Functional Interaction between Oct-1 and Retinoid X Receptor*

The retinoid X receptor (RXR) is a member of the nuclear hormone receptor superfamily and heterodimerizes with a variety of other family members such as the thyroid hormone receptor (TR),1 retinoic acid receptor, vitamin D receptor, and peroxisome proliferator-activated receptor. Therefore, RXR is supposed to play a key role in a ligand-dependent regulation of gene transcription by nuclear receptors. In this study, we have identified the octamer-binding transcription factor-1 (Oct-1) as a novel interaction factor of RXR. In vitro pull-down assays using RXR deletion mutants showed that the interaction surfaces were located in the region encompassing the DNA binding domain (C domain) and the hinge domain (D domain) of RXR. We also showed that RXR interacted with the POU homeodomain but not with the POU-specific domain of Oct-1. Gel shift analysis revealed that Oct-1 reduced the binding of TR/RXR heterodimers to the thyroid hormone response element (TRE). In transient transfection assays using COS1 cells, Oct-1 repressed the T3-dependent transcriptional activity of TR/RXR heterodimers, consistent with in vitro DNA binding data; however, transcriptional activation by Gal4-TR(LBD) (LBD, ligand binding domain), which lacks its own DNA binding domain but retains responsiveness to T3, was not influenced by Oct-1. These results suggest that Oct-1 functionally interacts with RXR and negatively regulates the nuclear receptor signaling pathway by altering the DNA binding ability of the receptors.

The retinoid X receptor (RXR) is a member of the nuclear hormone receptor superfamily and heterodimerizes with a variety of other family members such as the thyroid hormone receptor (TR), 1 retinoic acid receptor, vitamin D receptor, and peroxisome proliferator-activated receptor. Therefore, RXR is supposed to play a key role in a ligand-dependent regulation of gene transcription by nuclear receptors. In this study, we have identified the octamer-binding transcription factor-1 (Oct-1) as a novel interaction factor of RXR. In vitro pull-down assays using RXR deletion mutants showed that the interaction surfaces were located in the region encompassing the DNA binding domain (C domain) and the hinge domain (D domain) of RXR. We also showed that RXR interacted with the POU homeodomain but not with the POU-specific domain of Oct-1. Gel shift analysis revealed that Oct-1 reduced the binding of TR/RXR heterodimers to the thyroid hormone response element (TRE). In transient transfection assays using COS1 cells, Oct-1 repressed the T3-dependent transcriptional activity of TR/RXR heterodimers, consistent with in vitro DNA binding data; however, transcriptional activation by Gal4-TR(LBD) (LBD, ligand binding domain), which lacks its own DNA binding domain but retains responsiveness to T3, was not influenced by Oct-1. These results suggest that Oct-1 functionally interacts with RXR and negatively regulates the nuclear receptor signaling pathway by altering the DNA binding ability of the receptors.
The steroid/thyroid hormone receptor superfamily is a large group of related transcriptional factors that control cellular differentiation, development, and homeostasis by direct interaction with distinct cis-elements in target genes (1,2). This superfamily includes receptors for steroids, thyroid, vitamin D3, retinoids, and a large number of orphan receptors whose cognate ligands are still unknown (3). Members of the super-family are characterized by a highly conserved cysteine-rich DNA binding domain containing two zinc finger structures necessary for sequence-specific DNA interaction (4). The complex carboxyl-terminal region of the receptors contains ligand binding, receptor dimerization, and putative transcriptional activation function (5). Members of the superfamily regulate transcription by binding to response elements containing two or more copies (often degenerate) of the consensus motif AGGTCA (6,7). Recently it was shown that retinoic acid receptor, TR, VDR, and peroxisome proliferator-activated receptor form heterodimers with the RXR on bipartite hormone response elements composed of non-symmetrical head-to-tail tandem AGGTCA "half-sites" (8 -10). More than half of the orphan receptors have been shown to heterodimerize with RXR (3). Thus, RXRs are supposed to play a key role in ligand-dependent transcriptional activity of nuclear receptors.
