MOLECULAR AND CELLULAR BIOLOGY, Dec. 1999, p. 8033–8041 0270-7306/99/$04.00⫹0 Copyright © 1999, American Society for Microbiology. All Rights Reserved.
Vol. 19, No. 12
Multiple Components of the HSP90 Chaperone Complex Function in Regulation of Heat Shock Factor 1 In Vivo STEVEN BHARADWAJ, ADNAN ALI,
AND
NICK OVSENEK*
Department of Anatomy and Cell Biology, College of Medicine, University of Saskatchewan, Saskatoon, Saskatchewan, Canada S7N 5E5 Received 22 April 1999/Returned for modification 14 May 1999/Accepted 30 August 1999
Rapid and transient activation of heat shock genes in response to stress is mediated in eukaryotes by the heat shock transcription factor HSF1. It is well established that cells maintain a dynamic equilibrium between inactive HSF1 monomers and transcriptionally active trimers, but little is known about the mechanism linking HSF1 to reception of various stress stimuli or the factors controlling oligomerization. Recent reports have revealed that HSP90 regulates key steps in the HSF1 activation-deactivation process. Here, we tested the hypothesis that components of the HSP90 chaperone machine, known to function in the folding and maturation of steroid receptors, might also participate in HSF1 regulation. Mobility supershift assays using antibodies against chaperone components demonstrate that active HSF1 trimers exist in a heterocomplex with HSP90, p23, and FKBP52. Functional in vivo experiments in Xenopus oocytes indicate that components of the HSF1 heterocomplex, as well as other components of the HSP90 cochaperone machine, are involved in regulating oligomeric transitions. Elevation of the cellular levels of cochaperones affected the time of HSF1 deactivation during recovery: attenuation was delayed by immunophilins, and accelerated by HSP90, Hsp/c70, Hip, or Hop. In immunotargeting experiments with microinjected antibodies, disruption of HSP90, Hip, Hop, p23, FKBP51, and FKBP52 delayed attenuation. In addition, HSF1 was activated under nonstress conditions after immunotargeting of HSP90 and p23, evidence that these proteins remain associated with HSF1 monomers and function in their repression in vivo. The remarkable similarity of HSF1 complex chaperones identified here (HSP90, p23, and FKBP52) and components in mature steroid receptor complexes suggests that HSF1 oligomerization is regulated by a foldosome-type mechanism similar to steroid receptor pathways. The current evidence leads us to propose a model in which HSF1, HSP90 and p23 comprise a core heterocomplex required for rapid conformational switching through interaction with a dynamic series of HSP90 subcomplexes. to-trimer transitions, have not been elucidated. Since induction of protein unfolding and accumulation of damaged proteins is a common feature of various stress conditions, it has been suggested that Hsps, as molecular chaperones sensitive to the level of denatured proteins, provide a basis for HSF1 regulation. An autoregulatory role for Hsp/c70 is supported by numerous observations in various model systems (1, 3, 4, 7, 11, 13, 25, 31). However, a mechanism by which Hsp/c70 influences different steps of the HSF1 activation-deactivation pathway has not been clearly defined. Hsp/c70 may function to maintain HSF1 in its inactive monomeric state or to disassemble trimers following stress (5, 27, 37) and also, along with Hdj1, to repress the transcriptional activation domain during attenuation (42). However, since Hsp/c70 has not been linked to trimer formation or transcriptional upregulation in the initial phases of the activation pathway, it appears that other chaperones or cellular factors must be involved in modulating HSF1. HSP90 has emerged as a key factor in the regulation of HSF1. The initial evidence came from affinity chromatography experiments demonstrating HSF1 interaction with HSP90 in vitro (28) and later from in vitro reconstitution experiments showing HSF1 interaction with HSP90 and several associated cochaperones (29). Two recent reports provided more direct clues. Using an in vitro HeLa cell extract system, Zou et al. demonstrated enhanced threshold activation of HSF1 with the HSP90-binding agent geldanamycin, or with antibody immunodepletion of HSP90, and concluded that HSP90 acts as an HSF1 repressor (45). Using the Xenopus oocyte model system, our laboratory has shown that HSP90 is present in immune complexes with both active and inactive HSF1 and further that disruption of HSP90 in vivo with geldanamycin or microin-
The heat shock response is characterized by increased synthesis of heat shock proteins (Hsps) which prevent denaturation of cellular proteins under a variety of stress conditions. In multicellular organisms, this is directed at the transcriptional level by transient activation of the heat shock transcription factor HSF1 (reviewed in references 26 and 44). The stress activation and attenuation pathways involve a number of modifications to the HSF1 molecule. Under nonstress conditions, HSF1 exists as transcriptionally inactive non-DNA-binding monomers that, in response to various stress stimuli, assemble into active homotrimers capable of binding to heat shock elements (HSEs) in hsp gene promoters. These oligomeric changes involve dynamic inter- and intramolecular interactions between conserved hydrophobic heptad repeats. Additional steps include hyperphosphorylation, the functional impact of which is unclear, as well as activation of the transcriptional activation domain. Numerous studies have shown that trimer formation is tightly associated with increased HSE binding and that transcription is regulated independently of trimerization. Active trimers are eventually converted back to inactive monomers after an appropriate amount of Hsp expression or upon return to nonstress conditions. Although the peptide sequence requirements and oligomerization properties have essentially been determined, the signaling pathways through which various stress inducers activate HSF1, as well as the folding mechanisms controlling monomer* Corresponding author. Mailing address: Department of Anatomy and Cell Biology, College of Medicine, University of Saskatchewan, 107 Wiggins Rd., Saskatoon, SK, Canada S7N 5E5. Phone: (306) 966-4069. Fax: (306) 966-4298. E-mail:
[email protected]. 8033
8034
BHARADWAJ ET AL.
