Hsp90 Phosphorylation Is Linked to Its Chaperoning Function ASSEMBLY OF THE REOVIRUS CELL ATTACHMENT PROTEIN*

Studies on Hsp90 have mainly focused on its involvement in the activation of several families of protein kinases and of steroid hormone receptors. Little is known regarding the role of Hsp90 in the folding of nascent proteins. We previously reported that Hsp90 plays an active role in the posttranslational assembly of the C-terminal globular head of the reovirus attachment protein (cid:1) 1. We show here that Hsp90 becomes phosphorylated in this process. However, only the unphosphorylated form of Hsp90 is complexed with (cid:1) 1, suggesting that Hsp90 phosphorylation is coupled to the release of the chaperone from the target protein. Geldanamycin, which blocks (cid:1) 1 maturation by preventing the release of Hsp90 from (cid:1) 1, also inhibits Hsp90 phosphorylation. Taken together, these results demonstrate that Hsp90 phosphorylation is linked to its chaperoning function.

Cellular chaperones are a group of proteins whose major roles appear to be the prevention of target protein aggregation and the promotion of their correct folding and assembly (1)(2)(3). Chaperones such as Hsp70 and Hsp40 likely interact with a wide variety of polypeptides, and thus their involvement in the folding processes is believed to be of a general and universal nature. Another major molecular chaperone, Hsp90, appears to be more specific in terms of "client" selection and has been shown to interact mainly with proteins involved in transcription regulation and signal transduction pathways, such as steroid hormone receptors and protein kinases (4 -8). Apparently Hsp90 stabilizes these proteins and keeps them in a conformation amenable to activation under appropriate conditions. Although the extent of Hsp90 involvement in the general folding scheme of cytosolic proteins remains unclear, it has been recently suggested, from studies on Drosophila, that Hsp90 plays a major role in morphological evolution (9). It was proposed that Hsp90 normally suppresses morphogenic variations that become manifest upon Hsp90 impairment, leading to abrupt evolutionary changes.
Although it is generally accepted that Hsp90 coordinates with other chaperones to promote the folding and assembly of the target protein by a binding and release mechanism, precisely how these processes are regulated remains unclear. Recent evidence suggests that assembly of the glucocorticoid receptor-Hsp90 complex involves the sequential involvement of Hsp70 and Hsp90, both being ATP-dependent events (10). That ATP binding and hydrolysis are essential for Hsp90 function was also demonstrated using the progesterone receptor (11). ATP binding apparently also results in conformational changes in Hsp90 concomitant with its association with the cochaperone p23 (4). Another less well studied phenomenon relates to Hsp90 phosphorylation, which may represent yet another level of regulation of Hsp90 function. Hsp90 can be phosphorylated in vitro at two serine residues by casein kinase II (12) and at two threonine residues by DNA-dependent protein kinase (13). In addition, Hsp90 was shown to enhance the kinase activity of eukaryotic initiation factor 2␣ kinase but only after prior phosphorylation of Hsp90 by casein kinase II (14). In another study, treatment of cells with the serine/threonine phosphatase inhibitor okadaic acid led to hyperphosphorylation of Hsp90 and decreased association between Hsp90 and pp60 v-src , suggesting a link between Hsp90 phosphorylation and target protein interaction (15). The role of Hsp90 in the folding of newly synthesized protein has not been extensively probed. Hartson et al. (16) studied the folding of the lymphoid cell kinase p56 lck translated in vitro and demonstrated the association of Hsp90 with newly synthesized p56 lck molecules. It was further revealed that although the Src homology 2 domain folds independently of Hsp90, folding of the catalytic domain (C-terminal to the Src homology 2 domain) is Hsp90-dependent. Subsequent phosphorylation at the C-terminal domain of p56 lck correlated with stabilization of the kinase, which is no longer associated with Hsp90. Another study using the heme-regulated eukaryotic initiation factor 2␣ kinase translated in vitro shows that Hsp90 plays an obligatory role in this kinase acquiring and maintaining a conformation that is competent for transformation into an aggregation-resistant activable kinase (17).
