Components of the CCR4-NOT Complex Function as Nuclear Hormone Receptor Coactivators via Association with the NRC-interacting Factor NIF-1*

CCR4-NOT is an evolutionarily conserved, multicomponent complex known to be involved in transcription as well as mRNA degradation. Various subunits (e.g. CNOT1 and CNOT7/CAF1) have been reported to be involved in influencing nuclear hormone receptor activities. Here, we show that CCR4/CNOT6 and RCD1/CNOT9, members of the CCR4-NOT complex, potentiate nuclear receptor activity. RCD1 interacts in vivo and in vitro with NIF-1 (NRC-interacting factor), a previously characterized nuclear receptor cotransducer that activates nuclear receptors via its interaction with NRC. As with NIF-1, RCD1 and CCR4 do not directly associate with nuclear receptors; however, they enhance ligand-dependent transcriptional activation by nuclear hormone receptors. CCR4 mediates its effect through the ligand binding domain of nuclear receptors and small interference RNA-mediated silencing of endogenous CCR4 results in a marked decrease in nuclear receptor activation. Furthermore, knockdown of CCR4 results in an attenuated stimulation of RARα target genes (e.g. Sox9 and HoxA1) as shown by quantitative PCR assays. The silencing of endogenous NIF-1 also resulted in a comparable decrease in the RAR-mediated induction of both Sox9 and HoxA1. Furthermore, CCR4 associates in vivo with NIF-1. In addition, the CCR4-enhanced transcriptional activation by nuclear receptors is dependent on NIF-1. The small interference RNA-mediated knockdown of NIF-1 blocks the ligand-dependent potentiating effect of CCR4. Our results suggest that CCR4 plays a role in the regulation of certain endogenous RARα target genes and that RCD1 and CCR4 might mediate their function through their interaction with NIF-1.

with a near stoichiometric level of RNA polymerase II along with other pre-initiation factors such as TBP, SRB7, and TFIIH. The various distinct coactivator complexes cooperate with one another temporally at target gene promoters. Studies have indicated a sequential recruitment of coactivators in vivo, with chromatin modification (Swi/Snf and histone acetyltransferases) preceding the recruitment of TRAP/ DRIP complexes (48 -50).
The CCR4-NOT complex is a multicomponent, evolutionarily conserved global regulator of gene expression. The yeast CCR4-NOT complex exists in two forms of 1.0 and 1.9 MDa, the smaller complex consisting of nine core proteins: CCR4 (carbon catabolite repressor 4), NOT1-5, CAF1 (CCR4-associated factor 1), CAF40, and CAF130 (51,52). CCR4 was first identified in Saccharomyces cerevisiae as being necessary for non-fermentative gene expression (53). CCR4 and CAF1 have also been shown to be components of the major cytoplasmic mRNA deadenylase, with CCR4 as the catalytic component (54,55). The NOT proteins were isolated as repressors of transcription from promoters lacking a canonical TATA sequence (56). The NOT proteins play an important role in transcriptional regulation through their interaction with TFIID and other components of the Mediator complex (57,58).
Several human orthologs of subunits of the CCR4-NOT complex have been identified (59,60). The composition of the mammalian CCR4-NOT complex and the functionality of several components have also been conserved. For instance, CNOT2 (hNOT2) functions as a repressor of reporter gene activity, with a suggestion of the involvement of a histone deacetylase activity (61). A recent study identified CNOT1 as a ligand-dependent repressor of nuclear hormone receptor activity (62). CNOT6 (hCCR4) and the newly identified CNOT6L (another homolog of yCCR4) have both been shown to possess a 3Ј-5Ј deadenylase activity in vitro (63,64). The expression of a catalytically inactive CCR4 resulted in impeded deadenylation and decay of mRNA (65).
We have previously reported the cloning and characterization of NIF-1 (NRC-interacting factor 1) (35). NIF-1 enhances ligand-dependent transcriptional activation by nuclear hormone receptors through its interaction with the coactivator NRC. NIF-1 also exerts its activating function on other transcription factors such as c-Fos and c-Jun (35). To further understand the function of NIF-1 and to identify other interacting partners, we set up a yeast two-hybrid screen using full-length NIF-1 as bait. This screen identified several putative NIF-1 interactors. One such interesting clone was identified as RCD1 (required for cell differentiation), a component of the CCR4-NOT complex.
In this study, we report that NIF-1 associates with RCD1, and RCD1 as well as CCR4 function as nuclear hormone receptor coactivators. CCR4 is an AF-2-dependent coactivator and appears to play a vital and specific role in the regulation of certain RAR␣ target genes. Although these factors do not directly associate with nuclear receptors, our findings are consistent with a model that suggests that they mediate their effects through a NIF-1 coactivator complex.

