Evidence That Peroxisome Proliferator-activated Receptor α Is Complexed with the 90-kDa Heat Shock Protein and the Hepatitis Virus B X-associated Protein 2*

The peroxisome proliferator-activated receptor α (PPARα) is a ligand-inducible transcription factor, which belongs to the nuclear receptor superfamily. PPARα mediates the carcinogenic effects of peroxisome proliferators in rodents. In humans, PPARα plays a fundamental role in regulating energy homeostasis via control of lipid metabolism. To study the possible role of chaperone proteins in the regulation of PPARα activity, a monoclonal antibody (mAb) was made against PPARα and designated as 3B6/PPAR. The specificity of mAb 3B6/PPAR in recognizing PPARα was tested in immunoprecipitations using in vitrotranslated PPAR subtypes. The mAb 3B6/PPAR recognized PPARα, failed to bind to PPARβ or PPARγ, and is efficient in both immunoprecipitating and visualizing the receptor on protein blots. The immunoprecipitation of PPARα in mouse liver cytosol using mAb 3B6/PPAR has resulted in the detection of two co-immunoprecipitated proteins, which are heat shock protein 90 (hsp90) and the hepatitis B virus X-associated protein 2 (XAP2). The concomitant depletion of PPARα in hsp90-depleted mouse liver cytosol was also detected. Complex formation between XAP2 and PPARα/FLAG was also demonstrated in an in vitro translation binding assay. hsp90 interacts with PPARα in a mammalian two-hybrid assay and binds to the E/F domain. Transient expression of XAP2 co-expressed with PPARα resulted in down-regulation of a peroxisome proliferator response element-driven reporter gene activity. Taken together, these results indicate that PPARα is in a complex with hsp90 and XAP2, and XAP2 appears to function as a repressor. This is the first demonstration that PPARα is stably associated with other proteins in tissue extracts and the first nuclear receptor shown to functionally interact with XAP2.

PPARs are soluble transcription factors that are activated by a diverse class of lipophilic compounds (2). With the activation of PPAR, a concomitant induction of a number of genes that code for peroxisomal fatty acid metabolizing enzymes was observed in mouse liver. Among these, the peroxisomal enzyme AOx is the most broadly used indicator of peroxisome proliferator action. Transcription of the AOx gene is increased by exposure to the hypolipidemic peroxisome proliferator WY-14,643, and this effect is mediated by a PPRE located 570 base pairs upstream of the transcriptional start site (3). This PPRE contains a direct repeat of the sequence motifs TGACCT and TGTCCT, which is separated by a single nucleotide. A heterodimer of PPAR and RXR binds to PPREs located in upstream regulatory regions of various target genes and the RXR ligand, 9-cis-retinoic acid, increases PPAR/RXR transcriptional activity (4).
There are three PPAR subtypes, designated as PPAR␣ (NR1C1), -␤ (NR1C2), and -␥ (NR1C3); and each subtype is capable of binding to DNA after heterodimerizing with RXR (NR2B1) (5). Many of the genes regulated by PPAR␣ are involved in fatty acid metabolism. PPAR isoforms are activated by both exogenous and endogenous chemicals, including phthalates used as plasticizers, fibrate-type hypolipidemic drugs, and naturally occurring polyunsaturated fatty acids. PPAR␣ is highly expressed in the liver, kidney, and cardiac smooth muscle. However, most studies examining PPAR␣ function have been performed in the liver, where PPAR␣ is responsible for the tumor promotional properties of peroxisome proliferators such as DEHP. Following exposure to DEHP and other peroxisome proliferators, rodents demonstrate biological and biochemical responses such as peroxisome proliferation, increased microsomal fatty acid oxidation, increased hepatic hydrogen peroxide formation, hepatomegaly, hyperplasia, and subsequent neoplasia (6). The involvement of PPAR␣ in cell proliferation and tumor promotion has been examined using PPAR␣ knockout mice (7,8). In contrast to wild type mice, following exposure to chemicals such as clofibrate, DEHP, and WY-14,643, PPAR␣ knockout mice were unable to demonstrate the characteristic effects of peroxisome proliferators, which are hepatomegaly, peroxisome proliferation, and transcriptional activation of various target genes (7). PPAR␣ null mice were also refractory to the WY-14,643-induced replicative DNA syn-thesis (8). These results show that PPAR␣ is the major isoform required for mediating the responses resulting from the actions of peroxisome proliferators in liver. PPAR␣ is expressed in human liver at low levels, and humans appear to be insensitive to the carcinogenic response of peroxisome proliferators (6,9,10), although controversy exists and further studies are needed.
