HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Felts, S. J. Articles by Toft, D. O. Search for Related Content PubMed PubMed Citation Articles by Felts, S. J. Articles by Toft, D. O. Social Bookmarking                      What's this?

J Biol Chem, Vol. 275, Issue 5, 3305-3312, February 4, 2000

The hsp90-related Protein TRAP1 Is a Mitochondrial Protein with Distinct Functional Properties^*

Sara J. Felts^§, Barbara A. L. Owen, PhuongMai Nguyen^¶, Jane Trepel^¶, David B. Donner, and David O. Toft

From the  Department of Biochemistry and Molecular Biology, Mayo Graduate School, Rochester, Minnesota 55905, the ^¶ Medicine Branch, NCI, National Institutes of Health, Bethesda, Maryland 20892, and the  Department of Microbiology and Immunology and the Walter Oncology Center, Indiana University School of Medicine, Indianapolis, Indiana 46202

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The hsp90 family of molecular chaperones was expanded recently due to the cloning of TRAP1 and hsp75 by yeast two-hybrid screens. Careful analysis of the human TRAP1 and hsp75 sequences revealed that they are identical, and we have cloned a similar protein from Drosophila. Immunofluorescence data show that human TRAP1 is localized to mitochondria. This mitochondrial localization is supported by the existence of mitochondrial localization sequences in the amino termini of both the human and Drosophila proteins. Due to the striking homology of TRAP1 to hsp90, we tested the ability of TRAP1 to function as an hsp90-like chaperone. TRAP1 did not form stable complexes with the classic hsp90 co-chaperones p23 and Hop (p60). Consistent with these observations, TRAP1 had no effect on the hsp90-dependent reconstitution of hormone binding to the progesterone receptor in vitro, nor could it substitute for hsp90 to promote maturation of the receptor to its hormone-binding state. However, TRAP1 is sufficiently conserved with hsp90 such that it bound ATP, and this binding was sensitive to the hsp90 inhibitor geldanamycin. In addition, TRAP1 exhibited ATPase activity that was inhibited by both geldanamycin and radicicol. Thus, TRAP1 has functions that are distinct from those of hsp90.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

hsp90 (90-kDa heat shock protein) is a molecular chaperone best known for its association with a variety of signal transduction systems. hsp90 was first linked to signal transduction when pp60^v-src was found in a complex with hsp90 and pp50 (Cdc37) (1, 2). Shortly thereafter, hsp90 was found in complexes with steroid hormone receptors (3-5). The underlying theme in these studies and others more recent is that hsp90 is required for the stabilization of a "poised" complex of yet inactive molecules (see Refs. 6-8 for review). Inhibition of hsp90, either by the introduction of specific mutations in yeast (9) or Drosophila (10) or through the use of pharmacologic inhibitors such as geldanamycin (11-13) or radicicol (14), causes rapid destabilization of hsp90 complexes. The target protein, or "substrate", is then usually targeted for degradation. Thus, the intracellular signals mediated by that particular substrate are compromised. It has therefore been proposed that hsp90 is not required for the folding of most proteins, but instead plays a role in the maturation of a subset of proteins typically involved in signal transduction (9).

But, hsp90 molecules are highly conserved from bacteria to mammals, perhaps suggesting a broader range of biologic functions. hsp90 has been shown to act as a general molecular chaperone in several in vitro assays. hsp90 suppresses the aggregation of non-native proteins (15) and can also promote the refolding of substrates in cooperation with the hsp70 system (16, 17). The dependence on hsp90 may thus reflect the relative complexity of the organism. For example, bacteria lacking HtpG (the prokaryote homologue) appear to be fully functional (18); yet disruption of yeast hsp90 (Hsp82/Hsc83) is lethal, and several mutations have been identified that are temperature-sensitive for growth (19). hsp90 has been linked to the cell cycle machinery in yeast through its association with Wee1, Cdc2, and Cdc37 (20-23), but the molecular details that implicate hsp90 in cell viability are not clear. A broader function for hsp90 in Drosophila has been proposed through the work of Rutherford and Lindquist (24). These authors found that hsp90 has the potential to act as an evolutionary capacitor, allowing many otherwise severe phenotypes to be masked unless hsp90 function is compromised by mutation or usurped due to the demands of environmental stress.

