Retinoblastoma-binding protein 2 (Rbp2) potentiates nuclear hormone receptor-mediated transcription.

Retinoblastoma-binding protein 2 (Rbp2) was originally identified as a retinoblastoma protein (RB) pocket domain-binding protein. Although Rbp2 has been shown to interact with RB, p107, TATA-binding protein, and T-cell oncogene rhombotin-2, the physiological function of Rbp2 remains unclear. Here we demonstrate that Rbp2 not only binds to nuclear receptors (NRs) but also enhances the transcription mediated by them. Rbp2 interacts with the DNA-binding domains of NRs and potentiates NR-mediated transcription in an AF-2-dependent manner. Both the N-terminal and C-terminal domains of Rbp2 are critical for the transactivation activity of Rbp2 on NRs. The C terminus is the NR-interacting region. In addition, RB functions in maximizing the effect of Rbp2 on the transcription by NRs. These results suggest that Rbp2 is a coregulator of NRs and define a potential role for Rbp2 in NR-mediated transcription.

Rbp2 not only interacts with NRs but also enhances their transcription activity. The presence of Rbp2 is required for the transcription mediated by NRs, and RB is important for maximal activation. Both the C terminus and N terminus of Rbp2 play an important role in its activity.

EXPERIMENTAL PROCEDURES
Materials-pGem-4Z-Rbp2 was a gift from Deborah Defeo-Jones. Estrogen receptor (ER), retinoic acid receptor (RAR), retinoid X receptor, and vitamin D receptor (VDR) were from Shen Cai Lin, and VDRE-LUC was from Victor Yu of the Institute of Molecular and Cell Biology, Singapore. GST-YKT6 was a gift from Tao Zhang of the Institute of Molecular and Cell Biology. RSV-CBP plasmid was a gift from Richard H. Goodman. Constructs of Rb and E7 were from Paramjeet Singh. 17␤-Estradiol, all-trans-retinoic acid, vitamin D 3 , and dexamethasone were purchased from Sigma. Luciferase assay reagents and TNT-coupled reticulocyte lysate system were purchased from Promega. Mouse anti-HA and anti-VSVG were purchased from Roche Molecular Biochemicals.
Cell Culture, Transfection, and Luciferase Assay-HeLa cells and C33A cells were grown in Eagle's minimal essential medium supplemented with 10% fetal bovine serum. COS-7 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. Transient transfections were done in 6-well plates for the luciferase assays and in 10-cm plates for immunoprecipitation. The cells were transfected by using LipofectAMINE (Life Technologies, Inc.) according to the manufacturer's instructions. After transfection, the medium was changed to the one with 10% fetal bovine serum or with 10% charcoal-stripped serum, and the cells were induced with ligands for 18 h if necessary. Luciferase assays were done by using the Dual Luciferase Assay System developed by Promega.
Expression and Purification of GST Fusion Proteins-Escherichia coli BL21 (DE3) carrying the pGEX fusion constructs was grown in Luria-Bertani medium at 37°C until OD 0.8 -1 and was induced with 0.1 mM isopropyl-␤-D-thiogalactopyranoside overnight at room temperature. Bacteria were harvested by centrifugation and lysed by sonication in phosphate-buffered saline. The lysates were spun at 10,000 ϫ g for 30 min at 4°C, and the supernatant was mixed with 0.5 ml of glutathione-Sepharose 4B (Amersham Pharmacia Biotech) for 2 h at 4°C. The beads were washed four times with phosphate-buffered saline. The bound proteins were either stored at 4°C or eluted with 5 volumes of 10 mM reduced glutathione in 50 mM Tris (pH 8). The pure GST fusion proteins were stored at 4 or Ϫ20°C.
Protein-Protein Interaction in Vitro-Protein-protein affinity chromatography with purified GST-ER (amino acids 251-595), GST-RAR (amino acids 143-462), GST-VDR (amino acids 66 -427), GST-GR (amino acids 465-795), GST-RB (amino acids 373-792), and GST-YKT6 or GST alone bound to glutathione-Sepharose (5 g/25 l of resin), and 5 l of [ 35 S]methionine-labeled in vitro translated Rbp2 protein was done in the presence or absence of appropriate ligands in binding buffer containing 50 mM Tris-HCl (pH 7.6), 150 mM NaCl, 0.1% Nonidet P-40, 0.5 mM EDTA, 1 mM dithiothreitol, and protease inhibitors mixtures (Roche Molecular Biochemicals) in a total volume of 500 l at 4°C for 2 h. The resin was subsequently washed four times with 1 ml of binding buffer. Bound proteins were released in SDS-polyacrylamide gel electrophoresis (PAGE) sample buffer and then analyzed by electrophoresis under denaturing conditions and autoradiography.