Many steroid hormone receptors exist in a complex with the chaperone hsp90 and certain co-chaperones, such as immunophilins or immunophilin-like proteins. The highly abundant chaperone hsp90 aids in protein folding and modulates the function of a wide range of proteins. In particular, hsp90 stabilizes certain steroid hormone receptors by aiding the folding of the ligand binding domain of the receptor to a high affinity ligand binding conformation. There are two classes of immunophilins: one group designated as FKBPs, which bind drugs like FK506 and rapamycin; and the other, referred to as cyclophilins, which are known to bind cyclosporin A. The larger immunophilins, FKBP51, FKBP52, and CyP40, are associated with steroid hormone receptors (11). Mature untransformed glucocorticoid receptors and progesterone receptors are in a heterocomplex with hsp90, CyP40 or FKBP52, and p23 (12). The precise role of FKBP52 or CyP40 in mature steroid receptor complexes has been difficult to determine.
Besides steroid hormone receptors, many other proteins are in a complex with hsp90 including serine/threonine kinases such as Raf, tyrosine kinases such as v-Src, and tumor suppressor p53 (13)(14)(15)(16)(17). The AhR, a basic helix loop helix/Per-Arnt-Sim transcription factor, is also known to be associated with hsp90, and hsp90 appears to be required for the AhR to maintain a proper conformation to bind ligand (18). Mapping of the hsp90 binding site on the AhR reveals that there are two separate domains complexed with hsp90, corresponding to amino acids 1-166 and 289 -347 of AhR, the latter sequence contains part of the ligand binding domain (19). The ability of hsp90 to associate with different classes of receptors suggests the possibility that hsp90 could associate with PPAR␣.
Type II steroid hormone receptors (e.g. PPAR, RXR, pregnane X receptor, etc.) are presumed to be localized to the nucleus, and the inactive receptors have not been demonstrated to associate with the hsp90 co-chaperone complexes (20). We hypothesized that PPAR␣ could complex with hsp90 or other chaperones. One protein that has been found to associate with PPAR␣ is heat shock protein 72; however, the functional significance of this association has not been determined yet (21). The only other proteins reported to associate with PPAR␣ are co-activators such as PPAR␣-binding protein (designated PBP) (22), the steroid receptor co-activator 1 (23), p300 (24), peroxisome proliferator-activated receptor interacting protein (designated PRIP) (25), and nuclear receptor co-repressor (26). Thus, little is known about the regulation of PPAR␣ through protein-protein interactions. In this report PPAR␣ was found to associate with hsp90 and the immunophilin-like protein XAP2. In addition, evidence is provided that suggests XAP2 represses PPAR␣-mediated transcriptional activity.

EXPERIMENTAL PROCEDURES
Antibody Production-The rat PPAR␣ cDNA (obtained from Frank Gonzalez, National Cancer Institute, National Institutes of Health, Bethesda, MD) was subcloned into the pMALc2 plasmid by PCR, and the fusion protein PPAR/maltose binding protein was expressed in the bacterial strain BL21. The fusion protein was purified on amylose resin and the purified protein eluted with maltose. The fusion protein was cleaved with Xa protease, after cleavage PPAR␣ precipitated out of solution and was collected by ultracentrifugation. The precipitated protein was dissolved in performic acid as previously described, lyophilized, dissolved in 0.1 N NaOH, and neutralized with 0.1 N HCl (15). This protein was used as an antigen, and BALB/c mice were immunized as described (27). An ELISA was developed using the PPAR␣/maltose binding protein. Sera from mice was tested by ELISA, and positive mice were used in hybridoma production essentially as previously described (27). The wells that were positive in the ELISA assay were then screened on a Western blot of PPAR␣ mounted in a Miniblotter45 (Immunetics, Cambridge, MA). The hybridomas that were positive in this assay were cloned by limiting dilution and further characterized.