EXPERIMENTAL PROCEDURES
Yeast Two-hybrid cDNA Library from GH4C1 Cells-Poly(A ϩ ) RNA isolated from GH4C1 cells was used for the synthesis of cDNA with a Stratagene cDNA synthesis system. The cDNA was size-fractionated using Sephacryl S200. Fractions above 0.4 kb were pooled, precipitated with ethanol, and ligated with EcoRI-XhoI-digested pEG202, which constitutively expresses the cDNA as a fusion with the LexA DNA binding domain in S. cerevisiae. The pEG202-cDNA was transformed into Sure bacteria (Stratagene) by electroporation. This cDNA library contains ϳ1.2 ϫ 10 7 independent transformants, and the average insert size was estimated to be 1.5 kb.
Yeast Two-hybrid Screen-Full-length NIF-1 was cloned into the EcoRI site of pJGM vector (35) and was used as bait in a yeast two-hybrid screen. pJGM was derived from the parent vector pJG4 -5, as described earlier (24). All methods and transformation procedures have been described previously (24). The yeast strain EGY48 harboring the LacZ reporter pSH18 -34 and pJGM-NIF-1 was transformed with the GH4C1 pEG202 cDNA library. Transformants were directly screened on 5-bromo-4chloro-3-indolyl-␤-galactopyranoside (X-gal) synthetic defined minimal medium (SD)-galactose-raffinose plates lacking Trp, Ura, His, and Leu. The putative positive clones were then verified by replica plating each clone on Trp Ϫ , Ura Ϫ , His Ϫ , and Leu Ϫ X-gal SD-galactose-raffinose and Trp Ϫ , Ura Ϫ , His Ϫ , and Leu Ϫ X-gal-SD-dextrose plates. Yeast clones exhibiting a positive LacZ response on galactose-raffinose plates, but not on dextrose plates, were considered to be potential NIF-1-interacting clones. The putative cDNAs were then isolated, sequenced, and subjected to size determination and further analysis.
Plasmids-Expression plasmids for nuclear receptors and various reporters have been described previously (24). The cloning of pCDNA3-NIF-1 and pJGM-NIF-1 has also been described (35). The plasmid IRES-RCD1 (66) was generously provided by Dr. Hiroto Okayama (University of Tokyo). pGEX-RCD1 was cloned by inserting full-length RCD1 into the EcoRI site of pGEX-4T-1. The pLPC vector (which introduces a FLAG and an HA tag into the cloning sequence) was kindly provided by Dr. Jeffrey Ye. pLPC-RCD1 was generated by first introducing a 5Ј-BamHI and 3Ј-EcoRI restriction enzyme site into the coding sequence of RCD1. The BamHI-EcoRI fragment was then cloned in-frame into the pLPC vector. Primers: 5Ј-BamHI, CTACCAGATGGATCCCACAGCCTGGCGA-CGGCTGCGC; 3Ј-EcoRI, CTACCAGATGAATTCTCAC-TGAGGGGGCAGGGGGATACC. The plasmid pSG5-FLAG-CCR4 (67) was a kind gift from Dr. Laura Corbo (Unité INSERM, France). The yeast expression vectors for nuclear receptors have been described earlier (24,35). The yeast expression plasmid pEG202⌬PL-RCD1 was constructed by cloning the EcoRI insert from IRES-RCD1 into the pEG202⌬PL vector. The plasmid pEG202⌬PL-CCR4 was cloned by inserting the BamHI-SalI fragment of pSG5-CCR4 in the pEG202⌬PL multiple cloning site.
Cell Culture and DNA Transfection-HeLa cells were routinely maintained in Dulbecco's modified Eagle's medium containing 10% bovine calf serum supplemented with glutamine and antibiotics. MCF-7 cells were maintained in Dulbecco's modified Eagle's medium containing 10% fetal calf serum supplemented with glutamine and antibiotics. For the reporter studies, HeLa cells were seeded in 6-well plates at 150,000 cells per well, and the medium was replaced with Dulbecco's modified Eagle's medium containing 10% hormone-depleted bovine calf serum (68) before transfection. Transfections were performed using calcium phosphate coprecipitation. Various ligands such as T 3 for TR␣, TTNPB for RAR␣, and LG100153 for RXR␣ were used at concentrations of 1 M. Dexamethasone for GR and estradiol (E 2 ) for ER␣ were used at 0.5 M and 0.1 M, respectively. Typically, 300 -600 ng of chloramphenicol acetyltransferase (CAT) reporter plasmid and 150 -300 ng of the expression plasmids were used per sample unless otherwise indicated. All transfections were performed in duplicate or triplicate.
Yeast ␤-Galactosidase Assays-Various ligands, such as T 3 for TR␣, 9-cis-RA for RAR␣ and RXR␣, E 2 for ER␣, and deoxycorticosterone for GR were used at concentrations of 1 M. Yeast colonies were first grown exponentially in Ura Ϫ , His Ϫ , and Trp Ϫ SD-dextrose medium, washed, diluted to the appropriate density, and incubated in Ura Ϫ , His Ϫ , and Trp Ϫ SD-galactose-raffinose medium for 12 h followed by quantitation of ␤-galactosidase as described earlier (35). ␤-Galactosidase units are expressed as (A 420 ϫ 1000)/(minutes of incubation ϫ A 600 of the yeast suspension), where A 420 and A 600 are optical density at 420 and 600 nm, respectively.