The mechanisms by which nuclear hormone receptors regulate target gene transcription are currently under intensive investigation. The ligand-activated nuclear receptors may promote formation of preinitiation complex of the basal transcriptional apparatus and facilitate transcription by RNA polymerase II. Recently, a number of nuclear receptor-associated proteins have been identified that interact with receptors and regulate their transcriptional activities. A nuclear receptor corepressor, N-CoR or a related factor SMRT, binds to unliganded receptors and acts as a transcriptional silencer of nuclear receptors (11)(12)(13). It has been shown that N-CoR and SMRT recruit mSin3 and mRPD3 that possess histone deacetylase activity and make the chromatin transcriptionally inactive (14). When ligands bind to receptors, the co-repressor complex dissociates, and a co-activator complex containing N-CoA1/ SRC-1, CBP/p300, and p/CAF associates with the receptors (15)(16)(17)(18)(19)(20)(21). Interestingly, these complexes have histone acetyltransferase activity and make the chromatin unwind, resulting in the transcriptionally active state (22)(23)(24).
Although recent extensive studies have depicted a model of nuclear receptor action, it is still not enough to explain the divergent biological effect of nuclear hormone receptors in development, differentiation, and cell cycle regulation. Therefore, it is reasonable to speculate that a large number of novel factors may associate with nuclear receptors. By using a biochemical technique, we have identified the octamer-binding transcription factor-1 (Oct-1) as a novel interaction factor of RXR. Oct-1 is ubiquitously expressed and activates the octamer motif containing promoters which has been shown to be related to cell cycle regulation of the human histone H2B gene and the constitutive expression of small nuclear RNA genes (25)(26)(27). Oct-1 is a member of a family of transcription factors characterized by the presence of a bipartite DNA binding domain, the POU domain (28,29). This POU domain consists of two conserved regions, a POU-specific domain and a POU homeodomain. Both subdomains have a helix-turn-helix motif and act * This study was supported in part by a grant from the Ministry of Education, Japan. 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.

EXPERIMENTAL PROCEDURES
Isolation of Interacting Proteins-Rat GH3 cDNA library was constructed using T7 expression phage and screened by a full-length human RXR␣ as a probe. Isolated clones were subcloned into pGEM 3 and sequenced by Applied Biosystems 3300 autosequencer. [ 35 S]Methionine-labeled peptides were produced using the T7 TNT-coupled system (Promega), and their interactions with RXR␣ were confirmed by puldown experiment using matrix-bound glutathione S-transferase (GST)-RXR␣. About 1 ϫ 10 6 clones were screened, and one clone that corresponds to human Oct-1 cDNA containing amino acid residues 371-438 was confirmed as an interacting partner with RXR.
The Palx2 TK luciferase gene contains two copies of a palindromic TRE upstream of the thymidine kinase (TK) promoter in the PA3 luciferase vector (46). The rGH chloramphenicol acetyltransferase reporter plasmid that contains rat growth hormone promoter region spanning from Ϫ237 to ϩ8 from the transcription start site was a gift from D. D. Moore (see Ref. 47). To construct the mammalian expression vector for Gal4 DBD fusion protein, PCR-amplified ligand binding domain of TR␣1 was inserted in frame into BamHI-and SalI-cloning sites of the pM vector (CLONTECH). The following oligonucleotides were used to amplify the TR␣1 LBD: forward primer, 5Ј-atcggaattc ATGGC-CATGGACTTGGTTCT-3Ј, and reverse primer, 5Ј-gatcgtcgacTTA-GACTTCCTGATCCTCAA-3Ј. UASx4 TK luc reporter plasmid was gift from Dr. R. M. Evans (see Ref. 48).
