Estrogen Receptor β (ERβ1) Transactivation Is Differentially Modulated by the Transcriptional Coregulator Tip60 in a cis-Acting Element-dependent Manner*

Background: The interactions between estrogen receptor β (ERβ1) and different coregulators are responsible for the distinct functions of ERβ1. Results: Tip60 enhances ERβ1 transactivation at the AP-1 site but inhibits it at ERE sites. Conclusion: Tip60 is either a coactivator or a corepressor for ERβ1 in a regulatory element-dependent manner. Significance: Tip60 is the first multifaceted coregulator of the transcriptional activity of ERβ1 that has been identified. Estrogen receptor (ER) β1 and ERα have overlapping and distinct functions despite their common use of estradiol as the physiological ligand. These attributes are explained in part by their differential utilization of coregulators and ligands. Although Tip60 has been shown to interact with both receptors, its regulatory role in ERβ1 transactivation has not been defined. In this study, we found that Tip60 enhances transactivation of ERβ1 at the AP-1 site but suppresses its transcriptional activity at the estrogen-response element (ERE) site in an estradiol-independent manner. However, different estrogenic compounds can modify the Tip60 action. The corepressor activity of Tip60 at the ERE site is abolished by diarylpropionitrile, genistein, equol, and bisphenol A, whereas its coactivation at the AP-1 site is augmented by fulvestrant (ICI 182,780). GRIP1 is an important tethering mediator for ERs at the AP-1 site. We found that coexpression of GRIP1 synergizes the action of Tip60. Although Tip60 is a known acetyltransferase, it is unable to acetylate ERβ1, and its coregulatory functions are independent of its acetylation activity. In addition, we showed the co-occupancy of ERβ1 and Tip60 at ERE and AP-1 sites of ERβ1 target genes. Tip60 differentially regulates the endogenous expression of the target genes by modulating the binding of ERβ1 to the cis-regulatory regions. Thus, we have identified Tip60 as the first dual-function coregulator of ERβ1.

associated proteins onto other transcription factors (4,20). For example, AP-1 recruits CBP and p300, which bind to p160 coactivators. ERs then tether onto the transcriptional complex of AP-1 through the physical interaction with p160 coactivators (4,20). In short, the diverse actions of a nuclear receptor such as ER␤1 could depend largely on its interacting coregulators.
This study investigated the biological function of Tip60 on ER␤1 transactivation, particularly at the various cis-regulatory sequences and/or in the presence of different types of ligand. The dependence of histone acetyltransferase (HAT) domain activity in Tip60 was evaluated with a HAT domain mutant. Its interactions with other common coregulators such as SRC-1 and GRIP1 were determined. Moreover, the co-occupancy of ER␤1 and Tip60 at cis-regulatory elements of endogenous ER␤1 target genes and their differential regulation by Tip60 were evaluated. Here, we showed that Tip60 is a unique dualfunction coregulator of ER␤1 in a cis-acting element-dependent manner.

EXPERIMENTAL PROCEDURES
Cell Culture Conditions-HEK293 and DU-145 cells were grown in Eagle's minimal essential medium supplemented with 10% fetal bovine serum, L-glutamine. PC-3 cells were grown in F-12K medium supplemented with 10% fetal bovine serum (ATCC, Manassas, VA). All cells were grown in 1% penicillin/ streptomycin. The phenol red-free DMEM was supplemented with 10% charcoal-stripped fetal bovine serum (CSS) prior to the addition of ligands in experiments. Cells were grown at 37°C and 5% CO 2 .
Plasmids, siRNAs, and Recombinant Protein-Full-length ER␤1 and ER␣ were subcloned into pGBKT7 vector, whereas Tip60 was cloned into pACT2 vector (Clontech). ER␤1 and Tip60 were also cloned into pcDNA-HisMax (Invitrogen) or subcloned into pENTR entry vector (Invitrogen) and then transferred into destination vector pDEST40 through gateway cloning (Invitrogen). In addition, full-length ER␤1 and ER␣ were subcloned into the pGBKT7 vector, whereas Tip60 was cloned into the pACT2 vector (Clontech). SRC-1 and GRIP-1, gifts from Dr. Nancy Weigel (Baylor College of Medicine, Houston), were cloned into pcDNA3.1. ONTARGETplus SMARTpool 4 siRNAs specific to Tip60 were used for gene knockdown. ONTARGETplus nontargeting siRNA was used as the negative control (Thermo Scientific Dharmacon). Recombinant ER␤1 protein was purchased from Thermo Scientific Pierce.
To generate different domain-deleted ER␤1 constructs, a c-Myc tag was first added by PCR to the N terminus of the full-length ER␤1 coding sequence, which was cloned into pDEST40. We generated different domain-deleted ER␤1 by performing PCR with different sets of primers (Table 1) and using ER␤1-pDEST40 as the template.

TABLE 1 Primers used in the experiments of domain-deletion study of ER␤1 and site-directed mutagenesis of Tip60
F is forward, and R is reverse.

Domain deletion of ER␤1
Construction of ER␤1 Stably Expressed Cell Lines-Stably expressed cell lines were constructed according to the published data (33). Full-length ER␤1 or LacZ (negative control) was subcloned, respectively, into pLenti6 lentiviral vector by Multisite Gateway Cloning (Invitrogen) and transfected into 293FT for production of lentivirus. The titer of lentivirus was measured, and the multiplicity of infection of PC-3 cells was determined. Lentivirus-infected PC-3 cells were selected with blasticidin (10 g/ml) for 3 weeks. Quantitative reverse transcription (RT)-PCR, Western blot, and ␤-galactosidase assay were performed to confirm the stable expression of ER␤5 or LacZ.
In Vitro Coimmunoprecipitation (Co-IP)-T7 promoter and HA tag were added to the N terminus of the coding sequence of Tip60 by PCR. pGBKT7 vector containing the full-length of ER␤1, ER␣, and purified PCR product of Tip60 were, respectively, translated in vitro by the TNT T7-reticulocyte system (Promega, Fitchburg, WI) labeled with EasyTag EXPRESS 35 S protein labeling mix (PerkinElmer Life Sciences). Tip60 (10 l) and ER␤1 or ER␣ (each 10 l) proteins were mixed at 4°C for 1 h. Lysates were incubated with 20 l of EZview red anti-HA affinity gel (Sigma) at 4°C overnight with agitation. The samples were subjected to SDS-PAGE. The dried gel was exposed to x-ray film for 72 h, and an intensifying screen (Eastman Kodak) was used for signal enhancement. Films were scanned using the Odyssey Infrared Imaging System (LiCor Bioscience, Lincoln, NE).