Antibodies and Coimmunoprecipitation-Rabbit polyclonal antibody against RCD1 was a kind gift from Dr. Hiroto Okayama (University of Tokyo). Rabbit polyclonal antibody against CCR4 was a kind gift from Dr. Laura Corbo (Unité INSERM, France). The anti-NIF-1 (BL2739) and the anti-Sox9 (AB5535) antibodies were from Bethyl Laboratories and Chemicon International, respectively. The anti-Cyclin D1 antibody was a gift from Dr. Michele Pagano (New York University School of Medicine). The antibody against ␤-actin was from Abcam. The anti-HA antibody was from Roche Applied Sciences. For coimmunoprecipitation, pLPC-RCD1 (HA-tagged) and full-length NIF-1 were transfected into HeLa cells. The cells were lysed with buffer (50 mM Tris-HCl, pH 7.4, 1% Nonidet P-40, 0.25% sodium deoxycholate, 150 mM NaCl, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 mM Na 3 VO 4 , 1 mM NaF, and protease inhibitors), and the lysate was incu-bated with an anti-HA affinity matrix (Roche Applied Science). The immunoprecipitated proteins were then eluted from the beads with an HA peptide (Covance Inc.) and then analyzed by Western blotting using anti-NIF-1 antibody. To study the interaction between CCR4 and NIF-1, pSG5-CCR4 (FLAG-tagged) and full-length NIF-1 were transfected into 293T cells. Whole cell lysates were prepared in lysis buffer (20 mM Hepes, pH 7.9, 150 mM NaCl, 20 mM KCl, 0.2 mM EDTA, 10 M ZnCl 2 , 1 mM MgCl 2 , 0.1% Triton X-100, and protease inhibitors) as described previously (67). Equal amounts of the lysates were incubated with anti-FLAG affinity matrix (Sigma). The immunoprecipitated proteins were eluted with FLAG peptide (Sigma) and then analyzed by Western blotting using anti-NIF-1 antibody.
siRNA Transfection-siRNAs directed against RCD1, CCR4, NIF-1, and a non-targeting control (siControl) were obtained from Dharmacon. HeLa and MCF-7 cells were transfected with the siRNAs (final concentration, 25 nM and 40 nM, respectively) using Hiperfect siRNA transfection reagent from Qiagen. For the receptor reporter assays, the appropriate DNA constructs were transfected into HeLa cells (Lipofectamine) 24 h after the siRNA transfection. The CAT assay was performed 64 h after the initial siRNA transfection. To study the effect of the siRNAmediated knockdown on endogenous target genes, 40 nM of the siRNAs was transfected into MCF-7 cells using Hiperfect. The cells were lysed after appropriate incubation with ligand. Equal amounts of protein were then analyzed by SDSgel electrophoresis, followed by Western blotting for either Sox9 or cyclin D1.

RESULTS
Identification of RCD1 as an NIF-1 Interactor-A LexAtagged NIF-1 was found to be constitutively active in S. cerevisiae and could not be used to identify potential NIF-1 interactors from the previously described cDNA library (35). Therefore, a full-length NIF-1 expressed as a B42 fusion protein was used as bait against a GH4C1 LexA cDNA library in a yeast two-hybrid screen as described under "Experimental Procedures." One of the interacting factors was identified as RCD1. RCD1 was initially identified in Schizosaccharomyces pombe as being required for nitrogen starvation-induced sexual development (70). It plays a critical role in the yeast's commitment to differentiation. The ortholog of RCD1 in S. cerevisiae is CAF40, a component of the evolutionarily conserved CCR4-NOT complex (52). The human ortholog CNOT9/hRCD1 is a 299-amino acid protein with a recently characterized Armadillo repeat (71). It has been shown to be important for retinoic acid (RA)induced endodermal differentiation of F9 embryonic carcinoma cells as well as RA-regulated lung development in mice (66).
