The antiprogesterone RU486 stabilizes the heterooligomeric, non-DNA-binding, 8S-form of the rabbit uterus cytosol progesterone receptor

The salt-induced (0.3 M KCl) transformation of the non-transformed, heterooligomeric 8S-form of the rabbit uterus cytosol progesterone receptor (PR) was analyzed by density gradient ultracentrifugation (8S----4S conversion) and DNA-cellulose chromatography (non-binding----binding forms). After 1 h treatment at 2 C, greater than 90% of agonist (R5020 or Org2058)-PR complexes were transformed, contrary to antagonist (RU486)-PR complexes, which did not undergo any transformation. Thus, there is stabilization of the non-transformed receptor form by RU486 as compared to the effect of agonist binding. The hydrodynamic parameters of both agonist- and antagonist-bound non-transformed receptors were similar and the calculated Mr were approximately 283,000 and approximately 293,000, respectively. In both cases, purification indicated the presence of a 90-kD non-hormone-binding protein associated with the hormone binding unit(s). Transformation of RU486-PR complexes occurred after exposure to high salt at increased temperature and was correlated to the dissociation of the 90-kD protein from the receptor. Both agonist- and antagonist-bound transformed forms of PR had apparent similar affinities for DNA-cellulose. Molybdate-stabilized and KCl-treated RU486-PR complexes were more stable, as assessed by steroid binding, than the corresponding R5020-PR complexes, arguing in favor of a stabilizing effect of both the 90-kD protein and RU486 against inactivation. These cell-free experiments support the concept that RU486 in the rabbit uterus system stabilizes the 8S non-DNA binding, non-transformed form of the receptor at low temperature. The possibility that impaired dissociation of the heterooligomeric receptor form is involved in the antiprogesterone activity of RU486 is discussed.     

Evidence from pulse-chase labeling studies that the antiglucocorticoid hormone RU486 stabilizes the nonactivated form of the glucocorticoid receptor in mouse lymphoma cells

A pulse-chase labeling technique was used to determine the properties of glucocorticoid receptors occupied by the antiglucocorticoid hormone RU486 in S49.1 mouse lymphoma cells. Cells were pulse-labeled with [35S]methionine and then at the beginning of the chase, either no hormone (control), dexamethasone, or RU486 was added to cells. At 4 h into the chase, cytosol was prepared and receptors were immunoadsorbed to protein A-Sepharose using the BuGR2 antireceptor antibody. Immunoadsorbed proteins were resolved by gel electrophoresis and analyzed by autoradiography. The 90 kDa heat shock protein (hsp90) coimmunoadsorbed with receptors from control cells when protein A-Sepharose pellets were washed with 250 mM NaCl but not when protein A-Sepharose pellets were washed with 500 mM NaCl, indicating that hsp90-receptor complexes are disrupted by a high concentration of salt in the absence of molybdate. hsp90 coimmunoadsorbed with receptors from RU486-treated cells even when protein A-Sepharose pellets were washed with 500 mM NaCl, indicating that RU486 stabilizes the association of hsp90 with the glucocorticoid receptor. In contrast, hsp90 did not coimmunoadsorb with receptors from dexamethasone-treated cells, consistent with earlier evidence that hsp90 dissociates from the receptor when the receptor binds glucocorticoid hormone. Dexamethasone induced a rapid quantum decrease in the amount of normal receptor recovered from cytosol but did not induce a decrease in the amount of nuclear transfer deficient receptor recovered from cytosol, consistent with tight nuclear binding of normal receptors occupied by dexamethasone. In contrast, RU486 did not induce a quantum decrease in the recovery of normal receptors from cytosol, indicating that receptors occupied by RU486 are not tightly bound in the nuclear fraction. We conclude that the antiglucocorticoid hormone RU486, in contrast to the glucocorticoid hormone dexamethasone, stabilizes the association between the glucocorticoid receptor and hsp90. The decreased affinity of receptors occupied by RU486 for the nuclear fraction may be due to their association with hsp90 and may account for the failure of RU486 to exert agonist activity.     

