Turning Off Estrogen Receptor {beta}-Mediated Transcription Requires Estrogen-Dependent Receptor Proteolysis

Recent studies have shed light on the ligand-dependent transactivationmechanisms of nuclear receptors (NRs). When the ligand doseis reduced, the transcriptional activity of NRs should be downregulated.Here we show that a ubiquitin-proteasome pathway plays a keyrole in turning off transcription mediated by estrogen receptorß (ERß). ERß shows estrogen-dependentproteolysis, and its degradation is regulated by two regionsin the receptor. The N-terminal 37-amino acid-region is necessaryfor the recruitment of the ubiquitin ligase, i.e., the carboxylterminus of HSC70-interacting protein (CHIP), to degrade ERß.In contrast, the C-terminal F domain protects ligand-unboundERß from proteolysis to abrogate proteasome association.Suppression of CHIP by interfering RNA inhibited this switchingoff of receptor-mediated transcription when the ligand dosewas reduced. Our results suggest that after ligand withdrawal,the active form of the NR is selectively eliminated via ligand-dependentproteolysis to downregulate receptor-mediated transcription.

INTRODUCTION

Top

Introduction

Estrogen is a growth factor that stimulates cell growth anddifferentiation in diverse tissues. Dramatic changes in thelevels of estrogen in the blood occur during the normal humanmenstrual cycle. During the first 7 to 10 days, the level ofestrogen is low. In the middle of the cycle (days 11 to 15),it rises and then falls abruptly. The effects of estrogen aremediated through the estrogen receptors (ERs), ER ${alpha}$ and ERß,which function as ligand-induced transcription factors and belongto the nuclear receptor (NR) superfamily (2, 5, 12, 20, 31,33, 41, 49). When the estrogen level is high, estrogen bindsto ERs to activate the transcription of target genes. This inducesthe ligand-binding domain (LBD) to undergo a characteristicconformational change, whereupon the receptor dimerizes, bindsto DNA, and subsequently stimulates gene expression (10). ER ${alpha}$ is stimulated by two distinct activation regions, activationfunction 1 (AF-1) and AF-2, which are located in the LBD andexert ligand-dependent transcriptional activity. Crystal structureanalysis of ERs and other NRs has revealed the presence of 12conserved helices in their LBDs (46). The LBD forms a structuredescribed as a sandwich of 12 ${alpha}$ -helices (helices 1 to 12) witha central hydrophobic ligand-binding pocket. Helix 12, the mostC-terminal of these helices, has been identified as the criticalcore (AD core) of the AF-2 function of the receptor and playsan important role in coactivator binding to the ligand-boundreceptor (6, 19, 24, 25, 33, 35, 39, 43, 45, 55, 56).

When the level of estrogen is decreased, receptor-mediated transcriptionshould be downregulated. Because the response to estrogen istightly controlled, the downregulation of ER-mediated transcriptionshould also be tightly regulated. Although such regulation ofthe NRs has generated intense interest, the molecular mechanismsregulating these processes are not understood (38). To switchoff transcription, there are at least two possible mechanisms.The first involves the dissociation of ligands from the hormone-bindingpocket of the receptor. However, the dissociation rate of theligand from its receptor is very low, because the receptor locksthe hormone within its activation pocket and covers it withcoactivators (16). Another possible mechanism is the rapid degradationof active receptors. Several nuclear receptors, such as theER and those for progesterone, glucocorticoid, thyroid hormonereceptor, retinoid X, and retinoic acid, are ubiquitinated anddegraded in the course of their nuclear activities (8, 11, 13,26, 29, 32, 36, 47, 51, 57). These degradation steps might beinvolved in turning off NR-mediated transcription.

Evidence now suggests that proteasome-dependent degradationof receptors is necessary for the activation of these receptors(38). Moreover, this requires NR-mediated transcriptional activity,implying that the degradation and the transactivation of NRsare mutually independent processes. In the ubiquitin-proteasomepathway, proteins destined for degradation are conjugated bypolyubiquitin chains, in which one ubiquitin molecule is conjugatedby another through one of its seven lysine residues, typicallyK48. These polyubiquitin chains are then recognized by the regulatorycomplex of the 26S proteasome.

Here we show that there is a ubiquitin-proteasome pathway forERß that is not coupled to transactivation. To investigatethis, we purified the ubiquitin ligase complex for ERßand identified a protein complex containing the carboxyl terminusof heat shock cognate 70 (HSC70)-interacting protein (CHIP)(1). CHIP binds directly to the N-terminal 37-amino-acid regionof ERß and ubiquitinates it to induce ERßdegradation. In contrast, the C-terminal F domain inhibits thebinding of proteasome to the receptor and protects it from proteolysisin the absence of estrogen. Suppression of CHIP by interferingRNA (RNAi) inhibited the switching off of receptor-mediatedtranscription that occurs when the ligand dose is reduced. Ourresults suggest that ligand-dependent degradation selectivelyand rapidly eliminates the active form of a nuclear receptorto downregulate receptor-mediated transcription after ligandwithdrawal.

MATERIALS AND METHODS

Top

Materials and Methods

Expression vectors. The ER ${alpha}$ /ß and CHIP expression plasmids and their deletionmutants have been described previously (48). ERE-TATA-Luc andERE-TATA-LucCP were constructed by inserting three consensusestrogen response elements into a luciferase reporter plasmidencoding a TATA box, pGL3-Basic or phRG(R2.2)-Basic. pGL3-Basic,phRG(R2.2)-Basic, and pGL3-Control vectors were purchased fromPromega (Madison, WI).

Antibodies. Anti-FLAG-M2 mouse monoclonal antibody (Sigma, St. Louis, MO),anti-hemagglutinin (HA) mouse monoclonal antibody (Roche), anti-humanER ${alpha}$ antibody (Chemicon, Temecula, CA), anti-human ERßantibody (PPMX), and anti-glyceraldehyde-3-phosphate-dehydrogenasemouse monoclonal antibody (American Research Products, Inc.)were used at appropriate dilutions, according to the manufacturers'instructions. A rabbit polyclonal antibody directed againstCHIP was provided by T. Chiba (Tokyo Metropolitan Instituteof Medical Science).