NIF-1 Interacts with RCD1 in Vitro and in Vivo-To confirm the interaction between RCD1 and NIF-1 in vitro, a full-length GST-RCD1 fusion was expressed in E. coli and purified with glutathione-agarose beads. The 35 S-labeled NIF-1 was incubated with GST or GST-RCD1, and the samples were analyzed for binding as described under "Experimental Procedures." As shown in Fig. 1A, 35 S-NIF1 binds to GST-RCD1 but not to GST alone. The interaction between NRC-15 (fragment of NRC encompassing the NIF-1-interacting domain and the nuclear receptor-interacting domain), and NIF-1 has been well characterized and was used as a positive control (35). A yeast twohybrid interaction assay further confirmed the association between RCD1 and NIF-1. A LexA-tagged RCD1 and B42tagged NIF-1 were re-introduced into S. cerevisiae. The interaction between RCD1 and full-length NIF-1 was determined by measuring the activity of ␤-galactosidase in yeast extracts. As shown in Fig. 1B, a specific interaction between NIF-1 and RCD1 was confirmed. To determine whether RCD1 and NIF-1 interact in vivo, an influenza hemagglutinin (HA) epitopetagged RCD1 was coexpressed with NIF-1 in HeLa cells. The lysates were immunoprecipitated with HA antibody affinity matrix and eluted with an HA peptide, followed by SDS-gel electrophoresis and Western blotting with an antibody against NIF-1 (Fig. 1C). These results confirm the specific interaction between RCD1 and NIF-1 in mammalian cells.
RCD1 Enhances Ligand-dependent Transcriptional Activity of Nuclear Hormone Receptors-Our results in Fig. 1 indicate an interaction between RCD1 and NIF-1. To determine if RCD1 could affect nuclear receptor activity, we carried out an initial experiment by expressing TR␣ and its appropriate CAT reporter ⌬MTV-IR-CAT in HeLa cells along with increasing amounts of RCD1. RCD1 enhanced the ligand-dependent activation of TR␣ in a dose-dependent manner ( Fig. 2A). NIF-1 has been previously described as a potent cotransducer of nuclear receptors and was therefore used as a positive control (35).
RCD1 has been shown to be part of a RA-dependent complex that also contains the RARs, although a direct interaction between RAR␣ and RCD1 has not been confirmed (66). We sought to examine the effect of RCD1 expression on the liganddependent activation of RAR␣, RXR␣, ER␣, and GR (Fig. 2, B and C). Expression vectors for RCD1, nuclear receptors, and the respective CAT reporters were introduced into HeLa cells. RCD1 enhanced activation by both RAR␣ and RXR␣ in a liganddependent manner. The effect of RCD1 expression on the activation by GR was comparatively less than the other receptors (TR␣, RAR␣, and RXR␣) studied. No effect on ER␣ was observed. These results suggest that expression of RCD1 enhances ligand-dependent nuclear receptor activation, but with a preference for certain receptors.
CCR4, a Component of the CCR4-NOT Complex, Is a Nuclear Hormone Receptor Coactivator-hCCR4 protein, also termed as CNOT6, is an integral component of the eukaryotic CCR4-NOT complex (63). CCR4 and its associated factor, CAF1 have been implicated in transcription regulation in yeast and mammalian cells. Both CAF1 and CCR4 have been shown to enhance the activity of ER␣ in transient transfection studies (67,72). We decided to study the effect of CCR4 on liganddependent nuclear hormone receptor activation. Various nuclear receptors and their respective CAT reporters were expressed in HeLa cells along with CCR4. CCR4 enhanced the activation by TR␣, RAR␣, RXR␣, and GR in a ligand-de- interacts with NIF-1 in yeast. RCD1 was expressed as a LexA fusion protein and tested against a B42 fusion protein of NIF-1. The interaction was quantified by measuring the activity of ␤-galactosidase in yeast extracts. C, RCD1 associates with NIF-1 in mammalian cells. pLPC-RCD1 (HA-RCD1) was expressed in HeLa cells along with either pCDNA3-NIF-1 or pCDNA3 vector. Whole cell extracts were prepared 36 h later and were immunoprecipitated using an anti-HA affinity matrix. The bound proteins were eluted off the beads using a HA-peptide and processed for SDS-gel electrophoresis followed by Western blotting using an anti-NIF-1 antibody. pendent manner (Fig. 3). As noted previously, ER␣ activation was also increased by CCR4 expression.
CCR4 and RCD1 Do Not Directly Interact with Nuclear Hormone Receptors-A direct interaction between CCR4/RCD1 and nuclear hormone receptors could explain the increased activation as shown above (Figs. 2 and 3). To explore this possibility, we set up a yeast two-hybrid interaction assay between LexA fusions of nuclear receptors and B42-RCD1 or B42-CCR4. A B42 fusion of NRC-15 has been shown previously to interact directly with receptors and was therefore used as a positive control (24). RCD1, CCR4, and NRC-15 were found to be expressed at similar levels as assessed by Western blotting with an anti-HA antibody. However, as shown in Fig. 4, there was no interaction detected between RXR␣ and either RCD1 or CCR4. Additional interaction studies between RCD1 or CCR4 and TR␣-LBD, full-length TR␣, ER␣-LBD, full-length ER␣, RAR␣-LBD, and GR-LBD also showed no evidence for any direct interaction (data not shown).