Plasmid Production-The rPPAR␣ cDNA was subcloned into mammalian expression vector pCI (Promega Biotech, Madison, WI). A PCR primer was designed to amplify the 3Ј half of PPAR and a primer containing the 3Ј end of the cDNA plus a nucleotide sequence corresponding to the FLAG sequence, along with a stop codon, and a KpnI site. This PCR product was digested with BstEII and KpnI and subcloned into the pCI/rPPAR␣ vector digested with the same restriction enzymes. The SP163 enhancer sequence was amplified by PCR using primers containing BamHI sites with pcDNA4/HisMax C vector as a template, and was inserted into pBK/CMV digested with BamHI. The rPPAR␣/FLAG cDNA was subcloned from pCI/rPPAR␣/FLAG using XhoI and KpnI restriction sites into pBK/CMV/SP163. The vector pSG5/ mPPAR␣ was obtained from Jonathan Tugwood (Astra Zeneca), whereas pSG5/mPPAR␤ and pSG5/mPPAR␥ vectors were obtained from Paul Grimaldi (INSERM). The SP163 enhancer sequence was amplified by PCR using primers containing SacI and SpeI sites, with pcDNA4/HisMax C vector as a template, and was inserted into pBK/ CMV digested with SacI and SpeI. PCR primers were designed to amplify full-length mPPAR␣, -␤, and -␥ and to add a FLAG sequence to the 5Ј end of the cDNA. The forward and reverse primers contained SpeI and ClaI sites, respectively. The PCR products were digested with SpeI and ClaI and subcloned into the corresponding sites in pBK/CMV/ SP163. PPRE-driven reporter pAOx(x2)-luciferase, which is under the control of rat AOx promoter, was a gift from Dr. David Waxman (Boston University, Boston, MA). The vector pDJM/␤gal, which consists of ␤-galactosidase reporter gene under the control of the murine phosphoglycerate kinase 1 promoter, was kindly provided by M. W. McBurney (University of Ottawa, Ottawa, Canada). The construction of pCI/XAP2 was previously described (28).
PPAR␣ was fused to the GAL4 DNA binding domain cDNA in the vector pM (Clontech), and human hsp90 was fused to the VP16 activation domain cDNA in the vector pVP16. Domains of PPAR␣ were cloned into the pM vector as well as all the combination of domains using standard techniques. A GAL4-responsive reporter vector, pFR-luciferase (Stratagene), was used to assess transcriptional activity mediated by pM and pVP16 after transfection into cells. Transfection efficiency was assessed by co-transfection of the control Renilla vector, pRLTK (Promega).
Isolation of Tissue Cytosolic Extracts-Each tissue was quickly removed and minced in the presence of MENGM ϩ protease inhibitors (Sigma) at a 1/10 ratio (w/v) and homogenized with a Dounce homogenizer. The homogenate was centrifuged at 10,000 ϫ g for 20 min; the resulting supernatant was centrifuged at 100,000 ϫ g for 60 min. Cytosol was removed, carefully avoiding the upper lipid layer. Protein concentration of the cytosol was measured using the bicinchoninic acid (BCA) assay (Pierce).
In Vitro Translation-In vitro translations of PPAR subtypes and XAP2 were performed using a TNT rabbit reticulocyte lysate coupled transcription/translation kit, as described by the manufacturer (Promega Biotech).
