Visualization and Mechanism of Assembly of a Glucocorticoid Receptor (cid:1) Hsp70 Complex That Is Primed for Subsequent Hsp90-dependent Opening of the Steroid Binding Cleft*

A minimal system of five proteins, hsp90, hsp70, Hop, hsp40, and p23, assembles glucocorticoid receptor (GR) (cid:1) hsp90 heterocomplexes and causes the simultane-ous opening of the steroid binding cleft to access by steroid. The first step in assembly is the ATP-dependent and hsp40 (YDJ-1)-dependent formation of a GR (cid:1) hsp70 complex that primes the receptor for subsequent ATP-dependent activation by hsp90, Hop, and p23. This study focuses on three aspects of the GR priming reaction with hsp70. First, we have visualized the primed GR (cid:1) hsp70 complexes by atomic force microscopy, and we find the most common stoichiometry to be 1:1, with some complexes of a size ~1:2 and a few complexes of larger size. Second, in a recent study of progesterone receptor priming, it was shown that hsp40 binds first, leading to the notion that it targets hsp70 to the receptor. photometric detection of phenylthiocarbonyl derivatives of amino acids. Quantification was performed in reference to norleucine as an internal standard. The identity of the protein bands was confirmed by Western blotting membranes prepared in duplicate.

Glucocorticoid receptors (GR) 1 are recovered from hormonefree cells as large multiprotein complexes containing a dimer of hsp90 (for review see Refs. 1 and 2). Hsp90 binds to the ligand binding domain (LBD) of the GR, and the steroid binding activity of the GR is absolutely hsp90-dependent (3,4). When the GR is stripped of its associated hsp90, it immediately loses its ability to bind steroid, and steroid binding activity is regenerated when GR⅐hsp90 heterocomplexes are reformed by the hsp90/hsp70-based chaperone machinery (5,6). The steroids bind deep in a hydrophobic cleft that appears to be collapsed in the absence of ligand, such that the LBD must change its conformation to allow entry of ligand (7). The hsp90/hsp70-dependent chaperone machinery carries out the ATP-dependent opening of the binding cleft in the GR LBD such that it can be accessed by steroid. In addition to opening the steroid binding cleft, formation of a complex with hsp90 is accompanied by conformational changes that increase the sensitivity of the GR LBD to attack by thiol-derivatizing agents and trypsin (8 -10).
The multiprotein chaperone machinery that forms receptor⅐hsp90 heterocomplexes was originally identified in reticulocyte lysate (11,12). This machinery has been reconstituted (13), and a mixture of five purified proteins (hsp90, hsp70, 2 Hop, hsp40, and p23) is now used to achieve efficient receptor⅐hsp90 heterocomplex assembly (14,15). The chaperones hsp90 and hsp70 are both essential for opening the steroid binding cleft in the GR LBD, and hsp40, Hop (hsp70/hsp90 organizing protein), and p23 act as co-chaperones to increase the rate or extent of GR⅐hsp90 heterocomplex assembly (16). Hop binds independently to hsp90 and hsp70 to form an hsp90⅐Hop⅐hsp70 complex (17). Although Hop is not essential for GR⅐hsp90 heterocomplex assembly, when Hop is present to form a machinery, assembly is much faster (16). Hop does not act solely as a tie rod to bring the essential chaperones together; it also influences the conformational state and function of each protein (18). The hsp90⅐Hop⅐hsp70 complexes also contain small amounts of the hsp70 co-chaperone hsp40 (14), and together they form the hsp90/hsp70-based chaperone machinery. Once the machinery has assembled the GR⅐hsp90 heterocomplex, p23 binds dynamically (19) to the ATP-dependent conformation of hsp90 (20) and stabilizes its association with the receptor.
Stepwise assembly experiments with purified proteins are beginning to separate the assembly and cleft opening process into an ordered series of events in which two ATP-dependent steps have been resolved. In the first step, immunoadsorbed receptor that has been stripped of endogenous chaperones is incubated with purified hsp70 and hsp40 (YDJ-1) in the presence of ATP (5,21). This produces a receptor⅐hsp70⅐hsp40 complex that can be washed free of unbound hsp70 and hsp40 and then incubated with purified hsp90, Hop, and p23. In this first reaction, the receptor is "primed" to bind hsp90 and be acti-vated during the second incubation. Both the initial priming step with hsp70 and the second activating step with hsp90 are ATP-dependent and both are potassium-dependent, showing that there is a continuous requirement for the ATPase activity of hsp70 (5,22). Both the priming step and hsp90 binding are rapid, and it is the subsequent opening of the steroid binding cleft that is rate-limiting (22,23). In this work, we focus on three aspects of the GR priming reaction with hsp70.
In general, the stepwise assembly observations made with glucocorticoid and progesterone (PR) receptors have defined similar assembly mechanisms, but some differences in models of the priming step do exist. For example, by comparing Coomassie Blue-stained bands in the primed receptor complex, we found an hsp70/GR molar ratio of 1.1:1 (5), whereas Hernandez et al. (21) found an hsp70/PR molar ratio of ϳ2:1. Thus, we don't know whether the hsp70 is acting as a monomer or a dimer in the priming step. By amino acid analysis, we find here that the hsp70/GR molar ratio is ϳ1.3:1, and by atomic force microscopy of the primed complexes we see a distribution of several sizes, the most common of which reflects a 1:1 molar ratio.
A second difference relates to the role of hsp40 (we use the purified yeast homolog YDJ-1) in priming. We have shown that ATP-bound hsp70 binds directly to the hsp90-free, non-steroid binding GR, whereas hsp70 that has been converted to its ADP-bound conformation by preincubation with YDJ-1 does not (23). Thus, we have assumed that hsp70 is the first component of the chaperone machinery to contact the GR, and the role of YDJ-1 in priming is to promote the ATPase activity of GR-bound hsp70. In their studies with the PR, Toft and coworkers (21) found that YDJ-1 first binds with high affinity to the PR in a 1:1 molar ratio and is required for subsequent hsp70 binding. We show here that this added role of hsp40/ YDJ-1 in targeting the receptor for binding of hsp70 does not apply to the GR priming reaction.