jected antibodies activates HSF1, delays recovery, and inhibits HSF1-mediated transcription after heat shock (2). This evidence led us to suggest a model for HSF1 regulation in which HSP90 forms a complex with HSF1 and participates in modulating both monomer-trimer transitions and transcriptional activity (2). HSP90 functions as part of a large and dynamic heterocomplex along with a number of cochaperones including Hip (p48), Hop (p60), p23, Hsp/c70, and the immunophilins Cyp-40, FKBP51, and FKBP52. This multifunctional chaperone complex prevents aggregation of unfolded proteins under stressful conditions and also regulates the folding and maturation of a number of signal transduction molecules and receptors (for reviews, see references 12 and 36). Receptors for progesterone (PR) and glucocorticoid (GR) are well-characterized cellular targets of the HSP90 chaperone complex (reviewed in reference 35). HSP90-receptor interactions establish and maintain high-affinity hormone binding conformations through a stepwise assembly process involving sequential formation of an intermediate complex containing HSP90, Hsp/c70, Hip, and Hop and a mature complex containing HSP90, p23, and an immunophilin. The molecular details of HSP90 function and the specific roles of cochaperones in the steroid folding pathways are currently being investigated by several groups using the cell-free reticulocyte lysate model system. It had previously been hypothesized that assembly mechanisms similar to the steroid receptor pathways may also be involved in the folding of various HSP90 target substrates, including HSF1 (29). Here we use Xenopus oocytes as an in vivo model to investigate the functional roles of HSP90 cochaperones in HSF1 regulation under relevant physiological conditions. Immunotargeting experiments involving direct nuclear microinjection of antibodies against individual HSP90 chaperone complex proteins resulted in either activation of HSF1 under nonstress conditions or significant delay of attenuation. In addition, mobility supershifts of HSF1-HSE complexes were observed with antibodies against HSP90, p23, and FKBP52, suggesting these proteins assemble into a heterocomplex with active HSF1 trimers. The data provide evidence that multiple components of the HSP90 chaperone machinery participate in regulating HSF1 in vivo. MATERIALS AND METHODS Oocyte manipulations and stress treatments. Xenopus laevis frogs were purchased from Xenopus I (Ann Arbor, Mich.). Ovary portions were surgically removed from adult female frogs, and follicular cells were removed from oocytes by treatment in calcium-free OR2 buffer (82.5 mM NaCl, 2.5 mM KCl, 1 mM MgCl2, 1 mM NaH2PO4, 5 mM HEPES [pH 7.8], 10 mg of streptomycin sulfate per liter, 10 mg of benzylpenicillin per liter) containing 2 mg of collagenase (type II; Sigma) per ml for 3 h at 18°C. Oocytes were washed extensively, allowed to recover for 4 h in OR2 (as above, with 1 mM CaCl2) at 18°C, and maintained in OR2 during experimental treatments. Only stage VI oocytes were selected for experiments. Nuclei and cytoplasm were obtained under OR2 by scoring animal hemispheres with a needle and gently squeezing the equatorial region with watchmaker’s forceps. The nonshock temperature was 18°C; unless otherwise indicated, heat shock was performed at 33°C for 1 h. In all experiments, a minimum of 20 oocytes was used for each sample. Immunotargeting and protein overexpression. For immunotargeting experiments, antibodies were diluted (1:1) in sterile H2O immediately prior to microinjection. Antibody solutions (15 nl) were injected directly into the nucleus or cytoplasm as indicated. Injected oocytes were incubated for 30 min at 18°C and then further incubated at the nonshock temperature (18°C), heat shocked at 33°C, or heat shocked and then placed at control temperature during recovery for the indicated times. Protein extracts were prepared for gel mobility shift analysis and immunoblotting immediately after treatments. HSP90 monoclonal antibody (MAb) SPA 830 was from StressGen, Victoria, British Columbia, Canada; anti-HSF1 antiserum was provided by R. Morimoto, Northwestern University, Evanston, Ill.; anti-YY1 polyclonal antibody (PAb), anti-human IB-␣ PAb, and NFB (p65) PAb sc-372-G were from Santa Cruz Biotechnology, Santa Cruz, Calif.; anti-Hsp70 MAb clone 3a3 and Cyp-40 PAb PA3-023 were from Affinity BioReagents, Golden, Colo.; Hip (clone 2G6), Hop (clone f5), FKBP51
MOL. CELL. BIOL. (clone Hi51), and FKBP52 (clone Hi52c) MAbs were gifts from D. Smith, University of Nebraska, Omaha; p23 MAb clone JJ3 was a gift from D. Toft, Mayo Graduate School, Rochester, Minn. Anti-human PTP-1B antibody clone 15 was from Transduction Laboratories, Lexington, Ky.; anti-human CREB PAb was from New England Biolabs, Beverly, Mass. For overexpression experiments, capped mRNAs encoding human chaperones were synthesized in vitro, using T7 RNA polymerase (Pharmacia Biotech); 50 nl of mRNA solution (2 mg/ml) was injected directly into oocyte cytoplasm. Expression plasmids encoding wild-type Hip (33), Hop (9), Cyp-40 (22), FKBP51 (30), and FKBP52 (30) were gifts from D. Smith. Expression plasmid for human p23 (20) was a gift from D. Toft. Purified bovine HSP90 was a gift from R. Matts, Oklahoma State University, Stillwater; purified Hsp/c70 was from StressGen. For elevation of Hsps, 20 ng of each protein was injected directly into oocyte nuclei. Identical concentrations of bovine serum albumin (BSA) were used for controls. Protein extracts and gel mobility shift assays. Protein extracts were prepared by homogenizing oocytes in buffer C (50 mM Tris-Cl [pH 7.9], 20% glycerol, 50 mM KCl, 0.1 mM EDTA, 2 mM dithiothreitol, 10 g of aprotinin per ml, 10 g of leupeptin per ml) (14) in a Dounce homogenizer. Homogenates were transferred to Eppendorf tubes and spun for 5 min at 15,000 ⫻ g (4°C). The resultant supernatants were immediately frozen in liquid nitrogen and stored at ⫺80°C. Oocytes were homogenized in a volume of 10 l of buffer C per oocyte. DNA mobility shift assays were performed with HSE oligonucleotide probes as previously described (17). DNA-binding reaction mixtures contained 10 l of extract (1 oocyte equivalent by volume, or 20 g of soluble protein). Binding reactions were performed with 1 g of poly(dI-dC), 10 mM Tris (pH 7.8), 50 mM NaCl, 1 mM EDTA, 0.5 mM dithiothreitol, and 5% glycerol in a final volume of 20 l. Reaction mixtures were incubated on ice for 20 min and immediately loaded onto 5% nondenaturing polyacrylamide gels containing 6.7 mM Tris-Cl (pH 7.5), 1 mM EDTA, and 3.3 mM sodium acetate. Gels were electrophoresed for 2.5 h at 150 V, dried, and exposed to autoradiography with X-ray film (XAR; Kodak). Antibody recognition experiments were performed by adding antibodies directly into DNA-binding reactions or by mixing antibodies with oocyte extracts for 20 min on ice prior to DNA-binding reactions. Immunoblotting. Protein extracts were fractionated by sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDS-PAGE; 10% acrylamide gel) and electroblotted onto polyvinylidene difluoride or nitrocellulose membranes, and the blots were blocked (2 h at room temperature) in TBST (20 mM Tris-Cl [pH 7.6], 137 mM NaCl, 0.1% [vol/vol] Tween 20) containing 5% milk powder. Antibodies were diluted in TBST with 2.5% milk (1:5,000), and blots were incubated in primary antibody for 2 h at room temperature. Blots were washed in TBST and incubated with secondary antibody (horseradish peroxidase-conjugated goat anti-rabbit or anti-mouse immunoglobulin G) (Bio-Rad) diluted 1:10,000 in TBST– 2.5% milk for 2 h at room temperature. Blots were washed, and proteins were visualized by chemiluminescence (Renaissance system; Dupont NEN) and autoradiography with X-ray film (XAR; Kodak). CAT assays. Chloramphenicol acetyltransferase (CAT) assays were performed with 1 oocyte equivalent of whole-cell extract from uninjected or microinjected oocytes as previously described (2, 32). Antibody injections or heat shock treatments were performed 4 h after plasmid injections, and then oocytes were incubated at the nonshock temperature for 12 h to allow for CAT expression. A pool of at least 20 microinjected oocytes was used for each experimental treatment. The acetylated products were separated by thin-layer chromatography and visualized by autoradiography.