Cell culture and transfection. MDA-MB231 breast cancer cells and human embryonic kidney 293cells were routinely maintained in Dulbecco's modified Eagle'smedium (DMEM) supplemented with 10% fetal bovine serum (FBS).Twenty-four hours before transfection, the medium was changedto phenol red-free DMEM containing 4% charcoal-stripped FBS.Transfection was performed using PerFectin transfection reagent(Gene Therapy Systems) or TransFast transfection reagent (Promega)according to the manufacturers' protocols. The cells were treatedwith or without MG132 (10^–6 M) and estrogen (10^–8M). Twenty-four hours after the addition of estrogen, the cellswere harvested and analyzed by Western blotting using the appropriateantibodies.

Coimmunoprecipitation and Western blotting. The 293 cells were transfected with the appropriate plasmidsand lysed in TNE buffer (10 mM Tris-HCl [pH 7.8], 0.5% NonidetP-40 [NP-40], 0.15 M NaCl, 1 mM EDTA, 1 µM phenylmethylsulfonylfluoride [PMSF], 1 µg/ml aprotinin). Extracted proteinswere immunoprecipitated with antibody-coated protein A/G-Sepharose(Amersham) or anti-FLAG M2 agarose (Sigma). The bound proteinswere separated by sodium dodecyl sulfate-polyacrylamide gelelectrophoresis (SDS-PAGE), transferred onto polyvinylidenedifluoride membranes (Millipore), and detected with the appropriateantibodies and secondary antibodies conjugated with horseradishperoxidase. Specific proteins were detected using an enhancedchemiluminescence Western blot detection system (Amersham).

Luciferase assay. For luciferase assays, ERE-TATA-Luc plasmids were cotransfectedwith expression vector encoding ERß or its mutants.As a reference plasmid with which to normalize transfectionefficiency, the pRSVßGAL vector or the pRL-CMV vectorwas cotransfected in all experiments. Twenty-four hours aftertransfection, the culture medium was replaced with fresh mediumcontaining 0.2% FBS, either estrogen (10^–8 M) or ethanolvehicle was added, and the cells were incubated for an additional24 h. Cell extracts were prepared, and luciferase assays wereperformed following the manufacturer's protocol (Promega). ß-Galactosidaseactivity was measured to control for the efficiency of eachtransfection. Individual transfections, each consisting of triplicatewells, were repeated at least three times.

Protein purification. Glutathione S-transferase (GST) protein or an immobilized 37-amino-acidregion that we designated the "domain required for degradation"(DRD) fused to GST (GST-DRD) was preincubated for 1 h at 4°Cin GST-binding buffer (20 mM Tris-HCl [pH 7.9], 180 mM KCl,0.2 mM EDTA, 0.5 mM PMSF, and 1 mM dithiothreitol) containingbovine serum albumin (BSA; 1 mg/ml). Bead-immobilized proteinswere then incubated at 4°C for 6 to 10 h with 293 cell extracts.After the beads had been washed three times with GST buffer(GST-binding buffer with 0.1% NP-40), they were washed againwith GST buffer containing 0.2% N-lauroyl sarcosine. Proteinsbound to the DRD were eluted with 15 mM reduced glutathionein elution buffer (50 mM Tris-HCl [pH 8.3], 150 mM KCl, 0.5mM EDTA, 0.5 mM PMSF, 5 mM NaF, 0.08% NP-40, 0.5 mg/ml BSA,and 10% glycerol). Recombinant CHIP was expressed in Escherichiacoli and purified using Ni-agarose.

Immunofluorescence. The 293 cells were grown on poly-L-lysine-coated 12-well culturedishes and transfected with plasmids. Twenty-four hours aftertransfection, the cells were fixed with 4% paraformaldehydein phosphate-buffered saline (PBS) for 10 min and permeabilizedwith Triton buffer (50 mM Tris-HCl [pH 7.5], 0.5% Triton X-100,150 mM NaCl, 2 mM EDTA) for 15 min. The cells in each well wereblocked with PBS containing 1% BSA and 0.5% goat serum for 3h at 37°C. The cells were incubated with the appropriateantibody in PBS containing 1% BSA for 2 h at 37°C. Afterthe cells had been washed with PBS, they were incubated withAlexa-Fluor-488-conjugated goat anti-rat immunoglobulin G orAlexa-Fluor-594-conjugated goat anti-mouse immunoglobulin G(Molecular Probes) for 1 h at 37°C and washed with PBS.The samples were mounted with Vectashield mounting medium (VectorLaboratories) and analyzed using a Keyence Biozero BZ-8000.

Pulse-chasing. The 293 cells were transfected with ERß, and 48 hafter transfection the cells were labeled for 30 min at 37°Cwith [³⁵S]methionine (50 µCi/ml) in methionine-free DMEM.The cells were then washed twice and incubated in DMEM containing0.2% FBS for the indicated periods (chase). At each time point,cell lysates were immunoprecipitated with anti-ERßantibody. The immunoprecipitates were resolved by SDS-PAGE andvisualized by autoradiography. A phosphorimager was used toquantify the metabolically labeled ERß present ateach time point.

RNAi. The 293gp cells were transfected with pSINsi-hU6 vector (Takara)containing the target sequence of CHIP or LacZ (control) andwith vesicular stomatitis virus-encoded envelope protein G togenerate the small interfering RNA (siRNA) retroviral supernatant.MDA-MB-231 cells were transfected with retroviral supernatantin the presence of Polybrene (8 µg/ml). Twenty-four hoursafter transfection, the viral supernatant was replaced withfresh DMEM containing 10% FBS. The transfected cells were selectedwith G418 (1 mg/ml). The target sequences were 5'-GCACGACAAGTACATGGCGGA-3'for CHIP and 5'-GCTACACAAATCAGCGATT-3' for LacZ.