Yeast Two-hybrid Assays-ER␣-or ER␤1-pGBKT7 and Tip60-pACT2 were cotransformed into yeast strain Y187 through the polyethylene glycol/lithium acetate method with the use of the Yeastmaker yeast transformation system (Clontech). Procedures followed the manufacturer's protocol. The transformed yeast cells were grown on quadruple dropout (SD/ϪAdeϪHisϪLeuϪTrp) (QDO) agar with X-␣-galactosidase until the appearance of blue colonies.
Ni-NTA Purification of His-tagged Proteins-HEK293 cells were transfected with ER␤1 and Tip60. After a 24-h transfection, medium was added with 10 nM E 2 . Cells were lysed in lysis buffer (50 mM NaH 2 PO 4 , 300 mM NaCl, 10 mM imidazole, 0.1% Tween 20) containing complete EDTA-free protease inhibitor mixture (Calbiochem) followed by sonication. About 1 mg of total lysate was incubated with 20 l of Ni-NTA-agarose beads (Qiagen, Valencia, CA) at 4°C overnight. Washing and elution procedures followed the manufacturer's protocol. The samples were subjected to Western blot analysis. IRDye secondary antibody was used to detect the protein bands, and the Odyssey Infrared Imaging System (LiCor Bioscience) was used to detect the signals.
Mammalian Co-IP-HEK293 cells transfected with plasmids or ER␤1 stably expressed PC-3 cells were used. Medium was added with or without 10 nM E 2 as indicated. Cells were lysed in M-PER lysis buffer (Thermo Scientific Pierce) containing protease inhibitor mixture. Lysates were incubated with 2 g of Tip60 or ER␤1 antibody at 4°C overnight and then with protein G Dynabeads (Invitrogen) at room temperature for 1.5 h. The immunoprecipitates were subjected to Western blot analysis.
In the domain-deletion study, full-length and domain-deleted ER␤1 constructs were immunoprecipitated by EZview red anti-c-Myc affinity gel (Sigma). IgG XP isotype was used as negative control (Cell Signaling Technology).
Immunofluorescence Staining-HEK293 cells or ER␤1 stably expressed PC-3 cells were seeded on a round coverslip. HEK293 cells were transfected with ER␤1 and Tip60. Cells were fixed in 10% formalin and permeabilized with 1% Nonidet P-40. Normal chicken serum was used for blocking. Cells were incubated with rabbit ER␤ (H150) and goat Tip60 (N-17) at room temperature for 1 h followed by incubation with different fluorescenttagged secondary antibodies. DAPI (Sigma) was used for nuclear counterstaining. Prolong R Gold anti-fade reagent (Invitrogen) was used for signal enhancement. Fluorescent images were obtained with an Axiovert 200 M fluorescent microscope equipped with an AxioCam MRm camera and Axiovision 4.8 software (Carl Zeiss, Oberkochen, Germany).
Site-directed Mutagenesis-The acetylation-deficient mutant of Tip60, Tip60⌬HAT (Q377E/G380E), was generated with the use of the Stratagene QuikChange lightning site-directed mutagenesis kit (Agilent Technologies, Santa Clara, CA) as described in the protocol. Primers for mutagenesis were designed through the QuikChange primer design program (Agilent Technologies) ( Table 1). In brief, the mutant strand synthesis was done by PCR, and products were treated with the restriction endonuclease DpnI to digest the parental DNA. The mutated single-stranded DNA was converted to the duplex form in vivo through bacterial transformation. Plasmids were extracted and sequenced to confirm the mutations.
In Vitro and in Vivo Acetylation Assay-For the in vitro acetylation assay, HEK293 cells were transfected with either wild-type Tip60 (Tip60WT) or Tip60⌬HAT. Cells were treated with 3 M TSA and 5 mM nicotinamide for 6 h. Recombinant Tip60 was purified on the Ni-NTA column as described above, and the lysis and wash buffers were added with 1 M TSA and 5 mM nicotinamide, which are inhibitors of different deacetylase families. The Tip60-bound Ni-NTA column was resuspended in HAT buffer (50 mM Tris-HCl, pH 8, 10% glycerol, 100 M EDTA, 1 mM DTT, 1 mM PMSF, 10 mM sodium butyrate, 5 mM nicotinamide) with 500 M acetyl-CoA and 500 g of recombinant ER␤1. The mixture was incubated at 30°C for 1 h. Lysates were subjected to Western blot analysis.
For the in vivo acetylation assay, HEK293 cells were transfected with ER␤1, Tip60WT, or Tip60⌬HAT. Cells were treated with 3 M TSA and 5 mM nicotinamide for 6 h. Immunoprecipitation was performed with ER␤1 or Tip60 antibody, and the lysis and wash buffers were added with 1 M TSA and 5 mM nicotinamide. Lysates were subjected to Western blot analysis.
Luciferase Reporter Assay-Different luciferase reporter plasmids were used. The pt109-ERE3-Luc carrying 3ϫ vitellogenin ERE was provided by Dr. Craig Jordan (Fox Chase Cancer Center, Philadelphia). The pAP-1-Luc was purchased from Clontech. The C3 ERE-Luc, c-Fos ERE-Luc, progesterone receptor (PR) ERE-Luc, and pS2 ERE-Luc reporters were gifts from Dr. Carolyn Klinge (University of Louisville, Louisville, KY). NFB-Luc and pSp1 3 -Luc were provided by Dr. Francis Chan (University of Massachusetts Medical School, Worcester, MA). HEK293 cells were seeded on 24-well plates at 2.8 ϫ 10 5 in phenol red-free medium supplemented with 10% charcoal-stripped serum (CSS). Expression plasmids of ER␤1, GFP, or Tip60, together with luciferase reporter plasmids and ␤-galactosidase, were transiently transfected into cells. Different ligands, such as E 2 , DPN, GEN, EQ, DAI, API, TAM, RAL, ICI, and BPA were added to the medium after a 24-h transfection. Transactivation activities of ER␤1 were measured by using the Bright-Glo luciferase kit (Promega). Normalization of transfection efficiency was done by measuring ␤-galactosidase activity using the ␤-gal assay kit (Promega). Each independent experiment was carried out in technical triplicates.
Quantitative RT-PCR-Total RNA was extracted with TRIzol reagent (Invitrogen), and cDNA synthesis was done with SMART Moloney murine leukemia virus reverse transcriptase with poly(dT) primer following the manufacturer's protocols (Promega). Quantitative RT-PCR was performed with ABI7900 real time PCR system (Invitrogen). The sequences of primers used are summarized in Table 2.