The in vitro translation system has also been used extensively in our laboratory to reveal the mechanisms of folding and assembly of the reovirus cell attachment protein 1, a trimeric protein positioned at the 12 vertices of the icosahedral virion (18 -22). The 1 trimer is highly asymmetric, with an N-terminal fibrous tail that is anchored to the virion and a C-terminal globular head that interacts with the cell receptor (23)(24)(25)(26). Evidence from in vitro translation studies has revealed that these two structurally distinct domains are generated by separate trimerization events (18). The N-terminal fibrous tail is highly ␣-helical and contains an extended heptad repeat of hydrophobic residues, endowing this region with the intrinsic propensity to form a triple coiled-coil. During 1 biogenesis, assembly of three neighboring nascent chains occurs cotranslationally (i.e. on the polysome) at the N terminus (19,20). This process does not involve any chaperones or ATP and results in the generation of a loose triple coiled-coil. As the triplex moves down the polysome, chaperones (e.g. Hsp70) begin to interact with the emerging residues of the elongating C termini, preventing their misfolding and aggregation. Protein 1 leaves the polysome as a partially assembled trimer with some chaperones attached to the unassembled C termini. This type of 1 is known as the "unstable hydra form" because it migrates as monomers even under "nondissociating" conditions (without boiling) in SDS-PAGE 1 (18,19). Subsequent ATP-dependent release of chaperones presumably provides the opportunity for the loose coiled-coil to quickly snap together (tightening the coiled-coil), whereas the remaining portions of the three C termini are available for continued interaction with chaperones. This structure, with a stably assembled N terminus and an unassembled C terminus, is called "stable hydra form," and it migrates as a retarded trimer in SDS-PAGE under nondissociating conditions. Further ATP-dependent release and rebinding of chaperones leads to global assembly and folding of the C terminus, generating mature 1 with the characteristic lollipop-shaped structure (called the "mature compact form"), which migrates as an unretarded trimer in SDS-PAGE under nondissociating conditions. We have proposed that the involvement of two mechanistically distinct oligomerization events for the same molecule, one cotranslational and one posttranslational, may represent a common approach to the generation of oligomeric proteins in the cytosol (19).
Our previous demonstration that Hsp90 is involved in the posttranslational assembly of the 1 globular head (21) has led us to question the possible role of Hsp90 phosphorylation in this process. In the present study, we demonstrate that Hsp90 phosphorylation occurs during the posttranslational maturation of 1 and is coupled to the assembly of the globular head. However, only the unphosphorylated form and not the phosphorylated form of Hsp90 is complexed with immature 1, suggesting that Hsp90 phosphorylation is linked to the release of the chaperone from the target protein. Inhibiting this release using the benzoquinone ansamycin, geldanamycin (GA), also abrogates Hsp90 phosphorylation. These observations have led us to propose that Hsp90 phosphorylation is linked to its chaperoning function.

EXPERIMENTAL PROCEDURES
In Vitro Transcription and in Vitro Translation-The plasmids encoding the full-length and various truncated 1 products have been described previously (18,27). All transcripts were generated in vitro using the MEGAscript TM Transcription Kit (Ambion) for the Sp6 polymerase promoter. A typical transcription reaction involved incubation of 1 g of linearized plasmid DNA with the prescribed contents of the Ambion Transcription Kit (total final volume of 20 l) for 5 h at 37°C. The mRNA product was isolated by LiCl precipitation followed by cleanup with the Bio101, Inc. RNaid Kit. The purified mRNA was then resuspended in 0.1% diethyl pyrocarbonate-treated water to a final concentration of ϳ0.5 g/l and stored at Ϫ70°C for future use.
Transcripts were translated in vitro in rabbit reticulocyte lysate (Promega) according to the manufacturer's specifications. Typically, for analysis of translation products, 0.5-1.0 g of mRNA was incubated at 37°C with 7 Ci of [ 35 S]methionine (Amersham Pharmacia Biotech), 1 l of 1.0 mM methionine minus amino acids (Promega), and 20 l of rabbit reticulocyte lysate (Promega) for the duration described in the figure legends.