Effects of antiestrogen versus antiprogestin on transformed and nontransformed steroid receptors

In order to determine if different physicochemical properties exist among antihormone-receptor complexes, we have compared the interaction of the antiprogestin RU486 with progesterone receptor (PR) versus the triphenylethylene antiestrogen H1285 (4-(N,N-diethyl-aminoethoxy)-4'-methoxy-alpha-(p-hydroxyphenyl-alp ha'- ethylstilbene] with estrogen receptor (ER) from rabbit uterine tissue. Contrary to other reports, we observed no difference in the sedimentation properties of transformed PR (4S) when bound by the antagonist RU486 versus the progesterone agonist R5020 in either cytosol or DEAE partially-purified receptor preparations analyzed on sucrose gradients containing 0.3 M KCl. In addition, we found no difference in the sedimentation properties of these receptor preparations in the presence of 10 mM sodium molybdate: the nontransformed RU486-PR and nontransformed R5020-PR both sedimented as a 6S species. These same results were obtained when the receptor preparation and gradient analysis were performed in the absence of monothioglycerol. Likewise, there was no change in the sedimentation properties of the transformed PR when the receptor, partially purified in the absence of molybdate, was analyzed on sucrose gradients containing 10 mM sodium molybdate to prevent receptor alteration during centrifugation. From DNA-cellulose assays performed with partially purified PR in the absence of molybdate we determined that the 4S form of R5020-PR and RU486-PR is transformed receptor; whereas in the presence of molybdate, the 6S species is nontransformed. In contrast, we found a different pattern of sedimentation when comparing transformed antiestrogen-receptor complexes with transformed estrogen-receptor complexes. In this case, transformed H1285-ER sedimented as 6S and estradiol-ER sedimented as 4S. We conclude from these experiments that these two antihormones, RU486 and H1285, may have different mechanisms of action in their antagonism of steroid hormone action. Antiestrogen stabilizes the salt-transformed ER as a dimer while antiprogestin appears to permit dissociation of the oligomeric form of the receptor to the monomeric form.     

Steroid hormone-dependent interaction of human progesterone receptor with its target enhancer element

We investigated the requirement of steroid hormone for the specific binding of progesterone receptor to its cognate progesterone responsive element (PRE) in cell-free experiments. We prepared unfractionated nuclear extracts from human breast cancer (T47D) cells which are rich in progesterone receptors and used a gel retardation assay to monitor receptor-DNA complex formation. Exposure of receptor to either progesterone, R5020, or the antiprogestin RU38 486 in vivo or in vitro led to the formation of two protein-DNA complexes (1 and 2) which were not detected in nuclear extracts unexposed to hormone. Similar treatment with cortisol or estradiol failed to induce the formation of these complexes. The complexes were specific for PRE, since they could be competed efficiently in binding competition experiments by oligonucleotides containing PRE. A monoclonal antibody which recognizes both A and B forms of human progesterone receptor, interacted with both complexes 1 and 2 and shifted them to slower migrating forms. Another antibody which only recognizes the B form interacted with only complex 1 but not with complex 2, establishing that the complexes 1 and 2 were indeed formed by progesterone receptor forms B and A, respectively. We conclude from the above studies that in vivo or in vitro treatment of nuclear progesterone receptor with either progesterone or R5020 or RU38 486 alone can lead to detection of high affinity complexes formed between the PRE and the receptor present in unpurified nuclear extracts.     

Effects of the steroid antagonist RU486 on dimerization of the human progesterone receptor.