So the hypothesis is that hsp90 is a general chaperone required for the folding and stabilization of a number of cellular proteins that, because of the complexity of eukaryotes, just happen to be involved in cellular regulation. One caveat to this hypothesis is the possibility that hsp90 itself may have evolved to include other family members. A larger hsp90 family of proteins might allow for a second level of specificity toward a subset of cell regulatory proteins. Many eukaryotic proteins have molecular "cousins" that perform similar functions in parallel systems. We reasoned that this might be the case for hsp90. The cloning of TRAP1 (tumor necrosis factor receptor-associated protein 1), as a type I tumor necrosis factor receptor-associated protein (25), and hsp75, as a retinoblastoma (Rb)-binding protein (26), gave us the impetus to test our hypothesis. In this report, we describe our initial characterization of TRAP1/hsp75.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Fluorescence Microscopy-- PC-3-M cells, a human androgen-independent prostate cancer cell line (27), were grown on sterile coverslips and incubated with 500 nM MitoTracker Orange CMTMRos (Molecular Probes, Inc., Eugene, OR) for 30 min according to the manufacturer's instructions. Cells were then fixed with 3.7% formaldehyde in phosphate-buffered saline (PBS)¹ for 10 min and permeabilized with 0.2% Triton X-100 for 10 min at room temperature (28). Nonspecific binding sites were blocked by incubating the coverslip with 1% bovine serum albumin in PBS for 1 h at 4 °C. Primary antibody (TRAP1-6) was added to the coverslip and incubated for 1 h at 4 °C. The coverslip was washed two times for 2 min with PBS and incubated at 4 °C for 1 h with fluorescein isothiocyanate-conjugated goat anti-mouse immunoglobulin (Molecular Probes, Inc.). The coverslips were then washed two times for 2 min with PBS. Cells were stained with 0.4 µg/ml DAPI (Sigma) for 10 min at room temperature. The coverslip was then rinsed with PBS, rinsed quickly with water, air-dried, and mounted using Slow-Fade (Molecular Probes, Inc.). Confocal images were obtained using a Zeiss confocal microscope with a ×100 objective.

DNA Constructs-- TRAP1 cDNA was as described previously (25). Automated DNA sequencing added a G at position 1925 of the GenBank^TM sequence (accession number U12595), which shifted the codon frame. With this correction, the C termini of TRAP1 and hsp75 are identical. The TRAP1 coding sequence was subcloned into pET9a (Novagen, Madison, WI) for expression in Escherichia coli. To improve protein expression, the 5'-primer (5'-CTGACTGATCCATATGTTGCAGGCAGGACGACTGTTCAGC-3') included an N-terminal methionine within the context of the NdeI restriction site. Based on the study of Chen et al. (26), we judged the next amino acid to be leucine. The 3'-primer (5'-AGTCAGTCGGATCCTTATCAGTGTCGCTCCAGGGCCTTGAC-3') added a second stop codon and a BamHI restriction site for cloning. All DNA sequences were confirmed by automated DNA sequencing.

The TRAP1 sequence from Drosophila was obtained by BLAST search of the Drosophila genome data base. Two ESTs (accession numbers GM03281 and CK01417) were identified as potential TRAP1 sequences. A third EST (accession number GH01040) was found in subsequent BLAST searches and found to confirm the 5'-alignment of a TRAP1-like open reading frame that encompassed an N-terminal mitochondrial localization sequence. Oligonucleotide primers (5'-TGGCTACACTGTGCACACATGTGCGCTG-3' and 5'-TTAGTATTTCTCCAGGGCCCGCGATAGCAG-3') were designed to these ESTs, and polymerase chain reaction was used to amplify the DNA sequence from a premade Drosophila cDNA (CLONTECH, Palo Alto, CA). The reported sequence was established based on automated sequencing of the original polymerase chain reaction product and two independent clones in pGEMT-Easy (Promega, Madison, WI).

Peptide Sequence Analysis-- The amino terminus of cellular TRAP1 protein was determined using immunopurified TRAP1 from HeLa cells. The sample was resolved by SDS-PAGE and transferred to polyvinylidene fluoride membrane. The membrane-bound proteins were stained with Coomassie Blue. The TRAP1 band was excised and sequenced by gas-phase Edman degradation in the Mayo Protein Core Facility.

Proteins-- Human recombinant hsp90, p23, and Hop were expressed and purified as described previously (29). Recombinant TRAP1 was expressed in E. coli BL21(DE3) pLysS cells. Harvested cells were disrupted by sonication in 3 volumes of buffer (20 mM Tris, pH 7.4, 0.1 mM EDTA, and 10 mM monothioglycerol plus 0.1 mM leupeptin, 0.1 mg/ml bacitracin, 77 µg/ml aprotinin, 1.5 µM pepstatin, and 1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride). The centrifuged lysate was then fractionated on DEAE-cellulose with a 0-0.5 M KCl gradient elution. Pooled fractions, determined by SDS-PAGE, were dialyzed and subjected to Mono Q with a 0-1 M KCl gradient elution. Fractions containing TRAP1 were pooled and fractionated on a Superdex 200 16/60 column. TRAP1 analyzed for ATPase activity was further purified by chromatography on heparin-Sepharose with a 0.05-0.5 M KCl gradient elution. Pooled fractions were then dialyzed against 10 mM Tris, pH 7.4, 50 mM KCl, and 0.1 mM dithiothreitol and stored at 70 °C. The final preparations were at least 95% pure as determined by SDS-PAGE. Protein concentrations were determined by amino acid analysis.