RESULTS

Top

Results

An N-terminal 37-amino-acid region is essential for ERß degradation. The level of ERß protein was reduced by the additionof estrogen to the MDA-MB-231 breast cancer cell line. Thisdegradation was inhibited by the proteasome inhibitors MG132(Fig. 1A) and lactacystin (data not shown). In a ubiquitinationassay, ERß was ubiquitinated in both the presenceand the absence of estrogen (Fig. 1B). Thus, ERß isdegraded via ubiquitin-proteasome pathways. Preliminary biochemicalstudy showed that a single K48-linked polyubiquitin chain issufficient to target a substrate to the 26S proteasome. Therefore,we investigated the ubiquitin linage of the chains using a ubiquitinmutant lacking specific lysines, Ub(K48R). HA-tagged ubiquitin(HA-Ub) or Ub(K48R) was transfected into 293 cells, and thelevel of ERß protein and the ubiquitination statuswere determined by Western blotting using an anti-FLAG-M2 oran anti-HA antibody. Expression of Ub(K48R) reduced the degradation(Fig. 1C) and the ubiquitination status (Fig. 1D) of ERß,suggesting that the polyubiquitin chains of ERß arelinked via K48. These results indicate that whereas ERßis polyubiquitinated via K48 in ubiquitin in both the presenceand the absence of estrogen, the proteolysis is estrogen dependent.Our observations imply that ubiquitination in itself is notthe only way to regulate ERß.

View larger version (37K):
[in this window]
[in a new window]
FIG. 1. ERß is degraded through ubiquitin-proteasome pathways. (A) ERß is degraded in an estrogen-dependent manner. ERß-transfected MDA-MB-231 cells were cultured in the presence or the absence of estrogen (10^–8 M) and the proteasome inhibitor MG132 (10^–6 M). ERß level was analyzed by use of Western blots probed with the indicated antibody. (B) ERß is degraded via ubiquitin-proteasome pathways. ERß-transfected MDA-MB-231 cells were cultured in the presence or the absence of estrogen (10^–8 M) and MG132 (10^–6 M). ERß was immunoprecipitated (IP) using anti-ERß antibody. Upper panel: the ubiquitination status of ERß was analyzed by Western blotting (IB) and probed with an anti-ubiquitin antibody. Lower panel: immunoprecipitated ERß was detected by Western blotting and probed with an anti-ERß antibody. IgG, immunoglobulin G; M, molecular mass. (C) The estrogen-dependent degradation of ERß is inhibited by mutated ubiquitin Ub(K48R). FLAG-tagged ERß and DNA encoding HA-Ub (1,000 ng) or HA-Ub(K48R) (1,000 ng) were transfected into 293 cells in the presence or the absence of estrogen (10^–8 M). ERß was detected by use of Western blots probed with anti-FLAG-M2 antibody. (D) The polyubiquitin chain of ERß is linked via K48. FLAG-tagged ERß and DNA encoding HA-Ub (1,000 ng) or HA-Ub(K48R) (1,000 ng) were transfected into 293 cells in the presence or the absence of estrogen (10^–8 M) and MG132 (10^–6 M). FLAG-tagged ERß was immunoprecipitated with anti-FLAG-M2 antibody. Upper panel: the ubiquitination status of ERß was analyzed by use of Western blots probed with anti-HA antibody. Lower panel: immunoprecipitated ERß was detected by use of Western blots probed with anti-FLAG antibody. (E) The A/B domain is necessary for the estrogen-dependent degradation of ERß. DNA (500 ng) encoding the indicated FLAG-tagged ER deletion mutants was transfected into 293 cells. These cells were cultured in the presence or the absence of estrogen (10^–8 M) and MG132 (10^–6 M). To evaluate the protein levels of the ER mutants, Western blot analysis was performed using anti-FLAG-M2 antibody.

To investigate the molecular mechanism by which ERßis degraded, DNA sequences encoding truncated forms of ER ${alpha}$ orERß were transfected into 293 cells, and the proteinlevels were examined by Western blot analysis. ER ${alpha}$ ( ${Delta}$ 181), whichdoes not contain the A/B domain, still underwent ligand-dependentdegradation (Fig. 1E). In contrast, truncation of the A/B domainof ERß, forming ERß( ${Delta}$ 140), produced no ligand-dependentdegradation, suggesting that the A/B domain of ERßis necessary for its ligand-dependent degradation (Fig. 1E).

We then investigated the region responsible for the degradationof ERß. The levels of truncated ERß proteinin the presence and the absence of estrogen were examined byWestern blotting. As shown in Fig. 2A, ERß( ${Delta}$ 20), inwhich the 20 N-terminal amino acids are deleted, was elevatedin the presence of estrogen compared with wild-type ERß.ERß( ${Delta}$ 37) exhibited almost no degradation, indicatingthat the N-terminal 37 amino acids are essential for the ligand-dependentdegradation of ERß (Fig. 2A). To assess whether thisregion is sufficient for the ligand-dependent degradation ofERß, it was fused to ERß( ${Delta}$ 140), which doesnot contain the A/B domain and shows little ligand-dependentdegradation. In contrast to that of ERß( ${Delta}$ 140), thelevel of ERß( ${Delta}$ 38-140) protein, in which the 37-amino-acidregion is fused to ERß( ${Delta}$ 140), was reduced in an estrogen-dependentmanner (Fig. 2A). This reduction was abrogated by the additionof MG132, indicating that ERß( ${Delta}$ 38-140) is degradedby proteasomal pathways. Interestingly, the degradation of ERß(mC),which has three amino acid substitutions in the DNA-bindingdomain (C domain) and has almost no ability to bind DNA (30),was similar to that of wild-type ERß (Fig. 2A), indicatingthat binding to the DNA element is not necessary for ligand-dependentreceptor degradation. Similar results were obtained when othercell lines, such as MCF-7 or MDA-MB-231, were used (data notshown). From these observations, we concluded that the N-terminal37-amino-acid region of ERß is involved in its ligand-dependentdegradation and designated this region the DRD. We next determinedthe ubiquitination status of the truncated forms of ERß.The DRD deletion accumulated a monoubiquitinated form and showeda reduced level of the polyubiquitinated form (Fig. 2B), suggestingthat the DRD is required for the polyubiquitination of ERß.