Immunoprecipitations-In vitro translated samples were immunoprecipitated with the mAb 3B6/PPAR 2 pre-bound to Protein G-Sepharose. The hsp90 was immunoprecipitated from 300 g of C57BL/6N mouse liver cytosol using the mAb 3G3p90 pre-bound to 100 l of Protein L-Sepharose (Pierce). PPAR␣ was immunoprecipitated from 300 g of liver cytosol using mAb 3B6/PPAR pre-bound to goat antimouse IgG-agarose. Mouse IgM (Rockland, Gilbertsville, PA) and mouse IgG (Jackson Immunoresearch, West Grove, PA) bound resins were used in the hsp90 and PPAR control IP, respectively. In all IP, antibody binding to resin was carried out in PBS at 4°C with mixing. After 1 h, the PBS was removed and the resin was washed twice in PBS and then twice in MENGM buffer. In both hsp90 and PPAR IP, liver cytosol was immunoprecipitated in the presence of IP buffer, which consists of MENGM supplemented with 100 mM NaCl, 10 mg/ml bovine serum albumin (Fisher Scientific, Springfield, NJ), and 5 mg/ml ovalbumin (Sigma), for 1 1/2 h at 4°C with mixing. In hsp90 IP, the immunedepleted cytosol was isolated after centrifugation of the resin for 1 min at 113 ϫ g. Before SDS-PAGE analysis, all IP were washed five times with MENGM supplemented with 100 mM NaCl. Samples (either pellet 2 mAb 3B6/PPAR is available from Affinity Bioreagents (Golden, CO). or immune depleted cytosol) were then mixed with 2ϫ Tricine sample buffer, heated at 95°C for 5 min, and resolved by SDS-PAGE. After SDS-PAGE, proteins were transferred from the gel to a PVDF membrane (Millipore, Bedford, MA) at 16 V for 3 h at 4°C in a Genie blotter (Idea Scientific Co., Minneapolis, MN).
Analysis of PPAR␣ Interaction with hsp90 and XAP2-Co-immunoprecipitation of XAP2 and hsp90 with PPAR␣ was detected on membranes using anti-ARA9 (Novus Biological, Littleton, CO) and anti-hsp84/hsp86 (Affinity Bioreagents), respectively. Peroxidaseconjugated donkey anti-rabbit and goat anti-mouse antibodies were used as the secondary antibodies (Jackson Immunoresearch). Depletion of PPAR␣ in hsp90-depleted cytosol was visualized using mAb 3B6/ PPAR and peroxidase-conjugated Protein L. As a loading control, the membrane was probed for the 35-kDa protein, lactase dehydrogenase. To detect each protein, SuperSignal West Pico chemiluminescent substrate (Pierce) was used. To quantify the amount of proteins, antibodies were stripped from the blot using 0.1 M glycine, pH 2.5, and reprobed with primary antibody followed by incubation with 125 I-labeled secondary antibodies (Amersham Biosciences). The amount of PPAR␣ in hsp90-depleted cytosol relative to control was evaluated using a Cyclone phosphorimager (Packard, Meriden, CT).
In Vitro Translation Binding Assay-The constructs pCI/XAP2 and pBK/CMV/PPAR␣/FLAG were in vitro translated in the presence of [ 35 S]methionine using the TNT rabbit reticulocyte lysate system according to the instructions from the manufacturer. In vitro translated pCI/ XAP2 and pBK/CMV/PPAR␣/FLAG constructs were mixed and incubated for 1 h at 4°C. For the control, an in vitro translation, which is carried out without the receptor construct, was mixed with in vitro translated pCI/XAP2. After incubation, PPAR␣/FLAG was immunoprecipitated using anti-FLAG M2-agarose (Sigma). After a 1-h incubation with mixing, the IP were washed five times using MENGM ϩ 100 mM NaCl, samples were subjected to SDS-PAGE, and proteins were transferred from the gel to a PVDF membrane. Co-immunoprecipitated XAP2 was detected by autoradiography.
Mammalian Two-hybrid Assay-Mammalian two-hybrid studies were performed using the Matchmaker Mammalian-2-Hybrid system (Clontech). COS-1 cells were plated in 24-well plates and allowed to recover overnight. Transfection was carried out with LipofectAMINE reagent (Invitrogen) according to the protocol of the manufacturer. Each well was transfected with 200 ng of pM/PPAR␣ or pM containing various domains of PPAR␣, 200 ng of pVP16 or pVP16/hsp90, 100 ng each of pFR-luciferase and pRL-TK. Transfected cells were treated with 50 M WY-14,643 or Me 2 SO for 6 h and assayed for luciferase activity. Relative luciferase activity was corrected using the internal transfection control (pRL-TK) and the Dual Luciferase Kit (Promega). The corrected luciferase values were also corrected for changes in protein levels.
