Stoichiometry, abundance, and functional significance of the hsp90/hsp70-based multiprotein chaperone machinery in reticulocyte lysate.

Rabbit reticulocyte lysate contains a multiprotein chaperone system that assembles the glucocorticoid receptor (GR) into a complex with hsp90 and converts the hormone binding domain of the receptor to its high affinity steroid binding state. This system has been resolved into five proteins, with hsp90 and hsp70 being essential and Hop, hsp40, and p23 acting as co-chaperones that optimize assembly. Hop binds independently to hsp70 and hsp90 to form an hsp90.Hop.hsp70 complex that acts as a machinery to open up the GR steroid binding site. Because purified hsp90 and hsp70 are sufficient for some activation of GR steroid binding activity, some investigators have rejected any role for Hop in GR.hsp90 heterocomplex assembly. Here, we counter that impression by showing that all of the Hop in reticulocyte lysate is present in an hsp90.Hop.hsp70 complex with a stoichiometry of 2:1:1. The complex accounts for approximately 30% of the hsp90 and approximately 9% of the hsp70 in lysate, and upon Sephacryl S-300 chromatography the GR.hsp90 assembly activity resides in the peak containing Hop-bound hsp90. Consistent with the notion that the two essential chaperones cooperate with each other to open up the steroid binding site, we also show that purified hsp90 and hsp70 interact directly with each other to form weak hsp90.hsp70 complexes with a stoichiometry of 2:1.


From the Department of Pharmacology, The University of Michigan Medical School, Ann Arbor, Michigan 48109
Rabbit reticulocyte lysate contains a multiprotein chaperone system that assembles the glucocorticoid receptor (GR) into a complex with hsp90 and converts the hormone binding domain of the receptor to its high affinity steroid binding state. This system has been resolved into five proteins, with hsp90 and hsp70 being essential and Hop, hsp40, and p23 acting as co-chaperones that optimize assembly. Hop binds independently to hsp70 and hsp90 to form an hsp90⅐Hop⅐hsp70 complex that acts as a machinery to open up the GR steroid binding site. Because purified hsp90 and hsp70 are sufficient for some activation of GR steroid binding activity, some investigators have rejected any role for Hop in GR⅐hsp90 heterocomplex assembly. Here, we counter that impression by showing that all of the Hop in reticulocyte lysate is present in an hsp90⅐Hop⅐hsp70 complex with a stoichiometry of 2:1:1. The complex accounts for ϳ30% of the hsp90 and ϳ9% of the hsp70 in lysate, and upon Sephacryl S-300 chromatography the GR⅐hsp90 assembly activity resides in the peak containing Hopbound hsp90. Consistent with the notion that the two essential chaperones cooperate with each other to open up the steroid binding site, we also show that purified hsp90 and hsp70 interact directly with each other to form weak hsp90⅐hsp70 complexes with a stoichiometry of 2:1.
Unliganded steroid receptors exist in cytosols in a heterocomplex with the ubiquitous protein chaperone hsp90 1 (for review, see Ref. 1). Hsp90 binds to the ligand binding domain (LBD) of the receptors (1), and the glucocorticoid receptor (GR) LBD must be bound to hsp90 for the receptor to have high affinity steroid binding activity (2,3). The receptor⅐hsp90 heterocomplexes are assembled by a multiprotein chaperone system that was first studied in reticulocyte lysate (4,5). Both biochemical data (6) and data from GR mutants (7,8) support a model (3) in which the hydrophobic ligand binding cleft in the LBD is opened to access by steroid during heterocomplex as-sembly. The assembly system in reticulocyte lysate has been reconstituted (9), and a mixture of five purified proteins, hsp90, hsp70, 2 Hop, hsp40, and p23, is now used to achieve optimal receptor⅐hsp90 heterocomplex assembly (10,11).
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 (12). Hop binds independently to hsp90 and hsp70 to form an hsp90⅐Hop⅐hsp70 complex (13), and assembly proceeds faster when Hop is present to bring the two essential chaperones together (12). These complexes also contain small amounts of the hsp70 co-chaperone hsp40 (10), and together they form the hsp90/hsp70-based chaperone "machinery." The chaperone machinery can be prepared simply by mixing purified components, or it can be immunoadsorbed from reticulocyte lysate with a monoclonal antibody against Hop (10,14). When mixed with immunoadsorbed GR, the immunoadsorbed chaperone machinery converts the GR to its steroid binding form in an ATP-dependent manner (10,14). Once the machinery has assembled the GR⅐hsp90 heterocomplex, p23 binds dynamically (15) to the ATP-dependent conformation of hsp90 (16) and stabilizes its association with the receptor.