Chromatin Immunoprecipitation (ChIP) and Re-ChIP Assays-PC-3-ER␤1 cells were grown in CSS-containing medium supplemented with 10 nM E 2 . ChIP assays were performed as described previously (34), except for the use of magnetic beads (Dynabeads) for capturing antibodies (Invitrogen). In re-ChIP assays, DNA-containing magnetic beads were incubated in TE buffer with 10 mM dithiothreitol (DTT) to elute the immunoprecipitated DNA after the first ChIP assay. The second ChIP assay was performed with the purified DNA by the second antibody. The ChIP DNA was amplified by PCR with the ABI7900 real time PCR system. The sequences of primers used in the amplification are summarized in Table 2.
Statistical Analysis-The Student's t test of QuickCalcs (GraphPad Software, La Jolla, CA) was used for statistical analysis. p values calculated were two-sided, and values Ͻ0.05 were considered statistically significant.

ER␤1 Can Interact with Tip60 in Either the Absence or Presence of Estrogen-To show the physical binding between ER␤1
and Tip60, we performed in vitro coimmunoprecipitation. Tip60 translated in vitro was incubated with ER␣ or ER␤ in the presence of E 2 and immunoprecipitated with HA antibody. The translated Tip60 interacted with both ER␣ (Fig. 1A, lane 2) and ER␤1 (Fig. 1A, lane 6). To confirm the interactions in a cellular system, we cotransformed ER␤1, ER␣, or empty vector with Tip60 into yeast cells. We were surprised to find that Tip60 interacted with ER␤1 in the absence or presence of E 2 , as indicated by the growth of blue yeast colonies (Fig. 1B, left panel). Consistent with previous findings (29,32), ER␣-Tip60 interaction occurred only in the presence of E 2 (Fig. 1B, middle panel). To verify the interaction in a mammalian system, we transfected HEK293 cells with Tip60 or empty vector along with ER␤1, followed by immunoprecipitation (Fig. 1C). ER␤1 was coimmunoprecipitated with Tip60 in the absence and presence of E 2 (Fig. 1C, lanes 1 and 3). Their interaction was verified by reciprocal coimmunoprecipitation using ER␤1-specific antiserum. Tip60 was coimmunoprecipitated only when cells overexpressed ER␤1 and Tip60 (Fig. 1D, lane 1). However, no Tip60 was coimmunoprecipitated when the cells overexpressed only ER␤1 (Fig. 1D, lane 2) or Tip60 alone (Fig. 1D, lane 3). The interaction was also confirmed in a cell line with a high endogenous level of Tip60. A prostate cancer cell line, PC-3, with ectopic expression of ER␤1 (PC-3-ER␤1) was used (35). Tip60 was coimmunoprecipitated with ER␤1 in the absence or presence of E 2 (Fig. 1E, lanes 1 and 3).
We further determined the presence of ER␤1 and Tip60 in the same subcellular compartments. ER␤1 (red) was shown to be colocalized with Tip60 (green) (Fig. 1F) in the nucleus of HEK293 cells in the absence or presence of E 2 . Colocalization of the two proteins also was observed in PC-3-ER␤1 (Fig. 1G). These data show that ER␤1 physically interacts with Tip60 inside the nucleus in either the absence or presence of E 2 .
Tip60 Differentially Regulates ER␤1 Transactivation at ERE and AP-1 Sites-ER␤1 is a transcription factor controlling gene expression by either directly binding to consensus DNA sequences or tethering on other transcription factors (2,5,36). We were interested in investigating whether Tip60 enhances ER␤1 transactivation and whether the effect is dependent on a cis-regulatory element.

TABLE 2 Primers used in the experiments of quantitative RT-PCR and ChIP real time PCR
F is forward, and R is reverse.

Primers Sequences
We therefore transfected Tip60, ER␤1, and different luciferase reporter plasmids into HEK293 cells. Tip60 reduced ER␤1 transactivation at the vitellogenin-ERE site in the absence or presence of E 2 (Fig. 3A). Moreover, we verified its inhibitory effect at ERE sequences of different ER␤1-target genes. Tip60 inhibited ER␤1 transactivation at C3-ERE (Fig. 3B) and c-Fos-ERE sites (Fig. 3C) in the absence or presence of E 2 and also at pS2-ERE (Fig. 3D) and PR-ERE sites (Fig. 3E) in the absence of E 2 . To determine its mode of regulatory action, we showed that the inhibitory effect of Tip60 on ER␤1 transactivation was concentration-dependent (Fig. 3F). Tip60 decreased constitutive and E 2 -induced transactivation, and the fold change also was similar in the absence and presence of E 2 (Fig. 3F). Apart from directly binding to DNA sequences, ER␤1 can interact with coregulators to tether onto other transcription factors to activate the transcription. Tip60 enhanced ER␤1 transactivation at FIGURE 1. ER␤1 can interact with Tip60 in either the absence or presence of estrogen. A, Tip60 interacts with ER␤1 and ER␣ in vitro. ER␤1, ER␣, and HA-tagged Tip60 were translated in vitro and labeled with [ 35 S]methionine. The lysates were mixed and incubated with E 2 and then immunoprecipitated (IP) with HA antibody. The immunoprecipitates were resolved by SDS-PAGE and analyzed by autoradiography. B, ER␤1 interacts with Tip60 in yeast cells independent of E 2 . ER␤1, ER␣, or empty vector (pGBKT7) was transformed into yeast with Tip60. The transformed cells were grown on quadruple dropout agar (QDO) containing X-␣-galactosidase and DMSO or E 2 until the appearance of blue colonies. C, ER␤1 interacts with Tip60 in vivo. HEK293 cells were grown in CSS-containing medium and transfected with ER␤1 and His-tagged Tip60 before the addition of E 2 . Lysates were precipitated on an Ni-NTA column and immunoblotted (IB) with ER␤1 or Tip60 antibody. The samples were run on the same gel. D, ER␤1-Tip60 interaction was confirmed by reciprocal coimmunoprecipitation. Procedures were similar to those in C, except that lysates were immunoprecipitated with ER␤1 antibody. E, ER␤1 interacts with Tip60 in an E 2 -independent manner in a hormone-sensitive prostate cancer cell line, PC-3. ER␤1 stably expressed PC-3 cells (PC-3-ER␤1) were grown in CSS-containing medium before the addition of DMSO or E 2 . Lysates were immunoprecipitated by ER␤1 antibody and immunoblotted with ER␤1 or Tip60 antibody. F, ER␤1 colocalized with Tip60 with or without E 2 . HEK293 cells were grown in CSS-containing medium transfected with ER␤1 and Tip60 followed by the incubation of DMSO (vehicle) (upper panel) or E 2 (lower panel). G, ER␤1 colocalized with Tip60 in PC-3. PC-3-ER␤1 cells were grown in full-serum containing medium. F and G, antibodies to ER␤1 and Tip60 were used for immunostaining, and DAPI was used as the nuclear marker. The images in F and G were captured by a fluorescence microscope. Bar, 20 m.
