The tetratricopeptide repeat domain of protein phosphatase 5 mediates binding to glucocorticoid receptor heterocomplexes and acts as a dominant negative mutant.

We previously identified a protein-serine phosphatase designated PP5, based on the binding of its tetratricopeptide repeat (TPR) domain to the atrial natriuretic peptide receptor (Chinkers, M. (1994) Proc. Natl. Acad. Sci. U. S. A. 91, 11075-11079). We have now identified another protein complex to which PP5 is targeted through its TPR domain. A 90-kDa protein, identified as heat shock protein 90 (hsp90) by immunoblotting, specifically co-immunoprecipitated from COS-7 cell lysates with the FLAG-tagged TPR domain of PP5. hsp90 also co-immunoprecipitated with full-length FLAG-tagged PP5 overexpressed in COS-7 cells and with endogenous PP5 from untransfected COS-7 cells or rat brain. During gel filtration, PP5 and hsp90 comigrated in a high molecular weight complex. Since glucocorticoid receptors (GR) exist as large heterocomplexes containing hsp90 bound to TPR proteins, we hypothesized that PP5 might be associated with these complexes. Consistent with this hypothesis, PP5 specifically co-immunoprecipitated with GR from mouse L cell lysates. To test the functional importance of this TPR-mediated association in living cells, we used a dominant negative PP5 mutant consisting only of its TPR domain. The mutant inhibited GR-mediated transactivation by approximately 70% in transfected CV-1 cells. This is the first evidence that the TPR proteins in steroid receptor heterocomplexes may be required for signaling in vivo.

Steroid receptors acquire the ability to bind hormone only after they are assembled into a heterocomplex containing several other proteins (1,2). The first component of the receptor heterocomplexes to be identified was heat shock protein 90 (hsp90), 1 which is important for the assembly of functional receptors both in vitro and, in yeast, in vivo (3). Based largely on in vitro reconstitution studies, it appears that receptors form hsp90 complexes by an ATP-dependent mechanism that requires hsp70 and proteins designated p60 and p23 (4,5). When p60 dissociates from its binding site on hsp90, one of the large immunophilins, such as FKBP52 or CyP-40, can bind (6), forming the mature, native receptor heterocomplex containing hsp90, an immunophilin, p23, and often some hsp70. Upon binding of hormone, the receptor can dissociate from the complex and bind to DNA (1,2). The hormone-bound receptor can then activate transcription of specific genes, leading to a biological response. Although several of the proteins in steroid receptor heterocomplexes have been shown to be important for receptor assembly in vitro (1,2,4), in general their functions are poorly understood. In particular, the normal roles of the large immunophilins and of p60, each of which associates with hsp90 through a tetratricopeptide repeat (TPR) domain (6 -9), remain to be established.
TPR domains consist of a variable number of degenerate tandem 34-amino acid repeats (10). The original studies describing the TPR motif predicted that each repeat would consist of two amphipathic alpha helices separated by a turn, and that the hydrophobic faces of these helices could interact with each other to mediate protein-protein interactions (11,12). TPR domains have been described in proteins having a variety of functions in a variety of subcellular compartments and appear to function as targeting domains, mediating specific proteinprotein interactions (reviewed in Ref. 10). While indirect evidence consistent with the model in which TPR domains interact with each other has been presented (13,14), more evidence has accumulated implicating TPR domains in binding to non-TPR proteins (10). Examples include the interactions between the TPR domains of human and yeast peroxisomal import proteins with the type 1 peroxisomal targeting signal (15,16), between the TPR domain of the Cyc8 corepressor and several associated proteins (17), and between the TPR domain of the immunophilin FKBP52 and hsp90 (7).
We and others have isolated cDNA clones encoding a proteinserine phosphatase designated PP5 (18 -20). PP5 is distinctive in that it contains an amino-terminal TPR domain that we originally isolated in a yeast two-hybrid screen, based on its ability to bind to the atrial natriuretic peptide (ANP) receptor (20). Since desensitization of the ANP receptor is associated with its dephosphorylation (21), PP5 may be the enzyme responsible for ANP receptor desensitization. In some cultured cells, PP5 is present at higher concentrations in the nucleus than in the cytoplasm (19). The biological role of PP5 in the nucleus or cytoplasm remains to be elucidated. Since the specificity of protein-serine phosphatases is largely controlled by their subcellular targeting (22), and since TPR domains may mediate protein targeting (10), biological functions of PP5 might be identified by determining to which proteins its TPR domain binds.