RESULTS Antibodies against HSP90, p23, and FKBP52 recognize heat-activated HSF1. In previous work, we demonstrated that HSP90 associates in a complex with active HSF1 and showed evidence that HSP90 could regulate some of the dynamic changes that occur during heat shock and attenuation (2). HSP90 is known to interact with a number of protein cofactors in the assembly and maturation process of cellular substrates such as PR and GR. Therefore, we were interested in testing the hypothesis that these additional constituents of the HSP90 chaperone machine participate in controlling the major structural and oligomeric transitions that control the HSE-binding activity of HSF1. Gel mobility supershift experiments were used to examine if antibodies against known components of the HSP90 chaperone complex could recognize heat-induced HSF1-HSE complexes. Antibodies against each of the chaperones (HSP90, Hip, Hop, Hsp/c70, p23, Cyp-40, FKBP51, and FKBP52) were mixed (individually) with aliquots of a crude extract of heatshocked oocytes in DNA-binding reactions with labeled HSE (Fig. 1A). Clear mobility supershifts of HSF1 were observed
VOL. 19, 1999
FIG. 1. Recognition of the heat-shock-activated DNA-binding form of HSF1 in gel mobility supershift assays with antibodies against components of the HSP90 chaperone complex. (A) Antibodies (Ab) against HSP90 and the array of cochaperones (indicated above the lanes) were incubated with identical aliquots of a heat-shocked oocyte extract in DNA-binding reactions with 32P-labeled HSE probe, and the electrophoretic migration of HSF1-HSE complexes was analyzed. (B) Same as above except that oocytes were stressed with 70 mM salicylate for 1 h. The position of the nonsupershifted heat-induced HSF1 complex is indicated at the left.
with HSP90, p23, and FKBP52 antibodies (compare lane 1 with lanes 2, 3, and 8). Antibodies against Cyp-40, Hip, Hop, Hsp/c70, and FKBP51 each caused slight but highly reproducible shifts in the mobility of the HSF1 complex (compare lane 1 with lanes 4 to 7 and lane 9). A similar supershift pattern was seen in with the same antibodies in experiments with salicylateshocked oocytes (Fig. 1B). Thus, HSP90, p23, and the immunophilin FKBP52 can be identified as components of a heterocomplex with the stress-activated DNA-binding form of HSF1. In these experiments, chaperone interactions with HSF1 were determined using the functional HSE-binding assay and observed through retardation of complexes after electrophoresis. The specificity of this approach was illustrated by the observation that control NFB and YY1 transcription factor antibodies had no effect on the HSF1 complex (Fig. 1B, lanes 11 and 12). As expected, HSF1 antibody completely obliterated the bandshift (lane 10). We presume the DNA-bound heterocomplexes contain HSF1 trimers, since it is well established that HSF1 binds HSE in the trimeric conformation. The comparatively weak supershifts observed with antibodies against Hip, Hop, Hsp/c70, Cyp-40, and FKBP51 suggest these proteins are largely absent from active HSF1 heterocomplexes. However, we note that physical interactions between HSF1 and most of the HSP90 cochaperones have previously been observed in vitro (29). Nuclear and cytoplasmic localization of HSP90 and cochaperone proteins in oocytes. Given its apparent association with components of the HSP90 chaperone system, we next wished to test the hypothesis that HSF1 is regulated by a chaperone assembly process similar to that of the steroid receptors. Since we know HSF1 is entirely a nuclear protein in Xenopus oocytes (24), it followed that components of a stress regulatory mechanism must also be nuclear, at least at some points in the activation-deactivation process. We previously reported that HSP90, although predominantly cytoplasmic, is detectable in the oocyte nucleus (2). To assess the nuclear-cytoplasmic distribution of accessory components of the HSP90 chaperone
REGULATION OF HSF1 BY HSP90 COMPLEX
8035
machine, extracts of manually enucleated oocytes from both nonshocked and heat-shocked oocytes were analyzed by Western blotting (Fig. 2). Abundant levels of each of the chaperones Hip, Hop, Hsp/c70, p23, Cyp-40, FKBP51, and FKBP52 were detected in intact oocytes. As for HSP90, the majority of each protein was in the cytoplasmic fraction, but small amounts were also detected in the nuclear fractions of both control and stressed cells. It was important to demonstrate that the nuclear extracts used in these assays were devoid of cytoplasmic contamination. Control blots (Fig. 2, bottom) showed that a nuclear marker protein, PCNA, was absent from cytoplasmic fractions. The cytoplasmic marker protein IB was not detected in nuclear fractions even after loading the equivalent of 10 nuclei. In contrast, the chaperone proteins were detected in extract equivalent to a single nucleus, indicating they were likely of nuclear origin. Although IB was absent even after overloading nuclear samples, we cannot strictly exclude the possibility that chaperone proteins were actually associated with the outer nuclear membrane. Since the entire array of HSP90 chaperone proteins appear to be present in the nuclei of both nonshocked and heatshocked cells, it is possible that the interactions between HSF1 and chaperone components implied by supershift assays (Fig. 1) actually occur in vivo. Interestingly, it appeared that heat shock elicited some redistribution of Hip and Hsp/c70 into the nucleus.
FIG. 2. Components of the HSP90 chaperone complex are present in the oocyte nucleus. Oocytes were incubated under nonshock (NS) or heat shock (HS) conditions, and proteins from single intact oocytes or from the nuclear and cytoplasmic fractions were separated by SDS-PAGE. Chaperone components were detected with corresponding antibodies (see Materials and Methods) by immunoblotting. In these blots, nuclear lanes contained extract approximately equivalent to one nucleus, except for the IB blot, in which nucleus lanes contained the equivalent of 10 nuclei.
8036
BHARADWAJ ET AL.
FIG. 3. Effects of elevated levels of HSP90 and Hsp/c70 on the HSE-binding activity of HSF1 under nonshock conditions, during heat shock, and during recovery. The levels of Hsp/c70 (top) and HSP90 (middle) were elevated by direct microinjection of purified proteins into nuclei. Similar amounts of BSA were injected into nuclei (bottom), and then oocytes were incubated at nonshock (NS) temperatures, heat shocked (time zero), or heat shocked and allowed to recover at the nonshock temperature (for times indicated at the top). Left, gel mobility shift assays comparing HSF1 activity in oocytes with elevated Hsps and identically treated uninjected controls; right, immunoblots showing relative levels of HSP90 or Hsp/c70 in uninjected (U) and injected (Inj.) oocyte nuclei.