the AP-1-response element (Fig. 3G). Tip60 increased ER␤1 transactivation more significantly in the absence of E 2 than in the presence of E 2 . The transcriptional regulation by Tip60 required ER␤1 expression because cells transfected with only Tip60 showed very little luciferase activity (data not shown). In contrast, no regulatory effect of Tip60 was observed at NFBand Sp1-binding sites (Fig. 3, H and I). Differential regulation of ER␤1 transactivation at vitellogenin ERE and AP-1 sites was also observed in the different prostate cancer cell lines PC-3 (Fig. 3, J and K) and DU-145 (Fig. 3, L and M). These data suggest that Tip60 enhances ER␤1 transactivation at the AP-1 site but reduces the transactivation at different ERE sites.
Various Ligands Modulate the Regulatory Effects by Tip60 on ER␤1 Transactivation-Because we found that the regulatory effect of Tip60 at AP-1 site was reduced by E 2 , we sought to determine whether various ligands could influence its regulation. This was especially relevant because transcriptional activ-ity of ER␤1 responds differently depending on ligands and binding sites (36). We tested five categories of chemicals, estrogens (E 2 and DPN), phytoestrogens (GEN, EQ, DAI, and API), selective estrogen receptor modulators (SERMs) (RAL and TAM), antiestrogen (ICI), and an endocrine disruptor (BPA). As with previous findings (8,36,37), ER␤1 transactivation at the ERE site was enhanced in the presence of estrogens or phytoestrogens but was suppressed in the presence of TAM, RAL, and ICI (Fig. 4A). In stark contrast, SERMs and antiestrogen stimulated the transactivation at the AP-1 site, whereas estrogens and phytoestrogens inhibited ER␤1 transcriptional activity (Fig. 4B). Moreover, we examined the regulatory effect by Tip60 in the presence of these ligands. The transcriptional inhibition by Tip60 persisted at the ERE site in response to all of the ligands except DPN, GEN, EQ, and BPA (Fig. 4A). In contrast, all the estrogens and phytoestrogens except apigenin downregulated the enhancement of ER␤1 transactivation by Tip60 at the AP-1 site (Fig. 4B). SERMs could not further up-regulate the effect of Tip60, whereas ICI was the only ligand that increased Tip60 enhancement over that of the control (Fig. 4B). Hence, we suggest that various ligands differentially modulate the regulatory effects by Tip60 at ERE and AP-1 sites.
ER␤1 Cannot Be Acetylated by Tip60 and Preferentially Interacts with Unacetylated Tip60-Tip60 was found to acetylate different transcription factors, such as androgen receptor, p53, c-Myc, and ataxia telangiectasia mutated (ATM) (24 -26, 38). To examine whether ER␤1 can be acetylated by Tip60, we performed an in vitro acetylation assay. The structural domains of the Tip60 wild-type (Tip60WT) and the mutation sites of its HAT-defective mutant (Tip60⌬HAT) (Q377E/G380E) are shown in Fig. 5A. His-tagged Tip60WT and Tip60⌬HAT proteins were purified on Ni-NTA columns. Recombinant ER␤1 protein and purified Tip60 were incubated with acetyl-CoA. Consistent with the finding by another group (36), we found that Tip60WT, but not Tip60⌬HAT, was able to auto-acetylate in vitro (Fig. 5B, lanes 1 and 3). However, ER␤1 could not be acetylated by either Tip60WT or Tip60⌬HAT as shown by the absence of signal when pan-acetyl-lysine antibody was used (Fig. 5B, lanes 3 and 4).
Next, we verified the results in vivo. Either Tip60WT or Tip60⌬HAT was expressed simultaneously with ER␤1. The cells were incubated with TSA and nicotinamide to maximize the level of acetylation. Immunoprecipitation was performed with either ER␤1 or Tip60 antibody to isolate different populations of protein complex. Tip60WT, but not Tip60⌬HAT, was able to auto-acetylate in vivo, as shown in the input lysate (Fig.  5C, right panel). Immunoprecipitation was first performed with ER␤1 antibody to isolate ER␤1 complexes that may or may not contain Tip60. Although Tip60 was coimmunoprecipitated with ER␤1, no acetylation of ER␤1 or Tip60 was detected (Fig.  5C, left panel). Similarly, Tip60 antibody was then used in the pulldown assay to isolate two populations of Tip60 complexes, including the one with or without ER␤1. As expected, Tip60 and ER␤1 were isolated simultaneously. Although there was no acetylation of ER␤1, auto-acetylation of Tip60WT was detected in the immunoprecipitate (Fig. 5C, middle panel), and its unacetylated form may interact preferentially with ER␤1. To constructs and Tip60 is represented by "ϩ" and "Ϫ" signs. "ϩϩϩ" represents the strongest interaction, and "Ϫ" represents no interaction. AF-1, activation function 1; AF-2, activation function 2. B, HEK293 cells were grown in CSScontaining medium and transfected with Tip60 and different domain-deleted ER␤1 constructs. Lysates were immunoprecipitated (IP) with c-Myc antibody. Immunoglobulin IgG was used as the negative control. The immunoprecipitates were immunoblotted (IB) with c-Myc or Tip60 antibody. Asterisks denote the positions of ER␤1 and its mutants.
conclude, ER␤1 is not acetylated by Tip60 in vitro or in vivo and may preferentially interact with the unacetylated form of Tip60.