In some experiments, 5 g of GST-ER(A/B), GST-ER(DBD), and GST-ER(LBD) were incubated in the presence and absence of 17␤estradiol for 30 min at room temperature, followed by the addition of 1 mg of cell lysate and incubated for another 2 h at 4°C. Complexes were then centrifuged, washed three times in binding buffer, separated by SDS-PAGE, and Western-blotted with anti-Rbp2 antibodies.
Immunoprecipitation-An equal amount of each cDNA encoding HA-ER and VSVG-Rbp2 was introduced with LipofectAMINE (Life Technologies, Inc.) in COS-7 cells in 10-cm plates. Cells were induced with 17␤-estradiol for 16 h before harvesting at 48 h posttransfection and lysed with 1 ml of lysis buffer containing 150 mM NaCl, 50 mM Tris-Cl (pH 7.6), 0.2 mM EDTA, 0.1% Nonidet P-40, 1 mM dithiothreitol, and protease inhibitors mixture (Roche Molecular Biochemicals). Cell extracts were centrifuged at 14,000 for 30 min at 4°C, and the supernatant was collected and precleared with protein G-Sepharose (Amersham Pharmacia Biotech) for 4 h. Precleared extracts were incubated with 30 g of HA monoclonal antibody and 60 l of a 50% slurry of protein G-Sepharose overnight. Immunoprecipitated HA-tagged protein complexes were washed five times with lysis buffer and eluted by boiling in 2ϫ sample buffer. Eluted proteins were subjected to SDS-PAGE with a 7.5% polyacrylamide gel, and proteins were then transferred to nitrocellulose membrane (Amersham Pharmacia Biotech) and were probed either with anti-VSVG or anti-ER (Santa Cruz Biotechnology) and then with horseradish peroxidase-conjugated goat anti-mouse or anti-rabbit secondary antibody (Pierce). Binding was detected by supersignal chemiluminescent substrate (Pierce).
Binding of Rbp2 to the ER-DNA Complex-The binding assay was performed as described (35). Briefly, 30 pmol of 45-base pair doublestranded DNA fragment containing two copies of ERE labeled with biotin at the 5Ј end of one strand was conjugated to streptavidin beads (Pierce). In vitro translated 35 S-labeled ER was bound to the EREconjugated beads in binding buffer containing 20 mM Hepes (pH 7.9), 50 mM KCl, 10% glycerol, and 0.5 mM EDTA in the presence or absence of ligand at room temperature for 1 h. The beads were subsequently washed with the binding buffer and incubated with in vitro translated 35 S-labeled Rbp2 with the same buffer in the presence and absence of ligand at room temperature for 1 h. The beads were washed, and the bound proteins were resolved in SDS-PAGE and subsequently subjected to autoradiography.

Rbp2
Interacts with NR-Visual examination of Rbp2 amino acid sequence led to the identification of two LXXLL motifs (residues 725-729 and 945-949). To examine the potential association of Rbp2 with NRs, ER (amino acids 251-595) was fused to GST to generate GST-ER and incubated with 35 Slabeled Rbp2 in the presence or absence of 17␤-estradiol. 35 S-Labeled Rbp2 was retained on GST-ER beads in a ligandindependent manner (Fig. 1A, lanes 4 and 5) but not on GST or GST-YKT6 beads (40), a protein involved in vesicle transport (Fig. 1A, lanes 3 and 12). 35 S-Labeled Rbp2 was retained by the GST-RB pocket (amino acids 373-792) (Fig. 1A, lane 13), serv-ing as a positive control. GST fusion proteins containing RAR (amino acids 143-462), GR (amino acids 465-795), and VDR (amino acids 66 -427) were also produced and tested for interaction with 35 S-labeled Rbp2. All these have the ability to retain Rbp2 in the presence and absence of ligands, although with different efficiencies (Fig. 1A, lanes 6 -11). These results suggest that Rbp2 may directly interact with NRs in vitro in a ligand-independent manner.