Antibodies and Immunoprecipitations-- The mouse monoclonal antibodies against TRAP1 (TRAP1-6 and TR-1) were raised against the purified recombinant protein using conventional methods as described previously (30). Specific antibodies to hsp90 (H9010), p23 (JJ3), and Hop (F5) have been reported previously (31-33). To control for nonspecific interactions between proteins and immunoglobulin-protein A resins, a control immunoprecipitation, consisting of protein A-Sepharose and progesterone receptor antibody (PR22), was included in all assays. For immunoprecipitations of protein complexes formed in vitro, antibody-protein A-Sepharose resins were prepared by incubating the appropriate amount of antibody with the resin in 100 mM potassium phosphate, pH 8.0, for 1 h at room temperature with gentle rocking. The antibody-resin complexes were pelleted and equilibrated with the assay buffer and placed on ice until needed.

Protein Binding Assays-- TRAP1-p23 or TRAP1-Hop binding was assayed using 10 µg of each protein under conditions previously shown to be optimal for hsp90-p23 or hsp90-Hop complex formation in vitro (25 mM Tris, pH 7.5, 50 mM KCl, 1 mM dithiothreitol, and 5 mM MgCl₂ with the addition of 20 mM sodium molybdate and 5 mM ATP for p23 assays; see Ref. 29). The protein mixtures were incubated at 30 °C for 30 min and then added to antibody-protein A resins as indicated in the figure legends. The immunoprecipitations were incubated on ice for 1 h with occasional mixing and then washed four times with 1 ml of binding buffer. The proteins were eluted by boiling in SDS sample buffer, resolved by SDS-PAGE, and visualized by Coomassie Blue staining.

Progesterone Receptor Reconstitution-- Chicken oviduct cytosol preparation and progesterone receptor (PR) isolation were performed as described previously (34, 35). PR complexes were formed by incubation of antibody-resin-bound PR with 20 µg of hsp70, 20 µg of hsp90, 5 µg of Hop, 2 µg of Ydj-1, and 5 µg of p23 in a final volume of 200 µl of incubation buffer (10 mM Tris, pH 7.5, 50 mM KCl, 5 mM MgCl₂, 2 mM dithiothreitol, and 0.01% Nonidet P-40) plus 5 mM ATP for 30 min at 30 °C with frequent mixing. Where indicated, 10, 20, or 30 µg of TRAP1 was included along with or in place of hsp90. Following incubation, the mixtures were chilled on ice and supplemented with a mixture of 5 nM [1,2-³H]progesterone (1.9 TBq/mmol; NEN Life Science Products) and 100 nM unlabeled progesterone. The samples were incubated for 1 h at 4 °C with constant mixing to allow for hormone binding. Unbound hormone was then removed by washing three times with 1 ml of cold incubation buffer. The antibody-resin samples were resuspended in 1 ml of incubation buffer. Aliquots (100 µl) were removed for scintillation counting of bound progesterone. The remaining resins were pelleted and analyzed by SDS-PAGE.

ATP-Sepharose Binding-- ATP-Sepharose binding assays were conducted as described previously (34). Purified TRAP1 or hsp90 (5 µg) was incubated in a final 200 µl of buffer (10 mM Tris, pH 7.4, 50 mM KCl, 0.01% Nonidet P-40, and 1 mM dithiothreitol) for 30 min at 30 °C. Some samples also contained 20 mM Na₂MoO₄, 5 mM ATP, or various amounts of geldanamycin or Me₂SO control. The samples were washed with binding buffer. The bound protein was eluted with SDS sample buffer, resolved by SDS-PAGE, and visualized by Coomassie Blue staining.

ATPase Assays-- ATP hydrolysis was measured directly by the conversion of radiolabeled ATP to ADP. TRAP1 (5 µM) was incubated with 1 µM [-³²P]ATP (10 mCi/ml, 3000 Ci/mmol; NEN Life Science Products) at 30 °C for 0-480 min in buffer containing 40 mM Hepes-KOH, pH 7.4, and 2 mM MgCl₂. For ATP concentrations exceeding 1 µM, unlabeled ATP was added accordingly from a 100 mM stock of Mg-ATP, pH 7.4. Inhibition by geldanamycin or radicicol was performed with 5 µM TRAP1, 50 µM ATP, and various amounts of inhibitor added from a concentrated Me₂SO stock. Reactions were stopped by the addition of an equal volume of stop solution containing unlabeled AMP, ADP, ATP, and EDTA (12 mM each). Each reaction (2.5 µl) was spotted onto polyethyleneimine-cellulose plates (Selecto Scientific Inc., Suwanee, GA) and resolved by ascending thin-layer chromatography in 0.5 M LiCl and 2 N formic acid for 1 h at room temperature (36). The plates were air-dried, and the radioactive spots corresponding to ADP and ATP were quantified by the STORM-840 phosphoimaging system (Molecular Dynamics, Inc., Sunnyvale, CA). The ratio of ADP to ATP was used to calculate the percentage of ATP hydrolyzed. A blank consisting of all the reaction components added to the stop solution at reaction time 0 was run at each concentration of ATP assayed. Each reaction was analyzed in duplicate at three different time points depending upon the ATP concentration. The data plotted in Fig. 6 are the mean results of three independent experiments.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

TRAP1 Localizes to Mitochondria-- The manner in which TRAP1 was cloned would suggest that its cellular targets are the tumor necrosis factor receptor, a transmembrane cell-surface protein, and Rb (a primarily nuclear protein). Chen et al. (26) showed previously that TRAP1 was cytoplasmic and associated with Rb only during mitosis or heat shock, times when one of the two partners might move out of its usual locale. Our initial attempts to detect TRAP1 in various tissues or cultured cells suggested that TRAP1 was more difficult to extract than hsp90, which is mostly cytoplasmic (37). In subsequent experiments, we found that mild detergent would release TRAP1 from cells, suggesting that TRAP1 was associated with some membrane fraction.