The Hop (p60) of rabbit reticulocyte lysate (17) is the homolog of a human protein cloned by Honoré et al. (18) and the non-essential yeast heat shock protein Sti1 (19). Unlike hsp70 and hsp90, Hop alone does not possess any chaperone activity in protein refolding assays (20,21). Hop contains multiple tetratricopeptide repeats (TPR), with separate TPR domains determining its binding to hsp70 and hsp90. The N-terminal TPR1 domain binds to the C terminus of hsp70, and the central TPR2 domain binds to a TPR acceptor site in the C terminus of hsp90 (13,(22)(23)(24)(25)(26). In addition to bringing hsp70 and hsp90 together into the machinery for opening the steroid binding cleft in the GR LBD (14), studies of purified Hop⅐hsp90 interaction show that Hop inhibits hsp90 ATPase activity (27,28). Inasmuch as hsp90 ATPase activity is required to generate steroid binding activity (29) and the presence of Hop in the purified five-protein system accelerates the rate at which GR steroid binding activity is generated (12), inhibition of hsp90 ATPase activity may not be a critical component of Hop function in the activation of steroid binding sites by the hsp90⅐Hop⅐hsp70 machinery.
In 1992, we purified (by ammonium sulfate precipitation and molecular sieve chromatography) a high molecular mass complex from rabbit reticulocyte lysate that contained hsp90 and * This work was supported by National Institutes of Health Grant DK31573. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  hsp70 and had a low ability to assemble GR⅐hsp90 heterocomplexes with steroid binding activity (30). A factor that was separated from the complex during the ammonium sulfate step was required for efficient heterocomplex reconstitution (30), and this factor was identified as p23 (31). Subsequently, heterocomplexes containing hsp90, Hop, and hsp70 were immunoadsorbed from reticulocyte lysate with monoclonal antibodies against hsp90 (32) or Hop (14) and shown to assemble GR⅐hsp90 heterocomplexes with steroid binding activity. This multiprotein chaperone complex, originally called a "foldosome" (32), is now known as the hsp90/hsp70-based chaperone machinery. Yeast (Saccharomyces cerevisiae) contains similar hsp90(hsp82)⅐Hop(Sti1)⅐hsp70(Ssa) complexes (33), and mutation of sti1 results in decreased GR activation of a reporter gene in vivo (34).
These observations suggest that the hsp90⅐Hop⅐hsp70 chaperone machinery plays a role in GR⅐hsp90 heterocomplex assembly under cell-free conditions by reticulocyte lysate and by yeast in vivo, but Hop is not essential in either case. Immune depletion of Hop from reticulocyte lysate, for example, reduces its ability to generate GR steroid binding activity by 50% (12), and Sti1 mutant yeast still have one-third the GR activity of wild-type yeast at high hormone concentration (34). Recently, studies utilizing purified proteins for GR⅐hsp90 assembly have caused some investigators to totally repudiate the notion of the hsp90/hsp70-based chaperone machinery (35). In a purified protein assembly system, GR steroid binding activity can be generated with a combination of the core chaperones hsp90 and hsp70 without Hop or hsp40 (12,35). Also, GR⅐hsp90 heterocomplexes can be assembled by purified proteins in a two-step procedure that does not involve prior formation of the hsp90⅐Hop⅐hsp70 assembly machinery (36,37). However, we have shown that the rate of GR⅐hsp90 heterocomplex assembly is accelerated markedly when Hop is added to a mixture of purified hsp90, hsp70, YDJ-1, and p23 (12). Because Hop binds to hsp70 and hsp90, it co-purifies with both, and fastidious purification procedures are required to assure Hop-free preparations of these chaperones (12). It is likely that Hop contamination of the purified chaperones accounts for the difference between those who have concluded that Hop has no effect (35) and both ourselves (9,12,14) and the Toft laboratory (11) who have shown that the presence of Hop in a purified five-protein system optimizes receptor⅐hsp90 heterocomplex assembly.