We report here that the TPR domain of PP5 binds to hsp90 and that this binding mediates an association of PP5 with glucocorticoid receptor (GR) heterocomplexes. We further show that a truncated PP5 containing only the TPR domain acts as a dominant negative mutant, blocking GR signaling. To our knowledge, this is the first evidence that steroid receptorassociated TPR proteins are required for hormonally induced transcriptional activation in living cells.

EXPERIMENTAL PROCEDURES
Tissue Culture-COS-7 cells and CV-1 cells were maintained in DME containing 10% fetal bovine serum. L929 mouse fibroblasts (L cells) were grown in monolayer culture in DME containing 10% iron-supplemented calf serum. Sf9 cells were maintained in Grace's medium supplemented with lactalbumin hydrolysate, Yeastolate (Life Technologies, Inc.), and 10% fetal bovine serum.
Engineering of FLAG-tagged PP5 and FLAG-tagged TPR Domain of PP5 for Expression in Mammalian and Insect Cells-The 5Ј end of rat PP5 was amplified by polymerase chain reaction using a 5Ј primer that encoded an EcoRI site, an initiation codon, the FLAG epitope (DYKD-DDDK), the first five amino acids of PP5, and a 3Ј primer complementary to nucleotides 815-832 of our rat PP5 cDNA clone (20). The polymerase chain reaction product was subcloned and sequenced from its 5Ј end to the NheI site at position 392 of the original cDNA, to ensure the absence of mutations. The 5Ј end of the FLAG-tagged PP5 was then exchanged, as an XhoI/NheI fragment, with the corresponding 5Ј fragment of the original cDNA clone. The resulting construct was cloned into the EcoRI sites of the pCMV6 vector (for expression in mammalian cells) or the pVL1393 vector (for baculovirus construction).
During the course of these studies, it became clear that the sequence at the 5Ј end of our original rat cDNA clone (20) was derived from a chimeric cDNA, and that the sequence published by Becker et al. (18) corresponded to the authentic 5Ј-coding sequences of PP5 (23). Our original cDNA clone encoded the sequence QGY immediately after the initiation codon, whereas the correct sequence is AEGERTECAEP-PRDEPP. We used oligonucleotide-directed mutagenesis to correct the amino-terminal coding sequence of the FLAG-tagged PP5. The oligonucleotide 5Ј-AAGGACGACGATGACAAGGCCGAAGGCGAACGTACTG-AATGTGCTGAACCTCCTCGAGACGAACCTCCTGCCGAAGGCACT-CTGAAGC-3Ј was used to change the amino-terminal sequence of the FLAG-tagged construct from MDYKDDDDKQGYAEGTLK, to MDYKDDDDKAEGERTECAEPPRDEPPAEGTLK (underlined residues constitute the FLAG epitope). Sequencing was performed to confirm the mutagenesis and the absence of unwanted mutations. This corrected construct, pCMV6-FLAG-PP5, was then used for the mammalian cell expression studies described below.
The FLAG-tagged PP5 construct in Bluescript SK(ϩ) (with the corrected amino terminus) was truncated at a unique HindIII site to remove sequences encoding the carboxyl portion of PP5. Blunt-ending with Klenow fragment and insertion of a linker encoding an NheI site and a stop codon resulted in a cDNA encoding the FLAG-tagged TPR domain of PP5, terminating at Leu 181 . Few non-TPR residues were present in the FLAG-TPR construct; the TPR domain begins at Ala 28 and ends at Leu 165 (using the numbering of residues described in Becker et al. (18)). The cDNA fragment encoding the FLAG-tagged TPR domain was excised as an EcoRI/SalI fragment and cloned into the corresponding sites of the pCMV6 vector (24) for mammalian cell expression.