HAT Activity of Tip60 Is Not Involved in the Regulation of ER␤1 Transactivation at AP-1 and ERE Sites-Acetylation of androgen receptor (AR) by Tip60 is essential for up-regulating the transactivation of AR at AR-response elements (24). The inability of Tip60 to acetylate ER␤1 infers that its HAT activity may not be important for regulating ER␤1 activity. Luciferase reporter assays were performed to determine ER␤1 activity with the overexpression of Tip60WT and Tip60⌬HAT. Western blot analysis showed that their expression was similar (Fig.  6A). At the vitellogenin-ERE site, the two Tip60 proteins were equally effective in reducing ER␤1 transactivation (Fig. 6B). However, Tip60⌬HAT enhanced ER␤1 transactivation to a greater extent than Tip60WT did at the AP-1 site (Fig. 6C). To further determine the significance of the HAT activity of Tip60 to the transcriptional activity of ER␤1, we used a HAT inhibitor, anacardic acid, which inhibits Tip60-dependent acetylation (39). Similar to the results shown in Fig. 6, B and C, enhancement of ER␤1 transactivation by Tip60 was up-regulated at the AP-1 site in the presence of anacardic acid (Fig. 6E), but no change was observed at the ERE site (Fig. 6D). The results suggest that HAT activity of Tip60 is not required to regulate ER␤1 transactivation at ERE or AP-1 site.
Tip60 Interacts with GRIP1 to Enhance ER␤1 Transactivation at the AP-1 Site Synergistically-The p160 SRC family consists of three homologous members, SRC1, GRIP1, and SRC3 (40 -42). Of these, SRC1 and GRIP1 are coactivators of ERs at the AP-1 and ERE sites (4,43). Because Tip60 enhanced ER␤1 transactivation at the AP-1 site but diminished transactivation at different ERE sites, we investigated whether Tip60 has any combinatorial effect with p160 coactivators. We overexpressed different combinations of Tip60, SRC1, and GRIP1 together with ER␤1 and determined the regulation of ER␤1 transactivation by these proteins at the ERE and AP-1 sites. SRC1 enhanced ER␤1 transactivation in the absence of E 2 , whereas Tip60 and GRIP1 reduced ER␤1 transactivation in the absence or presence of E 2 . The effect of inhibition persisted when Tip60 and GRIP1 were overexpressed simultaneously (Fig. 7A). At the AP-1 site, all three coregulators were able to enhance ER␤1 transactivation with or without E 2 (Fig. 7B). In the absence of E 2 , coexpression of Tip60 and GRIP1 had the strongest stimulatory effect on the transactivation. Interestingly, overexpression of SRC1 abolished the synergistic effects of Tip60 and GRIP1 (Fig. 7B). To further investigate the synergistic effect of Tip60 and GRIP1 on ER␤1 transactivation at the AP-1 site, we performed luciferase assays with different ratios of GRIP1 and Tip60 plasmids. Consistent with the results in Fig. 7B, coexpression of GRIP1 and Tip60 resulted in a greater enhancement of ER␤1 transactivation than expression of GRIP1 alone, whereas a 1:1 ratio of GRIP1 and Tip60 plasmids resulted in the greatest enhancement at the AP-1 site (Fig. 7C). Next, an immunoprecipitation experiment was used to determine whether ER␤1, Tip60, and the two p160 coactivators are involved in a transcriptional complex. Tip60 interacted with ER␤1, GRIP1, and SRC1 (Fig. 7D). To conclude, ER␤1, Tip60, GRIP1, and SRC1 are able to form a multiprotein complex, whereas Tip60 and GRIP1 synergistically enhance ER␤1 transactivation at the AP-1 site.
Tip60 Differentially Regulates ER␤1 Target Genes by Modulating ER␤1 Binding to the cis-Regulatory Regions Possessing the ERE or AP-1 Site-In our study, Tip60 either enhanced or reduced ER␤1 transactivation at the AP-1 or ERE site. To inves-

. Tip60 differentially regulates ER␤1 transactivation at ERE and AP-1 sites but has minimal effect on other transcription factor-binding sites.
A-E, Tip60 reduces ER␤1 transactivation at various ERE sites. ER␤1 was transfected with GFP or Tip60 together with pCMV-␤-gal and vitellogenin ERE (A), C3 ERE (B), c-Fos ERE (C), pS2 ERE (D), or progesterone receptor ERE (E) reporter plasmids into HEK293 cells grown in CSS-containing medium. F, inhibition of ER␤1 transactivation by Tip60 is concentration-dependent. ER␤1 was transfected with different amounts of GFP and Tip60 together with pCMV-␤-gal and vitellogenin ERE reporter plasmid. Different ratios of plasmids of Tip60 to GFP were transfected. G-I, Tip60 enhances ER␤1 transactivation at AP-1 sites but has minimal effect on other transcription factor-binding sites. ER␤1 was transfected with GFP or Tip60 together with pCMV-␤-gal and reporter plasmids containing the binding site of AP-1 (G), NFB (H), or Sp1 (I) into HEK293 cells grown in CSS-containing medium. J-M, Tip60 reduces ER␤1 transactivation at ERE site but increases its transactivation at AP-1 site in different PCa cell lines. ER␤1 was transfected with GFP or Tip60 together with pCMV-␤-gal and reporter plasmids containing vitellogenin ERE (J and L) or AP-1-binding site (K and M) into PC-3 or DU-145 cells grown in CSS-containing medium. After the transfection, HEK293, PC-3, and DU-145 cells were added with DMSO or E 2 . Relative luciferase activity was determined and normalized with the ␤-gal activity. Results were the average of three independent experiments. All data are represented as mean Ϯ S.D. The statistical significance of the difference in luciferase activity between the overexpression of GFP and Tip60 in the presence of DMSO or E 2 is shown as follows: *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001.  AUGUST 30, 2013 • VOLUME 288 • NUMBER 35 tigate whether ER␤1 target genes are differentially regulated by Tip60, we determined their gene expressions in ER␤1 or LacZ stably expressed PC-3 cells (PC-3-ER␤1/PC-3-LacZ) after the knockdown of Tip60. The ectopic expression of ER␤1 and the efficiency of Tip60 knockdown were confirmed by quantitative RT-PCR (Fig. 8, A and B) and Western blotting (data not shown). We found that the expressions of CXCL12 and cyclin D2 were drastically increased in PC-3-ER␤1 compared with the control (PC-3-LacZ) (Fig. 8, C and D). Moreover, their expressions were differentially regulated with the knockdown of Tip60 in PC-3-ER␤1 cells. Expression of CXCL12 was further up-regulated (Fig. 8C), whereas that of cyclin D2 was downregulated after Tip60 depletion (Fig. 8D). The cis-regulatory sequence of the CXCL12 gene was found to have an ERE site (44), and sequence analysis revealed a predicted AP-1-binding site at the upstream region of the cyclin D2 gene (data not shown). In ChIP assays, ER␤1 and Tip60 were significantly recruited to the respective investigated regions (Fig. 8, E and F). Moreover, the co-occupancy of ER␤1 and Tip60 on the respective cis-regulatory regions of CXCL12 and cyclin D2 was confirmed in the re-ChIP assay (Fig. 8G). Similar results were observed in the reciprocal re-ChIP assay (data not shown). Next, we investigated the molecular mechanism of differential regulation of ER␤1 target genes by Tip60. Upon the depletion of Tip60, the recruitment of ER␤1 to the cis-regulatory region of CXCL12 was significantly enhanced, whereas the recruitment of ER␤1 to the investigated region of cyclin D2 was decreased (Fig. 8H). Collectively, our results showed that Tip60 differen-tially regulates the expression of ER␤1 target genes by modulating the binding of ER␤1 to their respective cis-regulatory regions.