Cell Culture and Transient Transfection and Reporter Assays-COS1 cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 100 units/ml penicillin G, and 0.25 mg/ml streptomycin at 37°C in 5% CO 2 . Transfection was done in COS1 cells using the standard calcium phosphate procedure. Typically, 0.25 g of Palx2-driven luciferase reporter was co-transfected with 100 ng of the indicated expression vectors. Cells were incubated for 12 h, and the medium on the cells was replaced with fresh medium and 10 -7 M T3 was added. Cells were harvested after 12 h. ␤-Galactosidase activity was measured by the method previously described using chlorophenol red-␤-D-galactopyranoside as a substrate (49). Luciferase assays were performed using the PicaGene Luciferase Assay System (Toyo Inki, Tokyo, Japan). Luciferase activity was determined using Lumat LB9501 (Berthold Japan K.K., Tokyo, Japan) and was expressed as relative light units normalized to the ␤-galactosidase activity. Chloramphenicol acetyltransferase activities were measured as described previously (50) and quantitated by PhosphorImager. Each transfection was conducted in triplicate, and data represent the mean Ϯ S.D. of more than three individual experiments.
In Vitro Transcription and Translation-Coupled transcription and translation of Oct-1/2 RXR were carried out using a T7 TNT in vitro transcription/translation kit (Promega) according to the manufacturer's instructions.
Overnight cultures of Escherichia coli BL21 carrying the recombinant GST-RXR␣ and GST control plasmids were diluted 100-fold, cultured for 5-6 h, and then induced with 0.1 mM isopropyl ␤-D-thiogalactopyranoside. After another 3 h, bacteria were collected and then washed with PBS. Pellets were suspended in PBS containing 1% (v/v) Triton X-100 and were then sonicated. Debris was removed by centrifugation. The fusion protein or the GST control protein was bound to glutathione-Sepharose (Amersham Pharmacia Biotech) and extensively washed with PBS containing 1% (v/v) Triton X-100. Matrix-bound proteins were used for interaction experiments.
GST Pull-down Assay-10 l of GST-Sepharose beads containing 2-5 g of GST recombinant proteins were incubated with [ 35 S]methionine-labeled proteins for 1 h at 4°C. Complexes were then centrifuged, washed three times in gel shift buffer, and separated by SDS-polyacrylamide gel electrophoresis. Radiolabeled signals were visualized and quantified using a PhosphorImager (Fuji BAS 1500). DNase I (5 units/ ml) was added to the reaction in experiments shown in Fig. 1B.
Gel Retardation Assay-Synthetic oligonucleotides representing each strand of the sequences were purified by polyacrylamide gel electrophoresis, eluted, and annealed. Double-stranded oligonucleotides were radiolabeled with dCTP (Ͼ3300 Ci/mmol; ICN, Costa Mesa, CA) by fill in reactions using Klenow large fragment DNA polymerase. Radiolabeled probes (10 fmol, 20,000 -30,000 cpm) were then incubated with binding proteins in 30 ml of reaction mixture containing 10 mM KPO 4 , pH 8.0 buffer, 1 mM EDTA, 80 mM KCl, 1 g of poly(dI-dC), 1 mM dithiothreitol, 0.5 mM MgCl 2 , 5 g of bovine serum albumin, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 2 mg/ml aprotinin, 1 mM leupeptin, 1 mM pepstatin. These reactions were incubated for 30 min at room temperature and analyzed on a 5% non-denaturing polyacrylamide gel in TAE buffer. Electrophoresis was performed at a constant voltage of 200 V at 4°C in the same buffer.