The membrane localization appeared to contradict the work of Chen et al. (26), which showed TRAP1 (as hsp75) to be cytoplasmic and perinuclear. We therefore prepared PC-3-M cells for immunofluorescence (Fig. 1). Our TRAP1 antibodies (both TRAP1-6 (Fig. 1C) and TR-1 (data not shown)) clearly recognized a protein with broad distribution throughout the cytoplasm (i.e. not nuclear; compare Fig. 1C to DAPI-stained nuclei in Fig. 1B), and this cytoplasmic immunofluorescence was particulate in appearance. Fig. 1A shows the distribution of MitoTracker-stained mitochondria. All three color channels are combined in Fig. 1D. The dominating presence of yellow color indicates that the TRAP1 immunofluorescence (green) and mitochondrial localization (red) are virtually identical.

View larger version (31K): [in this window] [in a new window]   Fig. 1.   Localization of TRAP1 to mitochondria in PC-3-M cells by fluorescence microscopy. A, staining of mitochondria with MitoTracker dye; B, staining of nuclei with DAPI; C, immunofluorescence using a monoclonal antibody to TRAP1 (TRAP1-6); D, tricolor composite of images A-C.

TRAP1 Has a Mitochondrial Localization Sequence and Is Conserved in Drosophila-- The initial reports on the cloning of TRAP1 and hsp75 by two independent yeast two-hybrid screens suggested that these hsp90-like proteins differed in their carboxyl termini (25, 26). A more careful analysis of the cDNA sequence of TRAP1 uncovered an error in the reported sequence (see "Experimental Procedures"), making the TRAP1 protein sequence the same as hsp75. However, there were additional discrepancies at the 5'-end of the cDNA sequences. Neither cDNA contained a consensus translation initiation site. We used BLAST searches of EST data bases to find additional cDNA sequence data. Analysis of a short EST (accession number AA171881) corresponding to this 5'-region of TRAP1/hsp75 verified the approximate length of the hsp75 open reading frame, but suggested that the fifth amino acid of the deduced hsp75 protein sequence should be changed from a leucine (CTG) to a methionine (ATG). This codon is in the context of a Kozak consensus sequence and is therefore likely to be the start of TRAP1/hsp75 translation.

Since our immunofluorescence and cell fractionations showed that this protein (which we call TRAP1) was mitochondrial, we analyzed the amino-terminal residues for features characteristic of proteins destined for the mitochondria (38). Indeed, the first 50 or so residues of TRAP1 are lacking in acidic amino acids and are enriched in arginine, serine, and leucine residues. In addition, PepsortII analysis (39) predicted TRAP1 to be 91.3% mitochondrial and for the mitochondrial leader peptide to be cleaved at GRLFSTQ. Using the start site predicted by Lee and co-workers (leucine at position 7; Ref. 26) (Fig. 2A), we expressed human TRAP1 in E. coli and generated monoclonal antibodies to the purified protein. TRAP1 was then immunopurified from HeLa cell lysates and analyzed by N-terminal sequencing. This showed unambiguously that the amino terminus of the mature endogenous protein was STQTAED (underlined in Fig. 2A), one amino acid away from the predicted cleavage site. In addition, Western blotting of subcellular membrane fractions detected a slightly larger precursor protein in mitochondrial preparations (data not shown). These observations are consistent with the conclusion that TRAP1 is mitochondrial and is targeted to the mitochondria by way of its N-terminal presequence.

View larger version (57K): [in this window] [in a new window]   Fig. 2.   Amino acid alignment of human and Drosophila TRAP1 and human hsp90. A, the sequence of human TRAP1 is shown along with the deduced amino acid sequence of the Drosophila clone (TRAP1dr). Identical amino acids are shaded. The amino terminus of the mature form of TRAP1, based on N-terminal sequencing of immunopurified endogenous TRAP1, is underlined. The corrected C terminus is also underlined. B, a schematic of the three proteins is shown to illustrate domains of conservation. Both TRAP1 proteins are ~35% identical to human hsp90, with an overall similarity of 50%. Four domains of human hsp90 (hsp90h) are shown: the ATP-binding domain, charged region, middle domain, and dimerization (C-terminal) domain. Both TRAP1 proteins lack the charged region, but three of these domains are conserved in TRAP1. The percent identities and similarities (in parentheses) to hsp90 are as indicated.