DISCUSSION
Estrogen signaling is mediated primarily by ER␣ and ER␤1, whereas ER␤1 is able to activate a distinct set of target genes and also to antagonize ER␣ transactivation (45)(46)(47)(48)(49). Although ERs share many common coregulators, the differential interaction between the coregulatory proteins and ERs may be responsible for their distinct functions (8). In this study, Tip60 was found to interact with ER␤1 in the absence or the presence of E 2 . Tip60 either enhances or inhibits ER␤1 transactivation, depending on the cis-regulatory sites. Moreover, Tip60 and GRIP1 enhance the transactivation at the AP-1 site synergistically. We also showed that ER␤1 is not acetylated by Tip60 and thus that the regulation of ER␤1 activity by Tip60 is independent of its HAT activity. In addition, Tip60 is able to differentially control the endogenous expression of ER␤1 target genes possessing the ERE or AP-1 site by modulating ER␤1 binding to the respective cis-regulatory regions. On the basis of these data, we suggest that ER␤1 transactivation is differentially regulated by Tip60 in a regulatory element-dependent manner.
Tip60 is an interacting partner of some hormone receptors, including ERs, AR, and PR. Their interactions were shown to require the presence of respective agonists (32). In this study, we found that the binding of Tip60 to ER␤1 does not require ligands and that the strength of the interaction is similar in the  (Tip60⌬HAT). B, ER␤1 is not acetylated by Tip60 in vitro. His-tagged wild-type of Tip60 (Tip60WT) or Tip60⌬HAT was transfected, respectively, into HEK293 cells, and Tip60 proteins were purified on an Ni-NTA column. Recombinant ER␤1 protein and Tip60 were incubated in HAT buffer containing acetyl-CoA. The immunoprecipitates (IP) were immunoblotted (IB) with acetyl-lysine, ER␤1, or Tip60 antibody. Asterisk denotes the nonspecific band that appeared when the blot was immunoblotted with pan-acetyl-lysine antibody. C, ER␤1 is not acetylated by Tip60 in vivo and preferentially interacts with unacetylated Tip60. Tip60WT or HAT was transfected with ER␤1 into HEK293 cells. Lysates were immunoprecipitated with either ER␤1 (left panel) or Tip60 (middle panel) antibody. Immunoglobulin IgG was used as the negative control. The immunoprecipitates were immunoblotted with acetyl-lysine, ER␤1, or Tip60 antibody.
absence or presence of E 2 . The discrepancy between our finding and that from another group may be due to our use of different ER␤1 sequences and interaction assays. Gaughan et al. (32) used a construct containing only LBD of ER␤1 in the mammalian two-hybrid assay. We used the full-length ER␤1, which is more biologically relevant in terms of protein folding, to show the interaction in yeast two-hybrid assays, in vitro and in vivo coimmunoprecipitation, and subcellular localization studies in different cell lines. It is not uncommon for ligand-independent interactions to occur between ER␤1 and coregulators. For example, phosphorylation of ER␤1 leads to ligand-independent recruitment of SRC1 (48), and GRIP1 is also recruited by unliganded ER␤1 (8,10,20). Both coactivators stimulated unliganded ER␤1 transactivation (8,10). Our data suggest that Tip60 interacts with ER␤1 regardless of E 2 presence.
The interaction of Tip60 with LBD of ER␣ in a ligand-dependent manner is well documented (29,30,32). The distinct mechanisms of recruiting Tip60 by ER␣ and ER␤1 imply that they may have different domains interacting with Tip60. We performed domain deletion of ER␤1 followed by immunoprecipitation to show that the hinge domain of ER␤1 is the interacting region. Although ERs interact with the most coactivators and corepressors at either or both N and C termini (50), they also bind to some coregulators at the hinge domain. L7/SPA interacts with the hinge domain of ER␣ and enhances transactivation of antagonist-occupied ER␣ at the ERE site (51). ER␣ also binds to PGC-1 at its hinge domain in a ligand-independent manner (52). Although the hinge domain of ERs is not as well characterized, it has been shown to affect protein degradation and activity of ER␤1 (53,54), ER␣ tethered-mediated AP-1 transactivation (55), and the functional synergy between AF-1 and AF-2 of ERs (56). Because AF-1 and AF-2 domains are responsible for E 2 -independent and E 2 -dependent activation of the transactivation of ERs (50), we speculate that the atypical interaction interface between ER␤1 and Tip60 at the hinge domain may contribute to the unique regulation of ER␤1 activity by Tip60.
Tip60 functions as a coregulator of many transcription factors (57). Hence, we determined its role in the regulation of ER␤1 transactivation by the luciferase assay and used reporter constructs with different cis-regulatory sequences of the target genes of ER␤1. Tip60 reduced ER␤1 transactivation at different ERE sequences, such as vitellogenin-, C3-, c-Fos-, pS2-and PR-EREs. Moreover, the inhibitory action of ER␤1 transactivation by Tip60 is concentration-dependent but E 2 -independent. Our results imply that Tip60 can inhibit transcription of certain ER␤1-regulated genes possessing ERE sites. In contrast, Tip60 increased the expression of some estrogen-regulated ER␣ target genes containing EREs (29,30). Because ER␤1 antagonizes ER␣-dependent transcription through hetero-dimerization (50), Tip60 may be a key factor in determining the antagonism between ERs. ER␤1 also interacts with other transcription factors to mediate the transcription through tethering. We showed that Tip60 did not regulate ER␤1 transactivation at either the NFB or the Sp1 site but that it drastically increased the transactivation at the AP-1 site. Moreover, the enhancement by Tip60 was more drastic in the absence of E 2 . It is not surprising for a coregulator to show dual regulation of the activity of transcription factors. GRIP1 acts as a coactivator of ER␣ at ERE and AP-1 sites (4,8) but inhibits the activity of E 2 -bound ER␣, which tethers on c-Jun and NFB at TNF␣ promoter (58). In addition, GRIP1 is either a coactivator or a corepressor of glucocorticoid receptor in a hormone-response element-dependent manner (59). Our study not only shows that the regulation of ER␤1 transactivation by Tip60 occurs in an E 2 -independent manner but also provides evidence that it can enhance or inhibit the transactivation at the AP-1-response element or ERE, respectively.