CCR4 Functions through the Ligand-binding Domain of Nuclear Hormone Receptors-The LBDs of nuclear hormone receptors encompass a ligand-dependent activation function 2 (AF2 domain) through which a number of coactivators (e.g. p160/SRCs and NRC/NCoA6) mediate their effects (5). Our results show that CCR4 is a strong coactivator of nuclear hormone receptors. To determine if CCR4 functions through the ligand binding domain of the receptors, the LBDs fused to the Gal4 DNA binding domain were used in lieu of full-length receptors in reporter assays. Various Gal4 DBD-LBD constructs were expressed in HeLa cells along with a Gal4-responsive CAT reporter. Expression of CCR4 resulted in enhanced ligand-dependent activation of the LBD constructs (Fig. 5A). Although the expression of RCD1 enhanced the liganddependent transcriptional activation of TR␣, RAR␣, and RXR␣, its effect on the Gal4-LBD constructs of these receptors was less pronounced than that found with CCR4 (Fig. 5B).
Down-regulation of CCR4 Expression Decreases Nuclear Hormone Receptor Activation-Overexpression of RCD1 and CCR4 enhances ligand-dependent activation of nuclear hormone receptors (Figs. 2 and 3). To better explore the physiological role of endogenous RCD1 and CCR4 in nuclear receptor activation, an RNA interference approach was used. HeLa cells were transfected with siRNAs designed specifically against RCD1 and CCR4 mRNAs. The decrease in the protein levels of RCD1 and CCR4 was analyzed 60 h later by Western blotting (Fig. 6A). To study the effect of knockdown of RCD1 and CCR4 on nuclear receptor activation, expression vectors for various nuclear receptors and their corresponding CAT reporters were introduced into HeLa cells 24 h post siRNA transfection, and the CAT activity was subsequently analyzed 40 h later.
The transcriptional activation of TR␣, RAR␣, and RXR␣ was significantly inhibited by the knockdown of endogenous CCR4, whereas ER␣ and GR activation was affected to a lesser degree (Fig. 6, B and C). This suggests a selective role for CCR4 in nuclear hormone receptor activation. The decrease in RCD1 expression using RCD1-directed siRNA had less of an effect on nuclear receptor activation than that observed with the knockdown of CCR4. The ligand-dependent transcriptional activation of the respective reporter genes by TR␣, RAR␣, RXR␣, Increasing amounts (400 ng and 800 ng) of RCD1 were used. T 3 was used at 1 M. B, ⌬MTV-IR-CAT reporter was transfected into HeLa cells with either an RAR␣ or an RXR␣ expression vector along with RCD1. The ligands for RAR␣ and RXR␣, TTNPB, and LG100153, respectively, were each used at 1 M. C, expression vectors for GR and ER␣ were transfected into HeLa cells along with RCD1 and the corresponding reporter plasmids MMTV-CAT and ⌬MTV-ERE-CAT, respectively. The ligand for GR, dexamethasone, was used at 500 nM, and E 2 , the ligand for ER␣, was used at 125 nM. Experiments were carried out in triplicate. Error bars represent Ϯ S.D. ER␣, and GR was reduced, but by a lesser degree (Fig. 7A and data not shown). The siRNA-mediated knockdown of RCD1 also did not significantly affect the CCR4-dependent enhancement of nuclear receptor activation (as shown in Fig. 7B).

NIF-1 Knockdown Results in Decreased Nuclear Hormone
Receptor Activation-RCD1 has been shown to interact with NIF-1 (Fig. 1). To further study the effect of endogenous NIF-1 on nuclear receptor activation, we utilized an siRNA against NIF-1 mRNA to markedly reduce endogenous NIF-1 protein levels in HeLa cells (Fig. 8A). The effect of NIF-1 silencing on ligand-dependent activation of TR␣, RAR␣, and GR was subsequently analyzed. Nuclear hormone receptors and their reporter constructs were expressed in HeLa cells in which NIF-1 expression had been attenuated. The results suggest that the transcriptional activity of the receptors was considerably inhibited when NIF-1 was knocked down (Fig. 8B). These results concur with the ability of NIF-1 to enhance ligand-de-pendent activation of nuclear receptors as shown previously (35).

Effect of Silencing of RCD1 and CCR4 on Endogenous Nuclear Receptor Target Genes-Although the transfection results (Figs. 2-4)
indicate that RCD1 and CCR4 enhance nuclear receptor activation, the effect of the knockdown of CCR4 and RCD1 on endogenous target genes of nuclear hormone receptors would more clearly elucidate their physiological role. MCF-7 cells have been well documented to be both ER-and RAR-responsive (73,74). We therefore examined the effect of silencing of RCD1 and CCR4 on the expression of Sox9 (an RAR␣ target gene) and Cyclin D1 (an ER␣ target gene) in MCF-7 cells.