Expression of Recombinant Proteins-To express the fusion proteins with GST, PCR-amplified full-length RXR␣ cDNA or truncated fragments were inserted in frame into BamHI-and EcoRI-cloning sites of the pGEX-2TK vector (Amersham Pharmacia Biotech). Overnight cultures of E. coli BL21 carrying the recombinant GST fusions or GST control plasmid were diluted 100-fold, cultured for 5-6 h, and then induced with 0.1 mM isopropyl ␤-D-thiogalactopyranoside. After another 3 h, bacteria were collected and washed with PBS. Pellets were suspended in PBS containing 1% (v/v) Triton X-100 and sonicated. Debris was removed by centrifugation. The fusion protein or the GST control protein was bound to glutathione-Sepharose (Amersham Pharmacia Biotech) and extensively washed with PBS containing 1% (v/v) Triton X-100. Matrix-bound proteins were used for interaction experiments.
Interaction Experiments-In vitro translated 35 S-labeled proteins (1-2 l) were incubated for 20 min at room temperature with glutathione-Sepharose (10 l) preloaded with GST fusion or GST control protein in 250 l of binding buffer (20 mM Tris-Cl, pH 7.8, 100 mM NaCl, 10% glycerol, 1 mM dithiothreitol, 1 mM EDTA, 1 mM phenylmethanesulfonyl fluoride, 1 mM leupeptin, 1 mM pepstatin, 2 mg/ml aprotinin) in the presence or absence of 10 -6 M of T3. After extensive washing with binding buffer, bound proteins were eluted in 25 l of Laemmli sample buffer, boiled for 10 min, and resolved by sodium dodecyl sulfatepolyacrylamide gel electrophoresis (10%) followed by autoradiography. The results of the in vitro reactions and the amount of 35 S-labeled protein bound by GST fusions were visualized and quantified using a PhosphorImager (Fuji BAS 1500).

RESULTS
Isolation of Oct-1 as a Interacting Protein with RXR-In order to identify the interacting protein with RXR, we used biochemical methods. Rat GH3 cell cDNA library was screened with GST fusion containing full-length human RXR␣ as a probe. Positive clones were transcribed by T7 RNA polymerase, translated into 35 S-labeled peptides using [ 35 S]methionine, and used for pull-down experiments with GST-RXR␣ to confirm the interactions. One positive clone was identified, and the nucleotide sequences were determined. Sequence homology searches in GenBank TM revealed that the isolated clone contained the POU homeodomain of rat Oct-1 cDNA, which was highly homologous to the human Oct-1 amino acid residues 371-438.
Oct-1 Interacts with CD Domain of the RXR-To examine the interaction between Oct-1 and RXR, we used the matrix-bound fusion protein of glutathione S-transferase with RXR (GST-RXR) for in vitro pull-down assay. As shown in Fig. 1A, [ 35 S]methionine-labeled in vitro translated Oct-1 interacted with GST-RXR but not GST alone, and GST-RXR did not retain any of the in vitro translated control luciferase protein. As shown in Fig. 1B, addition of DNase I in the binding mixture did not alter the association of Oct-1 to matrix-bound RXR, indicating that the protein-protein interaction was not due to the presence of contaminating DNA. Reciprocal pull-down experiment was performed to confirm the interaction between RXR and Oct-1. As shown in Fig. 1C, matrix-bound GST-POU domain (3rd lane) and POU homeodomain (5th lane) specifically retains [ 35 S]methionine-labeled RXR, whereas GST alone (lane 2) or GST-POU-specific domain (4th lane) did not.
We next examined the specific domains in RXR that interact with Oct-1. A series of deletion mutants of GST fusion proteins representing overlapping portions of RXR ( Fig. 2A) were expressed in bacteria, purified, and used to bind 35 7). It was not surprising that the highly conserved DNA binding domain of nuclear receptors could also serve as a site for binding of co-regulator proteins. We identified the nuclear protein Oct-1 as a binding protein for the CD regions of RXR.