We have shown previously that a short segment of the GR that lies at the extreme amino terminus of the LBD is required for LBD⅐hsp90 heterocomplex assembly and steroid binding activity (24). This segment lies at the rim of the ligand binding cleft of the receptor (25), and mutations within this segment alter both steroid binding activity and transcriptional activity of the GR (25,26). Although we know that GR LBD mutants lacking this segment do not form heterocomplexes with hsp90, we do not know whether it is the first hsp70-dependent step or the second hsp90-dependent step of assembly that is affected. By using the two-step assembly protocol, we show that a fusion protein lacking the requisite segment for hsp90 binding (GST/ 554C) forms a primed 554C⅐hsp70 complex that can bind hsp90/Hop, but hsp90 dissociates during the second assembly step, and no steroid binding activity is regenerated. This suggests that it is the second assembly step with hsp90 that is defective when the segment at the rim of the ligand binding cleft is not present.

Materials
Untreated rabbit reticulocyte lysate was purchased from Green Hectares (Oregon, WI). [6, H]Dexamethasone (40 Ci/mmol) and 125 Iconjugated goat anti-mouse and goat anti-rabbit IgGs were obtained from PerkinElmer Life Sciences. Protein A-Sepharose, goat anti-mouse, and goat anti-rabbit horseradish peroxidase conjugate, and monoclonal anti-GST antibody (clone GST-2) were from Sigma. Dulbecco's modified Eagle's medium was from BioWhittaker (Walkersville, MD). The BuGR2 monoclonal IgG antibody against the GR was from Affinity Bioreagents (Golden, CO). The AC88 monoclonal IgG against hsp90, the N27F3-4 anti-72/73-kDa hsp monoclonal IgG (anti-hsp70), and anti-hsp40 polyclonal IgG were from StressGen Biotechnologies (Victoria, British Columbia, Canada). TransFast transfection reagent was purchased from Promega (Madison, WI). Scanning probes were from Veeco Metrology (Sunnyvale, CA). Escherichia coli expressing YDJ-1 was a gift from Dr. Avrom Caplan (Mount Sinai School of Medicine). E. coli expressing Hop was kindly provided by Dr. David Smith (Mayo Clinic, Scottsdale, AZ). Hybridoma cells producing the FiGR monoclonal IgG against the GR were generously provided by Dr. Jack Bodwell (Dartmouth Medical School).

Methods
Transformation and Plasmid Purification-Construction of pMTGST520C and pMTGST554C, which contain GST and a thrombin cleavage site linked to the amino-terminal rat GR sequence 520 -795 (pMTGST520C) or 554 -795 (pMTGST554C), was described previously (24). The constructs were transformed into BL21 competent cells (Stratagene) selected on LB plates containing 50 g/ml ampicillin and grown in LB medium. The plasmid DNAs were extracted and purified using the Qiagen Mega kit.
Transient Transfection of GST/Fusion Protein-COS-7 cells were grown as monolayer cultures in 162-cm 2 culture flasks in Dulbecco's modified Eagle's medium with 10% fetal calf serum. The cultures were grown to ϳ50% confluency, washed, and incubated for 1 h with 5 ml of serum-free medium containing 25 g of plasmid DNA and 75 l of TransFast transfection reagent. The transfection medium was replaced with regular medium, and the cell cultures were continued for 48 h.
Expression of Mouse GR in Sf9 Cells and Cytosol Preparation-Mouse GR was expressed in Sf9 cells, and cytosol was prepared as described previously (5). Briefly, Sf9 cells were grown in SFM900 II serum-free medium (Invitrogen) supplemented with Cytomax (Kemp Biotechnology, Rockville, MD) in suspension cultures maintained at 27°C with continuous shaking. Cultures were supplemented with 0.1% glucose at infection of baculovirus and 24 h post-infection as described by Srinivasan et al. (27) and harvested by centrifugation. Cell pellets were washed in Hanks' buffered saline solution, resuspended in 1.5 volumes of HEM buffer (10 mM Hepes, 1 mM EDTA, and 20 mM sodium molybdate, pH 7.4) with 1 mM phenylmethylsulfonyl fluoride and 1 tablet of Complete-Mini protease inhibitor mixture per 3 ml of buffer, and ruptured by Dounce homogenization. The lysate was then centrifuged at 100,000 ϫ g for 30 min, and the supernatant, referred to as "cytosol," was collected, aliquoted, flash-frozen, and stored at Ϫ70°C.
Cultures of mouse fibroblast L929 (L cells) and transiently transfected COS-7 cells were harvested by scraping into Hanks' buffered saline solution. Cell pellets were washed, suspended, and ruptured, and cytosol was prepared using the same technique described above for the Sf9 cells.
Immunoadsorption of GR-Receptors were immunoadsorbed from aliquots of 50 l of Sf9 or 100 l of L cell cytosol by rotation for 2 h at 4°C with 18 l of protein A-Sepharose precoupled to 9 l of FiGR ascites suspended in 200 l of TEG (10 mM TES, pH 7.6, 50 mM NaCl, 4 mM EDTA, 10% glycerol). Prior to incubation with reticulocyte lysate or purified proteins, immunoadsorbed receptors were stripped of associated hsp90 by incubating the immunopellet for an additional 2 h at 4°C with 350 l of 0.5 M NaCl in TEG buffer. The pellets were then washed once with 1 ml of TEG buffer followed by a second wash with 1 ml of Hepes buffer (10 mM Hepes, pH 7.4).
GR⅐hsp90 Heterocomplex Reconstitution-For single step assembly of GR⅐hsp90 heterocomplexes, FiGR immunopellets containing GR stripped of chaperones were incubated with the five-protein assembly system (20 g of purified hsp90, 15 g of purified hsp70, 0.6 g of purified human Hop, 6 g of purified p23, 0.125 g of purified YDJ-1) adjusted to 55 l with HKD buffer (10 mM Hepes, pH 7.4, 100 mM KCl, 5 mM dithiothreitol), containing 20 mM sodium molybdate and 5 l of an ATP-regenerating system (50 mM ATP, 250 mM creatine phosphate, 20 mM magnesium acetate, and 100 units/ml creatine phosphokinase). The assay mixtures were incubated for 20 min at 30°C with suspension of the pellets by shaking the tubes every 2 min. At the end of the incubation, the pellets were washed twice with 1 ml of ice-cold TEGM buffer (TEG with 20 mM sodium molybdate) and assayed for steroid binding capacity and for receptor-associated proteins.