Although assembly with the purified chaperones is useful for studying the mechanism by which GR⅐hsp90 heterocomplexes are formed and how steroid binding sites are created, the purified system does not replicate the conditions of reticulocyte lysate. The concentrations of hsp90 and hsp70 in the purified system, for example, are several times those of reticulocyte lysate. Here, we revisit the reticulocyte lysate system to determine the relative abundance of hsp90, hsp70, and Hop and the amount of each protein that is present in the chaperone machinery. We find that all of the Hop, ϳ30% of the hsp90, and ϳ9% of the hsp70 in rabbit reticulocyte lysate exist in hsp90⅐Hop⅐hsp70 heterocomplexes. When the reticulocyte lysate is separated into two pooled fractions by molecular sieve chromatography, 92% of the ability to activate GR steroid binding activity resides in the large M r peak of hsp90, which contains 92% of the Hop. A second pool of fractions containing the unbound hsp90 (38%) and hsp70 (53%) and 8% of the Hop has a low GR activating activity. This suggests that in reticulocyte lysate the Hop-containing chaperone machinery plays the major role in GR⅐hsp90 heterocomplex assembly, with free hsp90 and hsp70 playing a considerably lesser role. We have determined by both native gel electrophoresis and cross-linking that the chaperone machinery in reticulocyte lysate possesses an hsp90:Hop:hsp70 stoichiometry of 2:1:1. We also show by cross-linking that purified hsp90 and hsp70 bind weakly to each other in the absence of Hop, producing an hsp90⅐hsp70 complex with a stoichiometry of 2:1. This is consistent with the notion (37) that hsp90 and hsp70 interact directly with each other while the steroid binding cleft is being opened by either the chaperone machinery or by the purified chaperones without Hop.

Materials
Untreated rabbit reticulocyte lysate was purchased from Green Hectares (Oregon, WI). [6, H]Triamcinolone acetonide (38 Ci/mmol) and 125 I-conjugated goat anti-mouse IgG were obtained from PerkinElmer Life Sciences. Sephacryl S-300 was from Amersham Pharmacia Biotech. Protein A-Sepharose, goat anti-mouse horseradish peroxidase conjugate, and molecular weight markers used for non-denaturing gels were from Sigma Chemical Co. The BuGR2 monoclonal IgG antibody against the GR and the 3G3 monoclonal IgM against hsp90 were from Affinity Bioreagents (Golden, CO). The AC88 monoclonal IgG against hsp90 and the N27F3-4 anti-72/73-kDa hsp monoclonal IgG (anti-hsp70) were from StressGen (Victoria, BC, Canada). Escherichia coli expressing YDJ-1 was a gift from Dr. Avrom Caplan (Mount Sinai School of Medicine). The DS14F5 monoclonal IgG against Hop and E. coli expressing Hop were 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).

FIG. 1. Quaternary structures of hsp90 and hsp70.
A, immunoblots of purified rabbit hsp90 and hsp70 that were electrophoresed under non-denaturing conditions. B, immunoblots of uncrosslinked (Ϫ) or glutaraldehyde cross-linked (ϩ) hsp90 and hsp70 electrophoresed under denaturing conditions. 90 and 70 refer to monomers of hsp90 and hsp70; and 90⅐90 and 70⅐70 refer to the homodimers. C, immunoblots of hsp90, hsp70, and Hop in 10-l aliquots of reticulocyte lysate electrophoresed under non-denaturing conditions. hsp90/hsp70-based Chaperone Machinery Methods Immunoadsorption of GR-Mouse GR was expressed in Sf9 cells, and cytosol was prepared as previously described (36). Receptors were immunoadsorbed from 50-l aliquots of Sf9 cytosol by rotation for 2 h at 4°C with 14 l of protein A-Sepharose precoupled to 7 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 lysate subfractions, 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. The pellets were then washed once with 1 ml of TEG followed by a second wash with 1 ml of Hepes buffer (10 mM Hepes, pH 7.4).
Glucocorticoid Receptor Heterocomplex Reconstitution-For assembly of GR⅐hsp90 heterocomplexes, FiGR immunopellets containing GR stripped of chaperones were incubated with 40 l of reticulocyte lysate or with 40 l of a lysate subfraction from Sephacryl S-300 chromatography plus 6 g of purified p23 and 0.4 g of purified YDJ-1. Incubation volumes were adjusted to 50 l with HKD buffer (10 mM Hepes, 100 mM KCl, 5 mM dithiothreitol, pH 7.35) 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.