Oct-1 Inhibits the Binding of TR/RXR Heterodimer to Hormone Response Element-The above results were of interest because the DNA binding domain of RXR has been reported to be involved in the formation of TR/RXR heterodimers on direct repeat DNA element. To characterize further interaction between Oct-1 and receptor heterodimers, we performed gel mobility shift assays using bacterially expressed and purified TR and RXR. The results shown in Fig. 3 indicated that addition of increasing amounts of bacterially expressed and purified GST-Oct-1 POU domain attenuated the binding of RXR/TR heterodimers to TRE (lanes 4 and 5), whereas addition of GST alone did not alter the binding (lane 3). The DNA binding activity of TR/RXR heterodimers was decreased by addition of GST-Oct-1 POU domain. These results showed that association of Oct-1 to RXR inhibited the RXR/TR heterodimers from binding to DNA elements.
Oct-1/2 Represses the Hormone-dependent Transcriptional Activity-To test the possible role of Oct-1 and Oct-2 in T3-dependent transcriptional activation, we performed transient transfection experiments in COS1 cells. Full length Oct-1 or Oct-2 expression vectors or empty expression vectors were cotransfected with a luciferase reporter plasmid containing two copies of the TR response element into COS1 cells. After 12 h, cells were harvested, and luciferase activities were determined. As shown in Fig. 4, both Oct-1 and Oct-2 repressed the T3-dependent transcriptional activity. In the presence of T3, coexpression of Oct-1 decreased the TR activity by approximately 40% and Oct-2 by approximately 30%. Co-expression of Oct-1/2 did not influence the amount of expression of TR when determined by T3 binding assay. Furthermore, Oct-1/2 had little effect on cytomegalovirus promoter (data not shown). These data suggest that Oct-1 can function as a co-repressor for the T3-dependent transcriptional activity of the TR/RXR heterodimers. We next examined the effect of Oct-1 expression on naturally occurring T3-responsive promoter. As shown in Fig.  4B, rat growth hormone promoter, which was one of a well characterized T3-responsive promoter, was inhibited in a similar manner, suggesting the physiological relevance of the inhibitory role of Oct-1 in T3-dependent transcription.
Oct-1/2 Also Interact with TR and VDR-Because highly conserved DNA binding domain of RXR was involved in the interaction, it is likely that other nuclear receptors could also interact with Oct-1. We next examined the interaction of other nuclear receptor family members with Oct-1 and Oct-2. As we expected, both 35 S-labeled Oct-1 and Oct-2 associated to matrix-bound TR and VDR as well as RXR (Fig. 5A). In order to confirm the heterodimerization ability of GST fusion proteins with RXR, 35 S-labeled RXR was incubated with matrix-bound GST fusions (lanes 11-15). Significant associations were detected with GST-TR and GST-VDR, whereas only weak association was detected with GST-RXR, consistent with previous data showing preference of heterodimerization of RXR. We next tested whether Oct-1 can interact with irrelevant DNA-binding transcription factor AP-2. As shown in Fig. 5B, 35 S-labeled Oct-1 did not bind to the GST-AP-2, whereas significant amounts of 35 S-labeled Oct-1 associated with GST-RXR, suggesting the specific interaction of Oct-1 with nuclear receptors.
Oct-1 Did Not Influence the Transcriptional Activation by Gal4-TR(LBD) Fusion Protein-We further investigated the effect of Oct-1 on transcriptional activation by Gal4-TR(LBD) which lacks its own DNA binding domain but retains responsiveness to T3. As shown in Fig. 6, Oct-1 did not influence the transcriptional activation by Gal4-TR␣(LBD) fusion protein on upstream activating sequence luciferase reporter. These data are consistent with in vitro results showing that Oct-1 interacted with nuclear receptors via their DNA binding domain and inhibited their DNA binding activity. The DNA binding domain of nuclear receptor was required for the inhibitory effect of Oct-1.

DISCUSSION
In this study, we have examined the interaction between RXR and POU domain of Oct-1. Our results indicated that Oct-1 interacted with nuclear receptors by direct protein-protein interaction and influenced the ligand-dependent transcriptional activity of nuclear receptors.