For two-step assembly, stripped GR immune pellets were mixed with hsp70, YDJ-1, and the ATP-regenerating system in a final volume of 55 l adjusted with HKD buffer and incubated for 5 min at 30°C. Following this priming reaction, pellets were washed once with TEG buffer and once with 10 mM Hepes buffer. The immunopellets were then incubated for 20 min at 30°C with the purified hsp90, Hop, and p23 adjusted to 55 l with HKD buffer containing 20 mM sodium molybdate and the ATP-regenerating system. At the end of the second incubation, the pellets were washed twice with 1 ml of ice-cold TEGM buffer and assayed for steroid binding capacity and for receptor-associated proteins.
Assay of Steroid Binding Capacity-Immune pellets to be assayed for steroid binding were incubated overnight at 4°C in 50 l of HEM buffer plus 100 nM [ 3 H]dexamethasone. Samples were then washed three times with 1 ml of TEGM buffer and counted by liquid scintillation spectrometry. The steroid binding is expressed as counts/min of [ 3 H]dexamethasone bound/FiGR immunopellet prepared from 100 l of cell cytosol.
For supernatants to be assayed for steroid binding, a 50-l aliquot of supernatant was incubated overnight at 4°C in 50 l of HEM buffer plus 100 nM [ 3 H]dexamethasone. Samples were mixed with dextrancoated charcoal and, after centrifugation, counted by liquid scintillation spectrometry. The steroid binding is expressed as counts/min of [ 3 H]dexamethasone bound/100 l of cell cytosol.
Gel Electrophoresis and Western Blotting-Immune pellets were resolved on 12% SDS-polyacrylamide gels and transferred to Immobilon-P membranes. The membranes were probed with 0.25 g/ml BuGR or 1 g/ml AC88, anti-hsp70, anti-hsp40, and anti-GST. The immunoblots were then incubated a second time with the appropriate 125 I-conjugated or horseradish peroxidase-conjugated counterantibody to visualize the immunoreactive bands.
Protein Concentration by Amino Acid Analysis-Stripped GR and hsp70-primed GR immunopellets were washed twice with TEG buffer, resolved on SDS-polyacrylamide gels, and transferred to Immobilon-P membranes. The membranes were stained with Coomassie Blue, and protein bands corresponding to the GR and hsp70 were excised and hydrolyzed in 6 N HCl at 110°C for 24 h. The hydrolysate was analyzed using a model 420H hydrolyzer/derivatizer from Applied Biosystems (Foster City, CA) with photometric detection of phenylthiocarbonyl derivatives of amino acids. Quantification was performed in reference to norleucine as an internal standard. The identity of the protein bands was confirmed by Western blotting membranes prepared in duplicate.
Peptide Construction and Competition-A peptide consisting of rat GR sequence 407-423 (Ser-Val-Phe-Ser-Asn-Gly-Tyr-Ser-Ser-Pro-Gly-Met-Arg-Pro-Asp-Val-Ser) was synthesized at the University of Michigan Protein Structure Core Facility using fluorenylmethoxycarbonyl (Fmoc) solid phase methodology. Synthesis was performed on a Symphony peptide synthesizer (Protein Technologies Incorporated, Woburn, MA). Peptide purity and structure were confirmed by high performance liquid chromatography and mass spectrometry. The peptide was dissolved in Hepes buffer, pH 7.4, to a final concentration of 100 mM and stored at Ϫ70°C. Immunopellets of GR and associated proteins bound to FiGR-protein A-Sepharose were washed once with TEG buffer and once with Hepes buffer, and the GR and GR-bound proteins were released into the supernatant by incubating the pellets with 1 mM of the peptide in 50 l of HEM buffer for 5 min at 30°C. At the end of the incubation, 16.7 l of HEM buffer was added, and the tubes were vortexed and centrifuged, and 50 l of the supernatants were collected for analysis.
Atomic Force Microscopy-All measurements were made with a Nanoscope IIIa Extended Multimode Scanning Probe Microscope from Digital Instruments (Santa Barbara, CA) using a 120 ϫ 120 m "JV" scanner and were conducted under fluid in tapping mode. NP-S oxidesharpened silicon nitride probes with a spring constant of 0.32 newton/m and a cantilever length of 100 m were used for imaging. Nominal tip radius of curvature for tips was reported to be Յ40 nm by the manufacturer. Fresh mica substrates were prepared by cleaving the mica surface immediately preceding sample deposition. Ten l of GR and associated proteins released from the FiGR-protein A-Sepharose pellet by peptide competition was deposited on the mica substrate and allowed to adsorb for 2 min, and 120 l of buffer was added to the fluid cell to ensure tip immersion. Scan parameters were optimized for each sample, with a typical tapping drive frequency of 9 kHz and a scan rate of 1 Hz. Multiple images for each sample were obtained with scan sizes ranging from 500 ϫ 500 nm to 2 ϫ 2 m. Because the theoretical width of the proteins is significantly smaller than the nominal tip radius, the height of each identified particle is used as a measure of its relative size. This makes it possible to limit the error associated with tip convolution artifacts. Heights of all identified particles were determined with Digital Instruments off-line section analysis routines.
Protein Purification-Hsp90 and hsp70 were purified from rabbit reticulocyte lysate by sequential chromatography on DE52, hydroxyapatite, and ATP-agarose as described previously (28). Human p23 was purified from 10 ml of bacterial lysate by chromatography on DE52 followed by hydroxyapatite chromatography as described (29). For purification of YDJ-1, bacterial sonicates were cleared by centrifugation, and YDJ-1 was purified by sequential chromatography on DE52 and hydroxyapatite as described previously (14). The bacterial expression of YDJ-1 has been described (30) as has the expression of human Hop (13).
Purification of human Hop was carried out in a similar manner by sequential chromatography on DE52 and hydroxyapatite. In all cases, the protein-containing fractions were identified by immunoblotting, and fractions from the final purification step were pooled, concentrated by Amicon filtration, dialyzed against HKD buffer, flash frozen, and stored at Ϫ70°C.