Transient Transfections and Luciferase Reporter Assays-COS-1 cells were transfected at 80% confluence in six-well tissue culture plates. The LipofectAMINE transient transfection procedure was employed according to the instructions from the manufacturer (Invitrogen). Each transfection included 500 ng of receptor (or pCI), 500 ng of PPRE-driven reporter construct pAOx(x2)-luciferase, 200 ng of the internal transfection control vector pDJM/␤gal, and 0 -1000 ng of pCI/ XAP2. The vector pCI was used to maintain the total amount of plasmid at 2 g of DNA/well. Transfected cells were harvested 24 h after the start of the transfection, and PPRE-driven luciferase reporter activity was assayed with a luciferase assay system (Promega Biotech) using a Turner TD-20e luminometer.

The mAb 3B6/PPAR Is Capable of Immunoprecipitating and
Visualizing PPAR␣-To meaningfully study the ability of PPAR␣ to associate with other proteins, a highly specific antibody that was efficient at immunoprecipitating native PPAR␣ needed to be produced. To achieve this goal, we chose to produce monoclonal antibodies to bacterially expressed PPAR␣. Three different clones that were positive in the initial screen using an ELISA procedure were tested for the ability to recognize PPAR␣. Among the three monoclonal antibodies examined, only 3B6/PPAR specifically recognized PPAR␣ (Fig. 1A). The specificity of mAb 3B6/PPAR in both immunoprecipitating and visualizing mouse PPAR␣ was tested using in vitro translated mouse PPAR constructs. Among the three isoforms of PPAR, mAb 3B6/PPAR only immunoprecipitated PPAR␣. Rat PPAR␣ can also be recognized by mAb 3B6/PPAR (Fig. 1B). The mAb 3B6/PPAR was also the only antibody capable of recognizing PPAR␣ on membranes (Fig. 1C). The ability of mAb 3B6/ PPAR to visualize PPAR␣ in mouse tissue extracts was tested using the cytosolic extracts from various C57BL/6N mouse tissues. The mAb 3B6/PPAR recognized PPAR␣ in liver, kidney, heart, and lung mouse cytosolic extracts (Fig. 1D). These data demonstrate that a monoclonal antibody has been made that is not only highly specific for PPAR␣ but also works well in immunoprecipitations and visualization of PPAR␣ on membranes.
FIG. 1. The mAb 3B6/PPAR is able to both immunoprecipitate and visualize PPAR␣ on protein membranes. A, the ability of mAb 3B6/PPAR to immunoprecipitate PPAR␣ was tested using in vitro translation lysate containing [ 35 S]methionine and mPPAR␣, mPPAR␤, or mPPAR␥. Three different clones that were positive in the initial screen using an ELISA procedure, and protein blots of PPAR␣ fusion protein were tested. B, the ability of mAb 3B6/PPAR to immunoprecipitate rat PPAR␣ was tested using in vitro translation lysate containing [ 35 S]methionine and rat PPAR construct. C, mouse PPAR␣, mPPAR␤, and mPPAR␥ expression constructs were in vitro translated in the presence of [ 35 S]methionine, subjected to SDS-PAGE, transferred to a PVDF membrane, and probed with mAb 3B6/PPAR and a secondary antibody peroxidaseconjugated Protein L. The presence of radioactivity was visualized by autoradiography (lower panel). The protein blot probed with mAb 3B6/PPAR was visualized using a secondary antibody peroxidase-conjugated Protein L (upper panel). D, the ability of mAb 3B6/PPAR to visualize PPAR␣ in cytosolic extracts of various mouse tissues. Approximately 150 g of protein from mouse liver, kidney, heart, and lung cytosol was subjected to SDS-PAGE followed by immunoblotting. PPAR␣ was detected using mAb 3B6/PPAR and peroxidase-conjugated Protein L.