We then asked if TRAP1 is conserved and similarly targeted to mitochondria in lower eukaryotes. BLAST searches of Saccharomyces cerevisiae data bases failed to find anything significant. However, similarities were obtained for both N- and C-terminal regions of the mature protein from BLAST searches of the Drosophila EST data base. There was less homology in the presequence domain. We used the regions of highest conservation to design oligonucleotide primers for amplification of a Drosophila cDNA. The deduced peptide sequence is aligned with human TRAP1 and hsp90 in Fig. 2A and diagrammed in Fig. 2B.

Overall, Drosophila TRAP1 is 54% identical and 66% similar to the mature form of human TRAP1. Each TRAP1 is also 50% similar to human hsp90. The N-terminal ATP-binding domains are the most highly conserved, being ~60% similar to hsp90. Both TRAP1 proteins lack the charged domain present in hsp90 (amino acids 212-275). In addition, the C termini are poorly conserved (31-39% similar) and lack the terminal EEVD sequence present in classic hsp90 molecules. Although our initial Drosophila clone was designed to encode only the mature protein sequence, analysis of additional Drosophila ESTs predicted an upstream arginine-rich mitochondrial import presequence and a conventional translation initiation site (Fig. 2A). Thus, Drosophila TRAP1 is likely to be mitochondrial and produced from a precursor protein.

TRAP1 Does Not Bind to hsp90 Co-chaperones-- In the cell, hsp90 functions as part of multiprotein complexes with other chaperones, in particular its co-chaperones p23 and Hop (p60). In vitro, hsp90 and p23 will form a complex in an ATP- and temperature-dependent manner (40). hsp90 and Hop will readily form a complex in vitro, irrespective of the nucleotide present (41-43). We therefore tested whether TRAP1 would bind to either of these co-chaperones in vitro. In Fig. 3, each set of three lanes examined the binding of the proteins indicated. Binding was tested by the ability of antibodies to either test protein to coprecipitate the other protein following incubation of the purified proteins at 30 °C. In all cases, antibody to PR was used to control for nonspecific binding of the test proteins to the immunoglobulin-protein A resin. The results in Fig. 3 show that TRAP1 did not bind to p23 in vitro (compare lanes 2 and 3 with lanes 5 and 6). In addition, TRAP1 did not bind Hop in vitro (compare lanes 8 and 9 with lanes 11 and 12). Immunoprecipitations using cell lysates also failed to detect an interaction between TRAP1 and p23, Hop, or a third hsp90-binding protein, CyP40 (data not shown). Thus, it appears that TRAP1 is not likely to function with the typical hsp90 partners. This conclusion would be consistent with TRAP1 having functions specific to a different compartment of the cell. However, the ability to interact with hsp90 co-chaperones in a stable complex does not immediately negate the possibility that TRAP1 might interact with these co-chaperones in a transient manner to facilitate protein folding.

View larger version (83K): [in this window] [in a new window]   Fig. 3.   TRAP1 does not bind in vitro to the hsp90 co-chaperones p23 and Hop. Purified TRAP1 was incubated with p23 lanes 1-3) or Hop (lanes 7-9) under conditions favorable for hsp90-p23 (lanes 4-6) or hsp90-Hop (lanes 10-12) binding. After a 30-min incubation, the protein mixtures were placed on ice and added to protein A-Sepharose pellets to which specific monoclonal antibodies had been bound. After 1 h on ice, the immunoprecipitates were washed free of unbound protein and resolved by SDS-PAGE. A photograph of the Coomassie Blue-stained gel is shown. Positions of the various test proteins as well as immunoglobulin heavy (HC) and light (LC) chains are indicated.

TRAP1 Does Not Function to Reconstitute Progesterone Receptor Complexes-- To test the ability of TRAP1 to facilitate protein folding, we asked whether TRAP1 could chaperone the in vitro reconstitution of hormone-binding activity of purified progesterone receptor, a classic model system for hsp90 function. Both the glucocorticoid receptor and PR have been shown to lose hormone-binding activity when stripped of chaperone proteins. However, this activity can be restored upon reconstituting receptor complexes in vitro (35, 44). This is an ATP-dependent assembly process that requires five proteins: hsp70 and its co-chaperone hsp40 (DNA-J), Hop, hsp90, and p23. To test TRAP1 activity in this system, PR was isolated from chick oviduct using antibody-affinity resin. The associating proteins were stripped away with high salt, and then aliquots of resin-bound receptor were added to protein mixtures containing purified hsp70, Ydj-1, Hop, p23, and hsp90 and/or TRAP1. After a 30-min incubation at 30 °C, a portion of each reconstitution assay was tested for hormone binding. The rest was washed free of unbound protein, and the bound proteins were resolved by SDS-PAGE.