Stoichiometry of the Primed GR⅐hsp70
Complex-The priming and subsequent reactivation of the GR in the two-step assembly protocol we will use throughout this work is illustrated in Fig. 1. When the stripped GR (Str) is incubated with the five-protein purified system (PS) in a single step, twice as much steroid binding activity is generated as in the two-step protocol (cf. conditions 3 and 11). In two-step assembly, priming with hsp70 is absolutely required for binding of hsp90 and generation of steroid binding activity in the second step (cf. conditions 9 and 11). During the second incubation with hsp90, Hop, and p23, the majority of the hsp70 in the primed GR⅐hsp70 complex dissociates (cf. conditions 7 and 11).
We have used L cell cytosol as a source of GR in all of the stoichiometry experiments (Figs. [1][2][3][4][5], because ϳ100% of the L cell GR is bound to hsp90, and nearly complete GR⅐hsp90 reassembly and activation of steroid binding activity is achieved when the stripped GR is incubated with the purified assembly system (19). In contrast, only about 13% of the overexpressed mouse GR in Sf9 cytosol is bound by hsp90 and in the steroid binding state (5), and the remainder is in receptor aggregates (31) that we have not been able to activate to the steroid binding state with the purified five-protein system. In Stripped L cell GR immune (I) or nonimmune IgG (NI) pellets were incubated with the purified five-protein system in a single reaction (1-step) or by sequential addition (2-step). For one-step reactivation, immunopellets were incubated for 20 min at 30°C with the five-protein purified system (PS) and the ATP-regenerating system or with buffer alone (Str). For two-step reactivation, immunopellets were incubated for 5 min at 30°C with YDJ-1 and the ATP-regenerating system in the presence (ϩ70) or absence (Ϫ70) of hsp70. The pellets were washed once with TEG buffer and once with Hepes buffer (Prime) and incubated for 20 min at 30°C with hsp90, Hop, p23, and the ATP-regenerating system (Prime ϩ React). The proteins in the immunopellets were washed four times with TEGM buffer and resolved by SDS-gel electrophoresis and immunoblotting. Duplicate pellets were washed twice with TEGM buffer and assayed for steroid binding activity. Fig. 2, the stripped L cell GR was primed with hsp70; the GR and hsp70 bands were resolved ( Fig. 2A), and the bands were excised and quantitated by amino acid analysis. As shown in Fig. 2B, the molar ratio of hsp70 to GR in the primed complex was ϳ1.3:1.
Bacterial DnaK and its mammalian mitochondrial and cytoplasmic homologs self-associate in solution into dimers, trimers, and probably higher oligomers, with the equilibrium between these forms being dependent upon the concentration of the chaperone and the equilibrium being shifted toward the monomer by ATP or binding of peptide substrate (32)(33)(34). Under concentrated conditions, our purified hsp70 is predominantly dimeric on native gel electrophoresis with some trimers being detectable, yet under dilute conditions in reticulocyte lysate, the free hsp70 behaves as a monomer on molecular sieve chromatography (6), as did our purified hsp70 (data not shown). The molar ratios of hsp70 to receptor in the primed GR (ϳ1.3:1) and in the primed PR (ϳ2:1) (21) suggest that priming involves the interaction of a limited number of hsp70 molecules at a limited site on the receptor. However, such stoichiometry measurements may reflect an average for multiple complexes of different composition, and additional information can be gained by direct visualization of the primed GR⅐hsp70 complexes.
Visualization of the Primed GR⅐hsp70 Complexes by Atomic Force Microscopy-To visualize the primed complex, we wanted to release the GR from the FiGR-protein A-Sepharose immunopellet by competition with an epitope peptide. In the experiment of Fig. 3, native GR⅐hsp90 heterocomplexes were immunoadsorbed from L cell cytosol to demonstrate that the GR with its chaperones can be released as functional (i.e. steroid-binding) heterocomplexes. As the concentration of epitope peptide is increased, more GR is released from the pellet into the supernatant (Fig. 3B), and at 1 mM peptide, half of the GR is released and in the steroid-binding state (Fig. 3C). Fig. 4 shows control (no GR), stripped GR, and primed GR⅐hsp70 after the receptor was released from immunopellets with epitope peptide. The primed GR is released into the supernatant with its accompanying hsp70 (Fig. 4A). Fig. 4, B-D,   FIG. 3. Release of functional GR from immunopellet. A, illustration of release of GR⅐hsp90 heterocomplex. B, dissociation of immunoadsorbed GR from FiGR-protein A-Sepharose pellet. GR was immunoadsorbed from 100-l aliquots of cell cytosol and washed once with TEG buffer and once with Hepes buffer. Immunopellets were incubated for 5 min at 30°C with 0 -1000 M FiGR epitope peptide in HEM buffer, vortexed, and centrifuged. The supernatants were removed, and the pellets were washed four times with TEGM buffer. Proteins in the supernatants (S) and pellets (P) were resolved by SDS-gel electrophoresis and immunoblotted. HC, FiGR heavy chain. C, steroid binding activity of released GR. Immunopellets were prepared as in B and incubated for 5 min at 30°C with 0 -1000 M peptide in HEM buffer, vortexed, and centrifuged. The supernatants were removed, and the pellets were washed four times with TEGM buffer. Steroid binding activity of the supernatants (solid line) and pellets (dashed line) was assayed.