View larger version (38K):
[in this window]
[in a new window]
FIG. 2. DRD-mediated degradation is involved in the downregulation of ERß-mediated transcription. (A) The N-terminal 37-amino-acid region is essential for estrogen-mediated ERß degradation. DNA sequences (500 ng) encoding the indicated FLAG-tagged ERß deletion mutants were transfected into 293 cells. These cells were cultured in the presence or the absence of estrogen (10^–8 M) and MG132 (10^–6 M). To evaluate the protein levels of the ERß mutants, Western blot analysis was performed using an anti-FLAG-M2 antibody. (B) DRD-regulated ubiquitination of ERß. DNA (500 ng) encoding the indicated FLAG-tagged ERß mutants and DNA (1,000 ng) encoding HA-tagged ubiquitin were transfected into 293 cells in the presence or the absence of estrogen (10^–8 M). Upper panel: FLAG-tagged ERß mutants were immunoprecipitated (IP) with anti-FLAG-M2 antibody. The ubiquitination status of ERß was analyzed by use of Western blots probed with anti-HA antibody. Lower panel: immunoprecipitated ERß mutants were detected by use of Western blots probed with an anti-FLAG-M2 antibody. I.B., Western blotting.

Purification and identification of an E3 ubiquitin ligase that specifically binds to the DRD of ERß. To investigate why the DRD induces ubiquitination, we examinedtwo possibilities. First, the DRD may contain a lysine residuethat is ubiquitinated by a ubiquitin ligase specific for ERß.Second, there may be a ubiquitin ligase that specifically recognizesand binds to the DRD to ubiquitinate ERß.

To test the first possibility, we introduced an amino acid substitutionat the lysine residue in the DRD. ERß(K4A), in whichthe lysine residue in the DRD is replaced by alanine, showedligand-dependent degradation similar to that of the wild type(Fig. 2A), implying that the lysine residue in the DRD is nota target for ubiquitination. Therefore, we investigated theother possibility.

We generated a GST-fused 37-amino-acid region (GST-DRD) withwhich to purify proteins that specifically bind to the DRD.Whole-cell extracts prepared from 293 cells were incubated witheither GST-DRD or GST. The bound proteins were precipitatedwith glutathione beads and separated using SDS-PAGE (Fig. 3A).Mass fingerprinting methods revealed that the proteins thatspecifically bound to the DRD were the carboxyl terminus ofCHIP, HSP70, and HSP90 (Fig. 3A).

View larger version (37K):
[in this window]
[in a new window]
FIG. 3. The DRD associates with a protein complex containing the carboxyl terminus of HSC70-interacting protein (CHIP) and HSC/HSP70. (A) Purification and identification of DRD-interacting proteins. Whole-cell extracts (WCE) prepared from 293 cells were incubated with GST-DRD or with GST as the control. Bound proteins were precipitated with GST-conjugated beads, eluted with glutathione, and subjected to SDS-PAGE followed by silver staining. The fractions eluted from GST and GST-DRD are shown. Proteins eluted from the DRD peptide were examined by mass spectrometry. (B) CHIP binds the DRD. Recombinant CHIP was incubated with GST-DRD or GST. The bound proteins were precipitated with glutathione beads. CHIP bound to GST-DRD was separated by SDS-PAGE. (C) Interaction between ERß and CHIP in vivo. ERß-transfected MDA-MB-231 cells were lysed and immunoprecipitated (I.P.) with anti-ERß antibody in the presence or the absence of estrogen (10^–8 M). The precipitates were Western blotted (I.B.) and probed with antibodies directed against CHIP and HSP70. IgG, immunoglobulin G. (D) DRD is necessary for the interaction between ERß and CHIP. ERß or ERß ${Delta}$ DRD was transfected into 293 cells. These cells were lysed and immunoprecipitated with anti-FLAG-M2 antibody. The precipitates were Western blotted and probed with antibodies directed against CHIP and HSP70. (E) CHIP induces the ubiquitination of ERß. DNA (500 ng) encoding FLAG-tagged ERß and DNA (1,000 ng) encoding HA-tagged ubiquitin were transfected into 293 cells with or without DNA (50 ng) encoding CHIP in the presence or the absence of estrogen (10^–8 M). FLAG-tagged ERß or ERß( ${Delta}$ DRD) was immunoprecipitated using anti-FLAG-M2 antibody. The ubiquitination status of ERß was analyzed by use of Western blots probed with anti-HA antibody.

CHIP has E3 ubiquitin ligase activity mediated by its carboxyl-terminalU-box domain and has the ability to bind to chaperones HSP/HSC70via its tetratricopeptide repeat (TPR) domain (7, 48). A pull-downassay showed that recombinant CHIP bound to the GST-DRD (Fig.3B). A coimmunoprecipitation assay demonstrated specific bindingof CHIP and HSP70 to ERß in both the absence and thepresence of estrogen (Fig. 3C). ERß ${Delta}$ DRD showed no abilityto bind CHIP (Fig. 3D). Therefore, we tested whether CHIP enhancesERß ubiquitination. When ERß was coexpressedwith CHIP, smeary bands of ubiquitin-conjugated ERßproducts were observed (Fig. 3E). In contrast, CHIP expressionhad no effect on the ubiquitination status of ERß( ${Delta}$ DRD).Thus, CHIP appears to mediate the ubiquitination of ERß.