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 (10 mM Hepes, pH 7.35, 1 mM EDTA, 20 mM molybdate) plus 50 nM [ 3 H]triamcinolone acetonide. Samples were then washed three times with 1 ml of TEGM and counted by liquid scintillation spectrometry. The steroid binding is expressed as counts per minute of [ 3 H]triamcinolone acetonide bound/FiGR immunopellet prepared from 50 l of Sf9 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 for GR or 1 g/ml AC88 for hsp90. The immunoblots were then incubated a second time with the appropriate 125 I-conjugated or horseradish peroxidase-conjugated counterantibody to visualize the immunoreactive bands. For electrophoresis under non-denaturing conditions, 10 l of reticulocyte lysate was mixed with 50 l of detergent-free buffer (312 mM Tris-HCl, pH 6.8, 50% glycerol, 0.05% bromphenol blue), and proteins were resolved on a 7.5% polyacrylamide gel, followed by Western blotting. The immunoblots were probed with AC88 for hsp90, 1 g/ml N27F3-4 for hsp70, or 0.1% DS14F5 mouse ascites for Hop. Molecular weight markers for non-denaturing gels were bovine serum albumin, monomer (66,000) and dimer (132,000), and Jack bean urease, trimer (272,000) and hexamer (545,000).
Glutaraldehyde Cross-linking of Purified Proteins-For cross-linking of purified hsp90 or hsp70, 15 g of purified protein was incubated for 1 h at room temperature with 0.8 mM glutaraldehyde in a final volume of 50 l adjusted with HKD buffer. The cross-linking was terminated by adding 12 l of 0.5 M Tris, pH 8.0, and continuing incubation for 30 min at room temperature. Proteins were resolved by SDS-polyacrylamide gel electrophoresis on 6% gels followed by Western blotting.
Relative Abundance of Proteins in Heterocomplexes and Reticulocyte Lysate-For determining stoichiometry of hsp90⅐Hop⅐hsp70 heterocomplexes, 50-l aliquots of reticulocyte lysate were immunoadsorbed to 18 l of protein A-Sepharose prebound with 0.5 l (ϳ7 g) of DS14F5 antibody against Hop or non-immune mouse IgG. The samples were rotated at 4°C for 2 h, and immunopellets were washed three times with 1 ml of TEG buffer. Relative amounts of hsp90 and hsp70 in immunoadsorbed Hop complexes were estimated by resolving the immune pellet proteins on 12% SDS-polyacrylamide gels and staining with Coomassie Blue. Ratios of hsp90 to hsp70 were determined by scanning multiple stained bands. For cross-linking, Hop immune pellets were suspended in 50 l of HKD buffer and incubated with 0.4 mM glutaraldehyde as described above.
To determine the concentration of Hop, hsp90 and hsp70 in reticulocyte lysate, aliquots of lysate were electrophoresed on SDS-polyacrylamide gels that also contained various amounts of purified hsp90, hsp70, and Hop to provide a standard curve for each protein. Immunoblots were prepared and probed with monoclonal IgGs against each protein, followed by incubation with 125 I-labeled anti-IgG counterantibody. Samples and purified standards were then excised and counted to permit calculation of the concentration of each protein.
Protein Purification-hsp90 and hsp70 were purified from rabbit reticulocyte lysate by sequential chromatography on DE52, hydroxyapatite, and ATP-agarose as described previously (38). Human p23 was purified from 10 ml of bacterial lysate by chromatography on DE52 followed by hydroxyapatite chromatography as described previously (39). 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 (10). The bacterial expression of YDJ-1 has been described (40) as has the expression of human Hop (9). 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.
Fractionation of Reticulocyte Lysate-Reticulocyte lysate (500 l) was applied to a column (1.5 ϫ 113 cm) of Sephacryl S-300, and the column was eluted with HKD buffer. Aliquots (100 l) of each 2.5-ml fraction were assayed for hsp90, hsp70, and Hop by SDS-polyacrylamide gel electrophoresis and immunoblotting, and total protein was assayed by the Bradford method. The indicated fractions were pooled and contracted to a volume of 300 l by Amicon filtration. Aliquots (40 l) of each contracted pool were assayed for GR⅐hsp90 heterocomplex reconstitution as described above.