Recent data show that nuclear receptor co-activators such as SRC-1 and CBP/p300 recruit histone acetyltransferase, p/CAF, and pCIP to DNA-bound nuclear receptors and lead to the unfolding of the DNA-core histone complex (22)(23)(24)51). In addition to co-activators, nuclear receptor co-repressor N-CoR and related factor SMRT, which were initially discovered through their ability to bind to unliganded nuclear receptors, recruit histone deacetylase (mSin3 and mRPD3) and result in the condensation of the chromatin structure to repress the basal transcription (14,12,52). Co-activator and co-repressor interact with the ligand binding domain and hinge domain of nuclear receptors, respectively. So far, little is known about potential role of the DNA binding region of nuclear receptors on transcriptional regulation. We report here the identification of the transcription factor Oct-1/2 as a binding protein for the DNA binding domain (DBD) of RXR and influence on liganddependent transcriptional activity of nuclear receptors. The highly conserved DBD of nuclear receptors could also serve as a site of interaction for co-regulator proteins, suggesting that Oct-1/2 function is analogous among the members of the nuclear receptors. Our results demonstrated that the POU homeodomain of Oct-1, which is known to form a complex with the herpes simplex virus transactivator VP16 (53), was involved in the interaction with RXR. POU homeodomains have been  5, 10, and 15) bound to glutathione-Sepharose beads. 10% input from the 35 S-labeled proteins is indicated (lanes 1, 6, and 11). B, 35  shown to be involved in protein-protein interaction as well as sequence-specific DNA binding. A number of transcription factors have been identified to interact with the POU domains of Oct-1 and/or Oct-2, e.g. TBP, TFIIB, HMG2, the lymphoidspecific transcriptional co-activator OBF-1 (30 -34). Recently, it has been reported that glucocorticoid receptor interacts with the POU domain of Oct-1 and modulates the ligand-dependent transcriptional activity of glucocorticoid receptor (54,55). Furthermore, Budhram et al. (56) also reported that the POU domains of Brn-3a and Brn-3b, which belong to the POU family of transcription factors, interact with estrogen receptor (ER) and regulate transcriptional activity of ER. It also has been demonstrated that POU transcription factor Pit-1 interacts with TR to activate growth hormone gene transcription (57) and with ER to modulate expression of the prolactin promoter in pituitary cells (58,59).
Herein, our results showed that the Oct-1 POU domain interacted with DBD of RXR and prevented TR/RXR heterodimers from binding to TRE. These results were consistent with transient transfection experiments showing that co-transfection of Oct-1 expression vector significantly inhibited the ligand-dependent activation by nuclear receptors. An inhibiting effect of Oct-1 could not be observed on thyroid hormone responsiveness mediated by Gal4-TR(LBD) fusion that lacks its DNA binding domain but retains its T3 responsiveness. We concluded that Oct-1 negatively regulates nuclear receptor transcriptional activity by competitive binding to DBD of receptors with DNA elements.
Since Segil et al. (60) reported that phosphorylation of the POU homeodomain of Oct-1 correlates with DNA binding and cell cycle regulation of gene transcription, it is tempting to speculate that phosphorylation of POU domain of Oct-1 may regulate the transcriptional activity of nuclear receptors via altering the interaction with RXR. As expression and activity of Oct-1 are differentially regulated in development, differentiation, and cell cycle regulation, nuclear receptor signaling is also controlled via the Oct-1 pathway at these physiological conditions. Furthermore, recent data showed that transcriptional activity of Pit-1 is determined by a regulated balance between a co-repressor complex that contains N-CoR/SMRT, mSin3, and histone deacetylase and a co-activator complex that contains the CBP and p/CAF (61). Since these co-repressors and coactivators also interact with nuclear receptors, complex and promiscuous cross-talk between POU transcription factors and nuclear receptors can be expected. It would also be of interest to examine the role of nuclear receptors in the transcriptional activity of Oct-1 on the octamer-binding site.