The results are shown in Fig. 4. Fig. 4A (lane 1) shows that, in the absence of hsp90 or TRAP1, hsp70, Ydj-1, and Hop were bound to the receptor. This complex was unable to bind hormone (Fig. 4B). The addition of hsp90 (lane 2) changed the composition of the PR complex to include not only hsp90, but also p23 at the expense of hsp70 and Ydj-1. In addition, this complex was able to bind hormone. The addition of various amounts of TRAP1 (lanes 3-5) had little effect on this hsp90-dependent reconstitution of hormone binding. In addition, TRAP1 failed to bind to the PR complex or to effect receptor maturation on its own (lanes 6-8). Note that a minor protein band at the position of TRAP1 was observed in Fig. 4A, but that this was not dependent on the dosage of TRAP1 and was also present in the absence of TRAP1 (lane 1). Thus, not only is TRAP1 incapable of stably associating with hsp90 co-chaperones (Fig. 3), it also is unable to interact in some productive transient manner with these proteins, or with the PR substrate, to chaperone PR maturation.

View larger version (43K): [in this window] [in a new window]   Fig. 4.   TRAP1 is not a chaperone for progesterone receptor reconstitution. Chicken PR was immunopurified from chick oviduct cytosol; stripped of coprecipitating proteins; and incubated with purified chaperone proteins hsp70, Ydj-1, Hop, p23, and hsp90 and/or TRAP1 to reconstitute an active receptor complex. A, Coomassie Blue-stained gel of assembled PR complexes. The position of each protein component is indicated; the asterisks indicate the positions of the A and B forms of PR. B, binding of [3H]progesterone to the corresponding PR complexes. Lane 1, intermediate complex formed in the absence of hsp90; Lane 2, mature complex formed with the addition of hsp90; lanes 3-5, addition of 10, 20, and 30 µg of TRAP1, respectively, in the presence of 10 µg of hsp90; lanes 6-8, addition of 10, 20, and 30 µg of TRAP1, respectively, in the absence of hsp90; lane 9, TRAP1 plus antibody-resin control. HC and LC, immunoglobulin heavy and light chains, respectively.

TRAP1 Binds ATP and Is Sensitive to Geldanamycin-- Several laboratories have now shown that the amino-terminal region of hsp90 binds ATP and that this region of hsp90 is also the target of the inhibitor geldanamycin (34, 45, 46). Since the amino terminus of TRAP1 is the region of highest homology to hsp90, we tested the ability of purified recombinant TRAP1 to bind to a novel ATP-Sepharose used previously by Grenert et al. (34). The results are shown in Fig. 5A. The amounts of TRAP1 and hsp90 proteins loaded onto the resin pellets are shown in lanes 11 and 12. The amounts of these proteins that bound to the ATP-Sepharose are shown in lane 1 (for TRAP1) and lane 6 (for hsp90). These data show that, like hsp90, TRAP1 bound ATP and that this binding was not affected by the presence of molybdate (lanes 2 and 7), an agent known to stabilize some hsp90 interactions. ATP binding was also equivalent at 30 and 4 °C (compare lanes 1 and 5 and lanes 6 and 10). In addition, binding to ATP-Sepharose was inhibited by the presence of free ATP (lanes 3 and 8) or geldanamycin (lanes 4 and 9).

View larger version (20K): [in this window] [in a new window]   Fig. 5.   TRAP1 binds to ATP-Sepharose and is sensitive to geldanamycin. Purified recombinant TRAP1 or hsp90 was incubated with ATP-Sepharose resin at 30 °C for 30 min. After washing, the bound proteins were eluted with SDS sample buffer and resolved by SDS-PAGE. The gel was stained with Coomassie Blue. In A, lanes 1 and 6 show the amount of TRAP1 or hsp90 protein that bound. Lanes 11 and 12 show the amount of TRAP1 or hsp90 added to each incubation. Sodium molybdate (Molyb; 20 mM; lanes 2 and 7), free ATP (5 mM; lanes 3 and 8), or geldanamycin (GA; 10 µg/ml; lanes 4 and 9) was added as indicated. In lanes 5 and 10, the incubations were performed on ice. In B, TRAP1 or hsp90 was bound to ATP-Sepharose in the presence of various concentrations of geldanamycin. The percent of protein bound was quantitated by densitometric scanning of the Coomassie Blue-stained gel. Error bars represent the variation obtained in two separate experiments.

Because geldanamycin is generally thought to be a specific inhibitor of hsp90, we tested various concentrations of GA for inhibition of TRAP1-ATP binding relative to hsp90-ATP binding. The results in Fig. 5B show that, although the degree of binding to ATP was lower for TRAP1 compared with hsp90, the relative sensitivities of these proteins to geldanamycin were similar. We also found that ADP was a better competitor for TRAP1-ATP-Sepharose binding than was ATP, paralleling the behavior of hsp90 (Ref. 34; data not shown).

TRAP1 Is an ATPase Sensitive to Geldanamycin and Radicicol-- For many years, the notion that hsp90 was an ATPase was controversial. However, more recent data have demonstrated that, although weak, hsp90 does indeed have the ability to hydrolyze ATP (41, 42). Because TRAP1 was able to bind to the same ATP affinity resin as hsp90 (Fig. 5), we tested its ability to hydrolyze ATP. As shown in Fig. 6, TRAP1 does possess ATPase activity. Incubation of purified TRAP1 with [-³²P]ATP resulted in the production of [-³²P]ADP. The curve in Fig. 6A fits the equation  = V_max[S]/(K_m + [S]). From this we determined a K_m for ATP of 33 µM and a k_cat of 0.1/min at pH 7.4 and 30 °C. These results are similar to those obtained with yeast hsp90 (K_m  100 µM and k_cat  0.1/min) (47).