FIG. 2. Stoichiometry of the primed GR⅐hsp70 complex determined by amino acid analysis. A, priming of the GR. Stripped L cell GR immune pellets were incubated for 5 min at 30°C with buffer alone (Str) or hsp70, YDJ-1, and the ATP-regenerating system (P). The immunopellets were washed four times with TEG buffer, and proteins were resolved by SDS-gel electrophoresis and electrotransferred to Immobilon-P membrane. Bands of GR, hsp70, and FiGR heavy chain (HC) were visualized with Coomassie Blue. B, calculated stoichiometry of the primed GR⅐hsp70 complex. Immune pellets prepared as in A were washed four times with TEG buffer, and proteins were resolved by SDS-gel electrophoresis and electrotransferred to Immobilon-P membrane. GR and hsp70 were visualized with Coomassie Blue, and the protein bands were excised. The amount of protein in the GR bands (open bars) and hsp70 bands (solid bars) was measured by amino acid analysis. The bars represent average ratio of protein to that of the untreated, stripped GR Ϯ S.E. for three experiments. Note that the amount of hsp70 used in the priming reaction yields the maximum level of primed GR⅐hsp70 complex.
shows the atomic force microscopy images of the no GR control, the stripped GR, and the primed GR⅐hsp70 complex, respectively. Fig. 4, E and F, shows sample primary data in the form of cross-sectional height analysis of Fig. 4, C and D, respectively, at the plane indicated by the arrows. The heights of over 300 individual particles were determined, and a summary is presented in Fig. 5A, where the solid line in the main panel represents the stripped GR distributing as a sharp peak at 2.1 nm height and the line in the inset represents hsp70 distributing in a sharp peak at 1.6 nm. In contrast, the primed GR distributes in multiple peaks (dashed line). The 1st peak represents hsp70-free GR, and the 2nd and 3rd peaks distribute around ϳ3.5 nm and ϳ5.4 nm height, respectively.
Using the nadir between each peak marked by the dashed line in Fig. 5A, the particle heights were grouped in bins in . Pellets were stripped with 500 mM salt and incubated for 5 min at 30°C with buffer alone (Str) or hsp70, YDJ-1, and the ATP-regenerating system (Ϫ, Prime). Immunopellets were washed once with TEG buffer and once with Hepes buffer and incubated for 5 min at 30°C with 1 mM epitope peptide in HEM buffer, vortexed, and centrifuged. The supernatants were removed, and the pellets were washed four times with TEGM. Proteins in the supernatants (S) and pellets (P) were resolved by SDSgel electrophoresis and immunoblotted. HC, FiGR heavy chain. B-D, immune pellets were prepared as in A and competed with 1 mM peptide. Ten l of supernatants from no cytosol control (B), stripped GR (C), and primed GR⅐hsp70 (D) were overlaid on a mica substrate, and adherent particles were visualized by atomic force microscopy. Images are of 500 ϫ 500 nm scans tilted 30°toward the observer to aid definition, and height scale is from 0 (black) to 10 nm (white). Heights of individual particles (*) were measured by determining the local minimum and maximum of each peak. E, representative cross-sectional height analysis of C. Cross-section indicated by the arrow and dashed line in C. F, representative cross-sectional height analysis of D. Cross-section indicated by the arrow and dashed line in D.
5B to determine the relative abundance of particles in each peak. The number of particles in each bin, their mean size, and percentage of the total number of particles imaged are presented in Table I. Because there are no standard particles of known mass and height, we have used the masses and particle heights of the GR and hsp70 to predict the composition of the heterocomplex peaks. The most common heterocomplex peak (3.0 -4.5 nm height) is consistent with particles containing one molecule of GR (2.1 nm) plus one molecule of hsp70 (1.6 nm). A somewhat less common peak at 4.6 -6.3 nm height is roughly consistent with particles containing one molecule of GR plus two molecules of hsp70. Each of these broad peaks may have some particles that contain YDJ-1 as well. In addition to these two major GR⅐hsp70 peaks, there are larger peaks, possibly reflecting trimers and higher oligomers of hsp70 bound to the GR. Fig. 5, C-E, shows close-up images of a stripped GR of 2.1 nm (C) and primed GR⅐hsp70 complexes of 3.2 nm (D) and 5.3 nm (E) height. Fig. 5F presents one of the images of purified hsp70 used to obtain the hsp70 particle height distribution shown in the inset of Fig. 5A.
We have shown previously that the primed GR⅐hsp70 complexes are heterogeneous with respect to their nucleotide binding state, in that ϳ1/3 of the receptor-bound hsp70 is in the ADP-and ϳ2/3 is in the ATP-dependent state (22). From atomic force microscopy analysis (Fig. 5A), we see that the primed GR⅐hsp70 complexes are heterogeneous with respect to their size as well, with complexes consistent with 1:1 and 2:1 molar ratios of hsp70 to GR being predominant. It is not known whether all of the GR⅐hsp70 peaks or only one of the peaks shown in Fig. 5A are in a state that is primed to bind hsp90/ Hop in the second step reaction. Thus, it seems inappropriate at this time to form rigid models of the priming step based on stoichiometry data that require the action of specifically a monomer or a dimer of hsp70 on the receptor. It is possible that only one molecule of hsp70 is bound directly to and interacting productively with the receptor, and that under these conditions of assembly with purified proteins the receptor-bound hsp70 can associate with other molecules of hsp70 to form dimers and trimers, much as purified hsp70 and its homologs have been shown to do in solution (32)(33)(34). It is perhaps important to note that analysis of both the native and assembled hsp90⅐Hop⅐hsp70 machinery has revealed a molar ratio of 1 hsp70 per hsp90 dimer (6,18). If the preassembled machinery can interact with the receptor to produce receptor⅐hsp90 heterocomplexes, then it is reasonable to predict that the initial binding to the receptor and the priming step is by one molecule of hsp70 interacting with one molecule of receptor.
YDJ-1 Does Not Target Hsp70 to the GR in Single-step Assembly-In addition to their co-chaperone role in promoting the ATPase activity of hsp70, some members of the J-domain family of proteins bind to substrate proteins, leading to the speculation that they also serve to recruit hsp70 to the substrate (reviewed in Ref. 35). To understand the mechanism of receptor priming and, ultimately, receptor⅐hsp90 heterocomplex assembly, it is important to determine whether it is hsp70 or hsp40 that first contacts the receptor. Either hsp70 or hsp40 or both must recognize a common topological feature on the receptor and on other structurally unrelated client proteins that are assembled into heterocomplexes with hsp90. It has been proposed that the general topological feature that is recognized in proteins that are in their properly folded, least energy state is the region where hydrophobic surfaces of the protein interior merge with their hydrophilic exterior (2), this region being the opening of the steroid binding cleft in the case of the GR (see Ref. 26

and references therein).