Effect of chaperone overexpression on HSF1 in vivo. In the next experiments, we were interested in determining, under relevant physiological conditions, the potential for each component of the HSP90 chaperone machine to function as modulators of HSF1. The first strategy was to overexpress individual proteins by microinjection of in vitro-synthesized mRNAs into oocytes and then assess any subsequent effects on the HSE-binding function of HSF1 in vivo. It was possible to attain approximately twofold increases in the amount of cochaperones with cytoplasmic injection of mRNA (see Fig. 4, right), but this approach failed to elevate Hsp/c70 and HSP90 levels (data not shown), probably because these are already very abundant proteins. We solved this problem by microinjecting purified Hsp/c70 and HSP90 directly into oocyte nuclei, after which we could detect substantial elevation of their respective levels (Fig. 3, right). HSF1 activity was then examined in oocytes in which each component of the HSP90 chaperone complex was individually elevated, either in nonshocked cells, immediately following heat shock, or at several time points during recovery. In each experiment, uninjected oocytes from the same ovary were used as parallel controls, since we have often observed that the recovery profile of HSF1 can vary between cell batches from different females. HSF1 remained inactive under nonstress conditions after twofold elevation of each chaperone, and the relative amounts of induced HSF1HSE complexes were similar to those in uninjected controls immediately following heat shock (Fig. 3 and 4). Elevation of HSP90 and Hsp/c70 resulted in a slightly accelerated rate of recovery relative to uninjected controls (Fig. 3). Elevating Hip and Hop resulted in marked acceleration in the rate of recovery, but no significant effect on recovery was observed after overexpression of p23 (Fig. 4B). Interestingly, each of the immunophilins Cyp-40, FKBP51, and FKBP52 delayed recovery relative to uninjected controls (Fig. 4B). With each chaperone, the same pattern was consistently observed in replicate experiments, regardless of the time frame of recovery in controls (data not shown). Despite contrasting results with immunophilins and other chaperones, the simplest interpretation of these experiments is that multiple components of the HSP90 machine are involved in the process of trimer disassembly
MOL. CELL. BIOL.
during attenuation. Nascent Hip, Hop, and microinjected Hsps could have increased the efficiency of trimer disassembly either directly via their independent chaperone activities or by promoting the formation of additional active chaperone complexes. Alternatively, it is possible that nascent immunophilins had disrupted endogenous chaperone complexes, normally required for efficient trimer disassembly, through competitive interactions with the tetratricopeptide-binding site of HSP90, as has been reported to occur in vitro (8, 10). Note that in controls with injected BSA, HSF1 activity was not affected under nonshock conditions, immediately after heat shock, or during recovery (Fig. 3). Thus, the effects on HSF1 observed in these experiments appeared to be a specific result of increased chaperone concentration or activity rather than simply the presence of additional proteins. HSF1 activity is affected by immunotargeting chaperones in vivo. An alternative strategy to overexpression of HSP90 chaperone components is to disrupt or deplete their activities in vivo. Direct microinjection of specific antibodies into Xenopus oocytes provides a unique opportunity to examine the function of components of HSP90 chaperone complexes in vivo. This is particularly useful because specific chaperone activities can be suppressed in a relatively short time frame, avoiding cell growth and viability problems that would be inevitable consequences of genetic or antisense knockouts of key chaperones. We have previously shown that microinjection of specific antibodies is an effective means to target specific protein activities in oocytes. For example, injected antibodies against HSP90 induced HSF1 activity under nonshock conditions, demon-
FIG. 4. Effects of elevated levels of components of the HSP90 chaperone machine on HSF1. The levels of Hip and Hop (A) and the immunophilins Cyp40, FKBP51, and FKBP52 or p23 (B) were increased by microinjection of corresponding mRNAs, and HSF1 was compared with uninjected controls by gel shift assay (left). Immunoblots showing relative levels of cochaperones in mRNAinjected oocytes (Inj.) and uninjected (U) controls are shown on the right.
VOL. 19, 1999
strating a role in repressing HSF1 (2), a result which was consistent with immunodepletion of HSP90 in an in vitro model (45). To target the array of HSP90 chaperones, we used antibodies that are known to interact with endogenous Xenopus proteins (Fig. 2 to 4). Since HSF1 is a nuclear factor, chaperone antibodies were microinjected directly into oocyte nuclei, and the presence of each antibody was confirmed by Western blotting (Fig. 5C). Immediately following heat shock, the amounts of HSF1-HSE complexes were not significantly different in uninjected and antibody-injected cells (Fig. 5A). Thus, antibody-directed disruption of chaperones appeared to have little or no effect on the efficiency of trimer assembly. However, activation of HSF1 under nonstress conditions was seen after immunotargeting both HSP90 and p23 (Fig. 5A). This effect was consistently reproduced in repeats of this experiment and can be seen more clearly in the right panels of Fig. 5A. This stress-independent activation appeared to be a specific result of immunotargeting HSP90 and p23, since HSF1 was not induced under similar nonstress conditions by antibodies against any of the other chaperones (Fig. 5A) or by antibodies against a number of control proteins (Fig. 5B). Induction of HSF1 in the absence of stress with antibodies against p23 and HSP90 suggest these chaperones normally function to inhibit trimerization of HSF1 in unstressed cells. This result, along with the apparent presence of these chaperones in active HSF1 heterocomplexes, suggests that HSP90 and p23 might remain associated with HSF1 in a semipermanent heterocomplex throughout the activation-deactivation pathway. The second major observation in the experiment shown in Fig. 5A was a significant delay in attenuation during recovery from heat shock relative to uninjected controls. Retention of HSF1-HSE complexes occurred in cells injected with antibodies against HSP90, Hip, Hop, p23, FKBP51, and FKBP52 but not with Cyp-40 and Hsp/c70 antibodies. A series of injection controls involving nuclear injection of antibodies against NF-B, IB, CREB, and PTP-1B showed no discernible effect on HSF1, either under nonstress conditions, upon heat shock, or during recovery (Fig. 5B). In addition, injection of antibodies against HSP90-associated chaperones into oocyte cytoplasm did not activate HSF1 under nonstress conditions and had no effect on the amount of HSF1-HSE complexes formed immediately following heat shock or during recovery relative to uninjected controls (data not shown). Since neither the microinjection procedure itself nor the presence of antibodies alone influenced the HSE-binding properties of HSF1, the delay in recovery can be attributed to specific antibody-directed disruption of individual components of the chaperone complex (HSP90, Hip, Hop, p23, and immunophilins) in nuclei. Either this was through disruption of the activity of these components physically linked to HSF1-HSP90 heterocomplexes or, alternatively, chaperones not directly associated in HSF1 were disrupted by nuclear antibodies. In either case, the results of immunotargeting experiments suggest that HSP90 and p23 function to repress HSF1 and that multiple components of the HSP90 chaperone machine function in regulating the attenuation process in vivo. A potential problem in these experiments is that depletion of chaperone activity by immunotargeting in oocyte nuclei might itself be sensed as a cellular stress. Hence, both the activation of HSF1 under control conditions seen with HSP90 and p23 antibodies and the delay in attenuation with HSP90, Hip, Hop, p23, FKBP51, and FKBP52 antibodies could have been through mimicking the effects of cellular stress. It was therefore necessary to determine if targeting of these key cellular chaperones actually induced a stress response in oocytes. We tested for this by monitoring transcription from a micro-
REGULATION OF HSF1 BY HSP90 COMPLEX
8037
FIG. 5. Effect on HSF1 after immunotargeting of HSP90 chaperone components with microinjected antibodies. (A) Left, gel mobility shift assay of uninjected or antibody (Ab)-injected oocytes that were incubated at nonshock temperatures (NS), heat shocked for 1 h (time zero), or heat shocked and then allowed to recover at the nonshock temperature (for times indicated at the top). Right, gel mobility shift assay showing HSF1 induction upon HSP90 or p23 antibody injection. (B) Gel mobility shift assay showing the effect of microinjected antibodies against NF-B, IB, PTP-1B, and CREB on HSF1. (C) Immunoblots showing presence of antibodies in injected oocytes. Antibodies were detected with the appropriate secondary antibodies.