We previously reported, using a coimmunoprecipitation assay, that the B form (PR-B) of the human progesterone receptor from T47D human breast cancer cells dimerizes in solution with the A receptor (PR-A) and that the extent of dimerization correlates with receptor binding activity for specific DNA sequences [DeMarzo, A.M., Beck, C.A., O?ate, S.A., & Edwards, D.P. (1991) Proc. Natl. Acad. Sci. U.S.A. 88, 72-76]. This suggested that solution dimerization is an intermediate step in the receptor activation process. The present study has tested the effects of the progesterone antagonist RU486 on solution dimerization of progesterone receptors (PR). As determined by the coimmunoprecipitation assay, RU486 binding did not impair dimerization of receptors; rather, the antagonist promoted more efficient solution dimerization than the progestin agonist R5020. This enhanced receptor dimerization correlated with a higher DNA binding activity for transformed receptors bound with RU486. RU486 has been shown previously to produce two other alterations in the human PR when compared with R5020. PR-RU486 complexes in solution exhibit a faster sedimentation rate (6 S) on salt-containing sucrose density gradients than PR-R5020 complexes (4 S), and PR-DNA complexes have a faster electrophoretic mobility on gel-shift assays in the presence of RU486. We presently show that the 6 S PR-RU486 complex is a receptor monomer, not a dimer. The increased sedimentation rate and increased mobility on gel-shift assays promoted by RU486 were also observed with recombinant PR-A and PR-B separately expressed in insect cells from baculovirus vectors. These results suggest that RU486 induces a distinct conformational change both in PR monomers in solution and in dimers bound to DNA. We also examined whether conformational changes in PR induced by RU486 would prevent a PR polypeptide bound to RU486 from heterodimerization with another PR polypeptide bound to R5020. To evaluate this, PR-A and PR-B that were separately bound to R5020 or RU486 in whole cells were mixed in vitro. PR-A-RU486 was capable of dimerization with PR-B-R5020, and this was demonstrated for heterodimers both formed in solution and bound to specific DNA. The capability to form heterodimers in vitro raises the possibility that the antagonist action of RU486 in vivo could in part be imposed in a dominant negative fashion through heterodimerization between one receptor subunit bound to an agonist and another bound to RU486.     

Novel antiprogestins Org 31806 and 31710: interaction with mammalian progesterone receptor and DNA binding of antisteroid receptor complexes.

We have examined steroid binding parameters and transformation of calf uterine progesterone receptor (PR) liganded with progestins (progesterone and R5020) and the newly synthesized antiprogestins (Org 31806 and 31710). Species specificity analysis indicated that [3H]R5020 binding in the chicken oviduct cytosol could be eliminated in the presence of 100-fold excess radioinert progesterone and R5020 but not Org 31806 and 31710. In the calf uterine cytosol, the progestins and the antiprogestins appeared to interact with the same PR as revealed by the displacement of [3H]R5020 by all of the above steroids. When the extent of [3H]R5020 binding was examined in the presence of different concentrations of radioinert steroids, the relative affinity with which these compounds interacted with the uterine PR was found to be comparable. A 23 degrees C incubation of cytosol transformed the progestin-bound PR complexes increasing their binding to DNA-cellulose from 5 (0 degrees C, nontransformed) to 35%. Under these conditions, 20% Org 31710- and RU486-occupied PR complexes bound to DNA-cellulose whereas only 10% Org 31806-receptor complexes were retained by the resin. Transformation (23 degrees C) of cytosol receptor caused a loss of the larger 8 S form and an increase in the smaller 4 S form. In its unliganded state or when it was complexed with R5020 or the antiprogestins, incubation of PR at 23 degrees C led to dissociation of the receptor-associated 90 kDa heat-shock protein (hsp90). The PR-hsp90 association was stabilized in the presence of 10 mM iodoacetamide when the ligand binding site was occupied by Org 31806 and 31710. The R5020-receptor complexes, however, allowed release of hsp90 under the above transforming conditions. Our results indicate that although Org 31806 and 31710 show no affinity for the avian PR, these steroids interact with the mammalian PR. We propose that the reported antiprogestational effects of Org 31806 and 31710 are mediated via their interaction with PR which appears similar to one that exists between PR and RU486.     