CHIP participates in the degradation of ERß. ERß is known to localize predominantly in the nucleus.Although a large proportion of the CHIP protein was localizedin the cytoplasm, a small amount was observed in the nucleusby immunostaining (Fig. 4A) and Western blot analysis (Fig.4B, left panel). To test whether CHIP-mediated ubiquitinationinduces the degradation of ERß, DNA encoding ERßwas transfected into 293 cells with or without DNA encodingCHIP. Western blot analysis revealed that the steady-state levelof ERß decreased in both the nucleus and the cytoplasmwhen CHIP was produced (Fig. 4B, left panel). ERßin MDA-MB-231 cells also decreased with the production of CHIP(Fig. 4B, right panel). We further determined CHIP's functionby producing MDA-MB-231 cells in which endogenous CHIP expressionwas suppressed by the introduction of a siRNA complementaryto sequences present in the CHIP mRNA. The introduction of thesiRNA vector into MDA-MB-231 cells suppressed the expressionof CHIP mRNA (data not shown) and protein (Fig. 4C, left panel)and the accumulation of ERß protein (Fig. 4C, rightpanel). In contrast, a control vector failed to alter the CHIPor ERß protein levels (Fig. 4C, right panel). Thus,CHIP appears to be involved in estrogen-dependent ERßdegradation. To confirm this, pulse-chase experiments were performed.In the presence of estrogen, the half-life of ERßexceeded 8 h, but it had a half-life of about 4 h when CHIPwas expressed (Fig. 4D).

View larger version (19K):
[in this window]
[in a new window]
FIG. 4. CHIP enhances the degradation of ERß. (A) CHIP protein was observed in both the cytoplasm and the nucleus. The 293 cells were transiently transfected with plasmids encoding HA-tagged CHIP and FLAG-tagged ERß. Mounted cells were examined by immunofluorescence microscopy as described in Materials and Methods. (a) The distribution of CHIP in the cell is shown. (b) The distribution of ERß is shown. (c) Merged images of panels a and b. (B) CHIP expression reduced the steady-state level of ERß. DNA (50 ng) encoding HA-tagged CHIP was cotransfected into 293 or MDA-MB-231 cells with DNA (500 ng) encoding FLAG-tagged ERß in the absence or the presence of estrogen (10^–8 M). The level of ERß protein was examined by use of Western blots probed with anti-FLAG (293) or anti-ERß (MDA-MB-231) antibody. (C) ERß accumulated after the siRNA-mediated suppression of endogenous CHIP. Plasmids encoding a siRNA specific for CHIP or the control vector were introduced into MDA-MB-231 cells. These cells were selected with G418. Levels of CHIP and ERß proteins were assessed by immunoblotting the whole-cell lysates with specific antibodies, as indicated. (D) CHIP facilitates the degradation of ERß. 293 or MDA-MB-231 cells were transfected with DNA (50 ng) encoding CHIP or with DNA (500 ng) encoding FLAG-tagged ERß in the presence of estrogen (10^–8 M). Cells were pulse-labeled with [³⁵S]methionine and then chased for the indicated times in medium containing unlabeled methionine. ³⁵S-labeled ERß in the immunoprecipitate was quantified by phosphorimaging, and the levels in control cells (closed circles) and CHIP-expressing cells (open circles) were plotted relative to the amount present at time zero.

To confirm that the DRD is necessary for CHIP-dependent ERßdegradation, DNA sequences encoding truncated forms of ERßwere transfected into 293 cells with or without DNA encodingCHIP. CHIP expression enhanced the degradation of ERß,ERß(mC), ERß(DRD; ${Delta}$ 140), and ERß(AF-1),all of which contain the DRD, but did not enhance the degradationof ERß( ${Delta}$ DRD) or ERß( ${Delta}$ 140), neither of whichcontains the DRD (Fig. 5A). To test the specificity of thiseffect, we created constructs in which the TPR and U-box domainsof CHIP were deleted ( ${Delta}$ TPR and ${Delta}$ U-box, respectively). CHIP bindsto HSP/HSC70 via its TPR motif and displays E3 ubiquitin ligaseactivity mediated by its U-box domain. Although the expressionof these truncated proteins was similar to that of full-lengthCHIP (data not shown), deletion of either of these domains abolishedthe effects of CHIP on ERß (Fig. 5B). Such a requirementfor a TPR motif indicates that CHIP may have to interact withHSP/HSC70 to promote ERß degradation. The functionalrequirement for the U-box confirms that CHIP regulates ERßubiquitination.

View larger version (30K):
[in this window]
[in a new window]
FIG. 5. CHIP regulates DRD-mediated ubiquitination and degradation. (A) The DRD is necessary for CHIP-dependent ERß degradation. DNA (500 ng) encoding the indicated FLAG-tagged ERß mutants was transfected into 293 cells, with or without DNA encoding CHIP. These cells were cultured in the presence or the absence of estrogen (10^–8 M). To evaluate the levels of ERß mutant proteins, Western blot analysis was performed using anti-FLAG-M2 antibody. (B) Both the TPR and U-box domain of CHIP are necessary for ERß degradation. DNA (50 ng) encoding CHIP, CHIP ${Delta}$ TPR, or CHIP ${Delta}$ Ubox was transfected into 293 cells with or without DNA (500 ng) encoding FLAG-tagged ERß. Levels of ERß protein were examined by use of Western blots probed with anti-FLAG-M2 antibody. (C) The CHIP-dependent ERß degradation is not affected by ER ${alpha}$ . DNA (500 ng) encoding FLAG-tagged ERß was transfected into 293 cells with or without DNA (500 ng) encoding FLAG-tagged ER ${alpha}$ and DNA (50 ng) encoding CHIP. Levels of ERß and ER ${alpha}$ protein were examined by use of Western blots probed with anti-FLAG-M2 antibody.

In a previous report, we showed that CHIP binds ER ${alpha}$ in the absenceof estrogen (48). Since ER ${alpha}$ and ERß heterodimerizereadily in the presence of each other, we examined the effectof cointroducing ER ${alpha}$ on CHIP-dependent degradation of ERß.As shown in Fig. 5C, the degradation of ERß was notaffected by coexpression of ER ${alpha}$ .