PPAR␣ Is in a Complex with hsp90 and XAP2-The chaperone protein hsp90 was immunoprecipitated from C57BL/6N mouse liver cytosol using the anti-hsp90 mAb 3G3p90, which has previously been shown to be efficient at immunoprecipitating hsp90 complexes (26). Approximately 50% depletion of PPAR␣ and ϳ40% depletion of hsp90 was obtained, compared with the control IP ( Fig. 2A). This result suggests that a significant proportion of PPAR␣ is complexed with hsp90. A recurring theme for many proteins that complex with hsp90 is the presence of hsp90 co-chaperone proteins, such as FKBP52, CyP40, XAP2, and p50 cdc37 . Thus, it was logical to screen for the presence of these previously characterized hsp90 co-chaperones on protein blots. Immunoprecipitation of PPAR␣ with the mAb 3B6/PPAR has yielded two co-immunoprecipitated proteins, identified as hsp90 and XAP2 (Fig. 2B). FKBP52 and CyP40 were not detected in a complex with PPAR␣ (data not shown). In the rabbit reticulocyte lysate system, XAP2 was in vitro translated in the presence of [ 35 S]methionine and was able to bind to in vitro translated PPAR␣/FLAG (Fig. 3). These data demonstrate that PPAR␣ from C57BL/6N mouse liver exists in a complex with hsp90 and XAP2.
PPAR␣ Interacts with hsp90 in a Mammalian Two-hybrid Assay-hsp90 can interact with PPAR␣ in cells as determined using a mammalian two-hybrid assay in COS-1 cells (Fig. 4). A reduced level of activity was observed with VP16/hsp90 expression in the presence of a WY-14,643 when compared with the level of activity in the presence of VP16 expression (control). This result suggests that hsp90/VP16 binds to the liganded GAL4-PPAR␣ and partially represses transcriptional activity (Fig. 4). Thus, ligand binding alone appears not to result in hsp90 dissociation from PPAR␣.
hsp90 Interacts Predominantly with the E/F Domain of PPAR␣-The ability of a given domain of PPAR␣ to interact with hsp90 was tested using a mammalian two-hybrid assay in COS-1 cells. The D and E/F domains, when expressed in the absence of other domains of PPAR␣, are capable of interacting with hsp90 (Fig. 5). However, the D domain in combination with other domains appears not to significantly interact with hsp90, suggesting that binding to the D domain in the context of the full-length PPAR␣ may not be significant. Thus, it appears that the E/F domain is the primary domain of PPAR␣ mediating hsp90 interaction.
XAP2 Represses Reporter Gene Activity Mediated by PPAR␣-In a COS-1 cell-based reporter assay, transient expression of XAP2 with PPAR␣ resulted in a decrease in PPREdriven luciferase reporter activity (Fig. 6A). PPAR␣ activity   FIG. 3. Binding of in vitro translated XAP2 to PPAR␣. Both XAP2 and PPAR␣/FLAG were in vitro translated and mixed, and PPAR␣ was immunoprecipitated using anti-FLAG M2-agarose, subjected to SDS-PAGE, and transferred to a PVDF membrane. PPAR␣ and co-immunoprecipitated XAP2 were detected by autoradiography.  2. PPAR␣ is in a complex with hsp90 and XAP2. A, depletion of PPAR␣ in hsp90-depleted cytosol. The hsp90 from mouse liver cytosol was immunoprecipitated using mAb 3G3p90 pre-bound to Protein L-agarose. After incubating 300 g of mouse liver cytosol with mAb 3G3p90 pre-bound to Protein L-agarose, the cytosol was isolated and subjected to SDS-PAGE followed by immunoblotting. Mouse IgM bound to Protein L-agarose was used in the control experiment. The 35-kDa protein lactase dehydrogenase was used as a loading control. B, PPAR␣ from mouse liver cytosol was immunoprecipitated using mAb 3B6/ PPAR pre-bound to goat anti-mouse IgG. Mouse IgG pre-bound to goat anti-mouse IgG-agarose was used in the control experiment. PPAR␣ was immunoprecipitated, subjected to SDS-PAGE, transferred to a PVDF membrane, and probed using mAb 3B6/PPAR and peroxidaseconjugated Protein L. The presence of hsp90 and XAP2 was detected using standard techniques.
was induced by WY-14,643 and was significantly suppressed by expression of XAP2 in a dose-dependent manner. The constitutive activity of PPAR␣ was also moderately repressed by XAP2. In contrast, the transient expression of FKBP52 did not affect either the ligand-inducible activity or the constitutive activity of PPAR␣ (Fig. 6B). This result suggests that XAP2 represses PPAR␣ transcriptional activity and this effect is unique to XAP2.