8038
BHARADWAJ ET AL.
FIG. 6. Effect of microinjected chaperone antibodies on HSF1-mediated transcription in unstressed cells. Oocytes were injected with hsp70-CAT and antibodies (Ab.) against each component of the HSP90 complex as indicated. CAT expression in antibody-injected oocytes was compared to that in similarly treated nonshocked (NS) or heat-shocked (33°C, 1 h; HS) samples without injected antibody (⫺). CAT assays were performed as described in Materials and Methods.
injected reporter construct (hsp70-CAT) under control of the hsp70 promoter, which is known to be switched on by heat shock and various other stresses (2, 6, 23) (Fig. 6). If immunodepletion of chaperones elicited a stress response, we expected to observe induction of HSF1-mediated transcription from the hsp70 promoter. In controls, CAT activity was induced by heat shock relative to unstressed oocytes, but no activation was observed in antibody-injected oocytes (Fig. 6). This experiment was repeated several times with different batches of oocytes yielding similar results (data not shown). In addition, Hsp levels were unaffected by nuclear injection of any of the chaperone antibodies (data not shown). It therefore appears that immunodepletion of each chaperone was not perceived as a cellular stress condition. Thus, the effects of delayed attenuation and activation of HSF1 observed in immunotargeting experiments were probably through direct disruption of HSP90 chaperone complexes that modulate HSF1. DISCUSSION Here we tested the hypothesis that constituents of the HSP90 chaperone machine participate in the regulation of HSF1 under both nonstress and heat shock conditions. Previous work identifying HSP90 as a suppressor of HSF1 in HeLa cell extracts (45) and a modulator of HSF1 activities in Xenopus oocytes (2) prompted us to examine the possibility that HSP90 cochaperones also function in HSF1 regulatory pathway. Key predictions were that cochaperones, if present in nuclei, could be detected in the HSF1-HSP90 heterocomplex and that disruption of their activities in vivo would have an impact on the DNA-binding activity of HSF1. Our results were generally consistent with these predictions. All of the HSP90 chaperone complex proteins, although predominantly cytoplasmic, were present at low levels in oocyte nuclei before and after heat shock (Fig. 2). Antibodies against HSP90, p23, and FKBP52 significantly supershifted HSF1 complexes in functional HSE-binding assays (Fig. 1). Therefore, we identify HSP90, p23, and the immunophilin FKBP52 as components of a mature heterocomplex with the stress-activated form of HSF1. Knockout or overexpression of each of the cochaperones had profound impacts on the HSE-binding activity of HSF1, either during attenuation or prior to heat shock. Taken together, our results provide strong evidence that dynamic changes in oligomerization states of HSF1 are chaperoned by components of the HSP90 chaperone machine. We conclude that HSP90 and associated cochaperones are required for folding of HSF1. In the present analysis, cochaperone interactions with HSF1 were detected with the functional HSE-binding assay and retardation of DNA-binding complexes. HSP90, p23, and FKBP52 appeared to interact stably with HSF1 under these assay conditions. In contrast, antibodies against Hip, Hop,
MOL. CELL. BIOL.
Hsp/c70, Cyp-40, and FKBP51 caused only slight shifts of the HSF1-HSE complex. This finding suggests that these latter components of the HSP90 chaperone machine are absent from active HSF1 heterocomplexes. However, previous in vitro reconstitution experiments have demonstrated HSF1 interaction with HSP90, Hsp/c70, Hop, p23, and each of the immunophilins (29), and so the existence of larger heterocomplexes should not yet be excluded. Putative interactions between HSF1 trimers and Hip, Hop, Hsp/c70, or the other immunophilins could be more transient and hence difficult to detect, or their antibody interactions could be sterically hindered in a larger heterocomplex. Although the current data suggest that HSF1 associates only with HSP90, p23, and FKBP52, it is important to consider that the electrophoretic mobility shift assay/supershift technique selects for cochaperones present in active DNA-binding complexes and excludes different HSF1chaperone complexes that may exist at different points of the activation-deactivation profile. In overexpression experiments (Fig. 3 and 4), we found that elevation of HSP90, Hsp/c70, Hip, and Hop decreased the time of deactivation to the non-DNA-binding state during recovery. Our findings with Hsp/c70 are consistent with previous results obtained with mammalian and Drosophila cells after elevation of Hsp/c70 (27, 37), although this effect has not been reported before with elevated levels of HSP90, Hop, and Hip. A simple interpretation is that these chaperones increased the efficiency of trimer disassembly directly through their independent chaperone activities. But it was confusing that the immunophilins (Cyp-40, FKBP51, and FKBP52) had the contrary effect of delaying recovery relative to uninjected controls. Since HSP90 chaperone complexes are highly dynamic, it is conceivable that nascent chaperones actually exerted their effects through remodeling of existing chaperone complexes in vivo. Consequently, Hip, Hop, and the Hsps may have accelerated recovery by promoting formation of additional active chaperone complexes. This could explain the opposite effects seen with immunophilins, which could have disrupted preexisting chaperone complexes through competitive interactions, thus delaying recovery. An intriguing possibility is that nascent immunophilins competed with Hop in vivo for the tetratricopeptide-binding site of HSP90, as has been demonstrated in vitro (8). Thus the delay of attenuation with immunophilins might be akin to the delay seen with injected Hop antibodies (Fig. 5A). It is difficult to fully explain the results of overexpression experiments because little is known about the behavior and function of HSP90 cochaperones in vivo, and we do not know how nascent cochaperones may have shifted the equilibrium of HSP90-containing complexes. Also, we cannot exclude the possibility that overexpressed cochaperones indirectly affected HSF1 complexes either by aiding or blocking normal cellular signaling and recovery. However, these results, combined with detection of cochaperones in HSF1 heterocomplexes, imply a more direct functional link between the HSP90 chaperone machine and regulation of HSF1 trimer disassembly. Antibody-directed knockouts of HSP90 cochaperones in vivo also showed profound effects on HSF1 activity. In considering the results of immunotargeting experiments, it is important to note that microinjected chaperone antibodies did not activate expression from the hsp70 promoter (Fig. 6). This implies that effects on HSF1 were manifested by specific targeting of chaperone complexes that function to regulate HSF1 rather than by mimicking a stress response. It is difficult, however, to rule out the possibility that general signaling pathways required for HSF1 activity were disrupted by immunodepletion of key chaperones in vivo. It was interesting that the HSE-binding activity of HSF1 was
VOL. 19, 1999