To determine whether Rbp2 and NRs can physically interact in mammalian cells, COS-7 cells were transfected with HAtagged ER and VSVG-tagged Rbp2 expression vectors. Cell lysates were first immunoprecipitated with anti-HA antibody, and the bound proteins were subsequently analyzed by immunoblotting with anti-VSVG antibody to detect Rbp2 that was coimmunoprecipitated. Rbp2 protein was clearly detected in the immunoprecipitates pulled down by anti-HA antibody, which recognized HA-tagged ER, from cells transfected with both ER and Rbp2, but not in the nontransfected cells (Fig. 1B, upper panel). The interaction between Rbp2 and ER is again ligand-independent. The membrane blot was stripped and immunoblotted again with anti-ER antibody to show the presence of ER protein in the anti-HA immunoprecipitates pulled down by anti-HA antibody (Fig. 1B, bottom panel). These results indicate that Rbp2 indeed interacts with ER in the cells.
The LXXLL Motifs of Rbp2 Are Not Required for Its Interaction with NRs-Rbp2 contains two LXXLL motifs in the central region (residues 725-729 as NR box 1 and residues 945-949 as NR box 2). LXXLL motifs are found in a large variety of LBD/ AF2-interacting proteins (39). We next examined whether these motifs are responsible for the observed interaction of Rbp2 with NRs. Various Rbp2 mutants that contain mutations in the LXXLL motifs were tested. The LXXLL motifs were altered to LXXAA, a change influencing the binding of coactivators to receptors. Three LXXAA mutants of Rbp2 were constructed as shown in Fig. 2A. Rbp2M1 has an LXXAA mutation in the first NR box (amino acids 725-729). Rbp2M2 contains a LXXAA mutation in the second NR box (amino acids 945-949), whereas Rbp2M12 has both the NR boxes mutated. These mutants were compared with wild type Rbp2 for interaction with ER in in vitro GST pull-down experiments. GST-ER interacts with Rbp2 and each of the Rbp2 mutants with comparable efficiencies (Fig.  2B). These results indicate that the LXXLL motifs of Rbp2 are not required for its interaction with NRs.
Rbp2 Interacts with the DBDs of NRs-Since the LXXLL motifs of Rbp2 are not required for binding to NRs (Fig. 2B), it is possible that Rbp2 binds to regions of NR other than LBD. To delineate the Rbp2-binding domain of ER, different regions of ER were expressed as fusion proteins with GST to produce GST-ER(A/B), GST-ER(DBD), and GST-ER(LBD), as shown in Fig. 3A. These various fusion proteins were used to examine their ability to retain Rbp2 from HeLa cell lysates. Proteins retained by these beads were analyzed by immunoblotting. As shown, Rbp2 was bound by GST-ER(DBD) and GST-RB (positive control) affinity beads (Fig. 3B, lane 4 and 7) but not by GST, GST-ER(A/B), or GST-ER(LBD) (Fig. 3B, lanes 2, 3, 5,  and 6). Similarly, the DNA-binding domain of RAR binds Rbp2 (data not shown). These results indicate that Rbp2 interacts with the DBDs of NRs.
Since Rbp2 interacts with the DBDs of NRs, it is important to examine whether Rbp2 could form a ternary complex with NR bound to DNA. To test this possibility, a binding assay using ER that had been bound to the estrogen-responsive element (ERE) conjugated to agarose beads was performed. Fig.  3C shows that in vitro translated 35 S-labeled ER bound to ERE beads in a dose-dependent manner (lanes 5-10). Similar to the reported ligand-independent interaction of RAR and retinoid X receptor with RARE beads in vitro (35), interaction of ER with immobilized ERE was independent of ligand (lanes 5-10). The ER-ERE beads were then incubated with the indicated amount of in vitro translated 35 S-labeled Rbp2 in the presence or absence of ligand. The bound proteins were detected by autoradiography. Rbp2 was retained together with ER on the ERE beads in a dose-dependent manner (Fig. 3C, lanes [5][6][7][8][9][10]. No direct interaction of Rbp2 with ERE beads was observed (lane 4), suggesting that observed retention of Rbp2 was mediated by ER. RAR or Rbp2 was not retained by ERE beads (lanes [11][12][13].