The F domain regulates ERß degradation negatively. Our results showed that whereas CHIP binds and ubiquitinatesERß in both the presence and the absence of estrogen,the degradation of ERß is estrogen dependent. To investigatethis discrepancy, DNA sequences encoding truncated forms ofERs were transfected into 293 cells and the protein levels wereexamined. The AD core domain is necessary for estrogen-dependentdegradation of ER ${alpha}$ (29, 48). Consistent with this observation,ER ${alpha}$ ( ${Delta}$ AD) was stabilized by ligand binding and thus accumulatedin response to estrogen (Fig. 6A). Interestingly, the deletionof the AD core of ERß resulted in a dramatic reductionin ERß in the absence of estrogen (Fig. 6A). Thisobservation raises the possibility that the C-terminal regionof ERß is necessary for the stabilization of ERß.To test this possibility, DNA sequences encoding deletion mutantsof the C terminus of ERß were transfected into 293cells and levels of expressed protein were determined by Westernblot analysis. When the F domain was deleted [ERß( ${Delta}$ F)],a dramatic reduction in ERß protein was observed inthe absence of estrogen (Fig. 6B). This reduction in ERßwas inhibited by MG132, indicating that the F domain suppressesthe proteasome-dependent degradation of ERß. The degradationof ERß( ${Delta}$ F) required the DRD, because ERß( ${Delta}$ DRD; ${Delta}$ F)was not degraded in the absence of estrogen. In contrast, thedegradation of ERß( ${Delta}$ 140; ${Delta}$ F) was enhanced by fusion withthe DRD [ERß(DRD; ${Delta}$ 140; ${Delta}$ F)] in both the presence andthe absence of estrogen (Fig. 6B). These results suggest thatthe F domain suppresses the DRD-dependent degradation of ERßin the absence of estrogen.

View larger version (42K):
[in this window]
[in a new window]
FIG. 6. The F domain suppresses DRD-mediated proteolysis of ERß. (A) AD core deletion in ERß increased the degradation of ligand-unbound receptor. DNA sequences (500 ng) encoding the indicated FLAG-tagged ER deletion mutants were transfected into 293 cells. These cells were cultured in the presence or the absence of estrogen (10^–8 M) and MG132 (10^–6 M). To evaluate the protein levels of the ER mutants, Western blot analysis was performed using anti-FLAG-M2 antibody. (B) F domain protects ERß from DRD-mediated degradation. DNA sequences (500 ng) encoding the indicated FLAG-tagged ERß deletion mutants were transfected into 293 cells cultured in the presence or the absence of estrogen (10^–8 M) and MG132 (10^–6 M). To evaluate the protein levels of the ERß mutants, Western blot analysis was performed using anti-FLAG-M2 antibody.

In contrast to what was seen for ERß, the deletionof the F domain in ER ${alpha}$ did not affect its degradation. To studythis, we generated a chimeric ERß, ERß ( ${alpha}$ F),in which the F domain was substituted with an ER ${alpha}$ F domain. However,there is almost no homology between the ER ${alpha}$ F and ERßF domains, and the F domain of ER ${alpha}$ was able to protect ERßfrom proteolysis in the absence of estrogen (Fig. 6B).

Next, we determined the ubiquitination statuses of ERßand ERß( ${Delta}$ F). That of ERß( ${Delta}$ F) was similar tothat of full-length ERß (Fig. 7A), indicating thatthe F domain does not regulate the ubiquitination step of ERß.Thus, the F domain may inhibit the accession of 26S proteasometo ERß structurally. Therefore, we evaluated the bindingbetween ERß and proteasomes. FLAG-tagged ERßor ERß( ${Delta}$ F) was transfected into 293 cells with SUG-1,which is one of the components of regulatory subunit of 26Sproteasome. A coimmunoprecipitation experiment revealed thatdeletion of the F domain enhanced association between ERßand SUG-1 (Fig. 7B). In addition, SUG-1 efficiently reducedERß( ${Delta}$ F) rather than ERß (Fig. 7C). Thus,in the absence of estrogen, the F domain appears to abrogatethe binding of the 26S proteasome to ERß to protectthe receptor from proteolysis.

View larger version (43K):
[in this window]
[in a new window]
FIG. 7. F domain abrogates the association of proteasomes with ERß. (A) Deletion of the F domain does not alter the ubiquitination status of ERß. DNA sequences (500 ng) encoding the indicated FLAG-tagged ERß mutants and DNA (1,000 ng) encoding HA-tagged ubiquitin were transfected into 293 cells in the presence or the absence of estrogen (10^–8 M). Upper panel: FLAG-tagged ERß mutants were immunoprecipitated (I.P.) with anti-FLAG-M2 antibody. The ubiquitination status of ERß was analyzed by use of Western blots probed with anti-HA antibody. I.B., Western blotting. Lower panel: immunoprecipitated ERß mutants were detected by use of Western blots probed with anti-FLAG antibody. (B) The F domain abrogates the association of SUG-1 with ERß. DNA sequences (500 ng) encoding either ERß or ERß( ${Delta}$ F) were transfected into 293 cells. FLAG-tagged ERß or ERß mutants were immunoprecipitated with anti-FLAG-M2 antibody. The precipitates were Western blotted and probed with an antibody against SUG-1. T.F., transfection. (C) The F domain protects ERß from SUG-1-dependent degradation. DNA sequences (100 ng) encoding SUG-1 were cotransfected into 293 cells with or without DNA sequences (500 ng) encoding FLAG-tagged ERß or ERß( ${Delta}$ F). The levels of ERß or ERß( ${Delta}$ F) protein were examined by Western blotting and probed with anti-FLAG antibody.