DISCUSSION
To visualize possible co-immunoprecipitated proteins complexed with PPAR␣, an antibody that is highly specific to PPAR␣ needed to be produced. Because of the high level of sequence similarity between the PPAR subtypes, most antibodies have been produced as peptide antibodies to unique amino acid sequences. We instead took the approach of making monoclonal antibodies to bacterially expressed full-length PPAR␣ and screening potential clones for their specific recognition of PPAR␣ relative to PPAR␤ and PPAR␥. This approach has led to the isolation of one highly specific PPAR␣ mAb. The results in Fig. 1 indicate that mAb 3B6/PPAR only recognizes PPAR␣ and is efficient at both immunoprecipitating and visualization of PPAR␣ on membranes. The mAb 3B6/PPAR is also able to efficiently immunoprecipitate PPAR␣ from mouse liver cytosolic extracts (Fig. 2B) and thus should be useful to test for the presence of PPAR␣-associated proteins.
PPARs belong to the type II steroid receptor family (20), which also includes members such as RXR and thyroid hormone receptor. These receptors are generally considered to be localized to the nucleus and appear not to be bound to other proteins in inactive complexes, as has been extensively described for type I receptors (e.g. glucocorticoid receptor, progesterone receptor, androgen receptor). Although oncogenic v-Erb A (NR1A1), a derivative of the thyroid hormone receptor, stably interacts with hsp90 in the cytoplasm (29), endogenous type II receptors are not associated with hsp90. For example, neither RXR nor thyroid hormone receptor is associated with hsp90 (30,31). Although PPAR␣ interacts with hsp72 in vitro (21), the association of a hsp90 molecular chaperone with PPAR␣ has not been previously established. Co-immunoprecipitation with the hsp90 mAb 3G3p90 has demonstrated that hsp70, p60, FKBP52, p50 cdc37 (32), as well as AhR (19) and the glucocorticoid receptor (33), all associated with this molecular chaperone. We utilized this approach, and our data suggest that a strong association exists between hsp90 and PPAR␣ in mouse liver cytosol (Fig. 2). Thus, the results in this report are the first evidence that a type II receptor is bound to a hsp90/co-chaperone complex.
The molecular chaperone hsp90 appears to aid in folding and maintenance of the appropriate conformation of the ligand binding domain of certain soluble receptors. In addition, hsp90 prevents protein aggregation as well as aids in the stability and function of a wide variety of client proteins (11). During heat stress hsp90 exhibits both elevated levels and a heat stress protective function (34). hsp90 is composed of an N-terminal dimerization domain and a charged domain, which binds calmodulin and functions in intramolecular interactions, as well as a C-terminal dimerization and TPR-binding domain (11). Both the N-terminal and C-terminal domains of hsp90 can associate with co-chaperones, whereas the middle domain binds the client proteins such as steroid hormone receptors (35,36). Receptors bound to hsp90 are considered to be inactive, FIG. 5. Determination of the PPAR␣ domain that interacts with hsp90. COS-1 cells were transfected with the appropriate plasmids to express each PPAR␣ domain or combination of domains fused to the GAL4 DNA binding domain. Cells were lysed and assayed for luciferase activity. Activity was corrected for transfection efficiency and expressed as -fold increase in activity relative to values obtained with expression of VP16. Induction ratio is expressed as the ratio of induction between pVP16 and pVP16-hsp90. Twenty-four h after the start of transfection, cells were lysed and PPRE-driven reporter activity was measured. Lower panel, COS-1 cells were co-transfected with PPAR␣/ FLAG and increasing concentrations of FKBP52. Twenty-four h after the start of transfection, cells were lysed and PPRE-driven reporter activity was measured. Relative luciferase activity was corrected relative to protein levels. unable to either dimerize or bind DNA. Upon ligand binding, hsp90 dissociates from the complex leading to receptor transformation to the DNA binding state in the nucleus. In addition, hsp90 appears to play a major role in receptor trafficking between cytoplasm and nucleus. Untransformed progesterone and glucocorticoid receptors exist in heterotetrameric 8 -9 S complexes containing one subunit of the steroid binding receptor and two subunits of the hsp90 with a hsp90 co-chaperone protein (e.g. FKBP52) (37). It is logical to hypothesize that hsp90 serves similar functions bound to PPAR␣, as has been observed with other steroid receptors. hsp90 binds near the ligand binding domain of the glucocorticoid receptor (11), and evidence in this report would indicate that hsp90 is also complexed near the ligand binding domain of PPAR␣. These data suggest that hsp90 may influence the ligand binding pocket of PPAR␣. Interestingly, because most of the hsp90 resides in the cytoplasm, its association with a nuclear receptor found predominantly in the nucleus would appear to require that a significant amount of hsp90 is present in the nucleus, which has previously been observed (38). Although the subcellular localization of PPAR␣-bound hsp90 is currently not clear, it is reasonable to hypothesize that the PPAR␣-hsp90 complex may also exist in the nucleus.