These results indicate that Rbp2 could form a ternary complex with ER bound to DNA.
The C-terminal Domain of Rbp2 Is Critical for Binding to NRs-Since the LXXLL motifs of Rbp2 are not required for binding to NRs, a series of deletion constructs of Rbp2, tagged with VSVG, were generated to delineate the NR-binding domain (Fig. 4A). GST-ER(DBD) was used to pull down interacting proteins from COS-7 cells transfected with the various deletion constructs of Rbp2. The deletion constructs were expressed efficiently in COS-7 cells as shown in Fig. 4B, upper panel. The binding results shown in Fig. 4B, lower panel, established that the C-terminal region (amino acids 1379 -1722) of Rbp2 is able to associate with ER(DBD) (Fig. 4B, lower panel,  lane 5). To map further the receptor interaction domain, the C-terminal region (amino acids 1379 -1722) was subcloned into two smaller fragments, and each retained the ability to interact with ER independently, although with apparently less efficiency (Fig. 4B, lanes 6 and 7). We conclude that Rbp2 has at least two separable receptor interaction domains located at its C-terminal region, and these two regions act together for optimal interaction with NRs.
Rbp2 Is Involved in the Transactivation Function of NRs in Vivo-The ability of Rbp2 to interact with NRs prompted us to test whether Rbp2 could stimulate transcription mediated by NRs. HeLa cells were cotransfected with expression vectors for ER, Rbp2, and ERE-LUC reporter in the presence or absence of 17␤-estradiol, and luciferase activity was measured. Fig. 5A showed that Rbp2 significantly increased ER-mediated induction of the ERE-LUC reporter gene activity. To determine whether Rbp2 stimulates transcription activation of other NRs, similar transfection experiments were carried out in HeLa cells. As shown, the transactivation activities of RAR (Fig. 5B), VDR, and GR (Fig. 5D) activity are also increased by Rbp2 at different degrees. These results indicate that Rbp2 is an NR coactivator.
To investigate the requirement of Rbp2 for NR-mediated transcription, we used an antisense approach to determine whether depletion of Rbp2 would prevent NRs from activating NR-dependent transcription. The antisense construct of Rbp2(Rbp2AS) was generated and shown to reduce effectively the amount of Rbp2 protein in cells by up to 90% (Fig. 5C, upper  panel, lane 3). The antisense construct of Rbp2 significantly inhibited the transcriptional functions of ER and RAR, either in the presence or absence of ligands (Fig. 5, A and B). The inhibition of transcription by Rbp2AS on NRs does not seem to be due to a general effect on transcription because transcription mediated by SV40 early promoter was not significantly affected by the expression of Rbp2 antisense (Fig. 5E). HeLa cells were transfected with SV40-LUC reporter plasmid, along with the expression vectors for Rbp2 or Rbp2AS, and luciferase activity was measured. Rbp2 enhanced SV40-LUC activity marginally (lanes 2 and 3), but Rbp2AS did not show any inhibitory effect on SV40-LUC activity (lanes 4 and 5). The suppression of transcription activity of ER by Rbp2AS is probably due to an inhibitory effect of Rbp2AS on effectors that may play a role in the transcription in the absence of ligand. These results demonstrate that Rbp2 is required for NR activation.
RB Is Necessary for Maximal Potentiation of ER-mediated Transcription by Rbp2-Since Rbp2 was originally identified as an RB pocket-binding protein, it is possible that the function of Rbp2 may depend on the presence of RB. To determine whether RB plays a role in Rbp2-enhanced NR transactivation, an RB-negative cell line, C33A, was chosen to test the requirement of RB on Rbp2 function. Fig. 6 showed that without the presence of RB, Rbp2 slightly potentiated the transcription activity of ER. Similarly, RB alone also potentiated ER-mediated transcription slightly. With the presence of RB and Rbp2 together, the effect on ER-mediated transcription was dra-  35 S-labeled ER or RAR were bound to ERE-conjugated agarose beads in the presence (ϩ) or absence (Ϫ) of ligands 17␤-estradiol or all-trans-retinoid acid (ATRA). The beads were washed and subsequently incubated with the indicated amount of in vitro translated, 35 S-labeled Rbp2 in the presence and absence of ligands. The bound proteins were detected by autoradiography. The in vitro translated, 35 S-labeled products of RAR, Rbp2, and ER used for binding were loaded in lanes 1-3. Extracts added to the ERE-conjugated beads without receptors are shown in lane 4. matically enhanced. E7 protein, a viral protein shown to interact with RB and inhibit its function, was used to demonstrate further the requirement of RB in Rbp2 function. The synergistic effects of RB and Rbp2 on ER function, as well as the ones by RB or Rbp2 alone, were significantly reduced in the presence of E7 protein (Fig. 6). These data indicate that RB is required for the maximal potentiation of NR-mediated transcription by Rbp2.