Receptor degradation is involved in the downregulation of transcription of the gene for ERß after ligand withdrawal. Recent evidence indicates that receptor degradation is coupledto transactivation and that the transcriptional activity ofthe gene for ER ${alpha}$ is abrogated by the addition of MG132. The transcriptionalactivity of ERß was also reduced by MG132 (Fig. 8A).Thus, it is likely that receptor degradation is necessary forthe ligand-dependent transactivation of the ERß gene.To investigate the association between CHIP-mediated receptordegradation and transactivation, we evaluated the transcriptionalactivities of ERß deletion mutants by use of a transientreporter assay. As shown in Fig. 8B, ERß( ${Delta}$ DRD) exhibiteda transcriptional activity higher than that of full-length ERß.However, ERß( ${Delta}$ 140) still showed transcriptional activity.When the DRD was fused to ERß( ${Delta}$ 140) to form ERß(DRD; ${Delta}$ 140),its transcriptional activity was reduced (Fig. 8B). Similarly,ERß( ${Delta}$ 140; ${Delta}$ F) exhibited higher transcriptional activitythan ERß( ${Delta}$ F), but fusion of DRD decreased its transcriptionalactivity (Fig. 8B). In agreement with these results, CHIP expressiondownregulated the transcriptional activity of the ERßgene (Fig. 8C). In contrast, CHIP siRNA enhanced transcriptionalactivity (Fig. 8C). The transcriptional activity of ERß( ${Delta}$ DRD)was not affected by expression of CHIP (Fig. 8C). Overall, ourresults indicate that CHIP-mediated degradation is not necessaryfor receptor-mediated transactivation. In addition, CHIP isable to ubiquitinate and to degrade ERß(mC) (Fig.5A), indicating that binding to the DNA element is not necessaryfor CHIP-mediated receptor degradation. These results show thatlarge amounts of liganded ERß protein are selectivelyand rapidly degraded independently of receptor-mediated transcriptionby CHIP.

View larger version (29K):
[in this window]
[in a new window]
FIG. 8. CHIP decreases transcriptional activity of ERß by degradation. (A) Proteasome-dependent degradation is necessary for the transactivation of ERs. DNA sequences (0.05 ng) encoding the ERß or ER ${alpha}$ were cotransfected into 293 cells with ERE-TATA-Luc (50 ng) and pRSVßGAL (50 ng) in the presence or absence of MG132 (10^–6 M), and the cell extracts were used in a luciferase assay. (B) DRD-mediated degradation is not coupled to ligand-dependent receptor transactivation. DNA sequences (0.05 ng) encoding the indicated ERß mutants were cotransfected into 293 cells with ERE-TATA-Luc (50 ng) and pRSVßGAL (50 ng), and the cell extracts were used in a luciferase assay. (C) CHIP production reduced the transcriptional activity of ERß. ERE-TATA-Luc (50 ng), pRSVßGAL (50 ng), and DNA sequences (0.05 ng) encoding either ERß or ERß( ${Delta}$ DRD) were cotransfected into 293 cells with or without DNA encoding CHIP and CHIP RNAi, and the cell extracts were used in a luciferase assay.

To investigate the biological significance of this "wastefulness,"we hypothesized that CHIP-mediated degradation selectively andrapidly degrades the active form of ERß to shut offERß-mediated transcription when the estrogen doseis reduced. To test this hypothesis, we measured the transcriptionalactivities of constructs encoding ERß and ERß( ${Delta}$ DRD)after estrogen withdrawal. DNA sequences encoding either ERßor ERß( ${Delta}$ DRD) were transfected into 293 cells togetherwith a plasmid in which the PEST sequence was fused to the luciferasegene (ERE-TATA-LucCP), and the luciferase activity was measuredat various time points after estrogen withdrawal. As shown inFig. 9A, the transcriptional activity of the ERß sequencewas reduced much faster than that of ERß( ${Delta}$ DRD). UsingMDA-MB-231 cells that stably express either control siRNA orCHIP siRNA, we tested whether CHIP is involved in the downregulationof ERß-mediated transcription after reduction of theligand dose. Plasmids containing the PEST-fused luciferase genewere transfected into MDA-MB-231 cells expressing either controlsiRNA or CHIP siRNA. ERß gene-mediated transcriptionwas downregulated faster in the MDA-MB-231 cells expressingcontrol siRNA than in the cells expressing CHIP siRNA (Fig.9A).

View larger version (29K):
[in this window]
[in a new window]
FIG. 9. CHIP switches off ERß-mediated transcription after the estrogen dose is reduced. (A) Downregulation of CHIP inhibits the turning off of ERß-mediated transcription after estrogen withdrawal. Upper panel: DRD-mediated degradation is involved in the downregulation of ERß-mediated transcription after estrogen withdrawal. The 293 cells were transfected with expression plasmids encoding FLAG-tagged ERß or ERß( ${Delta}$ DRD), together with ERE-TATA-LucCP (50 ng) and pRSVßGAL (50 ng). Twenty-four hours after transfection, the cells were treated with estrogen (10^–8 M) for an additional 24 h, after which the estrogen was withdrawn (time 0). These cells were harvested for luciferase assays and Western blotting at the indicated time points after estrogen withdrawal. Lower panel: CHIP is involved in the downregulation of ERß-mediated transcription when the ligand dose is reduced. Plasmids encoding a siRNA specific for CHIP or control vector were introduced into MDA-MB-231 cells. These cells were selected with G418. FLAG-tagged ERß (100 ng), ERE-TATA-LucCP (400 ng), and pGL3-Control (10 ng) were transfected into these cell lines. Twenty-four hours after transfection, the cells were treated with estrogen (10^–8 M) for an additional 24 h, after which the estrogen was withdrawn (time 0). These cells were harvested for luciferase assays and Western blotting at the times indicated after estrogen withdrawal. (B) Dose-response curve of the estrogen-induced CHIP-dependent degradation of the ERß. DNA (500 ng) encoding the FLAG-tagged ERß was transfected into 293 cells, with or without DNA encoding CHIP. These cells were cultured in the presence of the indicated concentration (conc.) of estrogen (from 10^–11 M to 10^–7 M). Levels of ERß protein were examined by use of Western blots probed with anti-FLAG-M2 antibody.

The serum estrogen concentration varies from 1.5 x 10^–10M to 1.8 x 10^–9 M during the menstrual cycle. Therefore,we investigated the dose response of the CHIP-dependent degradationof ERß. As shown in Fig. 9B, CHIP-dependent ERßdegradation was changed in the levels of estrogen in the bloodduring the menstrual cycle.

Overall, these results suggest that CHIP-mediated degradationselectively and rapidly eliminates the active form of the NRto downregulate receptor-mediated transcription after ligandwithdrawal.