Preparation of a Recombinant Baculovirus Expressing Epitopetagged PP5-A baculovirus for expression of FLAG-tagged PP5 was prepared by homologous recombination of pVL1393-FLAG-PP5 with BaculoGold linearized baculovirus DNA (Pharmingen) as described previously (25). Plaque-purified second passage virus stocks were used for infections.
Purification of Epitope-tagged PP5 from Sf9 Cells-A 100-ml suspension culture of Sf9 cells was infected with a recombinant baculovirus expressing FLAG-tagged PP5 at a multiplicity of infection of 3, then incubated for 3 days at 27°C. Cells were collected by centrifugation, washed once with 20 mM Hepes, pH 7.4, 150 mM NaCl, and sonicated in 5 ml of 20 mM Hepes, pH 7.4, 50 mM NaCl, 1 mM dithiothreitol, 20 mM benzamidine, 10 g/ml each aprotinin, leupeptin, and pepstatin. The homogenate was centrifuged at 12,000 ϫ g for 30 min at 4°C, and the supernatant was then centrifuged at 125,000 ϫ g for 30 min at 4°C. The supernatant from the second centrifugation was applied to a 1-ml MonoQ column at 4°C at a flow rate of 1 ml/min. Proteins were fractionated by elution with a 20-ml gradient of 50 -500 mM NaCl in 20 mM Hepes, 1 mM dithiothreitol. Peak fractions containing FLAG-tagged PP5, as determined by immunoblotting with a monoclonal antibody to the FLAG epitope (25), were pooled and further purified by immunoaffinity chromatography. The fractions were incubated batchwise with 1 ml of M2-agarose beads (IBI) for 1 h at 4°C, then transferred to a column and washed three times with 5 ml of Hepes-buffered saline containing 1 mM dithiothreitol. Epitope-tagged PP5 was then eluted by five repeated applications of 1 ml of Hepes-buffered saline, 1 mM dithiothreitol containing 200 g/ml FLAG peptide (DYKDDDDK) for 15 min each. Fractions were analyzed by immunoblotting using the M2 antibody, and peak fractions were pooled. The FLAG peptide was then removed by ultrafiltration. Approximately 200 g of purified protein were typically recovered from a 100-ml culture.
Preparation of Rabbit Antisera against PP5-After collection of preimmune serum, female New Zealand White rabbits were injected subcutaneously with 25-50 g of purified FLAG-tagged PP5. The antigen was mixed with Freund's complete adjuvant for the initial injection and with Freund's incomplete adjuvant for two booster injections given at 30-day intervals. Blood was collected and serum prepared 10 days postinjection.
Transfection-COS-7 cells in 10-cm plates were transfected with 10 g of pCMV6-FLAG-PP5, pCMV6-FLAG-TPR, or pCMV6 vector for control plates, using a DEAE-dextran procedure (26). CV-1 cells in 6-cm plates were transfected with 0.5 g of pSVL-GR (27), 2 g of the reporter plasmid PRE-PBL7 (28), 0.5 g of pRSV-luciferase (29) (to normalize for transfection efficiency), and 5 g of plasmid containing indicated amounts of pCMV6-FLAG-TPR (with the remainder of the 5 g made up of control pCMV6 vector), using a calcium phosphate procedure (30). The GR and CAT reporter plasmids were obtained from Dr. Stoney Simons (NIH). One day after transfection, CV-1 cells were rinsed with phosphate-buffered saline and incubated for 26 h in fresh growth medium, in the absence or presence of 100 nM dexamethasone.