RESULTS AND DISCUSSION
Quaternary Structures of hsp90, hsp70, and Hop-hsp90 forms homodimers, and dimerization is required for both receptor⅐hsp90 heterocomplex assembly in reticulocyte lysate (41) and hsp90 function iv vivo (42). Bacterial DnaK (43) and its mammalian mitochondrial (44) and cytoplasmic (45) hsp70/ hsc70 homologs self-associate in solution into dimers, trimers, and probably higher oligomers in a concentration-dependent equilibrium. We have examined, by both non-denaturing gel electrophoresis and cross-linking, the quaternary structure of the purified rabbit hsp90 and hsp70 used in the five-protein GR⅐hsp90 heterocomplex assembly system. The purified hsp90 is a mixture of monomer and dimer (Fig. 1A), and dimer can be recovered on denaturing gel electrophoresis after glutaraldehyde cross-linking (Fig. 1B). Few monomers of hsp90 are detected on non-denaturing gel electrophoresis of whole reticulocyte lysate (Fig. 1C). Thus, it appears that the hsp90 monomers are formed during purification, and the monomers account for 30 -50% of each purified preparation. Our purified hsp70 is predominantly dimeric on native gel electrophoresis (Fig. 1A), hsp90/hsp70-based Chaperone Machinery with some trimers being detectable, and the dimer can be seen on denaturing gel electrophoresis after glutaraldehyde crosslinking of the purified protein (Fig. 1B). Our purified, bacterially expressed Hop behaves as very large and diffuse aggregates on native gel electrophoresis, and we were not able to detect monomers, dimers, or discrete multimers (data not shown).
In Fig. 1C, aliquots of rabbit reticulocyte were submitted to gel electrophoresis under non-denaturing conditions and immunoblotted for hsp90, hsp70, and Hop, respectively. The hsp90 in reticulocyte lysate was detected in two major bands, with one being the homodimer (180-kDa band seen above the 132-kDa marker) and a second migrating at ϳ310 kDa. A portion of the hsp70 and all of the Hop in reticulocyte lysate also behave as a similar 310-kDa complex. Most of the hsp70 in reticulocyte lysate distributed on native gels as diffuse species of M r Ͼ 310, and no monomers or dimers of hsp70 were detectable by immunoblotting with the N27F3-4 antibody (Fig. 1C).
The 310-kDa Complex Contains hsp90, hsp70, and Hop-To ask if hsp90, hsp70, and Hop were present in the same complex, aliquots of reticulocyte lysate were immunoadsorbed with monoclonal antibody specific for Hop (Fig. 2), and the immu-noadsorbed supernatants were analyzed by native gel electrophoresis and immunoblotting. As shown in Fig. 2, immunodepletion of Hop yielded significant immunodepletion of the 310-kDa band of hsp90 and hsp70 without affecting the 180-kDa band of hsp90. Co-immunodepletion of all three proteins suggests that hsp90, hsp70, and Hop are present in the same ϳ310-kDa complex.
The migration of the complex at ϳ310 kDa under non-denaturing gel electrophoresis would be consistent with an hsp90: Hop:hsp70 stoichiometry of 2:1:1. In Fig. 3, aliquots of reticulocyte lysate were immunoadsorbed with the monoclonal antibody against Hop, the immune pellets were washed three times, and proteins in the immune pellets were detected by denaturing gel electrophoresis and staining with Coomassie Blue (Fig. 3A). The bands for hsp90, hsp70, and Hop were scanned in a densitometer to determine approximate ratios of each protein in the washed immune pellet. Fig. 3B presents the average ratios obtained from multiple separate immunoadsorptions like that of Fig. 3A. The ratio of hsp90 to hsp70 in the washed Hop immune pellets is about 2:1. But only 62% of the Hop is bound to hsp70 and the ratio of hsp90 to Hop is 1.24:1, consistent with 62% of the Hop being bound to dimers of hsp90.

hsp90/hsp70-based Chaperone Machinery
Thus, it appears that a significant amount of hsp70 and hsp90 are lost during washing of the Hop immune pellet.