The hsp90 co-chaperone proteins, which are found in a mature hsp90-receptor complex, include p23 and immunophilins Cyp40 and FKBP52. For example, CyP40 was associated with estrogen receptor (39), progesterone receptor (40), and the glucocorticoid receptor (41) complexes, whereas FKBP52 has been identified as a component of the estrogen receptor (39), progesterone receptor (42), and glucocorticoid receptor (43) oligomeric complexes. Both FKBP52 and CyP40 contain C-terminal TPR motifs, which are necessary for their interactions with the hsp90-receptor complex. FKBP52 is predominantly localized in the nucleus and may function by targeting receptors to this compartment, whereas CyP40 localizes in the nucleoli and may also function in protein trafficking. However, the exact role of immunophilins in receptor complexes is currently not clear. The hsp90 co-chaperone p23 was first discovered as a protein that complexes with purified receptors and appears to stabilize receptor-hsp90 association. Whether p23 also interacts with PPAR␣ remains to be tested. Interestingly, the molecular chaperone p50 cdc37 bound to hsp90 is involved in the folding of a number of protein kinases (13)(14)(15) and has recently been identified in a complex with the androgen receptor (44). This demonstrates the ability of hsp90 co-chaperones to interact with different classes of hsp90 client proteins, as has been observed here with XAP2 and PPAR␣.
In addition to hsp90, the immunophilin-like protein XAP2 co-immunoprecipitated with PPAR␣ (Fig. 2B). The ability of XAP2 to associate with PPAR␣ was also established in the reticulocyte lysate system (Fig. 3). XAP2 was originally discovered as a protein associated with the X protein of the hepatitis B virus (45). It contains three TPR motifs and regions of homology to immunophilins FKBP12 and FKBP52. However, unlike FKBP52, XAP2 is unable to bind to FK506 (46). XAP2 associates with the hepatitis B-virus protein X, the AhR (28,47), and the Epstein-Barr virus nuclear protein (EBNA-3) (48), which is a nuclear antigen that is necessary for B-cell transformation. Interestingly, XAP2 enhances AhR levels (35,46) and plays a role in the cytoplasmic localization of the dioxin receptor (49). XAP2, also known as ARA-9 and AIP, is highly expressed in spleen and thymus and minimally expressed in liver, kidney, and lung (28,50). There are low levels of XAP2 in the liver, in contrast to the high levels of PPAR␣ in this organ, which is logical from a biological standpoint when one considers that the liver is a target tissue for PPAR␣ activity. Several groups have reported XAP2 as a transcriptional activator for the AhR (28,47,51). In contrast, data presented here clearly suggest that XAP2 is a potent repressor of PPAR␣ activity. However, this result is consistent with the inhibition of hepatitis B virus X protein transcriptional activity by XAP2 (45). Interestingly, this effect has not been observed for FKBP52 in glucocorticoid receptor or progesterone receptor complexes (11). Future studies will examine whether or not XAP2 and/or hsp90 bind to other members of the PPAR family of receptors, as well as whether XAP2 plays a role in modulating tissue-specific activity of PPAR␣.