View larger version (16K): [in this window] [in a new window]   Fig. 6.   TRAP1 is a geldanamycin- and radicicol-sensitive ATPase. A, initial rates of ATP hydrolysis (i.e. the conversion of [32P]ATP to [32P]ADP) were measured as a function of ATP concentration using 5 µM purified TRAP1. The curve fits the equation  =Vmax[S]/(Km + [S]). Inset, the data are replotted using the Eadie-Hofstee equation, where the slope of the line equals 1/Km and the x intercept equals Vmax. B, shown is the inhibition of ATPase activity by various concentrations of geldanamycin () or radicicol (). Each point represents the mean of three independent experiments.

The benzoquinone ansamycin antibiotic geldanamycin and the macrocyclic antifungal antibiotic radicicol have both been shown to specifically inhibit hsp90 function in vivo and to compete with one another in binding to hsp90 or TRAP1 in vitro (48). Therefore, we tested whether geldanamycin and radicicol were inhibitors of TRAP1 ATPase activity. As shown in Fig. 6B, both antibiotics inhibited the ATPase activity of TRAP1 and had similar IC₅₀ values (~2 µM), but different kinetics. Our results are similar to the findings of Pearl and co-workers (47, 49), where geldanamycin and radicicol were shown to inhibit yeast hsp90 ATPase at low micromolar concentrations.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Our studies show that TRAP1 is an ATP-binding protein and an ATPase, thus securing its place among the hsp90-like proteins. These results also support previous observations that the N terminus of hsp90, the region of highest homology to TRAP1, is a novel ATP-binding domain (34, 46). Indeed, every amino acid shown by mutational analysis to be essential for ATP or geldanamycin binding and/or ATPase activity (e.g. Glu⁴⁶, Asp⁹², Gly⁹⁴, Gly¹⁴⁰, Gly¹⁴³, Gly¹⁴⁵, and Gly¹⁸²) (34, 50) is also conserved in TRAP1. As further testament to the homology between the amino termini of these two proteins, we show here that TRAP1-ATP binding is inhibited by concentrations of geldanamycin previously shown to inhibit hsp90-ATP binding.

The rate of ATP hydrolysis by TRAP1 and hsp90 is orders of magnitude less than that of some ATPases that share sequence and structural homology such as E. coli DNA gyrase and yeast topoisomerase II. However, this ATPase activity is similar to the E. coli mismatch repair enzyme MutL (51). Interestingly, the ATPase activity of TRAP1 is inhibited by both geldanamycin and radicicol, which have both been shown to specifically inhibit hsp90 function (52-54). Similarly, the ATPase activity of TRAP1 is inhibited by both geldanamycin and radicicol at concentrations shown to inhibit yeast hsp90. We estimate that the amount of TRAP1 in cell lysates is <10% the amount of hsp90 (data not shown). How this ratio is reflected in the total cell is uncertain. However, our data raise the possibility that perhaps some of the in vivo effects of geldanamycin or radicicol might be due to an effect on TRAP1. Final proof awaits the more difficult challenge of identifying some specific pathway that involves TRAP1, but not hsp90.

Despite its ATP-binding activity, TRAP1 does not bind the hsp90 co-chaperone p23. This was initially quite surprising, given that the formation of hsp90-p23 complexes requires that hsp90 be able to bind ATP (34). However, recent studies suggest that other regions of hsp90, more C-terminal, are also important for p23 binding. Some of these regions (e.g. amino acids 601-620) (55) are poorly conserved in TRAP1. We also show that TRAP1 does not bind the tetratricopeptide repeat-containing co-chaperone Hop. Again, it is the C terminus of hsp90 that appears important for the binding of Hop. Hartl and co-workers (56) showed that a glutathione S-transferase fusion protein containing amino acids 629-732 of hsp90 will bind Hop and hsp90-associating immunophilins in vitro. This region is only 42% similar between TRAP1 and hsp90, whereas the same regions in Drosophila and human TRAP1 proteins are 74% similar. Work by Smith and co-workers (55) also showed that the C-terminal EEVD sequence of hsp90, which does not exist in TRAP1, is critical for hsp90-Hop binding. Indeed, we have made a TRAP1/hsp90 chimera in which the C-terminal 66 amino acids of TRAP1 have been deleted and replaced with the C-terminal 103 amino acids of hsp90. This chimera (TRAPC90) will bind Hop in vitro, yet does not function in any hsp90 chaperoning assays tested (data not shown).