It is clear that both GR (16) and PR (15) heterocomplexes can be assembled with purified proteins in the complete absence of YDJ-1/hsp40. Thus, there is not an absolute requirement for YDJ-1 co-chaperone function or for it to target hsp70 to either receptor. However, when it is present in the purified system, YDJ-1 could be the first component interacting with the receptor to facilitate hsp70 binding. Indeed, Hernandez et al. (21) have shown with the five-protein assembly system that YDJ-1 binds first to the PR, and they have shown that YDJ-1 binds directly to the PR in the absence of hsp70 with a stoichiometry of one molecule of YDJ-1 to one molecule of receptor, remaining bound to the PR in a static manner during the assembly process, i.e. the PR⅐YDJ-1 complexes could be washed free of unbound YDJ-1 and then incubated with the other components of the system to yield steroid binding PR⅐hsp90 heterocomplexes (21). In contrast, we have shown that hsp70 in its ATP-bound form, but not in its ADP-bound form, binds directly to the purified GR in the absence of YDJ-1 (23). Thus, two models of the priming step of assembly have evolved, with the PR study strongly suggesting that the initial recognition of the site for heterocomplex assembly resides with the YDJ-1 component and the GR study suggesting that this recognition function resides with the hsp70 component of the five-protein system.
In Fig. 6, various amounts of YDJ-1 were present during one-step assembly of GR⅐hsp90 heterocomplexes. It should be noted that YDJ-1 is not seen in the final GR⅐hsp90 heterocom- plex until there is 1 g of YDJ-1 present (Fig. 6A), yet a maximum effect on steroid binding activity is achieved at 0.1 g of YDJ-1. Normally, 0.125 g of YDJ-1 and 15 g of hsp70 are present in our five-protein assembly system to yield a ratio of ϳ1.5 molecules of YDJ-1 per 100 molecules of hsp70, consistent with the co-chaperone acting catalytically to promote hsp70 ATPase activity.
In Fig. 7, we have incubated the GR with YDJ-1, and we show binding when 1 or 5 g of YDJ-1 are present but trace or no binding when it is present at 0.1 g (Fig. 7A), the concentration yielding maximum activity in one-step assembly. In Fig. 7B, the GR was first incubated with various amounts of YDJ-1, and the immunopellets were washed and then incubated with the purified protein system without YDJ-1. In the absence of any YDJ-1, receptor activation to the steroid binding state is about 50% of that of the full five-protein system. In contrast to the results with the PR (21), prebinding with YDJ-1 produced only a small increase in steroid binding activity when the GR was subsequently incubated with the four-protein system. The highest amount of YDJ-1 examined was 5 g, which is 50 times the amount required for maximal effect in the one-step assay (Fig. 6B) and is the same amount used in the study showing hsp40 binding as the first step in PR⅐hsp90 heterocomplex assembly (21).
YDJ-1 Does Not Target Hsp70 to the GR in Two-step Assembly-It has been noted for both the GR and the PR that receptor activation in the two-step protocol (i.e. first step with hsp70 and YDJ-1, followed by second step with hsp90, Hop, and p23) is more dependent upon the presence of hsp40 than is activation by a single incubation with four proteins Ϯ YDJ-1 (2, 21, 22). The reason for this is not clear, but it prompted us to examine the effect of various concentrations of YDJ-1 in the first step of a two-step assembly. As shown in Fig. 8, in this two-step protocol, 0.1 g of YDJ-1 yields considerable but not maximal activation. From the immunoblot, it can be seen that YDJ-1 is in the washed immunopellets that were primed when 1 or 5 g of YDJ-1 were present, but in contrast to observations with the PR (21), YDJ-1 binding to the GR immunopellet is not static and YDJ-1 dissociates during the subsequent assembly step with hsp90.
Because 0.1 g of YDJ-1 is sufficient for maximal activity in one-step assembly with the five-protein system (Fig. 6), it is possible that samples preincubated with 5 g of YDJ-1 have enough YDJ-1 in the washed immunopellet to support hsp70 priming of the GR. In the experiment of Fig. 9A, GR was preincubated with 0.1 or 5 g of YDJ-1, and the immunopellets were washed and incubated with hsp70 and the ATP-regenerating system. The primed immunopellets were then washed a second time and incubated with hsp90, Hop, and p23. It can be seen that there is a substantial amount of YDJ-1 in the immunopellet after the first incubation (Y) with 5 g of YDJ-1. Much of the YDJ-1 disappears during the second incubation, but it was nevertheless sufficient to promote priming as evidenced by a substantial increase in hsp70 in the primed immunopellet (P) versus samples that were preincubated without FIG. 6. YDJ-1 increases the reactivation of GR steroid binding activity when YDJ-1 is part of a one-step incubation with the purified five-protein system. A, stripped GR immune pellets were incubated with buffer (Str), reticulocyte lysate (RL), or the purified protein system containing 0 -10 g of YDJ-1 in the presence of the ATP-regenerating system for 20 min at 30°C. The immunopellets were washed four times with TEGM, and proteins were resolved by SDS-gel electrophoresis and immunoblotted. B, immunopellets prepared as in A were washed twice with TEGM and assayed for steroid binding activity. Conditions are as follows: lane 1, immunopellet incubated with buffer; lane 2, with reticulocyte lysate; lanes 3-7, with the purified protein system containing 0, 0.01, 0.1, 1, or 10 g YDJ-1. Bars represent average of three experiments Ϯ S.E. FIG. 7. YDJ-1 effect in two-step assembly is minimal. A, stripped GR immune (I) or nonimmune IgG (NI) pellets were incubated with 0 -5 g of YDJ-1 for 5 min at 30°C. The immunopellets were washed four times with TEGM, and proteins were resolved by SDS-gel electrophoresis and immunoblotted. B, stripped GR immune pellets were incubated with 0 -5 g of YDJ-1 for 5 min at 30°C. The immunopellets were then washed once with TEG and once with 10 mM Hepes and then incubated with the purified protein system without YDJ-1. Conditions are as follows: lane 1, stripped GR (Str); lane 2, stripped GR incubated with the five-protein system (PS), containing 0.1 g of YDJ-1; lanes 3-6, immunopellets prebound with the indicated amount of YDJ-1 and incubated in the second step with a four protein system lacking YDJ-1.
