5p was Tandem Affinity Tagged (TAP), immunoprecipitated using rabbit IgG agarose beads (Sigma), eluted with TEV protease, and Western blotted with anti-T7 monoclonal antibodies to detect Pop2 proteins. Mpt5p-TAP was detected in a Western blot using Peroxidase/anti-Peroxidase (Sigma). T7-tagged green fluorescent protein served as a negative control. C, Ccr4p deadenylase activity is critical for HO deadenylation in vivo. HO poly(A) tails were measured in wild-type (WT) and ccr4 deletion strains on denaturing polyacrylamide gels and Northern blotting. Poly(A) tail length is indicated on the left. Empty plasmid vector or plasmids expressing wild-type CCR4 or a catalytically inactive mutant form of CCR4 (CCR4 mt) were introduced in the ccr4 cells. D, T7-tagged wild-type CCR4 (CCR4-T7) and catalytically inactive CCR4 mutant (CCR4-T7 mt) proteins co-immunoprecipitate with Mpt5p. Mpt5p was Tandem Affinity Tagged (TAP), immunoprecipitated using rabbit IgG-agarose beads, eluted with TEV protease, and Western blotted with anti-T7 monoclonal antibodies to detect Ccr4 proteins. Mpt5p-TAP was detected in a Western blot using Peroxidase/anti-Peroxidase (Sigma). T7-tagged green fluorescent protein served as a negative control. Ccr4p. This Pop2p complex had no detectable deadenylase activity, even at high concentrations (Fig. 4A, lanes 15-17; supplemental Fig. S1A). Moreover, this complex, lacking Ccr4p, exhibited no deadenylation activity when mixed with Mpt5p (Fig. 4A, lanes 18 -20). Thus, Mpt5p does not activate Pop2p deadenylase activity and Ccr4p is critical for the deadenylase activity of the purified Pop2p complex.

DISCUSSION
In this report, we set out to discover how PUF proteins enhance deadenylation of target mRNAs. Mpt5p associates with the Ccr4p-Pop2p-Not complex through a direct proteinprotein interaction with the Pop2p subunit (13). By so doing, the deadenylase complex acts preferentially on those mRNAs to which Mpt5p is bound. We have shown here that Pop2p and Ccr4p are required for PUF-mediated deadenylation, both in vivo and in vitro. However, the roles of the two proteins differ dramatically.
The deadenylase activity of Pop2p is dispensable for PUF repression in vivo, for deadenylation of HO mRNA in vivo, and for PUF-stimulated deadenylation in vitro. Furthermore, purified, recombinant, yeast Pop2p, which was competent for binding to Mpt5p, possessed no deadenylase activity in vitro with or without Mpt5p (not shown). From these findings, we conclude that Pop2p does not catalyze PUF-stimulated deadenylation.
Although the enzyme activity of Pop2p is not required, the protein itself is essential for PUF-mediated deadenylation, both in vivo and in vitro. What role does Pop2p play? Ccr4p isolated from pop2 cells possesses nonspecific deadenylation activity, consistent with previous results (17,18). However, without Pop2p, Ccr4p is not stimulated by the PUF protein. We infer that Pop2p recruits the Ccr4p enzyme to the RNA. This may be achieved through a direct Pop2p-Ccr4p contact or could be indirect, through other subunits of the Ccr4p-Pop2p-Not complex (16). It is unlikely that contacts between the PUF protein and other subunits of the Ccr4p-Pop2-Not complex are critical, because deletion of Pop2p alone is sufficient to disrupt PUFmediated deadenylation (13). We do not detect direct contacts between Mpt5p and several other subunits of the Ccr4p-Pop2p-Not complex. 3 Furthermore, deletions of other subunits of the complex do not significantly influence HO deadenylation or repression (this study and Ref. 13). The most parsimonious interpretation is that Pop2p is physically required to mediate recruitment of Ccr4p to HO mRNA by Mpt5p.