The siRNA-mediated silencing of CCR4 and NIF-1 specifically attenuated the levels of endogenous Sox9 protein but had no detectable effect on the levels of endogenous ␤-actin (Fig.  9A). The knockdown of RCD1 also reduced Sox9 protein levels but to a lesser degree. The effect of siRNA-mediated silencing of CCR4 and RCD1 on Cyclin D1, an endogenous ER␣ target gene, was also analyzed. As noted above, the knockdown of CCR4 and RCD1 had a somewhat lesser effect on the ligand-dependent activation of ER␣ in transient transfection experiments, as compared with the other receptors studied ( Fig. 6 and data not shown). Consistent with these results, the expression of cyclin D1 was not affected by the siRNA-mediated knockdown of either RCD1 or CCR4 (as shown in Fig. 9B) in the E 2 -responsive MCF-7 cells.
Another well documented RAR␣-responsive gene is HoxA1 (75). Because there are no antibodies against HoxA1 available for Western blotting, we examined the effect of silencing of RCD1, CCR4, and NIF-1 on HoxA1 expression by a quantitative RT-PCR method. The decrease in RCD1 and CCR4 mRNA levels was first confirmed by semi-quantitative PCR (Fig. 10A). Western blotting confirmed the knockdown of NIF-1 in MCF-7 cells (Fig. 10A). The HoxA1 mRNA levels were initially analyzed by semi-quantitative RT-PCR to determine the relative importance of RCD1, CCR4, and NIF-1 (data not shown). These results were further confirmed and quantitated by realtime RT-PCR analysis (Fig. 10B, left panel). The silencing of both CCR4 as well as NIF-1 resulted in a decrease in the induction of HoxA1 mRNA in MCF-7 cells. In contrast, RCD1 does not appear to have as significant a role in the expression of endogenous HoxA1 by RAR␣. We also analyzed the endogenous Sox9 mRNA expression, both by semi-quantitative RT-PCR (data not shown) and real-time RT-PCR analysis (Fig. 10b,  right panel), and validated the effect of the knockdown of CCR4 and NIF-1.
CCR4 Associates with NIF-1 in Vivo in an RCD1-independent Manner-Our results indicate that both CCR4 and NIF-1 play a role in the regulation of Sox9 and HoxA1 (RAR␣ target genes) (Fig. 10). This raised the possibility of an association between  CCR4 and NIF-1. To address this question, FLAG-tagged CCR4 was coexpressed with NIF-1 in 293T cells. The protein lysates were initially immunoprecipitated with FLAG antibody affinity matrix. The FLAG-tagged CCR4 and its associated proteins were then eluted from the affinity matrix with a FLAG peptide. The samples were subjected to SDS-gel electrophoresis followed by Western blotting with an antibody against NIF-1. As shown in Fig. 11A, a specific interaction between CCR4 and NIF-1 was determined. The eluates were simultaneously Western blotted with an anti-FLAG antibody to confirm comparable expression levels of CCR4.
Because CCR4 and RCD1 are components of the CCR4-NOT complex and RCD1 has been shown to interact directly with NIF-1 (Fig. 1), we sought to determine if the association between CCR4 and NIF-1 occurs in an RCD1-dependent manner. 293T cells were transfected with either a control siRNA or an siRNA against RCD1 mRNA. The expression vectors for FLAG-tagged CCR4 and NIF-1 were then introduced into 293T cells 24 h post siRNA transfection. The whole cell lysates were immunoprecipitated and eluted as above, followed by SDS-gel electrophoresis and Western blotting with an anti-NIF-1 antibody. Our results indicate that the association between CCR4 and NIF-1 is independent of RCD1 (data not shown).
CCR4-enhanced Activation by RAR␣ Is NIF-1-dependent-An in vivo interaction between CCR4 and NIF-1 (Fig. 11A) and a comparable affect on the RAR␣ target genes studied (Sox9 and HoxA1) (Fig. 10) prompted us to further examine the associa-tion between CCR4 and NIF-1. MCF-7 cells were transfected with either an siRNA against NIF-1 or a corresponding control siRNA. An expression vector for CCR4 along with a RAR␣responsive CAT reporter was introduced 24 h after siRNA transfection and the CAT activity was subsequently analyzed 40 h later.
Ligand-dependent activation by endogenous RAR␣ was enhanced by CCR4, consistent with its function as a coactivator. The siRNA-mediated knockdown of NIF-1, however, inhibited the CCR4-dependent increase in transcriptional activation by RAR␣ (Fig. 11B). These results suggest that the activation by CCR4 requires the presence of NIF-1.

FIGURE 5. CCR4 mediates its effect through the LBDs of nuclear receptors.