Transformation of calf uterine progesterone receptor: analysis of the process when receptor is bound to progesterone and RU38486

Effects of different transforming agents were examined on the sedimentation characteristics of calf uterine progesterone receptor (PR) bound to the synthetic progestin [3H]R5020 or the known progesterone antagonist [3H]RU38486 (RU486). [3H]R5020-receptor complexes [progesterone-receptor complexes (PRc)] sedimented as fast migrating 8S moieties in 8-30% linear glycerol gradients containing 0.15 M KCl and 20 mM Na2MoO4. Incubation of cytosol containing [3H]PRc at 23 degrees C for 10-60 min, or at 0 degrees C with 0.15-0.3 M KCl or 1-10 mM ATP, caused a gradual transformation of PRc to a slow sedimenting 4S form. This 8S to 4S transformation was molybdate sensitive. In contrast, the [3H]RU486-receptor complex exhibited only the 8S form. Treatment with all three activation agents caused a decrease in the 8S form but no concomitant transformation of the [3H]RU486-receptor complex into the 4S form. PR in the calf uterine cytosol incubated at 23 or at 0 degrees C with 0.3 M KCl or 10 mM ATP could be subsequently complexed with [3H]R5020 to yield the 4S form of PR. However, the cytosol PR transformed in the absence of any added ligand failed to bind [3H]RU486. Heat treatment of both [3H]R5020- and [3H]RU486-receptor complexes caused an increase in DNA-cellulose binding, although the extent of this binding was lower when RU486 was bound to receptors. An aqueous two-phase partitioning analysis revealed a significant change in the surface properties of PR following both binding to ligand and subsequent transformation. The partition coefficient (Kobsd) of the heat-transformed [3H]R5020-receptor complex increased about 5-fold over that observed with PR at 0 degrees C.(ABSTRACT TRUNCATED AT 250 WORDS)     

Nuclear localization of two steroid receptor-associated proteins, hsp90 and p59

It has been proposed that the unliganded nontransformed form of steroid hormone receptor is a heterooligomer comprising, in addition to the hormone-binding subunit, two associated proteins: a heat shock protein of MW 90,000 (hsp90) and another protein of MW 59,000 (p59). Using monoclonal antibodies, we demonstrate immunocytochemically the presence of both hsp90 and p59 in cell nuclei of progesterone target cells of the rabbit uterus. While steroid receptors (e.g., progesterone receptors) appear to be exclusively nuclear, we find p59 predominantly in the cell nuclei and hsp90 in both the nucleus and the cytoplasm. In addition, Western blotting of high-salt extracts of nuclear proteins detects the presence of hsp90 and p59 in the nuclei of rabbit uterus. These observations are consistent with the presence of the untransformed heterooligomeric form of steroid hormone receptors in the nuclei of target cells.     

Characterization of R5020 and RU486 binding to progesterone receptor from calf uterus

We have examined and compared the binding characteristics of the progesterone agonist R5020 [promegestone, 17,21-dimethylpregna-4,9(10)-diene-3,20-dione] and the progesterone antagonist RU486 [mifepristone, 17 beta-hydroxy-11 beta-[4-(dimethylamino) phenyl]-17 alpha-(prop-1-ynyl)-estra-4,9-dien-3-one] in calf uterine cytosol. Both steroids bound cytosol macromolecule(s) with high affinity, exhibiting Kd values of 5.6 and 3.6 nM for R5020 and RU486 binding, respectively. The binding of the steroids to the macromolecule(s) was rapid at 4 degrees C, showing saturation of binding sites at 1-2 h for [3H]progesterone and 2-4 h for both [3H]R5020 and [3H]RU486. Addition of molybdate and glycerol to cytosol increased the extent of [3H]R5020 binding. The extent of [3H]RU486 binding remained unchanged in the presence of molybdate, whereas glycerol had an inhibitory effect. Molybdate alone or in combination with glycerol stabilized the [3H]R5020- and [3H]RU486-receptor complexes at 37 degrees C. Although the rate of association of [3H]RU486 with the cytosolic macromolecule was slower than that of [3H]R5020, its dissociation from the ligand-macromolecule complex was significantly slower than [3H]R5020. Competitive steroid binding analysis revealed that [3H]progesterone, [3H]R5020, and [3H]RU486 compete for the same site(s) in the uterine cytosol, suggesting that all three bind to the progesterone receptor (PR). Sedimentation rate analysis showed that both steroids were bound to a molecule that sediments in the 8S region. The 8S [3H]R5020 and [3H]RU486 peaks were abolished by excess radioinert progesterone, RU486, or R5020.(ABSTRACT TRUNCATED AT 250 WORDS)     