The modulation by ligands of ER␤1 signaling at different response elements has been well documented (36). We extensively investigated the effects of various steroidal compounds on the transcriptional regulation by Tip60. Consistent with the previous findings (36,60,61), we found that estrogenic compounds (E 2 and DPN) and phytoestrogens (GEN, EQ, DAI, and API) up-regulated ER␤1 transactivation at ERE, whereas SERMs (TAM and RAL) and antiestrogen (ICI) did the opposite. Surprisingly, DPN, GEN, and EQ abolished Tip60-mediated inhibition at the ERE site. Moreover, all estrogenic chemicals except apigenin significantly inhibited enhancement by Tip60 at the AP-1 site. The discrepancy may be due to differential conformational changes of ER␤1 through binding to dif- FIGURE 6. HAT activity of Tip60 is not necessary for regulation of the ER␤1 transactivation at AP-1 and ERE sites. A, expression of Tip60⌬HAT was similar to that of Tip60WT. Lysates were extracted and immunoblotted (IB) with Tip60 antibody. ␤-Actin was used as the loading control. B and C, HAT activity of Tip60 is not necessary for the regulation of ER␤1 transactivation at AP-1 and ERE sites. GFP, Tip60WT, or Tip60⌬HAT was transfected, respectively, with ER␤1, pCMV-␤-gal (B), AP-1(C), or vitellogenin-ERE reporter plasmids into HEK293 cells before the addition of E 2 . D and E, GFP or Tip60 was transfected, respectively, with ER␤1, pCMV-␤-gal (D), AP-1 (E), or vitellogenin-ERE reporters into HEK293 cells. After the transfection, DMSO or E 2 together with ethanol (vehicle) or anacardic acid (AnAc) was added as indicated. B-D, relative luciferase activity was determined as in Fig. 3. Results are the average of three independent experiments. Data are represented as mean Ϯ S.D. The statistical significance of the difference in luciferase activity between the overexpression of GFP and Tip60 in the presence of DMSO or E 2 is shown as follows: *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001. ferent estrogenic chemicals (62,63) and thus affect the formation of the ER␤1 transcriptional complex (62). Perhaps the binding of these compounds triggers the recruitment of other coactivators to counteract the Tip60-mediated inhibition (7). For example, GEN can recruit SRC1 isoforms to ER␤1 (64,65). Moreover, all estrogenic chemicals except apigenin significantly inhibited the enhancement by Tip60 at the AP-1 site. Although Fujimoto et al. (37) suggested that estrogens and phytoestrogens do not exert any regulatory effect on ER␤1-mediated AP-1 transactivation, previous findings and this study have clearly shown that estrogens or phytoestrogens repress the transactivation at the AP-1 site (36,66). It is tempting to speculate that these compounds reduce the potency of recruitment of coactivators, such as Tip60, by ER␤1 at AP-1 site. In contrast, ICI and SERMs were agonists of ER␤1-mediated AP-1 transactivation, but SERMs did not further up-regulate the enhancement by Tip60 as compared with the control. We suggest that SERMs cannot improve the potency of Tip60 recruitment by ER␤1. Another possible explanation may be that the binding of either Tip60 or SERMs causes a similar conformational change in ER␤1 that is favorable to tethering on the AP-1 site (67)(68)(69). Tip60 and SERMs are thus redundant to the enhancement of ER␤1 transactivation. To conclude, we showed that the differential regulation of ER␤1 transactivation by Tip60 at ERE and AP-1 sites is controlled through binding to different ligands.
Tip60 enhances the activities of certain transcription factors through acetylation (57). Thus, we sought to determine whether its regulation of ER␤1 activity is mediated through acetylation. We used different acetylation assays to illustrate that Tip60 is incapable of acetylating ER␤1. This is consistent with studies of other coregulators of ER␤1 that possess HAT activity, but none of them was found to acetylate ER␤1 (9,13,16,45,48). Moreover, acetylation of nuclear receptors is assumed to occur at a conserved motif "(K/R)XKK" (13), which is absent in ER␤1 (data not shown). These findings suggest that ER␤1 may not be post-translationally modified through acetylation.
In addition to acetylating its interacting partners, Tip60 can auto-acetylate to regulate its activity (71,72). In our in vivo acetylation assays, acetylation of Tip60 was detected only in the immunoprecipitation that used Tip60 antibody but not ER␤1 antibody, revealing that those Tip60 proteins in the ER␤1-Tip60 complex are probably unacetylated. The result implies that ER␤1 may preferentially interact with unacetylated Tip60, perhaps because auto-acetylation modifies the structure of Tip60 (71). Our study verified that HAT activity of Tip60 does not increase ER␤1 transactivation. In contrast, Tip60⌬HAT did not reduce but enhanced ER␤1 activity at the AP-1 site. The result was confirmed with the use of anacardic acid, which inhibits the HAT activity of Tip60 (37). The observation may be FIGURE 7. Tip60 interacts with GRIP1 to enhance ER␤1 transactivation at the AP-1 site synergistically. A and B, Tip60 and GRIP1 exert a synergistic effect on ER␤1 transactivation at the AP-1 site. Different combinations of GFP, Tip60, GRIP1, and SRC1 were transfected with ER␤1, pCMV-␤-gal, vitellogenin-ERE (A) or AP-1 reporter plasmids (B) as indicated. After the transfection, DMSO or 10 nM E 2 was added as indicated. C, synergistic effect of Tip60 and GRIP1 on the ER␤1 transactivation at the AP-1 site is concentration-dependent. GFP or different ratios of plasmids of Tip60 to GRIP1 were transfected. DMSO was added after the transfection. Relative luciferase activity was determined as in Fig. 3. Results are the average of three independent experiments. Data are presented as mean Ϯ S.D. The statistical significance of the difference in luciferase activity between overexpressing Tip60, GRIP1, and GFP is shown as *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001. D, Tip60 forms a multiprotein complex with p160 coactivators and ER␤1. HEK293 cells were transfected with Tip60, ER␤1, SRC1, and GRIP1 and grown in CSS-containing medium. Lysates were immunoprecipitated (IP) with Tip60 antibody. Immunoglobulin IgG was used as the negative control. The immunoprecipitates were immunoblotted (IB) with Tip60, ER␤1, GRIP1, or SRC1 antibody as indicated.
explained by the increased amount of unacetylated Tip60 that binds to ER␤1. In fact, HAT activity of Tip60 is not essential for the regulation of the activity of some transcription factors, such as CREB, STAT3, and PGC-1␣ (27,28,73). Our data indicate that ER␤1 transactivation is not regulated through HAT activity of Tip60. Furthermore, the receptor appears to interact preferentially with unacetylated Tip60.