N-terminal and C-terminal Domains of Rbp2 Are Critical for Potentiating NR-mediated Transcription-There is an ARID and three PHD motifs present in Rbp2, and these two types of motifs have been functionally defined to be involved in proteinprotein interactions (41,42). To characterize the importance of these motifs in the NR-mediated transcription by Rbp2, different deletion constructs of Rbp2, as shown in Fig. 7A, were generated to test for their abilities to activate ER-mediated transcription. All these constructs expressed the respective mutants efficiently in transfected cells (Fig. 7B). HeLa cells were transfected with ER, ERE-LUC, and the indicated deletion constructs individually in the presence or absence of ligand, and luciferase activities were measured. Fig. 7C showed that full-length Rbp2 induced ER-mediated transcription as shown previously (Fig. 5A). Surprisingly, once the N-terminal domain (amino acids 1-324) was deleted, the transactivation activity of Rbp2 on ER-mediated transcription was reduced to half that of full-length Rbp2 (Fig. 7C, compare lanes 3-6 with  lane 2). Similarly, once the C-terminal part of Rbp2 was deleted (amino acids 1320 -1722), the transactivation activity of Rbp2 was lost (Fig. 7C, compare lanes 7-10 with lane 2). This is consistent with the above data that the C-terminal domain of Rbp2 is the NR interaction domain. The requirement of the N-terminal domain of Rbp2 suggests that interaction with NR alone is not enough for the observed effect of Rbp2 on NRmediated transcription. When Rbp2 was expressed with an increasing amount of either Rbp2(dN4) or Rbp2(dC4), the transactivation activities of Rbp2 on ER were reduced (Fig. 7C,  lanes 12 and 13), indicating that these deletion mutants can inhibit normal activity of full-length Rbp2 in a dominant-negative fashion. These results suggest that the N-terminal and C-terminal regions of Rbp2 are both important for the activity of Rbp2 on NRs.
The Potentiation of NR-mediated Transcription by Rbp2 Is Dependent on the AF-2 Function of NR-Rbp2 interacts with DBDs of NRs in a ligand-independent manner, and the trans- activation of Rbp2 on NR-mediated transcription is dependent on ligands. To confirm further that Rbp2 exerts its effect on the ligand-dependent NRs, the transactivation activity of Rbp2 on different deletion mutants of ER was performed. The different forms of ER were generated as shown in Fig. 8A. As expected, the full-length ER induced an ER-responsive reporter in a ligand-dependent manner (Fig. 8B, lane 1). The transcriptional activity of ER was enhanced further by coexpression of Rbp2 (Fig. 8B, lane 2). In the absence of ligand, Rbp2 has no significant effect on ER-mediated transcription. Interestingly, Rbp2 has a stronger transactivation activity on ER(CDEF), which is devoid of the N-terminal region A/B (Fig. 8B, lanes 3 and 4). In   FIG. 5. Rbp2 potentiates NR-mediated transcription. A, Rbp2 enhances ER-mediated transcription. HeLa cells were transfected with ERE-LUC reporter plasmid and ER expression construct (ER), along with a Rbp2 expression construct (Rbp2) or the pCIneo empty expression vector (vector), or Rbp2 and a Rbp2 antisense construct (Rbp2 AS) in the presence and absence of 17␤-estradiol as depicted. Renilla luciferase pRL-TK expression plasmid was used as a control for transfection efficiency. The dual luciferase assay system from Promega was used to measure luciferase activity. Reporter