Metabolic Labeling and Immunoprecipitation-COS-7 cells were incubated for 16 h in medium consisting of 9 parts methionine and cysteine-free DME and 1 part growth medium, containing 100 Ci/ml [ 35 S]methionine/cysteine (EXPRE 35 S 35 S protein labeling mix, DuPont NEN). Cells were then placed on ice, washed once with cold phosphatebuffered saline and once with 10 mM Hepes, pH 7.4, 1 mM EDTA, 20 mM sodium molybdate, 10 mM MgCl 2 (buffer A). Cells were then scraped into buffer A containing 2 mM phenylmethylsulfonyl fluoride and 10 g/ml each aprotinin, leupeptin, and pepstatin, and lysed by passage 30 times through a 25-gauge needle. The lysates were then subjected to centrifugation at 18,500 ϫ g for 5 min at 4°C, and the supernatant from this centrifugation was clarified by centrifugation at 100,000 ϫ g for 1 h at 4°C. After samples were normalized for total radioactivity, the clarified lysates were precleared by incubating for 1.5 h at 4°C with 20 l of goat anti-mouse-IgG beads (Sigma, for immunoprecipitation with monoclonal antibody M2 (IBI) directed against the FLAG epitope) or protein A-agarose beads (Pierce, for immunoprecipitations using anti-PP5 serum). After removing the beads by centrifugation, the precleared extracts were incubated for 1.5 h at 4°C with either 20 l of goat anti-mouse beads to which 2.25 g of M2 had been preadsorbed, or 20 l of protein A-agarose beads to which the IgG from 5 l of preimmune or anti-PP5 serum had been preadsorbed. Beads were then washed five times with 10 mM Hepes, pH 7.4, 1 mM EDTA, 20 mM sodium molybdate, 50 mM KCl, 10% glycerol, and immune complexes were released by heating in SDS sample buffer.
Samples were then analyzed by SDS-PAGE as indicated in the figure legends, followed by fluorography using sodium salicylate (31). Molecular weight markers were detected by staining with Coomassie Blue or (for immunoblots) by using Rainbow molecular weight markers (Amersham Corp.).
Immunoprecipitation and Immunoblotting-Anti-PP5 immunoprecipitates were prepared from unlabeled COS-7 cells using the methods described above, normalizing for protein before performing immunoprecipitations, and fractionated by SDS-PAGE using Rainbow molecular weight markers (Amersham Corp.) as standards. For analyzing the association of hsp90 with PP5 in rat brain, half of a rat brain was homogenized in 6 ml of ice-cold buffer A containing 10 g/ml each aprotinin and leupeptin, and 20 mM benzamidine, using three 15-s bursts with a Polytron homogenizer. The homogenate was centrifuged at 12,000 ϫ g for 10 min at 4°C. The supernatant from that centrifugation was further centrifuged at 100,000 ϫ g for 1 h at 4°C. Equal aliquots of the precleared supernatant were then subjected to immunoprecipitation with preimmune or anti-PP5 serum as described above, except using 40 l of protein A-agarose beads and 10 l of serum. Following SDS-PAGE of the washed immune complexes and blotting to nitrocellulose filters, immunoblotting to detect hsp90 was performed as described previously (25) using a 1:1000 dilution of monoclonal anti-body AC-16 (Sigma H1775) followed by incubation with peroxidaseconjugated secondary antibody and chemiluminescent detection.
For immunoprecipitation of glucocorticoid receptors from L cells, cells were harvested by scraping into Earle's balanced salt solution followed by a wash in the same buffer and centrifugation at 500 ϫ g. The washed cells were suspended in 1.5 volumes of 10 mM Hepes, 1 mM EDTA, 20 mM sodium molybdate, pH 7.4, and lysed by Dounce homogenization. Homogenates were centrifuged for 1 h at 100,000 ϫ g, and the supernatant fluid was used for immunoprecipitation. The BuGR2 monoclonal antibody against glucocorticoid receptors (Affinity Bioreagents, Golden, CO) or nonimmune mouse IgG was prebound to protein A-Sepharose pellets by incubating 40 l of a 20% slurry of protein A-Sepharose for 1 h at 4°C with 150 l of TEG buffer (10 mM TES, 50 mM NaCl, 4 mM EDTA, 10% (w/v) glycerol, pH 7.6), and 6 g of antibody, followed by centrifugation and washing with TEG. Glucocorticoid receptors were immunoadsorbed from 400 l aliquots of L cell cytosol by rotation for 2 h at 4°C with 8 l of protein A-Sepharose prebound with BuGR2 or nonimmune IgG followed by three washes with 1 ml of TEGM buffer (TEG plus 20 mM sodium molybdate). Immune complexes were fractionated by SDS-PAGE on a 12% gel, transferred to Immobilon-P membranes, and probed with 1 g/ml BuGR2 antibody for the glucocorticoid receptor, 1 g/ml of the monoclonal antibody AC88 for hsp90 (StressGen), or 1000-fold dilutions of anti-PP5 serum (above), an antiserum to the C-terminal peptide of CyP-40 (Affinity Bioreagents), or an antiserum designated UPJ56 raised against FKBP52 (32), provided by Dr. Karen Leach (The Upjohn Co.).