YDJ-1 or with 0.1 g of YDJ-1. That the hsp70 priming is functional is shown in Fig. 9B where it is clear that the presence of high levels (1 and 5 g) of YDJ-1 in the first incubation permit activation of steroid binding activity in the third incubation, whereas 0.1 g of YDJ-1 yields the same steroid binding activity as the control without YDJ-1. YDJ-1 is not stably associated with the GR immunopellet, and most of it disappears during the priming step with hsp70, and virtually all of the rest is lost during the reactivation step with hsp90, Hop, and p23 (Fig. 9A).
It seems clear that we can repeat the observation made with the PR, in that the GR can be preincubated with a high amount of YDJ-1 and that the YDJ-1 associated with the washed GR immunopellet will permit receptor activation in subsequent priming with hsp70 followed by reactivation with hsp90. However, the amounts of YDJ-1 required are 10 -50-fold higher than the amount (0.1 g) required for maximal activation in a single incubation with five proteins. When 0.1 g of YDJ-1 is present during priming with hsp70, it promotes priming (Fig.  8), consistent with a catalytic co-chaperone stimulation of hsp70 ATPase activity. However, when the GR is preincubated with 0.1 g of YDJ-1 and subsequently primed with hsp70, there is no YDJ-1 binding to the GR and priming is not promoted (Fig. 9). These observations lead us to conclude that YDJ-1 does not interact first with the GR to target hsp70 to the receptor.
It should be noted that there is a fundamental difference between the GR and the PR that likely explains the difference between our data and the data of Hernandez et al. (21). The GR requirement for hsp90 binding is all or none. In the absence of hsp90, the binding cleft is closed and there is no high affinity steroid binding, and in the GR⅐hsp90 complex, the cleft is open and there is steroid binding. In contrast, hsp90 can be removed from the PR with the binding cleft remaining open, and the receptor maintains its steroid binding activity as long as it is not heated. When it is heated, the hsp90-free PR rapidly loses steroid binding activity, and the hsp90/hsp70-based chaperone machinery is required to maintain binding (36). In the experiments of Hernandez et al. (21), the PR is incubated with YDJ-1 while the steroid binding activity is being inactivated. Thus, YDJ-1 is exposed to an open steroid binding cleft with the potential for interaction with some exposed hydrophobic cleft interior. This may be why a stable PR⅐YDJ-1 complex is formed. If the PR is first bound by progesterone, which keeps the binding cleft closed, YDJ-1 does not bind to the receptor (21). When hsp90 is stripped from the GR, the binding cleft closes, and it is this closed cleft form of the receptor that we have exposed here to YDJ-1. It is also this closed cleft form of the GR that is directly bound by the ATP-dependent conformation of hsp70 (23).
Priming Is Not Dependent on the Hsp90-binding Segment of the GR LBD-By using fusion proteins containing glutathione S-transferase (GST) and short amino-terminal truncations just before and at the beginning of the GR LBD that are otherwise FIG. 8. Increasing concentrations of YDJ-1 increase receptor reactivation when YDJ-1 is incubated with hsp70 as part of a two-step priming reaction. Stripped GR immune pellets were incubated with buffer (Str) or primed with hsp70 and 0 -5 g of YDJ-1 and an ATP-regenerating system for 5 min at 30°C (P). The pellets were washed once with TEG buffer and once with 10 mM Hepes, and one sample of each pair was incubated for 20 min at 30°C with hsp90, Hop, p23, and the ATP-regenerating system (R). The proteins in the immunopellets were washed four times with TEGM and resolved by SDS-gel electrophoresis and immunoblotting. Duplicate pellets were washed twice with TEGM and assayed for steroid binding activity, with the lanes for steroid binding corresponding to the immunoblot lanes above. P stands for primed, and R stands for reactivated.

FIG. 9. Only high concentrations of YDJ-1 increase receptor reactivation when YDJ-1 is incubated with stripped GR prior to priming with hsp70.
A, stripped GR immune pellets were incubated with buffer (Str) or 0-5 g of YDJ-1 (Y) for 5 min at 30°C. The immunopellets were washed once with TEG and once with 10 mM Hepes and primed with hsp70 and the ATP-regenerating system (P) for 5 min at 30°C. The immunopellets were washed once with TEG and once with 10 mM Hepes and incubated with hsp90, Hop, p23, and the ATPregenerating system (R) for 20 min at 30°C. The immunopellets were washed four times with TEGM, and proteins were resolved by SDS-gel electrophoresis and immunoblotted. B, steroid binding activity of GR following three-step reactivation as in A, where stripped GR was incubated with 0, 0.1, 1, or 5 g of YDJ-1 in the first step. Stripped GR (Str) was also incubated with 0.1 g of YDJ-1 as part of a one-step reactivation with the complete five-protein system (PS).
intact to the carboxyl terminus, we have shown that a sevenamino acid segment (positions 547-553) lying in helix 1 of the rat GR LBD is required for both LBD⅐hsp90 heterocomplex assembly and steroid binding activity (24). In this earlier study, we focused solely on hsp90, and we did not assay the binding of hsp70. With the subsequent demonstration of the role of hsp70 in priming the GR for binding of hsp90 (5), the question arises as to whether this amino-terminal segment is required for heterocomplex assembly with hsp90 because it is required for priming by hsp70.
Here we have used two fusion proteins to answer this question. The fusion protein GST/520C behaves like the wild-type GR with respect to hsp90 binding and steroid binding, whereas the fusion protein GST/554C does not form a heterocomplex with hsp90 or have steroid binding activity. As shown in Fig.  10, GST/520C expressed in COS-7 cells is in steroid binding heterocomplexes containing both hsp90 and hsp70 (lane 4), whereas the fusion GST/554C (lane 6) is not in complex with hsp90 and is bound by only a trace of hsp70. Fig. 11 shows that GST/520C is bound by hsp70 in the priming step of two-step assembly with the purified system, and in the second step, it enters a heterocomplex with hsp90 and is reactivated to the steroid binding state. In contrast, GST/554C binds hsp70 in the FIG. 10. Cytosolic hsp90 and hsp70 are endogenously bound to GST/520C but not to GST/554C. A, diagram of full-length rat GR and GST/GR LBD truncation mutants GST/520C and GST/554C. The amino-terminal activation domain (AD) and DNA binding domain (DBD) are indicated in the full-length GR, as is the hinge region (hatched) between the DBD and LBD. B, endogenous GST/LBD⅐hsp90 heterocomplexes and steroid binding activity. Cytosols were prepared from Mock (lanes 1 and 2) or transiently transfected COS-7 cells overexpressing GST/520C (520C, lanes 3 and 4) or GST/554C (554C, lanes 5 and 6). The GRs were immunoadsorbed from 100-l aliquots of cytosol with nonimmune (NI) or GST-2 antibody (I) in TEGM buffer. Immunopellets were washed four times in TEGM buffer and assayed for GST/LBD, hsp90, and hsp70 by Western blotting. Duplicate pellets were washed twice with TEGM buffer and assayed for steroid binding activity. The steroid binding data represent average values from three experiments Ϯ S.E.