gene activities are expressed as fold induction, relative to that achieved with ⌭RE-LUC alone in the absence of 17␤-estradiol. B, Rbp2 enhances endogenous RAR-mediated transcription. HeLa cells were transfected with ␤RARE-LUC reporter plasmid, along with Rbp2 or vector or Rbp2 and Rbp2AS in the presence and absence of all-trans-retinoid acids as depicted. Luciferase activities were measured as in A. C, Rbp2AS effectively represses the expression of Rbp2 in vivo. COS-7 cells were transfected with vector, Rbp2, or Rbp2 along with Rbp2AS. Lysate was prepared, and an equal amount of protein was subjected to SDS-PAGE and immunoblotted with rabbit anti-Rbp2 antibody (upper panel). To confirm equal loading, the filter was re-probed with anti-␤-actin antibody (bottom panel). D, Rbp2 activates GR-and VDR-dependent transcription. Left panel, HeLa cells were transfected with GRE-LUC reporter plasmid and GR expression construct, along with Rbp2 or vector in the presence and absence of dexamethasone. Right panel, HeLa cells were transfected with VDRE-LUC reporter plasmid, along with Rbp2 or vector in the presence or absence of vitamin D 3 . Luciferase activities were measured as in A. E, the effect of Rbp2 on SV40 promoter. HeLa cells were transfected with SV40-LUC, along with increasing amounts of Rbp2 or Rbp2AS, and luciferase activity was measured as in A.
contrast, neither the AF1 activity encompassing the ER regions A and B(ER(ABC)), and the DBD(ER(C)) nor the AF-2 mutant (ER(CDEF**)) of ER were stimulated by coexpression of Rbp2 (Fig. 8B, lanes 5-10). These results suggest that Rbp2 exerts its effect on the AF-2 function of NR, and this explains why the NR-mediated transcription by Rbp2 is ligand-dependent.
Since the N-terminal and C-terminal domains of Rbp2 behaved in a dominant-negative fashion on the transcription mediated by NRs (Fig. 7C) and Rbp2 potentiation on NR transactivation is AF-2-dependent, it is possible that these two domains of Rbp2 could exert a dominant-negative effect on other NR coactivators. To examine this possibility, we transfected HeLa cells with ER and ERE-LUC along with either Rbp2, CBP, Rbp2(dC4), Rbp2(dN4), or Rbp2 and CBP, or CBP with increasing amount of either Rbp2(dN4) or Rbp2(dC4), and luciferase activities were measured (Fig. 8) With the presence of both Rbp2 and CBP, the transcription activity of ER is increased dramatically (lane 4). When CBP was expressed with the increasing amounts of either Rbp2(dC4) (lanes 9 and 10) or Rbp2(dN4) (lanes 6 and 7), the transactivation activity of CBP on ER was reduced, indicating that the N-terminal and C-terminal domains of Rbp2 could also exert their dominant-negative effect on other NR coactivators and that Rbp2 could work together with other coactivators.

DISCUSSION
In this study, we have demonstrated that Rbp2 interacts with NRs and displays significant transcription activation activity toward them. The results also suggest that the presence of Rbp2 is required for the optimal transcription mediated by NRs. In addition, RB is necessary for the maximal potentiation of the transcription mediated by NRs. Molecular dissection of Rbp2 revealed that both the N terminus and C terminus of Rbp2 are necessary for its transcription activation activity. Deletion of either one of them resulted in a suppression of the transcription activity of Rbp2. These observations suggest that one possible function of Rbp2 is to act as a coregulator of NRs.