Mammalian progesterone receptor shows differential sensitivity to sulfhydryl group modifying agents when bound to agonist and antagonist ligands

Modulation of calf uterine progesterone receptor (PR), in relation to its binding to synthetic steroids with known agonist (R5020) and antagonist (RU486) properties, was studied in the presence of iodoacetamide (IA), N-ethylmaleimide (NEM), beta-mercaptoethanol (MER), and dithiothreitol (DTT). Pretreatment of uterine cytosol at 4 degrees C with NEM (4-10 mM) reduced the binding of [3H]RU486 to PR by 40%, but [3H] R5020 binding was completely abolished. Whereas IA (2-10 mM) treatment did not affect [3H]RU486 binding, [3H]R5020 binding was totally eliminated. DTT or MER increased the binding of both steroids slightly (15%). [3H]R5020- or [3H]RU486-receptor complexes (Rc) migrated in the 8 S region and were eliminated upon pretreatment with NEM. At 23 degrees C, DTT increased the amount of 4 S [3H]R5020-Rc, but had no effect on the [3H]RU486-Rc. In the control, [3H]RU486 binding to the 8 S PR could be competed with radioinert R5020 or RU486, but R5020 failed to compete in the presence of IA. The heat-treated [3H]R5020- and [3H]RU486-Rc showed reduced binding to DNA-cellulose in the presence of NEM and IA. The results of our study suggest that SH group modifications differentially influence the properties of mammalian PR complexed with either R5020 or RU486. In the presence of IA, the [3H]RU486-Rc remained in the 8 S form when incubated at 23 degrees C, indicating that RU486 binding causes conformational changes in PR which are distinct from those that result upon R5020 binding.     

Studies of a plasma membrane steroid receptor in Xenopus oocytes using the synthetic progestin RU 486

A steroid binding protein (Mr = 110,000) has previously been identified in the plasma membrane of Xenopus laevis oocytes by photoaffinity labeling with [3H]R5020. In order to further characterize this steroid receptor, the photoaffinity labeled receptor protein was solubilized with 0.1% Brij 35. The solubilized labeled receptor yielded an approximate mol. wt of 102,000 +/- 2,000 by sucrose density gradient centrifugation, suggesting that the solubilized receptor exists as a monomer. RU 486, a synthetic progestin antagonist for mammalian cytosolic receptor systems, inhibited up to 70% of [3H] R5020 photoaffinity binding to the 110,000-Dalton receptor with an IC50 of 5 microM and induced germinal vesicle breakdown (GVBD) with an EC50 of 9.0 +/- 0.6 microM. GVBD induced by RU 486 was slower than with progesterone, and RU 486 was less powerful than progesterone. Micromolar concentrations of RU 486 also potentiated GVBD induced by sub-optimal concentrations of progesterone or R5020. Furthermore, RU 486 inhibited oocyte plasma membrane adenylate cyclase with an apparent IC50 of 7.5 +/- 2.5 microM. The close correlation of the EC50 value for RU 486 induction of GVBD with the IC50 values for inhibition of [3H]R5020 photoaffinity labeling of the 110,000-Dalton receptor and inhibition of adenylate cyclase activity further supports the physiological significance of the oocyte plasma membrane steroid receptor.     