FIG. 11. Both GST/520C and GST/554C can be primed by hsp70, but only GST/520C can form a stable complex with hsp90 and generate steroid binding. Stripped GST/LBD immune (I) or nonimmune IgG (NI) pellets were incubated with the purified five-protein system in a two-step reactivation. Immunopellets were incubated with buffer (Str) or hsp70, YDJ-1, and the ATP-regenerating system for 5 min at 30°C. The pellets were washed once with TEG buffer and once with 10 mM Hepes (Prime), and one sample of each pair was incubated for 20 min at 30°C with hsp90, Hop, p23, and the ATP-regenerating system (React). The proteins in the immunopellets were washed four times with TEGM and resolved by SDS-gel electrophoresis and immunoblotting. Replicate pellets were washed twice with TEGM buffer and assayed for steroid binding activity. The steroid binding data represent average values from three experiments Ϯ S.E.
FIG. 12. Both GST/520C and GST/554C can form an initial complex with hsp90. GST/LBD immune pellets were stripped (Str) and incubated with buffer (Ϫ) or with hsp70, YDJ-1, and the ATP-regenerating system (ϩ) for 5 min at 30°C. The pellets were washed once with TEG buffer and once with Hepes buffer and incubated for 1 min at 30°C with hsp90 and Hop. The proteins in the immunopellets were washed four times with TEGM and resolved by SDS-gel electrophoresis and immunoblotting.
priming step, but it does not form a heterocomplex with hsp90 during the second step reaction.
The second step reaction can be divided into two parts: initially, there is rapid binding of hsp90 and Hop with the primed GR⅐hsp70 complex, and this is followed by slow, ATPdependent formation of the GR⅐hsp90 heterocomplex with opening of the steroid binding cleft (23,23). To determine whether GST/554C has been primed such that it can bind hsp90, we incubated primed and unprimed complexes for 1 min with hsp90 and Hop. As shown in Fig. 12, both the GST/520C and GST/554C fusion proteins bound hsp90 when they were primed with hsp70. It could be that this early (1 min) complex represents a GR⅐hsp70⅐Hop⅐hsp90 complex in which a Hop⅐hsp90 unit has bound to GR-bound hsp70 but not to the GR. During the long incubation (20 min) of Fig. 11, the hsp70 dissociates from GST/554C, as it does from GST/520C, but because hsp90 cannot engage with the GST/554C, hsp90 and Hop dissociate with the hsp70. However, insofar as we can assess the priming reaction by binding of hsp70 and by the ability of the GST/LBD⅐hsp70 complex to bind hsp90, GST/ 554C appears to prime normally. It therefore appears that the amino-terminal segment that lies at the rim of the ligand binding cleft of the GR is required for subsequent ATP-dependent and hsp90-dependent opening of the cleft with concomitant assembly of a GR⅐hsp90 heterocomplex.
Priming Reflects a Focal Chaperone Attack on the Receptor-The major common theme that emanates from the mechanistic studies of GR⅐hsp90 and PR⅐hsp90 heterocomplex assembly is the very focal nature of the process. Only the receptor LBD is involved (1), and the attack of the chaperone machinery within that domain focuses on the steroid binding cleft. When we visualize the GR⅐hsp70 complexes formed in the priming reaction, ϳ70% involve a molecule of GR binding to only one or two molecules of hsp70 (Table I), and Toft and co-workers (21) found that one molecule of YDJ-1 bound to one molecule of PR. This very limited chaperone binding is consistent with a focal site of attack.
Although we find that binding of YDJ-1 is not required for targeting hsp70 to the GR, the hsp70 co-chaperone may very well contact the receptor. The observations of Hernandez et al. (21) with the PR suggest that YDJ-1 is bound directly to the receptor when it interacts as a co-chaperone with receptorbound hsp70. Inasmuch as hsp40/YDJ-1 promotes the ATPase activity of hsp70 (37) and hsp70 ATPase activity is required for both GR priming by hsp70 and the subsequent activation of the primed receptor complex by hsp90 (5,22), it is reasonable to propose that stimulation of hsp70 ATPase activity is the principal function of hsp40/YDJ-1 in opening of the steroid binding cleft. Although hsp40/YDJ-1 acts on the receptor-bound hsp70, it may also interact directly with the receptor. DnaJ has been shown to have an affinity for unfolded proteins (38,39), and this chaperone function of hsp40/YDJ-1 would bind it to a hydrophobic region in the receptor, forming a ternary ATP⅐hsp70⅐LBD⅐hsp40 complex in which the J domain of hsp40 is interacting also with receptor-bound ATP⅐hsp70 to promote ATP hydrolysis. This has been proposed by Han and Christen (40) for the bacterial homologs DnaJ and DnaK bound to substrate. In such a ternary complex, the hsp40 would not be targeting hsp70 to the GR LBD, but the formation of a ternary complex may improve the efficiency of co-chaperone interaction of hsp40 with receptor-bound hsp70. The likely site for this focal attack is the region where the hydrophobic interior of the steroid binding cleft merges with the receptor surface.
Because the steroid binding cleft of the PR remains open after dissociation of hsp90, much more of the hydrophobic cleft interior is exposed to favor a stable binding of hsp40. This is not observed with the GR, because the binding cleft zips up tightly when hsp90 dissociates, leaving less hydrophobic surface exposed for interaction with hsp40 and making the ternary complexes less stable. The concept of a ternary complex is important for our mechanistic understanding of the receptor priming process, as it may improve the efficiency of hsp40 stimulation of hsp70 ATPase activity. However, the GR example shows us that hsp40 binding is not required to "target" hsp70 to the receptor.