There are two LXXLL motifs (NR box 1 and NR box 2) present in Rbp2. As LXXLL motifs have been implicated in interaction of several other proteins with NRs, its presence motivated us to investigate the role of Rbp2 in NR function. Our studies suggest that both LXXLL motifs are not important for Rbp2 to interact with NRs. The Phd program (39) shows that the secondary structure of NR box 2, but not NR box 1, is predicted to be ␣-helical. Mutation of either NR box 1, NR box 2, or both did not seem to affect the ability of Rbp2 to interact with NRs as well as their ability to potentiate the transcription mediated by NRs (data not shown). Heery et al. (39,43) shows that the LXXLL motif is ␣-helical. They have defined a minimal "core" LXXLL motif as an 8-amino acid sequence spanning positions Ϫ2 to 6 relative to the primary conserved leucine residues, with a hydrophobic residue at position Ϫ1 relative to the first conserved leucine, and a nonhydrophobic residue at position ϩ2 showing high affinity for steroid and retinoid receptors. Neither NR box 1 nor NR box 2 shows the abovementioned characteristics, consistent with experiments demonstrating that the LXXLL motifs of Rbp2 are not involved in interaction with NRs. Wild type Rbp2 and all the NR box mutants of Rbp2 interact with ER equally well in the absence of ligand, whereas the conserved functional LXXLL motif has been showed to bind to NR only in the presence of ligand. These observations suggest that the LXXLL motifs of Rbp2 are not the NR interaction domain of Rbp2.
Since Rbp2 interacts with NRs in the absence of ligands and LXXLL motifs are not important for ER-Rbp2 interaction, this prompted us to map the region of both ER and Rbp2 responsible for their interaction. Different regions of ER were used to delineate the Rbp2 interaction domain. The DBD domain, but not A/B or LBD, was shown to interact with Rbp2. The recruitment of Rbp2 to ER complexed to ERE-conjugated beads clearly indicates that Rbp2 could form a ternary complex with NR bound to DNA. The hinge region, C-terminal to DBD, may also contribute to its interaction with Rbp2, as the DBD alone was not sufficient to achieve optimal interaction. Although most of the coactivators that were identified interact with the LBDs of NRs, there are proteins capable of recognizing DBDs of NRs. For example, PCAF (35), MBP1 of Drosophila (44), and SNURF of mammalian (45) all interact with the DBDs of NRs and act as cofactors of NRs.
By using ER-DBD fused to GST as the basis for affinity chromatography, we have established that the C-terminal 378 residues of Rbp2 are essential and sufficient for interacting with NRs. Rbp2-ER interaction occurs in a ligand-independent manner, but Rbp2 enhances the transcription mediated by NRs in a ligand-and AF2-dependent manner. Both the N terminus and C terminus of Rbp2 are shown to be critical for the transactivation. Deletion of either one of them resulted in an inhibition of the transcription mediated by ER. The C terminus is the region that interacts with NRs and is expected to be essential for the Rbp2's enhancement of NR-dependent transcription. There is a DNA-binding domain termed ARID (for AT-rich interaction domain), first found in Drosophila dead ringer (dri) (46) and a PHD motif in the N terminus of Rbp2. The ϳ100residue ARID sequence is present in a series of proteins strongly implicated in the regulation of cell growth, development, and tissue-specific gene expression (41). Although ARID was originally found to bind preferentially to AT-rich sites, there are other proteins showing no sequence preference in its DNA binding activity (47). Our preliminary results suggest that Rbp2 binds to nonsequence-specific native DNA cellulose prepared from sheared thymus DNA. This demonstrates that the ARID family proteins may be involved in a wider range of DNA interaction.
PHD motifs are zinc finger-like motifs, speculated to mediate protein-protein interactions. These motifs are commonly found in transcription factors and are implicated in chromatin-mediated gene regulation (42). Thus, the N-terminal Rbp2 deletions were expected to have lost the ability to regulate some Rbp2associated functions in cells. This is manifested when increasing amounts of the N-terminal domain of Rbp2 were coexpressed with wild type Rbp2 in the transcription assays. The more the mutant was expressed, the less transactivation activ-FIG. 6. RB is necessary for maximal potentiation of NR-mediated transcription by Rbp2. RB-negative C33A cells were transfected with ERE-LUC reporter plasmid and ER, along with the indicated expression constructs or vector in the presence and absence of 17␤-estradiol. Luciferase activities were measured as described in Fig. 5. ity of Rbp2 was observed. This N-terminal domain acted in a dominant-negative fashion by probably competing away some of the factors that are crucial for the transcription recruited by Rbp2. Although the detailed mechanism remains to be established, we currently favor the model that the N-terminal domain of Rbp2 is involved in interaction with factors that are important for transcriptional activities of NRs.