activated under nonshock conditions after targeting HSP90 and p23 (Fig. 5). This reveals that both chaperones may function to maintain HSF1 in the repressed monomeric state. To our knowledge, our results are the first indication that p23 may function as a repressor of HSF1. This was likely a specific effect of diminishing specific target protein activity in our experiments because HSF1 was not activated by antibodies against any of the other chaperones or by different control antibodies. It is also interesting that HSF1 remained transcriptionally inactive (Fig. 6), and so disruption of HSP90 and p23 appeared to elicit only a partially activated conformation. This stressindependent partial activation is consistent with the idea that additional independent modifications of the HSF1 molecule are necessary for transcriptional activity. Our findings with HSP90 targeting in vivo are also in agreement with the results of HSP90-immunodepletion in the in vitro extract system described by Zou et al. (45) and have hitherto been observed in this laboratory (2). We note that activation of HSF1 seen here was not observed after p23 immunodepletion in HeLa cell extracts (45) and assume that this contradiction is attributable to differences in the respective model systems. A second major effect of immunotargeting was a significant delay in the rate of decline of HSE complexes during recovery, which was seen with HSP90, Hip, Hop, p23, FKBP51, and FKBP52 antibodies (Fig. 5). The retention of HSE binding is interpreted as evidence that these proteins function in the process of HSF1 trimer disassembly. Given the presence of HSP90, p23, and FKBP52 in active heterocomplexes, we can suggest that these cochaperones may function directly to unfold trimers, or at least to potentiate disassembly. It also appears that Hip, Hop, and other cochaperones influence this process, despite the apparent absence of these proteins from the active HSF1 heterocomplex. The immunotargeting and overexpression experiments described here were performed in vivo under relevant physiological conditions, but some inconsistencies were apparent in comparing the results of these assays with individual cochaperones. For example, elevation of Hsp/c70 and Cyp-40 either accelerated or delayed attenuation, but corresponding results were not obtained with injected antibodies. This could reflect differences in the ability of particular antibodies to interact with target proteins or diminish chaperone activities in vivo. This was likely the case with Hsp70, for which a role in HSF1 regulation has already been established (5, 27, 37, 42). Since the Cyp-40 homologue Cpr7 was shown to suppress the heat shock response in yeast (15), a similar role for Cyp-40 in higher organisms should not be discounted. Also of note is that microinjected antibodies against p23 significantly delayed recovery, but overexpression had no apparent effect on attenuation. This might suggest that exogenous p23 either did not acquire an active conformation in oocytes or failed to disrupt preexisting chaperone complexes. It is also possible that endogenous p23 associates stably in HSF1 complexes, precluding any effect of exogenous proteins. Regardless of these discrepancies, the combined data provide evidence that many of the different HSP90 cochaperones play a role in HSF1 regulation. The results reported here support the previous hypothesis of Nair et al. (29) that HSF1 is regulated similarly to steroid receptors. Numerous in vitro studies using the rabbit reticulocyte lysate system have revealed that PR and GR pathways involve at least eight different protein constituents of the HSP90 complex and that assembly of the ligand-binding conformation requires a complex and well ordered series of dynamic interactions (9, 33, 35). In these pathways, p23 is associated with HSP90, and both proteins are required for maintenance of the high-affinity hormone binding state (19,
REGULATION OF HSF1 BY HSP90 COMPLEX
8039
20). One possible model for HSF1 regulation is that HSP90 and p23 comprise a semipermanent core heterocomplex required for rapid interconversion between monomers and trimers and that oligomerization is controlled by a foldosome-type mechanism involving a dynamic series of HSP90 subcomplexes containing different cochaperones. This possibility is supported by our observations of both HSP90 and p23 in the active HSF1 heterocomplex, activation of HSF1 and delay of attenuation after antibody disruption of both molecules, and the apparent influence of multiple cochaperones on the DNA-binding activity of HSF1. The putative HSF1 core complex could be compared to HSP90 and p23 in telomerase complexes, in which both chaperones are required for assembly and remain associated with active enzyme through its cycling (18). Assemblydisassembly of HSF1 involves a dynamic equilibrium under different cellular conditions, and as with steroid receptors (43), this occurs constantly. This observation leads us to suggest that HSP90 and p23 act coordinately to maintain HSF1 monomers in an inactive state that is competent for rapid assembly into trimers. Both molecules could remain associated with HSF1 as it is assembled into trimers, and the data suggest that both could be directly involved in the process of trimer disassembly. But what are the other cochaperones doing in what is probably a dynamic milieu? Studies of steroid complex formation have shown it is a multistep process involving transient association with Hsp/c70, Hip, and Hop. It is remarkable that the components we have identified in active HSF1 heterocomplexes, HSP90, p23, and FKBP52 are precisely the same as in late stage or mature steroid receptor complexes (35). This finding supports the hypothesis that a similar foldosome-type chaperone process occurs with HSF1. Our model predicts that the putative HSF1 core complex interacts with various cochaperones in a stepwise manner, with HSP90 and p23 acting as organizers of intermediate complexes. This could involve addition and release of individual components or interaction with HSP90 subcomplexes containing different combinations of cochaperones. As in steroid folding, we suggest that Hsp/c70 and HSP90 participate in a highly coordinated manner, along with their respective interacting or organizing proteins, Hop and Hip, to modulate HSF1 folding. Hip, Hop, and Hsp/c70 probably interact transiently with the putative core heterocomplex as HSF1 molecules are chaperoned through oligomeric changes. This view is consistent with the known autoregulatory roles of both HSP90 and Hsp/c70 and could explain the very weak but reproducible supershifts of HSF1-HSE complexes that we observe with their respective antibodies (Fig. 1). The apparent interaction of FKBP52 with the stress-activated heterocomplex is, to our knowledge, the first indication that immunophilins could be involved in HSF1 regulation. Its putative role in this regard is consistent with known immunophilin function. Immunophilins are peptidyl-prolyl isomerase family proteins that mediate protein folding and are identified as targets of immunosuppressive drugs cyclosporin A (Cyp-40) and FK506 (FKBPs) (16, 41). These are tetratricopeptide repeat-containing proteins that interact with HSP90 complexes (30, 38, 39), but their precise roles in receptor complexes have not yet been elucidated. Our observation that antibody disruption of FKBP52 delayed attenuation suggests its main function is to potentiate trimer disassembly. Since immunophilins are also thought to be involved in targeted protein trafficking (34), we raise the intriguing prospect that FKBP52 mediates rapid relocalization of HSF1 to and from stress granules within the nucleus, a phenomenon that has recently been observed in HeLa cells (21). We have not determined whether FKBP52 remains associated with HSF1 mono-
8040
BHARADWAJ ET AL.