Ccr4p is the enzyme responsible for PUF-stimulated deadenylation; deletion of CCR4 blocked deadenylation in vivo and in vitro, as did mutation of the Ccr4p active site. Further support for the role of Ccr4p in PUF-stimulated deadenylation comes from our previous analysis of HO poly(A) tails in a strain deleted of both PUF proteins that regulate HO (see Fig. 2b in Goldstrohm et al.;Ref. 13). When MPT5 is deleted (along with the redundant PUF4), the poly(A) tails on HO mRNA are removed very slowly, an effect that mirrors deletion of CCR4 (see Fig. 1A). The Ccr4p-Pop2p-Not complex remains func-  A, in vitro deadenylation assays using T7-tagged, purified Pop2p complexes. Wild-type (Pop2p-T7) or catalytically inactive mutant Pop2p (Pop2p-T7 mt) complexes was isolated from POP2 (pop2) or CCR4 (ccr4) deletion strains. Following the indicated incubation times, the reaction products were separated by denaturing polyacrylamide electrophoresis. The HO substrate RNA, which contained 14 adenosines at its 3Ј-end, was radioactively labeled at its 5Ј-end and incubated with 10 ng of each Pop2p complex for the indicated amount of time, in minutes, without (Ϫ) or with (ϩ) recombinant, purified Mpt5p fused to glutathione S-transferase (GST-Mpt5p). Marker RNAs with no poly(A) tail (A 0 ) or with a 14-adenosine poly(A) tail (A 14 ) were included in lanes 1 and 2, respectively. B, in vitro deadenylation assays with T7-tagged, purified wild-type (Ccr4p-T7) or catalytically inactive mutant (Ccr4p-T7 mt). These Ccr4p complexes were isolated from a CCR4 deletion strain (ccr4) or a POP2 deletion strain (pop2). The 5Ј-end labeled HO substrate RNA with 14 adenosines at the 3Ј-end and was incubated with 10 ng of each Ccr4p complex for the indicated amount of time, in minutes, without (Ϫ) or with (ϩ) recombinant, purified GST-Mpt5p. The reaction products were analyzed by denaturing polyacrylamide electrophoresis. Marker RNAs with no poly(A) tail (A 0 ) or with a 14-adenosine poly(A) tail (A 14 ) were included in lanes 1 and 2, respectively. tional in this double PUF deletion strain, but the deadenylase is no longer efficiently recruited to the mRNA. Therefore, poly(A) removal proceeds at a slow basal rate.
Our findings suggest that deadenylation is not the only way in which Mpt5p can reduce expression of mRNAs. The importance of poly(A) tails for efficient translation and mRNA stability is well documented (1,3,4). PUF repression correlates with enhanced deadenylation and degradation of target mRNAs in a number of biological systems (i.e. Xenopus, Drosophila, Caenorhabditis elegans, and yeast; reviewed in Ref. 8). Ccr4p is stringently required for deadenylation, but ccr4 mutants exhibit only a modest effect on repression by Mpt5p in reporter gene repression assays (13). In contrast, pop2 mutants are impaired both in deadenylation and repression.
One possible explanation for these findings is that Mpt5p may repress translation independent of deadenylation. The Drosophila PUF protein, Pumilio, which normally stimulates deadenylation of target mRNAs, can repress translation of nonadenylated mRNAs, albeit with reduced efficiency (27). We suggest that PUF proteins recruit multiple activities to repress target mRNAs. Because Pop2p is required for PUF repression, we suggest that it serves as a bridge to recruit the multifunctional Ccr4p-Pop2p-Not complex. The complex not only removes poly(A) tails via its Ccr4p subunit but also may elicit translational repression through other mechanisms. Dhh1p is an intriguing candidate to mediate this repression: it co-immunoprecipitates with Mpt5p and has a well established role in inhibiting translation (28 -31). Co-recruitment of decapping factors Dcp1p and Dcp2p by Mpt5p (13), which associate with Pop2p (29), would further seal the fate of the target mRNA.
The versatility of regulatory mechanisms is echoed even within a single regulatory family, the PUF proteins. The interaction between PUF and Pop2 proteins is conserved through evolution (13). Yet not all PUF proteins necessarily repress in precisely the same manner as Mpt5p. Moreover, the human Pop2p ortholog, CNOT8, which can interact with Ccr4p (39), is active as a deadenylase (23) and binds to a human PUF protein (13). Thus, a single human PUF protein may recruit two active deadenylases, CNOT8 and human CCR4, to degrade target mRNAs. It will be of interest to determine whether the requirements for the Ccr4p deadenylase and a Pop2p protein bridge are universal among PUF proteins or are a theme with many variations.