For immunoadsorption of hsp90 from L cell cytosol (see Fig. 4B), cytosol was diluted 10-fold with 10 mM Hepes, 1 mM EDTA, 10% glycerol, and 100-l aliquots were immunoadsorbed to 10-l pellets of Actigel-ALD precomplexed with either nonimmune IgM or with 3G3 anti-hsp90 monoclonal IgM (Affinity Bioreagents). Immune pellets were washed three times in 1 ml of TEGM buffer, and both immune pellets and the immunoadsorbed cytosols were assayed by Western blotting for hsp90 and PP5.
Gel Filtration Chromatography-A rat brain cytosolic extract was freshly prepared essentially as described above, but in a reduced volume of a different buffer (20 mM Hepes, pH 7.4, 1 mM EDTA, 3 mM MgCl 2 , 50 mM KCl), and passed through a 0.2-m filter. 200 l of the filtered extract, containing approximately 5.6 mg of protein, were then loaded onto a 25-ml Superose 6 column equilibrated with lysis buffer. The column was previously calibrated with the following standards, using the same buffer: thyroglobulin (669,000), ferritin (440,000), catalase (232,000), ovalbumin (43,000). Samples were eluted with the same buffer at a flow rate of 0.2 ml/min, and 0.5-ml fractions were collected. 40 l of each fraction were analyzed by SDS-PAGE in 8.5% gels, followed by immunoblotting and detection with anti-PP5 (1:4000) or with monoclonal antibody AC-16 against hsp90 (1:1000), as described above.
Glucocorticoid-induced Transcriptional Activation-Tranfected CV-1 cells, incubated in the absence or presence of dexamethasone as described above, were washed twice with phosphate-buffered saline and collected by centrifugation at 4°C for 3 min at 7,500 ϫ g. Cells were resuspended in 100 l of 0.1 M potassium phosphate, pH 7.8, 1 mM dithiothreitol, and lysed by three cycles of freezing in an ethanol/dry ice bath and thawing in a 37°C water bath. After clarifying the lysate by centrifugation at 4°C for 5 min at 18,500 ϫ g, 85 l of the supernatant were saved and diluted for luciferase (29) and CAT (33) assays. To inactivate endogenous acetylating enzymes, cell extracts were heated at 65°C for 15 min before performing CAT assays. To correct for variations in transfection efficiency, CAT activity was normalized to luciferase activity. Luciferase and CAT activity were linear with respect to time and protein concentration. Background CAT activity from cells incubated in the absence of dexamethasone was subtracted from total activity in dexamethasone-stimulated cells to determine dexamethasone-induced activity.

RESULTS
hsp90 Co-immunoprecipitates with the Overexpressed TPR Domain of PP5-We have shown an interaction between the TPR domain of PP5 and the protein kinase-like domain of the ANP receptor in the yeast two-hybrid system (20). In an attempt to identify additional proteins interacting with the TPR domain of PP5, we overexpressed the FLAG-tagged TPR domain of PP5 in COS-7 cells and performed co-immunoprecipitation experiments. A prominent 90-kDa protein co-immunoprecipitated with the FLAG-tagged TPR domain of PP5 from [ 35 S]methionine-labeled cells (Fig. 1A, lane 2), using a monoclonal antibody to the FLAG epitope. The 90-kDa protein was co-immunoprecipitated in stainable amounts (not shown), indicating that it was a major cellular protein that interacted efficiently with the TPR domain. The monoclonal antibody to the FLAG epitope did not immunoprecipitate the 90-kDa protein in the absence of the FLAG-tagged TPR domain (Fig. 1A,  lane 1), indicating that its immunoprecipitation was due to its association with the TPR domain of PP5.