mers or if stress granules are present in heat-shocked oocyte nuclei. It is tempting to speculate that the sequential order of interactions or mechanistic function of chaperones in the HSF1 folding pathway is similar to those in steroid receptor folding. However, given the disparate physical characteristics and biological functions of HSF1, it is likely that its folding pathway is significantly different. At the very least, we have identified potential roles for HSP90 and p23 in HSF1 repression and for HSP90, Hsp/c70, Hip, Hop, p23, and immunophilins in trimer disassembly. This suggests a new paradigm for stress gene regulation in which neither HSP90 nor Hsp70, nor any of the cochaperones, functions independently to regulate activity. Rather, it is more likely that HSF1 regulation in vivo involves a series of coordinated interactions with multiple components of the large HSP90 chaperone complex. Our current model does not exclude potential regulatory roles of additional factors such as a recently identified human HSF1-binding protein, HSBP1 (40). A number of important questions remain to be addressed regarding HSP90-HSF1 heterocomplexes. For example, what are the stoichiometric relationships between different chaperone components and HSF1, and what are the dynamics of these interactions at each phase of the response? In addition, what are the specific functional roles of each chaperone in oligomerization, phosphorylation, and transcription? Another question is whether the HSF1-HSP90 chaperone complex comprises or contributes to the cellular stress-sensing mechanism. One idea is that an HSF1 heterocomplex containing key cellular chaperones could be directly responsive to the intracellular environment and transmit signals from stress-induced changes in the levels of denatured or misfolded proteins directly to the HSF1 molecule. Alternatively, HSP90-cochaperone complexes could act downstream of the stress-sensing mechanism to chaperone HSF1 oligomeric transitions. It is important to emphasize that the present work points to studies of the steroid folding pathways for clues to help unravel fundamental mechanisms of the stress response. Further studies using the Xenopus oocyte and other models can now be aimed at elucidating the dynamics and functional details of the HSF1HSP90 chaperone complex and its potential relationship to the cellular stress-sensing mechanism.
MOL. CELL. BIOL.
6. 7. 8.
9. 10. 11. 12. 13. 14. 15. 16. 17. 18.
19.
20. 21. 22.
ACKNOWLEDGMENTS We are grateful to D. Smith for antibodies and expression vectors for Hip, Hop, and immunophilins, R. Morimoto for HSF1 antibody, D. Toft for p23 reagents, A. Wolffe for hsp70-CAT, and B. Roesler for providing various antibodies. We thank R. O’Carroll, J. Xavier, P. Mercier, and G. Davies for advice and technical assistance. S.B. was supported by graduate scholarships from the Natural Sciences and Engineering and Medical Research Councils of Canada. A.A. was supported by a postdoctoral fellowship from the Health Services Utilization Research Council of Saskatchewan. This research was supported by a Medical Research Council operating grant to N.O. REFERENCES 1. Abravaya, K. A., M. Myers, S. P. Murphy, and R. I. Morimoto. 1992. The human heat shock protein hsp70 interacts with HSF, the transcription factor that regulates heat shock gene expression. Genes Dev. 6:1153–1164. 2. Ali, A., S. Bharadwaj, R. O’Carroll, and N. Ovsenek. 1998. HSP90 interacts with and regulates the activity of heat shock factor 1 in Xenopus oocytes. Mol. Cell. Biol. 18:4949–4960. 3. Ananthan, J., A. L. Goldberg, and R. Voellmy. 1986. Abnormal proteins serve as eukaryotic stress signals and trigger the activation of heat shock genes. Science 232:522–525. 4. Baler, R., W. J. Welch, and R. Voellmy. 1992. Heat shock gene regulation by nascent polypeptides and denatured proteins: hsp70 as a potential autoregulatory factor. J. Cell Biol. 117:1151–1159. 5. Baler, R., J. Zou, and R. Voellmy. 1996. Evidence for a role of hsp70 in the
23. 24. 25. 26. 27. 28. 29.
30.
regulation of the heat shock response of mammalian cells. Cell Stress Chaperones 1:33–39. Bharadwaj, S., A. Hnatov, A. Ali, and N. Ovsenek. 1998. Regulation of the DNA-binding and transcriptional activities of heat shock factor 1 is uncoupled in Xenopus oocytes pineapplea. Biochim. Biophys. Acta 1402:79–85. Boorstein, W. R., and E. A. Craig. 1990. Transcriptional regulation of SSA3, an hsp70 gene from Saccharomyces cerevisiae. Mol. Cell. Biol. 10:3262–3267. Carrello, A., E. Ingley, R. F. Minchin, S. Tsai, and T. Ratajczak. 1999. The common tetratricopeptide repeat acceptor site for steroid receptor-associated immunophilins and hop is located in the dimerization domain of Hsp90. J. Biol. Chem. 274:2682–2689. Chen, S., and D. F. Smith. 1998. Hop as an adaptor in the heat shock protein 70 (Hsp70) and Hsp90 chaperone machinery. J. Biol. Chem. 273:35194– 35200. Chen, S., W. P. Sullivan, D. O. Toft, and D. F. Smith. 1998. Differential interactions of p23 and the TPR-containing proteins Hop, Cyp40, FKBP52 and FKBP51 with Hsp90 mutants. Cell Stress Chaperones 3:118–129. Craig, E. A., and C. A. Gross. 1991. Is Hsp70 the cellular thermometer? Trends Biochem. Sci. 16:1472–1477. Csermely, P., T. Schnaider, C. Soti, Z. Prohaszka, and G. Nardai. 1998. The 90-kDa molecular chaperone family: structure, function, and clinical applications. A comprehensive review. Pharmacol. Ther. 79:129–168. DiDomenico, B. J., G. E. Bugaisky, and S. Lindquist. 1982. The heat shock response is self-regulated at both the transcriptional and posttranscriptional levels. Cell 31:593–603. Dignam, J. D., R. M. Lebovitz, and R. G. Roeder. 1983. Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res. 11:1475–1489. Duina, A. A., H. M. Kalton, and R. F. Gaber. 1998. Requirement for Hsp90 and a CyP-40-type cyclophilin in negative regulation of the heat shock response. J. Biol. Chem. 273:18974–18978. Fruman, D. A., S. J. Burakoff, and B. E. Bierer. 1994. Immunophilins in protein folding and immunosuppression. FASEB J. 8:391–400. Gordon, S., S. Bharadwaj, A. Hnatov, A. Ali, and N. Ovsenek. 