In this study, we found that ER␤1 activity was enhanced by Tip60 at the AP-1 site. The ER␤1-mediated transactivation requires the recruitment of CBP/p300 and p160 coactivators at the AP-1-response element (4), where ER␤1 interacts primarily with p160 coactivators (4,43,74,75). These observations urged us to investigate whether Tip60 interacts with p160 coactivators to regulate ER␤1 transactivation. We found that Tip60 interacted with SRC1 and GRIP1, although it only enhanced ER␤1 activity synergistically with GRIP1 at the AP-1 site. Moreover, expression of different amounts of GRIP1 and Tip60 always resulted in a greater enhancement of ER␤1 transactivation compared with expression of GRIP1 alone, revealing that they simultaneously act as coactivators of ER␤1 at the AP-1 site. It is interesting that SRC1 was not synergistic with the other two coregulators, implying that it may have other mechanisms regulating ER␤1 transactivation. Because ER␤1 interacts with Tip60 at its hinge domain and GRIP1 binds to AF-1 and AF-2 domains of the receptor (43), we therefore hypothesize that Tip60 and GRIP1 cooperate to modify the conformation of ER␤1, permitting more efficient tethering on the AP-1 site.
In addition, we showed that Tip60 modulates ER␤1 regulation of endogenous gene expression in prostate cancer cells. In our search for ER␤1-regulated genes (10, 11), CXCL12 (76) and FIGURE 8. Tip60 differentially regulates ER␤1 target genes possessing ERE or AP-1 sites at their cis-regulatory regions in PC-3 cells. A and B, expression of ER␤1 and Tip60 in ER␤1 and LacZ stably expressed PC-3 cells upon the knockdown of Tip60 was determined. PC-3-LacZ/-ER␤1 cells were grown in CSS-containing medium and transfected with nontargeting control siRNA (siNT) or siRNAs specific to Tip60 (siTip). E 2 was added after 24 h. Expression of ER␤1 (A) and Tip60 (B) was determined by quantitative RT-PCR. Human GAPDH was used as the housekeeping gene. C and D, Tip60 differentially regulates ER␤1 target genes. PC-3-LacZ/-ER␤1 cells were treated as described in A and B. Expression of CXCL12 (C) and cyclin D2 (D) was determined by quantitative RT-PCR. The results are the average of three independent experiments. All data are represented as mean Ϯ S.D. The statistical significance of the difference in gene expression between different treatments is shown as follows: *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001. E-G, ER␤1 and Tip60 are both recruited to the cis-regulatory regions of CXCL12 and cyclin D2. PC-3-ER␤1 cells were grown in CSS-containing medium added with E 2 . ChIP assays were performed with ER␤1 (E) or Tip60 antibody (F). G, re-ChIP assay was performed with Tip60 antibody followed by the second immunoprecipitation with ER␤1 antibody. The ChIP DNA was amplified by real time PCR for the target regions containing an ERE site of CXCL12 or an AP-1 site of cyclin D2. The genomic region of ER␤ isoform 5 (ER␤5) containing neither an ERE nor an AP-1 site was used as the negative control. The fold enrichment of recruitment of ER␤1 and/or Tip60 at the target regions is relative to respective IgG controls. The results are the average of two independent experiments. All data are represented as mean Ϯ S.D. The statistical significance of the difference in the recruitment between ER␤1 (and/or Tip60) and IgG is shown as follows: *, p Ͻ 0.05. H, Tip60 differentially regulates the recruitment of ER␤1 to the cis-regulatory regions of CXCL12 and cyclin D2. PC-3-ER␤1 cells were grown in CSS-containing medium added with E 2 and transfected with siRNAs (siNT or siTip) for 48 h. ChIP assays were performed with ER␤1 antibody. The procedures of the amplification of ChIP DNA were similar to those described in E-G. The results are the average of two independent experiments. Data are represented as mean Ϯ S.D. The statistical significance of the difference in the ER␤1 recruitment with or without the knockdown of Tip60 is shown as follows: *, p Ͻ 0.05. AUGUST 30, 2013 • VOLUME 288 • NUMBER 35 cyclin D2 (previously unknown) were the only two that we identified in this study that were regulated by both ER␤1 and Tip60. We found that upon the knockdown of Tip60, the expression of CXCL12 increased and that of cyclin D2 decreased. Interestingly, the promoter region of CXCL12 contains multiple EREs (44,76,77) and that of cyclin D2 harbors two AP-1 sites based on bioinformatics. In the ChIP and re-ChIP assays, ER␤1 and Tip60 were shown to co-occupy the investigated regions. Moreover, the depletion of Tip60 appeared to increase ER␤1 binding to the promoter of CXCL12 and decrease its recruitment to the promoter of cyclin D2. These results raise the possibility that Tip60 promotes the recruitment of ER␤1 to AP-1 site but reduces its ERE binding, a mechanism that likely contributes to the differential regulation of ER␤1-targeted gene expression.

Differential Regulation of ER␤1 Transactivation by Tip60
In conclusion, we showed that Tip60 modulates ER␤1 action in a regulatory element-dependent manner as exemplified by its opposing roles on ER␤1 transactivation at the ERE and AP-1 sites. Furthermore, its coregulatory action on ER␤1 appears to be E 2 -independent at both cis-elements, unlike its action on ER␣. Contrary to common belief, Tip60 action is not mediated by its HAT activity. Our data also suggest that the interaction between Tip60 and GRIP1 synergistically enhances ER␤1 tethering on the AP-1 site. Moreover, Tip60 can modulate the recruitment of ER␤1 to the promoters of CXCL12 and cyclin D2, harboring the ERE and AP-1 site, respectively. Collectively, these data put Tip60 into the category of a multifaceted coregulator in the ER␤1 context, similar to GRIP1 in the regulation of the activities of ER␣ and glucocorticoid receptor (4,8,58,59).