Antiprogestin-receptor complexes: differences in the interaction of the antiprogestin RU38,486 and the progestin R5020 with the progesterone receptor of human breast cancer cells

In order to understand the molecular basis for antiprogestin action, we have compared the interaction of the antiprogestin [3H]RU38, 486 (RU486) and the progestin [3H]R5020 with the progesterone receptor (PR). In both MCF-7 and T47D human breast cancer cells, we have observed marked differences in the sedimentation properties of the PR on high salt sucrose gradients: while the R5020-receptor complexes sediment at approximately 4 S (4.4 +/- 0.1 S), the RU486-receptor sediments as a prominent 6 S species as well as a 4 S species. This binding is abolished by excess unlabelled R5020, RU486 or progesterone, but is unaffected by excess unlabelled hydrocortisone or dexamethasone, indicating that both the 4 S and 6 S species represent the PR and not glucocorticoid receptor. Although the relative distribution of 4 S and 6 S forms is not altered by treatment with DNAse or RNAse, exposure to 10 mM thioglycerol or to 3 M urea results in conversion of the 6 S to the 4 S form, suggesting that disulfide bonds and hydrophobic interactions are important in maintaining the integrity of the 6 S form. These findings suggest that the 6 S antiprogestin complex is formed as a result of the interaction of PR units with each other or with a different protein. This change in receptor association state may be an important aspect of the antiprogestin activity of RU486.     

Demonstration of steroid hormone receptors and steroid action in primary cultures of rat glial cells

Primary cultures of rat glial cells were established from newborn rat forebrains. A mixed population of oligodendrocytes and astrocytes was obtained, as confirmed by indirect immunofluorescence staining with specific markers for each cell type. Receptors were measured 3 weeks after primary culture in glial cells cultured in the presence or not of 50 nM estradiol and we have identified progesterone, glucocorticoid, estrogen, and androgen receptors (PR, GR, ER and AR), but only PR was inducible by the estrogen treatment. This estrogen-induction of PR was more dramatic in glial cells derived from female offsprings than from males, as measured by binding studies and by immunohistochemical techniques with the KC 146 anti-PR monoclonal antibody. The antiestrogen tamoxifen inhibited the estrogen induction, but had no effect by itself on PR concentration. Specific binding sites for PR, GR, ER and AR were measured by whole cell assays after labeling cells with, respectively, [³H]R5020, [³H]dexamethasone, [³H]OH-tamoxifen or [³H]R1881. PR and GR were also analyzed by ultracentrifugation and after exposure of cells to agonists, both receptors were recovered from cytosol as a 9S form, and from the nuclear high-salt, tungstate ions-containing fraction as a 4–6S form. In contrast, when the antiprogestin- and antiglucocorticosteroid RU486 was used as a ligand, a non-activated 8.5S receptor complex was found for both receptors in this nuclear fraction. The 8.5S complex of the GR was further analyzed in the presense of specific antibodies and, in addition to GR, the presence of the heat shock protein hsp90 and of a 59 kDa protein was found.

During primary culture, the effects of progesterone (P) and estradiol (E₂) were tested on glial cell multiplication, morphology and differentiation. Cell growth was inhibited by P and stimulated by E₂. Both hormones induced dramatic morphologic changes in oligodendrocytes and astrocytes and increased synthesis of the myelin basic protein in oligodendrocytes and of the glial fibrillary acidic protein in astrocytes.     

Retinoic acid receptor belongs to a subclass of nuclear receptors that do not form "docking" complexes with hsp90.