The DBD of NR does not possess transcription activation function like AF1 and AF2. When tethered to DBD of Gal4, Rbp2 failed to stimulate transcription from promoters containing Gal4-binding sites (data not shown), implying that the transactivation by Rbp2 is not mediated by Rbp2 alone. It probably involved other protein-protein interactions. Rbp2 is a large molecule with several PHD and RING finger motifs. It is thus possible that the transactivation by Rbp2 is mediated by other proteins that are either recruited by Rbp2 or associated with Rpb2. In this way, Rbp2 may function as a possible bridging factor participating in the coordination of activities of transcriptional signal from sequence-specific upstream factors and RNA polymerase II-basal transcription machinery. The C-terminal domain of Rbp2 also acted in a dominant-negative manner. TBP has been shown to interact the C terminus of Rbp2 Constructs for the expression of fulllength and various deletion mutants of Rbp2 were transiently transfected into COS-7 cells. Cell lysates were prepared, and aliquots containing 100 g of proteins were subjected to SDS-PAGE and immunoblotted with anti-VSVG antibody. The lane numbers correspond to the numbers for the mutants shown in A. C, N-terminal and C-terminal domains of Rbp2 are important for the enhancement of ER-dependent transcription by Rbp2. To check the effect of the various deletion mutants, HeLa cells were transfected with ERE-LUC reporter plasmid and ER, along with the indicated constructs or vector in the presence and absence of 17␤-estradiol (column clusters [1][2][3][4][5][6][7][8][9][10][11]. To check the effect of the N-terminal fragment Rbp2(dC4) and the C-terminal fragment Rbp2(dN4), the cells were also transfected with ERE-LUC, ER, along with Rbp2, and increasing amounts of either Rbp2(dN4) or Rbp2(dC4) in the presence of 17␤-estradiol (column clusters 12 and 13). Luciferase activities were measured as in Fig. 5. (15). It is possible that the transactivation achieved by Rbp2 is through its ability to interact with NRs and TBP, via its Cterminal domain, and other factors via its N-terminal domain.
Since Rbp2 is an RB-binding protein, the observed effect of Rbp2 on NR might be dependent on RB, which has been shown to potentiate GR activity (48). We determined whether the loss of RB affected the activity of Rbp2 on the transcription mediated by NR. In the RB-negative cell line, C33A, Rbp2 enhances transcription mediated by ER. This suggests that RB may not be required for the basal transactivation of Rbp2 on ER. In the presence of RB, Rbp2 synergistically enhances the transcription mediated by ER, and this effect is suppressed by the presence of an RB inhibitory protein E7. This implies that RB is required for the optimal activity of Rbp2 on ER. Rbp2 was originally identified as an RB pocket domain-binding protein (12). Besides binding to the pocket domain of RB through its LXCXE motif (amino acids 1373-1377), Rbp2 also interacts with RB pocket domain through its non-T/E1A-binding site (amino acids 1547-1558) (15). RB has been shown to potentiate glucocorticoid receptor (GR)-mediated transcription through interaction of its pocket domain with the transcription activator hBRM (49), and the N-terminal domain of RB is essential for the potentiation of transcription by GR (49). hBRM was shown to interact with the ligand-binding domain of ER (50). Since RB interacts with Rbp2, NR (through hBRM), and potentiates NR-mediated transcription with Rbp2 or hBRM (48) synergistically, it is possible that Rb, Rbp2, and NR could form a complex in activating the transcription by NR maximally.
In summary, we have provided evidence suggesting that one possible role of Rbp2 is to act as a coregulator of NR-mediated transcription. We showed that Rbp2 interacts with the DBDs of NRs and enhances transcription mediated by them in an AF-2-dependent manner. The C terminus and N terminus of Rbp2 are both critical for its participation in enhancement of NRmediated transcription. In addition, we have shown that RB is critical for the maximal activation of NR by Rbp2. These results imply a scenario in which Rbp2 acts as a bridging factor to recruit and/or coordinate multiple protein-protein interactions that are important for NR-mediated transcription.  1-5, and 8).
To check the effect of the N-terminal fragment Rbp2(dC4) and the C-terminal fragment Rbp2(dN4) on the activity of NR coactivator CBP, the cells were transfected with ERE-LUC, ER, along with CBP, and increasing amounts of either Rbp2(dC4) or Rbp2(dN4) in the presence or absence of 17␤-estradiol (column clusters 6 and 7, and 9 and 10).