1997. Distinct stress-inducible and developmentally regulated heat shock transcription factors in Xenopus oocytes. Dev. Biol. 181:47–63. Holt, S. E., D. L. Aisner, J. Baur, V. Tesmer, M. Dy, M. Oullette, J. Trager, G. Morin, D. O. Toft, J. W. Shay, W. E. Wright, and M. A. White. 1999. Functional requirement of p23 and Hsp90 in telomerase complexes. Genes Dev. 13:817–826. Hutchison, K. A., L. F. Stancato, J. K. Owens-Grillo, J. L. Johnson, P. Krishna, D. O. Toft, and W. B. Pratt. 1995. The 23-kDa acidic protein in reticulocyte lysate is the weakly bound component of the hsp foldosome that is required for assembly of the glucocorticoid receptor into a functional heterocomplex with hsp90. J. Biol. Chem. 270:18841–18847. Johnson, J. L., T. G. Beito, C. J. Krco, and D. O. Toft. 1994. Characterization of a novel 23-kilodalton protein of unactive progesterone receptor complexes. Mol. Cell. Biol. 14:1956–1963. Jolly, C., Y. Usson, and R. I. Morimoto. 1999. Rapid and reversible relocalization of heat shock factor 1 within seconds to nuclear stress granules. Proc. Natl. Acad. Sci. USA 96:6769–6774. Kieffer, L. J., T. W. Seng, W. Li, D. G. Osterman, R. E. Handschumacher, and R. M. Bayney. 1993. Cyclophilin-40, a protein with homology to the P59 component of the steroid receptor complex. Cloning of the cDNA and further characterization. J. Biol. Chem. 15:12303–12310. Landsberger, N., M. Ranjan, G. Almouzni, D. Stump, and A. P. Wolffe. 1995. The heat shock response in Xenopus oocytes, embryos, and somatic cells: a regulatory role for chromatin. Dev. Biol. 170:62–74. Mercier, P. A., J. Foksa, N. Ovsenek, and J. T. Westwood. 1997. Xenopus heat shock factor 1 is a nuclear protein before heat stress. J. Biol. Chem. 272:14147–14151. Mifflin, L. C., and R. E. Cohen. 1994. hsc70 moderates the heat shock (stress) response in Xenopus laevis oocytes and binds to denatured protein inducers. J. Biol. Chem. 269:15718–15723. Morimoto, R. I. 1998. Regulation of the heat shock transcriptional response: cross talk between a family of heat shock factors, molecular chaperones, and negative regulators. Genes Dev. 12:3788–3796. Mosser, D. D., J. Duchaine, and B. Massie. 1993. The DNA-binding activity of the human heat shock transcription factor is regulated in vivo by hsp70. Mol. Cell. Biol. 13:5427–5438. Nadeau, K., A. Das, and C. T. Walsh. 1993. Hsp90 chaperonins possess ATPase activity and bind heat shock transcription factors and peptidyl prolyl isomerases. J. Biol. Chem. 268:1479–1487. Nair, S. C., E. J. Toran, R. A. Rimerman, S. Hjermstad, T. E. Smithgall, and D. F. Smith. 1996. A pathway of multi-chaperone interactions common to diverse regulatory proteins: estrogen receptor, Fes tyrosine kinase, heat shock transcription factor HSF1, and the aryl hydrocarbon receptor. Cell Stress Chaperones 1:237–250. Nair, S. C., R. A. Rimerman, E. J. Toran, S. Chen, V. Prapapanich, R. N. Butts, and D. F. Smith. 1997. Molecular cloning of human FKBP51 and comparisons of immunophilin interactions with Hsp90 and progesterone receptor. Mol. Cell. Biol. 17:594–603.
VOL. 19, 1999 31. Nunes, S. L., and S. K. Calderwood. 1995. Heat shock factor 1 and the heat shock cognate 70 protein associate in high molecular weight complexes in the cytoplasm of NIH-3T3 cells. Biochem. Biophys. Res. Commun. 213:1–6. 32. Ovsenek, N., and J. J. Heikkila. 1990. DNA sequence specific binding activity of the Xenopus heat shock transcription factor is heat inducible before the midblastula transition. Development 110:427–433. 33. Prapapanich, V., S. Chen, and D. F. Smith. 1998. Mutation of Hip’s carboxyterminal region inhibits a transitional stage of progesterone receptor assembly. Mol. Cell. Biol. 18:944–952. 34. Pratt, W. B., M. Czar, L. F. Stancato, and J. K. Owens. 1993. The hsp56 immunophilin component of steroid receptor heterocomplexes: could this be the elusive nuclear localization signal-binding protein? J. Steroid Biochem. Mol. Biol. 46:269–279. 35. Pratt, W. B., and D. O. Toft. 1997. Steroid receptor interactions with heat shock protein and immunophilin chaperones. Endocrine Rev. 18:306–360. 36. Pratt, W. B. 1998. The hsp90-based chaperone system: involvement in signal transduction from a variety of hormone and growth factor receptors. Proc. Soc. Exp. Biol. Med. 217:420–434. 37. Rabindran, S. K., J. Wisniewski, L. Li, G. C. Li, and C. Wu. 1994. Interaction between heat shock factor and hsp70 is insufficient to suppress induction of DNA-binding activity in vivo. Mol. Cell. Biol. 14:6552–6560. 38. Ratajczak, T., A. Carrello, P. J. Mark, B. J. Warner, R. J. Simpson, R. L.
REGULATION OF HSF1 BY HSP90 COMPLEX
39.
40. 41. 42. 43. 44. 45.
8041
Moritz, and A. K. House. 1993. The cyclophilin component of the unactivated estrogen receptor contains a tetratricopeptide repeat domain and shares identity with p59 (FKBP59). J. Biol. Chem. 268:13187–13192. Renoir, J. M., C. Radanyi, L. E. Faber, and E. E. Baulieu. 1990. The non-DNA-binding heterooligomeric form of mammalian steroid hormone receptors contains a hsp90-bound 59-kilodalton protein. J. Biol. Chem. 265: 10740–10745. Satyal, S. H., D. Chen, S. G. Fox, J. M. Kramer, and R. I. Morimoto. 1998. Negative regulation of the heat shock transcriptional response by HSBP1. Genes Dev. 12:1962–1974. Schreiber, S. L. 1991. Chemistry and biology of the immunophilins and their immunosuppressive ligands. Science 251:283–287. Shi, Y., D. D. Mosser, and R. I. Morimoto. 1998. Molecular chaperones as HSF1-specific transcriptional repressors. Genes Dev. 12:654–666. Smith, D. F. 1993. Dynamics of Hsp90-progesterone receptor binding and the disactivation model for steroid receptor complexes. Mol. Endocrinol. 7:1418–1429. Wu, C. 1995. Heat shock transcription factors: structure and regulation. Annu. Rev. Cell Dev. Biol. 11:441–469. Zou, Z., Y. Guo, T. Guettouche, D. Smith, and R. Voellmy. 1998. Repression of heat shock transcription factor HSF1 activation by hsp90 (HSP90 complex) that forms a stress-sensitive complex with HSF1. Cell 94:471–480.