A, the Gal4 PMC110-CAT reporter was cotransfected in HeLa cells with vectors expressing a fusion of the Gal4-DBD and the receptor LBDs with or without CCR4. Cells were incubated with or without the corresponding ligands at the concentrations indicated previously. The samples were assayed for CAT activity after 40 h. Error bars represent Ϯ S.D. B, effect of RCD1 on the LBDs of nuclear hormone receptors. The assay was conducted in a similar fashion, except that HeLa cells were transfected with RCD1. FIGURE 6. Effect of siRNA-mediated knockdown of CCR4 on the activation of TR␣, RAR␣, RXR␣, GR, and ER␣. A, knockdown of RCD1 and CCR4. SMART-poolா siRNA against RCD1 and CCR4 (siRCD1 or siCCR4, respectively) were transfected into HeLa cells at a concentration of 25 nM, and protein lysates were prepared 60 h post-transfection. A non-targeting control SMARTpoolா siRNA (siControl) was also included at the same concentration. The extent of knockdown was analyzed by immunoblotting using the indicated antibodies. B, HeLa cells were transfected with either siControl or siCCR4. 24 h later, these cells were transfected with expression vectors for TR␣, RAR␣, and RXR␣ along with the ⌬MTV-IR-CAT reporter. The cells were incubated with the corresponding ligands as indicated, and CAT activity was assayed 40 h post-DNA transfection. Error bars represent Ϯ S.D. C, similar to B except that the expression vectors for GR and ER␣ were transfected along with their respective reporter constructs.

DISCUSSION
The evolutionarily conserved CCR4-NOT complex controls gene expression both at the level of transcription initiation and of mRNA degradation. The mammalian CCR4-NOT complex contains the orthologs of yeast NOT proteins (CNOT1-CNOT4) and also CNOT6 (hCCR4), CNOT7 (hCaf1), CNOT8 (hCalif), CNOT9/RCD1 (hCAF40), and CNOT10 (hCAF130) proteins (63). Certain NOT proteins have been reported to function as repressors of transcription. A repressor domain called the NOT-box has been identified in both CNOT2 as well as CNOT3 (61). The SMRT/NCoR-HDAC3 corepressor complex is thought to function as a cofactor for transcriptional repression by CNOT2 (76). The CCR4-NOT complex is also involved in mRNA decay. CNOT6 (hCCR4) has been shown to play a role in mRNA deadenylation (64).
Various components of the mammalian CCR4-NOT complex have been reported to be involved in transcriptional activation by nuclear hormone receptors. mCAF1 was found to be an essential factor for spermatogenesis and has been shown to be a regulator for RXR␤ (77,78). CNOT6 (hCCR4) and CNOT7 (hCAF1) were previously shown to significantly enhance the ER␣-mediated activation of a luciferase reporter gene (67,72). Our results here show that RCD1 (CNOT9), another component of the CCR4-NOT complex, enhances the transcriptional activity of nuclear receptors in transient transfection assays. Furthermore, we show that CCR4 augments the liganddependent transcriptional activation of a wide variety of nuclear receptors, including ER␣.   NIF-1 was previously characterized, through its interaction with the coactivator NRC, as a factor that enhances nuclear receptor activation (25,35). Because NIF-1 does not directly interact with nuclear receptors, the modulation of ligandbound nuclear hormone receptors by NIF-1 appears to occur through the direct interaction between NRC and nuclear hormone receptors. In this study, we identified RCD1 as an NIF-1interacting factor through a yeast two-hybrid screen. RCD1 is highly conserved across species and has been shown to be critical for RA-induced differentiation of F9 cells. RCD1 was identified as a component of the DRF transcriptional complex in these cells, which consists of at least p300/CBP, ATF-2, and the RARs (66,79).
We report here that RCD1 interacts with NIF-1 in vivo and in vitro. Transient transfection experiments indicate that RCD1 enhances the activity of TR␣, RAR␣, and RXR␣, and to a lesser degree, ER␣ and GR. siRNA-mediated knockdown of endogenous RCD1 resulted in a nominal decrease in the activation of nuclear receptors, including TR␣, RAR␣, and RXR␣. The ligand-dependent activation of ER␣ and GR was minimally affected by the knockdown of RCD1. Our studies on endogenous target genes corroborated these findings for RAR␣ and ER␣. The expression of endogenous Sox9 (RAR␣ target gene) in MCF-7 cells was not significantly influenced by the siRNA-mediated silencing of RCD1, as confirmed both by Western blotting and quantitative PCR. The expression of cyclin D1, an ER␣ target gene, was also not affected by the knockdown of endogenous RCD1. These results are consistent with the findings of Winkler et al. (62) who reported that the silencing of RCD1 did not affect the mRNA levels of TFF1, another ER␣ target gene. On the other hand, a decrease in the levels of c-Myc mRNA was reported, suggesting a gene-and promoter-specific function of RCD1.
CCR4 has been shown to be involved both in transcription as well as in mRNA decay. The two distinct sites of CCR4 function (nucleus and cytoplasm) indicate a dual residence of the protein. In agreement with the above, we have found that GFP-CCR4 localizes in both the nucleus and the cytoplasm. 3 Our results here further the link between CCR4 and nuclear receptors. In addition to enhancing ER␣ function (67), we show that CCR4 positively affects liganddependent activation of other nuclear receptors, including TR␣, RAR␣, RXR␣, and GR. This was confirmed by both transient transfection experiments as well as the effect of siRNA-mediated knockdown of CCR4 on nuclear receptor activation.