To determine the stoichiometry of hsp90:Hop:hsp70 in the complex, Hop immune pellets prepared as in Fig. 3A were cross-linked with glutaraldehyde and resolved by denaturing gel electrophoresis and immunoblotting. In Fig. 4A, the blots were probed with the AC88 monoclonal antibody specific for hsp90, and the bands were developed with radiolabeled counterantibody. As shown in lane 8 (Fig. 4A), an immune pellet that was not exposed to reticulocyte lysate but was exposed to cross-linker yielded three major bands representing crosslinked immunoglobulin chains detected with the secondary anti-IgG antibody. In lanes 4 -6 of Fig. 4A, Hop immune pellets were prepared from reticulocyte lysate and treated with glutaraldehyde. These cross-linked samples reveal an hsp90-con- FIG. 5. Fractionation of reticulocyte lysate by Sephacryl S-300 chromatography. An aliquot (500 l) of reticulocyte lysate was passed through a column of Sephacryl S-300 in HKD buffer. Each fraction was assayed for total protein by the Bradford method (solid line of A 595 at top) and hsp90, hsp70, and Hop by immunoblotting. The immunoblots of hsp90, hsp70, and Hop were scanned with a densitometer, and the relative amount of protein was expressed as a percentage of the highest fraction to yield the curves shown at the bottom. The brackets define the column fractions that were pooled and contracted to form the fractions A and B, which were used for the assays of Fig. 6. Hb indicates hemoglobin, which elutes in the major protein peak.   7. Cross-linking of purified hsp90 and hsp70. Purified hsp90 and hsp70 and a mixture of both were treated for 1 h with 0.4 mM glutaraldehyde (X-linker), and the cross-linked products were electrophoresed under denaturing conditions. The three lanes on the left were immunoblotted with AC88 antibody against hsp90 and the three lanes on the right were blotted with anti-hsp70. The numbers refer to hsp90, hsp70, the homodimers, and hsp70 cross-linked to a dimer of hsp90. hsp90/hsp70-based Chaperone Machinery taining complex migrating at ϳ310 kDa that is indicated by a horizontal arrow. As shown in Fig. 4B, this complex is also detected when immunoblots are probed with antibodies specific for Hop (lane 2) and for hsp70 (lane 4) but not when they are exposed only to the anti-IgG counterantibody (lane 6). The mass of this complex is consistent with an hsp90:Hop:hsp70 stoichiometry of 2:1:1. A smaller complex that migrates just above the 213-kDa marker contains hsp90 (Fig. 4A, lanes 4 and  5) and Hop (Fig. 4B, lane 2), and it may reflect a dimer of hsp90 bound to Hop. As the concentration of glutaraldehyde increases, the amount of the smaller complex diminishes, whereas the ϳ310-kDa complex achieves a plateau at glutaraldehyde concentrations of 0.4 and 1.0 mM (Fig. 4A, lanes 4 -6). This suggests that hsp90⅐hsp90⅐Hop complexes are formed at the lower glutaraldehyde concentration and more hsp70 becomes cross-linked to Hop at higher concentrations to form hsp90⅐hsp90⅐Hop⅐hsp70 complexes.
Abundance of the hsp90⅐Hop⅐hsp70 Complex in Reticulocyte Lysate- Fig. 3C presents the concentrations of hsp90, hsp70, and Hop in reticulocyte lysate. These data were obtained by electrophorescing aliquots of reticulocyte lysate with known amounts of the purified proteins to yield a standard curve on the same gel. The protein concentrations were then determined by quantitative immunoblotting with 125 I-labeled counterantibody. The lysate contains 1 molecule of Hop for every 6.8 molecules of hsp90 and 11.6 molecules of hsp70. Because hsp90 is a homodimer and there is one TPR binding site per dimer (46), it seems that ϳ30% of the hsp90 in reticulocyte lysate is in complex with Hop. As shown in Fig. 5, when reticulocyte lysate is chromatographed through a column of Sephacryl S-300, all of the Hop elutes in a high M r peak that also contains hsp90. Immunoadsorption of hsp90 from this peak yields co-immunoadsorption of Hop (data not shown), showing that the two proteins are bound to each other.