The TPR domain of PP5 is most closely related to TPR domains found in several large immunophilins known to associate with hsp90 (20). These immunophilins have been shown to interact with hsp90 through their TPR domains (6 -8). Based on these observations and on the fact that hsp90 is an abundant protein, we hypothesized that the 90-kDa protein co-immunoprecipitating with the TPR domain of PP5 might be hsp90. Immunoblotting with a monoclonal antibody to hsp90 confirmed this hypothesis. hsp90 was co-immunoprecipitated with the TPR domain of PP5 from lysates of unlabeled COS-7 cells overexpressing this domain, but not from lysates of control cells (Fig. 1B, lanes 1 and 2).
hsp90 Co-immunoprecipitates with Overexpressed Fulllength PP5-Under the same conditions, hsp90 also associated with full-length PP5 (Fig. 1). FLAG-tagged PP5, overexpressed in COS-7 cells, was immunoprecipitated with a monoclonal antibody to the FLAG epitope, either from [ 35 S]methioninelabeled cells (Fig. 1A) or unlabeled cells (Fig. 1B). As for the isolated TPR domain, a prominent 90-kDa protein that reacts with a monoclonal antibody to hsp90 co-immunoprecipitated with FLAG-tagged full-length PP5 (Fig. 1, lanes 3). Thus, the association of the TPR domain of PP5 with hsp90 is not an artifact of removing the phosphatase domain.
Association of Endogenous hsp90 with Endogenous PP5-We next tested whether this association was merely an artifact of overexpressing PP5, or whether endogenous PP5 associated with hsp90. This required production of an antiserum to PP5. Recombinant rat PP5, tagged at its amino terminus with the FLAG epitope, was expressed in the baculovirus system. The protein was purified to homogeneity by MonoQ chromatography followed by immunoaffinity chromatography on a matrix containing a monoclonal antibody to the FLAG epitope. The purified protein (Fig. 2) was used to immunize rabbits, resulting in specific, high titer anti-PP5 sera. We performed coimmunoprecipitation experiments to test whether endogenous PP5 associated with hsp90 in untransfected COS-7 cells or in rat brain (Fig. 3). We found that hsp90 co-immunoprecipitated with endogenous PP5 from extracts of COS-7 cells (Fig. 3, A and B) or rat brain (Fig. 3B). Neither PP5 nor hsp90 was immunoprecipitated by preimmune serum (Fig. 3, A and B). Thus, endogenous PP5 forms a complex with hsp90 in both cultured cells and in a normal tissue.
PP5 and hsp90 Are Present in a High Molecular Weight Complex-We used another method, gel filtration chromatography, to confirm the association between PP5 and hsp90. Rat brain extracts were fractionated on a Superose 6 column calibrated with various molecular weight markers, and fractions were analyzed by SDS-PAGE and immunoblotting with antisera to PP5 or to hsp90 (Fig. 4A). While purified PP5 runs as a monomer (data not shown), much of the PP5 present in brain extracts comigrates with hsp90 in a peak of approximately 600 kDa. While most hsp90 migrated in a high molecular weight complex, free hsp90 was also seen in one of three experiments (data not shown). This result was consistent with the association between PP5 and hsp90 demonstrated in co-immunoprecipitation experiments, and suggested that both molecules were present in a large complex.
In the experiment of Fig. 4B, hsp90 was immunoadsorbed from mouse L cell cytosol with a monoclonal IgM antibody, and the immune pellets were rapidly washed and assayed for hsp90 and PP5 by immunoblotting. Immunoadsorption of essentially all of the hsp90 from cytosol (compare lane 2 with lane 1) resulted in co-immunoadsorption of nearly all of the PP5. This suggests that nearly all of the PP5 is in high molecular weight form because of its association with hsp90.
Since steroid receptors exist as large heterocomplexes containing hsp90 bound to TPR proteins (1, 2), we hypothesized that PP5 might be associated with these receptor complexes.