CCR4 had a more potent effect on the transcriptional activation of nuclear receptors than that observed for RCD1. This could be explained by differences in the relative abundance of these factors. Thus, if RCD1 acts through the CCR4-NOT complex and RCD1 is much more abundant than CCR4, the effect of RCD1 expression or knockdown will be less than that found with CCR4. Alternatively, it is possible that RCD1 and CCR4 function independently through other yet to be defined pathways in addition to the CCR4-NOT complex. It must be noted here that the CCR4 knockdown resulted in a greater decrease in the ligand-dependent activation of TR␣, RAR␣, and RXR␣ compared with ER␣ and GR. This selective function of CCR4 was further elucidated with studies on endogenous hormone responsive target genes. The knockdown of CCR4 resulted in a decrease in RAR-mediated stimulation of Sox9 protein levels but did not have an effect on the ERmediated stimulation of cyclin D1. However, estrogen also initiates many non-nuclear signaling events, including the activation of several signal transduction pathways (e.g. phosphatidylinositol 3-kinase/AkT, protein kinase C, p38/mitogen-activated protein kinase, and Janus tyrosine kinase/signal transducers and activators of transcription) (80). It has been shown previously that estrogen rapidly stimulates phosphatidylinositol 3-kinase and AkT in MCF-7 cells leading to an Src-dependent induction of cyclin D1 expression (81,82). Therefore, it is possible that this non-genomic expression of cyclin D1 is independent of RCD1 and CCR4 and thus is not affected by the siRNA-mediated knockdown of either of the two factors.  MCF-7 cells were transfected with siRNAs against RCD1, CCR4, and NIF-1 as indicated. The extent of knockdown of RCD1 and CCR4 was determined by semi-quantitative RT-PCR. 100 ng of the total RNA was used for each sample. GAPDH was used as an internal control. To study NIF-1 knockdown, protein lysates were prepared 60 h after siRNA transfection and immunoblotted using anti-NIF-1 antibody. B, real-time RT-PCR analysis of HoxA1 and Sox9. MCF-7 cells were transfected with siRNAs against RCD1, CCR4, and NIF-1 at a concentration of 40 nM as indicated. A control siRNA was included in the assay. The cells were incubated with 1 M all-trans-RA (AtRA) for 24 h before the cells were processed. Total RNA (1 g) was reverse-transcribed, and 1 l of each cDNA sample was used for the quantitative PCR analysis. All samples were assayed in triplicate. GAPDH was used as an internal normalizing control. The quantification was done using the comparative Ct method. Error bars represent Ϯ S.D.
Quantitative PCR analysis confirmed the effect of CCR4 silencing on Sox9 mRNA levels. HoxA1, another RAR␣-responsive gene in MCF-7 cells was also affected by the siRNA-mediated knockdown of CCR4. The silencing of RCD1 did not influence the mRNA levels of HoxA1, consistent with the results observed with Sox9. As mentioned previously, NIF-1 expression enhances ligand-dependent nuclear receptor activation (35). In keeping with the coactivator function of NIF-1, Sox9 and HoxA1 mRNA levels were both attenuated by siRNA-mediated silencing of NIF-1. This decrease in the RAR-mediated stimulation of Sox9 and HoxA1 by the knockdown of NIF-1 was comparable to that observed with the silencing of CCR4.
Although neither RCD1 nor CCR4 directly interact with nuclear receptors, they potentiate their ligand-dependent activities. CCR4 mediates this function through the LBDs of nuclear receptors. This suggests that the RCD1-CCR4 complex tethers to the promoters of target genes through the interaction with another factor(s) that bind to the liganded nuclear recep-tor LBD. One such model would involve the interaction of a CCR4-NOT-RCD1 complex with NIF-1, which in turn interacts with NRC, which has been shown to bind directly to the liganded LBD of nuclear receptors (35). Our results indicate an in vivo association between CCR4 and NIF-1, which occurs independent of the presence of RCD1. Consistent with this result, the silencing of RCD1 did not affect CCR4-dependent enhancement of nuclear receptor activation (Fig. 7B). In a similar vein, we show that NIF-1 is necessary for the coactivator function of CCR4, providing an additional link between the two factors. In further support of this model, a recent study suggested that the knockdown of NRC leads to a marked reduction of RAR-mediated stimulation of Sox9 in MCF-7 cells (83).
In summary, we report that RCD1 and in particular CCR4 function as nuclear receptor coactivators. RCD1 was isolated as an interactor of NIF-1, which has been shown to enhance nuclear receptor activation. Our results provide the first indication that CCR4 enhances the activity of a variety of nuclear receptors, other than ER␣. Our findings on the effect of CCR4 on HoxA1 expression, along with the previously reported role of RCD1 in RA-mediated differentiation (66), suggest a link between components of the CCR4-NOT complex and the growth and differentiation of mammalian tissues.