In contrast to hsp90 and Hop, which elute in discrete peaks, hsp70 is present in all of the fractions eluting between the void volume and a peak that likely represents the hsp70 homodimer, eluting just before the massive protein peak containing the hemoglobin. The proteins are greatly diluted as they proceed through the Sephacryl column, and this may favor some dissociation of complexes, as indicated by the asymmetry of the Hop peak. If we assumed that under more concentrated conditions all of the Hop is associated with one molecule of hsp70 (as in the 310-kDa complexes observed on electrophoresis under non-denaturing conditions (Figs. 1C and 2)), then the data of Fig. 3C are consistent with the prediction that ϳ9% of the total hsp70 in reticulocyte lysate is in the hsp90⅐Hop⅐hsp70 complex.
Functional Significance of the hsp90⅐Hop⅐hsp70 Complex for GR⅐hsp90 Assembly-The fractions eluted from the Sephacryl S-300 column were combined to form fractions A and B as indicated in Fig. 5, and each fraction was contracted to 60% of the volume of reticulocyte lysate applied to the column. Fraction A contained 61% of the hsp90, 40% of the hsp70, and 92% of the Hop. Fraction B contained 38% of the hsp90, 53% of the hsp70, and 8% of the Hop. Each fraction was then supplemented with purified YDJ-1, p23, molybdate, and an ATPregenerating system, and the mixtures were incubated with immunoadsorbed GR that had been stripped of hsp90. As shown in Fig. 6A, fraction A promoted assembly of GR⅐hsp90 heterocomplexes and possessed over 90% of the receptor-activating activity as assayed by steroid binding. Fig. 6B shows that supplementation with both YDJ-1 and p23 is required for optimal activity with fraction A.
The data of Fig. 6 argue rather strongly for the importance of the hsp90⅐Hop⅐hsp70 machinery in GR⅐hsp90 heterocomplex assembly and generation of steroid binding activity by the reticulocyte lysate system. Most cell-free protein folding assays using the cytosolic chaperones hsp90 or hsp70 have been carried out with one of the purified proteins in the absence of the other. For GR⅐hsp90 heterocomplex assembly, both hsp90 and hsp70 are essential, but their function together is optimized by their presence in the multiprotein machinery with Hop (12). Inasmuch as Hop is widely distributed in the animal kingdom from yeast to humans (17) and Hop homologs are present in plants (47), it is reasonable to consider hsp90 as functioning together with hsp70 in the context of the multiprotein machinery where the two essential chaperones are brought into physical association by Hop (14).
Hsp90 and hsp70 Bind to Each Other Directly-Using biochemical separation procedures, we have previously shown that purified mammalian (mouse) hsp90 does not bind to purified hsp70 unless a factor from reticulocyte lysate is present to facilitate heterocomplex formation (48). Smith et al. (17) showed that hsp90 and hsp70 co-purified from reticulocyte lysate with a 60-kDa protein, which is now known to be Hop (13). In contrast to the mammalian chaperones, the purified hsp90 homolog of Neurospora crassa forms a rather tight complex with purified hsp70 that can be detected by normal biochemical separation procedures (49), and this hsp90⅐hsp70 complex interacts with polypeptide substrate (50). In Fig. 7, we asked if purified rabbit hsp70 and hsp90 interacted with each other. In this experiment, the proteins were mixed together in the presence or absence of glutaraldehyde, and the cross-linked products were resolved by electrophoresis under denaturing conditions. In the presence of cross-linker, a complex was recovered from the protein mixture that migrated slower than the hsp90 homodimer at a molecular mass consistent with an hsp90⅐hsp70 complex with a stoichiometry of 2:1.
The data of Fig. 7 suggest that mammalian hsp90 and hsp70 do have a weak ability to interact with each other. Because hsp90 and hsp70 are the only two components of the fiveprotein assembly system required to open the steroid binding cleft in the GR LBD, it seems likely that they may interact with each other as they modify receptor conformation. We have shown that GR⅐hsp90 heterocomplexes can be assembled in a two-step procedure (36). In the first step, immunoadsorbed GR is incubated with hsp70, hsp40 (YDJ-1), and an ATP-regenerating system to form a "primed" GR⅐hsp70 complex that can be washed free of unbound hsp70 and incubated in a second step with purified hsp90, Hop, p23, and ATP. The sequence of the two steps cannot be reversed, both steps are ATP-dependent, and steroid binding activity is generated during the second step (36). Purified hsp90 does not bind to the GR alone, but it does bind to the primed GR⅐hsp70 complex (36), at which time the two chaperones may directly interact with each other in opening the steroid binding cleft (37).