In the formation of steroid receptor complexes, hsp90 has been shown to act in cooperation with its co-chaperones and the hsp70 chaperone system (6, 57, 58). The co-chaperone Hop has the ability to bind both hsp90 and hsp70 and appears to bring these two chaperones together. The co-chaperone p23 binds hsp90 and stabilizes hsp90 complexes. The ability of hsp90 to bind p23 is a requisite step for the full reconstitution of hormone-binding activity of steroid receptors in vitro. In support of our observation that TRAP1 does not bind to these hsp90 co-chaperones, we have found that TRAP1 does not inhibit the hsp90-dependent reconstitution of PR-hormone binding. In addition, TRAP1 cannot substitute for hsp90 in this system. In fact, we have also found that TRAP1 has very little activity in a luciferase refolding assay (29) that utilizes hsp70, hsp90, and Hop.² Because the specific chaperone functions of hsp90, as tested through the PR and luciferase assays, tend to require that hsp90 receives substrate from hsp70 and Hop, our results suggest that TRAP1 lacks these specific chaperone functions. However, TRAP1, as hsp75, has been shown to assist in the refolding of denatured Rb to a protease-resistant state (26). This type of passive chaperone activity may be similar to the ATP-independent chaperone activity characterized for hsp90 (59, 60). Perhaps future studies will determine the relative significance of this activity.

Although another hsp90 family member, Grp94, functions in the endoplasmic reticulum (61), this is the first demonstration of an hsp90-like protein in mitochondria. Mitochondria are known to contain the chaperones Hsp60, Hsp10, and a form of hsp70 (mitochondrial Hsp70/Ssc1). Mitochondrial hsp70 has now been shown to be important for protein import (62, 63). The chaperone proteins Hsp60 and Hsp10 appear to function as general chaperones for protein folding (64) and have more recently been shown to play important roles in apoptotic pathways (65, 66). Curiously, TRAP1 resembles bacterial hsp90, HtpG, which has an N-terminal ATPase domain and a very short charged domain and also lacks a C-terminal EEVD sequence. However, because the overall similarity of TRAP1 to HtpG is 49%, whereas the similarity of human hsp90 to HtpG is 54%, it is difficult to say whether TRAP1 is the eukaryotic homologue of HtpG.

Studies on hsp90 suggest that its ATP-binding and hydrolysis activities are linked to its ability to interact with substrate proteins and also with its co-chaperones p23 and Hop (50, 67, 68). If TRAP1 parallels hsp90 in its function, it is likely that TRAP1 interacts with its own set of chaperone-like molecules. So far, however, we have been unable to identify any TRAP1-specific co-chaperones. One reason for this may be that the cell lysis conditions required to extract TRAP1 disrupt its interactions with other proteins. Since TRAP1 appears to be localized to mitochondria, any TRAP1-specific co-chaperones might also be mitochondrial. The significance of TRAP1 is supported by its existence in a variety of cell types (25, 26) and in eukaryotes from Drosophila to humans. Although it is tempting to speculate that TRAP1 may function in a chaperone pathway involved in mitochondrial import or apoptosis, further studies are needed to clearly relate this protein to specific cellular targets.

	ACKNOWLEDGEMENTS

We thank Laura Blaisdell, Nancy McMahon, and Bridget Stensgard for excellent technical assistance; Sherry Linander for manuscript preparation; Leonard Neckers and members of the Toft laboratory for comments on the manuscript; and Robert Matts for assistance on data base searches. Assistance was also provided by Benjamin Madden and M. Cristine Charlesworth in the Mayo Protein Core Facility. Geldanamycin was obtained from the Drug Synthesis and Chemistry Branch, Developmental Therapeutics Program, NCI, National Institutes of Health. Radicicol was provided by Dr. Shiro Akinaga (Kyowa Hakko Kogyo Co., Ltd., Shizuaka, Japan). ATP-Sepharose was generously provided by Timothy Haystead. PC-3-M cells were a gift from James Kozlowski. HeLa cell pellets for protein analysis were obtained from the National Cell Culture Center (Minneapolis, MN).

	FOOTNOTES

^* This work was supported by National Institutes of Health Grants DK46249 (to D. O. T.) and CA73023 and CA67891 (to D. B. D.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBank^TM/EMBL Data Bank with accession number(s) AF115775.

^§ To whom correspondence should be addressed: Dept. of Biochemistry and Molecular Biology, Mayo Foundation, 200 First St. S. W., Rochester, MN 55905. Tel.: 507-284-3074; Fax: 507-284-2053; E-mail: felts.sara@mayo.edu.

² B. D. Johnson, A. Chadli, S. J. Felts, and D. O. Toft, manuscript in preparation.

	ABBREVIATIONS

The abbreviations used are: PBS, phosphate-buffered saline; EST, expressed sequence tag; PAGE, polyacrylamide gel electrophoresis; PR, progesterone receptor; DAPI, 4,6-diamidino-2-phenylindole.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				C. N. Cunningham, K. A. Krukenberg, and D. A. Agard Intra- and Intermonomer Interactions Are Required to Synergistically Facilitate ATP Hydrolysis in Hsp90 J. Biol. Chem., July 25, 2008; 283(30): 21170 - 21178. [Abstract] [Full Text] [PDF]