We have recently reported that, in contrast to the glucocorticoid receptor, the thyroid hormone receptor does not bind to hsp90 when the receptor is translated in rabbit reticulocyte lysate [Dalman, F. C., Koenig, R. J., Perdew, G. H., Massa, E., & Pratt, W. B. (1990) J. Biol. Chem. 265, 3615-3618]. All of the steroid receptors that are known to bind hsp90 are recovered in the cytosolic fraction when hormone-free cells are ruptured in hypotonic buffer. In contrast, unliganded thyroid hormone receptors and retinoic acid receptors are tightly associated with nuclear components. In this paper, we translated the human estrogen receptor and the human retinoic acid receptor in reticulocyte lysate and then immunoadsorbed the [35S]methionine-labeled translation products with the 8D3 monoclonal antibody against hsp90. The estrogen receptor is bound to hsp90, as indicated by coimmunoadsorption, but the retinoic acid receptor is not. Translation and immunoadsorption of chimeric proteins containing the DNA binding domain of one receptor and the N-terminal and COOH-terminal segments of the other show that the DNA binding finger region of the estrogen receptor is neither necessary nor sufficient for hsp90 binding. These observations suggest that there are two classes within the steroid receptor family. In one class (e.g., glucocorticoid, mineralocorticoid, sex hormone, and dioxin receptors), the receptors bind to hsp90 and remain in some kind of inactive "docking" mode until hormone-triggered release of hsp90 occurs. In the retinoic acid/thyroid hormone class, the unligated receptors do not bind to hsp90, and the receptors appear to proceed directly to their high-affinity nuclear acceptor sites without entering the "docking" state.     

Mineralocorticosteroid receptor of the chick intestine. Oligomeric structure and transformation

The binding of [3H]aldosterone in the chick intestine cytosol was analyzed in terms of affinity and specificity. In this tissue, aldosterone binds to the mineralocorticosteroid receptor, with a high affinity (Kd approximately 0.3 nM) and low capacity (approximately 50 fmol/mg protein), and to the glucocorticosteroid receptor. The selective labeling of the mineralocorticosteroid receptor was achieved by incubating the cytosol with [3H]aldosterone in the presence of RU 486. This synthetic steroid completely inhibited the binding of [3H]aldosterone to the glucocorticosteroid receptor and did not bind to the mineralocorticosteroid receptor. The oligomeric structure of the mineralocorticosteroid receptor was studied by using BF4, a monoclonal antibody which reacts with the 90-kDa heat shock protein (hsp 90), a nonhormone-binding component of nontransformed steroid receptors. The mineralocorticosteroid receptor sedimented at 8.5 +/- 0.4 S (n = 8) in a 15-40% glycerol gradient. This peak was shifted to 11.2 +/- 0.6 S (n = 5) after incubation with BF4, indicating that, in the cytosol, hsp 90 was associated with the mineralocorticosteroid receptor. Dissociation of the complex was observed on gradients containing 0.4 M KCl, as judged by the absence of displacement by BF4 of the 4.3 +/- 0.4 S (n = 10) peak. The effect of molybdate and tungstate ions, and of dimethyl pimelimidate, an irreversible cross-linking agent, on the stability of the hsp 90-receptor complex was investigated. Complexes recovered in the presence of 20 mM molybdate ions dissociated on gradients containing 0.4 M KCl (5.2 +/- 0.6 S (n = 4), whereas complexes prepared in the presence of 20 mM tungstate ions sedimented at 8.5 +/- 0.4 S (n = 7). Similarly, complexes prepared in the presence of molybdate ions dissociated during high pressure liquid chromatography (HPLC) gel filtration analysis performed in 0.4 M KCl (RS (Stokes radius) = 3.9 +/- 0.5 nm (n = 3) versus 7.3 +/- 0.2 nm (n = 3) in the presence of 20 mM molybdate ions), whereas complexes prepared in the presence of tungstate ions did not dissociate (RS = 6.9 +/- 0.2 nm (n = 3]. As observed for the tungstate-stabilized receptor, the cross-linked receptor dissociated neither on gradient containing 0.4 M KCl (9.5 +/- 0.1 S (n = 3] nor during HPLC performed in 0.4 M KCl (RS = 6.5 +/- 0.3 (n = 4]. Furthermore, the cross-linked receptor was more resistant to the inactivating effect of urea on aldosterone binding than the noncross-linked receptor prepared in the presence of either molybdate or tungstate ions.