PP5 Is Present in GR Heterocomplexes-Several different proteins, including hsp90 and the TPR-containing immunophilins FKBP52 and CyP-40, have been shown to be components of steroid receptor heterocomplexes (1,2). We tested whether PP5 was associated with GR heterocomplexes in a co-immunoprecipitation experiment (Fig. 5). Low salt extracts from L cells, prepared in the presence of sodium molybdate to stabilize receptor complexes (34), were subjected to immunoprecipitation using a monoclonal antibody to GR (lane 3) or a nonimmune control antibody (lane 2), followed by immunoblotting. Blots were probed with antibodies specific for the receptor itself, hsp90, PP5, FKBP52, or CyP-40. For comparison, an aliquot of the cytosol before immunoprecipitation was also subjected to SDS-PAGE and immunoblotting (Fig. 5, lane 1). As shown previously, hsp90, FKBP52, and CyP-40 co-immunoprecipitated with GR. In addition, PP5 was a prominent component of the receptor complexes.
The Expressed TPR Domain of PP5 Acts as a Dominant Negative Mutant-The above experiments demonstrated an association between PP5 and GR heterocomplexes, but did not address the possible function of this association. In order to test the functional importance of PP5 in GR heterocomplexes, we used a deletion mutant of PP5 containing the TPR domain, but lacking the phosphatase catalytic domain. The expressed TPR domain of PP5 should act as a dominant negative mutant by displacing endogenous PP5 from hsp90 in GR heterocomplexes. The TPR domain of PP5 was co-expressed with GR and a CAT reporter gene in CV-1 cells. We then examined the ability of dexamethasone to stimulate transcription of the reporter gene in the absence and presence of the PP5 mutant.
As shown in Fig. 6, the TPR domain of PP5 inhibited transactivation in a concentration-dependent manner, with dexamethasone-induced transcription inhibited by approximately 70% at the highest concentration of the TPR plasmid tested. Similar results were obtained in six separate experiments. This result strongly suggests a functional role for PP5 or other TPR proteins in GR signaling. At this time, we cannot exclude the possibility that the effect on transactivation is due to displacement of other TPR proteins, in addition to PP5, from GR heterocomplexes. To our knowledge, this is the first evidence for a functional role of TPR domains per se in GR signaling in vivo. DISCUSSION TPR domains are emerging as important mediators of protein-protein interactions and subcellular targeting (10). We show here that the recently identified protein-serine phosphatase, PP5, is able to bind to hsp90 through its TPR domain and thereby associate with GR heterocomplexes. Steroid receptor heterocomplexes, in addition to a single steroid receptor and a dimer of hsp90, were previously known to contain proteins that bind via TPR domains to hsp90, as well as additional accessory proteins. Immature receptor heterocomplexes contain the TPR protein p60, the mammalian homologue of the yeast TPR protein STI1, while mature complexes contain the TPR proteins FKBP51, FKBP52, or CyP-40, and an acidic protein designated p23 (5). The fact that PP5 associates with progesterone receptor heterocomplexes in co-immunoprecipitation experiments 2 suggests that the association of PP5 with steroid receptors may be a general phenomenon. hsp90 has been shown to be required for steroid receptor signaling in vivo (3), but the experiment presented here in Fig.  6 represents the first evidence that any of the TPR proteins that bind to hsp90 are required for signaling. The ability of the expressed TPR domain of PP5 to act as a dominant negative mutant, strongly inhibiting GR-mediated transcriptional activation, suggests a role for PP5 in GR signaling in vivo. PP5 could play a role by regulating the phosphorylation state of GR or of associated phosphoproteins. The mechanisms by which phosphorylation regulates steroid receptor activity and nucleocytoplasmic shuttling remain incompletely understood (35). While regulation of the phosphorylation state of steroid receptors or associated proteins by PP5 may be an important factor in controlling the activity of these receptors, further studies will be required to test whether this is the case. At least some of the TPR proteins appear to compete for binding to the same site on hsp90 (6,8). It is possible, therefore, that the TPR domain of PP5 inhibits signaling in vivo by displacing one or more of these other proteins rather than by displacing endogenous PP5 alone.
In summary, we have shown that PP5, a recently discovered protein-serine phosphatase, forms a complex with hsp90 via a TPR domain. This complex is in turn associated with GR heterocomplexes. The observation that the TPR domain of PP5 acts as a dominant negative mutant to block GR signaling strongly suggests a role in vivo for PP5, or perhaps more generally for the TPR-containing proteins that bind to the same site in hsp90 complexes. This is the first direct evidence that these steroid receptor-associated TPR proteins are important for signaling.