DISCUSSION

Top

Discussion

In this study, we have shown that ERß is ubiquitinatedand degraded in an estrogen-dependent manner. We identifiedtwo regions within ERß that are essential for theregulation of ERß ubiquitination and degradation.The N-terminal 37-amino-acid region, the DRD, is necessary forthe polyubiquitination of ERß. Thus, ERßseems to be constitutively exposed to attack by ubiquitin ligasesbecause it contains the DRD. The C-terminal F domain abrogatesDRD-dependent ERß degradation, and the ubiquitinationstatus of ERß( ${Delta}$ F) was similar to that of full-lengthERß. In addition, coimmunoprecipitation experimentsrevealed that the F domain inhibits the binding of unligandedERß to the regulatory subunit of 26S proteasome. Thus,the F domain may regulate not ubiquitination itself but thestep involving proteasomal degradation. When estrogen bindsto ERß, the conformation of the C-terminal region,including the F domain, is changed, and the protective effectof the F domain is abolished to induce degradation. Some membersof the NR superfamily have an F domain, but its function isunknown. Our observations provide the first evidence that theF domain of ERß is involved in the regulation of itsproteasome-dependent degradation.

The activation of NRs appears to be coupled to their degradationby the ubiquitin-proteasome pathway (3, 4, 8). Furthermore,several components of the ubiquitin-proteasome pathway, includingSUG1/TRIP1 (17, 27, 32), RSP5/RPF1 (23), E6-AP (37), and UBC9(42), have been identified as NR-interacting proteins. TBL1/TBLR1,which are components of the NR corepressor complex, also mediatecorepressor/coactivator exchange by the ligand-dependent recruitmentof the ubiquitin-19S proteasome complex (40). Thus, the ubiquitin-proteasomepathway appears to play an important role in regulating NR levelsand restricting the duration and magnitude of receptor activityin response to ligands. Furthermore, proteasome-dependent degradationof the steroid receptor requires transcriptional activity, suggestingthat receptor degradation and receptor transactivation are mutuallyinterdependent.

Several recent studies have focused on the involvement of theubiquitin-proteasome pathway in the estrogen-dependent transactivationof ER ${alpha}$ (14, 29, 36, 44, 48, 50). Moreover, ER ${alpha}$ and its coactivatorscycle onto and off of estrogen-responsive promoters in ligand-dependentmanners (35, 44, 45). In this process, ER ${alpha}$ is reportedly ubiquitinatedafter each round of transcription, facilitating its releasefrom the promoter, which may be essential for subsequent ER ${alpha}$ -mediatedtranscription. We have shown that there is a degradation pathwayfor ERß that is not coupled to transcription. UsingERß( ${Delta}$ DRD), we also have shown that transcription-uncoupledreceptor degradation is necessary to abolish transcription viathe proteolysis of the liganded receptor when the ligand doseis reduced. ERß( ${Delta}$ DRD) still exhibited slight ligand-dependentdegradation, indicating that there may be other degradationpathways for ERß. These pathways could be coupledto the transactivation of ERß.

To investigate the ubiquitin-proteasome pathway for ERß,we purified the ubiquitin ligase complex that specifically bindsto the DRD and identified a protein complex containing CHIP.CHIP was first reported to induce the ubiquitination of GR boundto HSP90 for proteasomal degradation (7). Recent observationshave indicated that CHIP targets a number of HSP70/90-associatedproteins for ubiquitination and degradation (15, 18, 21, 22,28, 34, 48, 52-54). We (48) and others (9) have previously shownthat CHIP binds unliganded ER ${alpha}$ as a protein complex containingHSP90, HSC70, HSP70, HSP40, BAG-1, and BAG-2, all of which possessor assist chaperone functions, and a DnaJ-like protein, KIAA0678.CHIP preferentially induces the hydrolysis of abnormal or mutantforms of ER ${alpha}$ , acting as an E3 ubiquitin ligase capable of distinguishingthe nonnative states from the native states of the receptorin vivo.

Our results obtained from the analysis of ERß suggestthat CHIP is involved both in the "quality control" of receptorproteins and in switching off receptor-mediated transcription(Fig. 10). The response of endocrine target tissues to hormonesis tightly regulated and is dependent upon circulating levelsof available hormonal ligands. When the dose of circulatingligand is reduced, receptor-mediated transcription should bedownregulated. The dissociation rate of estrogen from its receptoris very low because a coactivator complex contacts the proteinsurface above the ligand pocket (16). Thus, it is difficultto cause rapid downregulation by ligand dissociation from itsreceptor. Our results provide the first evidence that receptordegradation is involved in the cessation of transcription whenthe ligand dose is reduced. Several nuclear receptors are reportedto exhibit agonist-dependent degradation. Therefore, it is possiblethat other forms of receptor-mediated transcription are downregulatedby the same mechanism as ERß when the agonist doseis reduced.

View larger version (27K):
[in this window]
[in a new window]
FIG. 10. Schematic representation of the degradation pathways of ER ${alpha}$ and ERß. ER ${alpha}$ binds CHIP in the absence of estrogen. When the estrogen binds ER ${alpha}$ , CHIP dissociates. ERß binds CHIP in both the absence and the presence of estrogen. In the absence of estrogen, the degradation of ubiquitinated ERß is inhibited by the F domain. When the estrogen binds ERß, the conformation of the F domain is changed to recruit 26S proteasome.

ERß is subject to proteasome-mediated degradationin the presence of estrogen, which subsequently limits transcriptionaloutput of ERß regardless of whether estrogen is removedor persistently available. As shown in Fig. 8C, CHIP downregulatesthe ERß-mediated transactivation in the presence ofestrogen. Thus, the CHIP-dependent degradation is involved bothin the downregulation of ERß activity and in switchingoff ERß-mediated transcription.

FOOTNOTES

* Corresponding author. Mailing address: Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba Science City, Ibaraki 305-8572, Japan. Phone: 81-29-853-6632. Fax: 81-29-853-4605. E-mail: junny{at}agbi.tsukuba.ac.jp.

Published ahead of print on 28 August 2006.

These authors contributed equally to this work.

REFERENCES

Top