In Vitro Phosphorylation Assay-Typically, translation was carried out for 7-9 min at 37°C in prewarmed rabbit reticulocyte lysate. These lysates may or may not have been supplemented with [ 35 S]methionine, depending upon the experiment (see the figure legends). Two microliters of [␥-32 P]ATP was then added to each reaction (final radio-specific activity, 2 Ci/l), and incubation was continued for the indicated times. For posttranslational chase experiments, translation was halted by either centrifugation (to pellet ribosomes) or addition of cyclohexamide (final concentration, 30 g/ml). [␥-32 P]ATP was then added, and the reactions were incubated further at 37°C for the periods described in the figure legends. The reactions were routinely supplemented with the 20 mM leupeptin (Sigma) as protease inhibitor. Geldanamycin was prepared as a stock solution of 175 M in 20% Me 2 SO, which was added to the reaction mixture to a final concentration of 7 M at the onset of the chase (unless stated otherwise in figure legends).
In Vitro Phosphorylation of Hsp90 Proteins-Purified bovine Hsp90 was obtained from Stressgen Inc. Rabbit Hsp90 was partially purified from rabbit reticulocyte lysate according to the protocol of Iannotti et al. (28). Labeling of bovine Hsp90 was carried out by incubating 5 g of protein in 35 l of kinase buffer (20 mM Tris-HCl, pH 7.2, 20 mM KCl, 10 mM MgCl 2 , 60 mM NaCl, 10 mM sodium metabisulfate, 20 mM ␤-glycerophosphate, 6 mM EDTA, 6 mM p-nitro-phenyl-phosphate, and 1 mM dithiothreitol) containing 7 Ci of [␥-32 P]ATP and 1 unit of casein kinase II for 15 min at 37°C. The partially purified preparation of rabbit Hsp90 protein was similarly labeled but in the absence of added casein kinase II.
SDS-PAGE-Discontinuous SDS-PAGE was performed using the protocol of Laemmli (29). 10% SDS-PAGE was used in all experiments. The sample was incubated in protein sample buffer (final concentration, 50 mM Tris, pH 6.8, 1% SDS, 2% ␤-mercaptoethanol, 10% glycerol, and 0.01% bromphenol blue) for 30 min on ice (nondissociating conditions) or, alternatively, boiled for 5 min (dissociating conditions) before SDS-PAGE. Nondissociating SDS-PAGE was routinely used to identify the three species of 1 (unstable hydra form, stable hydra form, and mature compact form) present in the reactions (18,21).
Immunoprecipitation-The anti-Hsp90 antibody H90-10 was a generous gift from Dr. David Toft (Mayo Clinic, Rochester, MN). The 3G3 monoclonal antibody was purchased from Affinity BioReagents, Inc. Polyclonal antibodies directed against the full-length 1 protein and the N-terminal half of 1 have been described previously (27). Immunoprecipitations were carried out on in vitro translations that were performed as outlined above. Typically, translation reactions were arrested by the addition of 5 volumes of TEM buffer (20 mM Tris, pH 7.4, 5 mM EDTA, 10 mM ammonium molybdate, 50 mM NaCl) containing 20 mM p-nitrophenyl-phosphate (in the case of phosphorylation reactions). To this was added the primary antibody at a dilution of between 1:35 and 1:150, as empirically determined for each antibody. For primary antibodies of the IgG class (anti-1 FL, anti-N 1, H90-10), the samples were incubated on ice for 1 h, and 20 l of protein-A acrylic beads (Sigma) preswollen in TEM buffer was then added. For immunoprecipitations using the 3G3 antibody (IgM class), the protein A acrylic beads were preadsorbed on ice for 1 h with IgM chain-specific rabbit anti-mouse IgG (Jackson Immuno-research) prior to addition to the diluted lysate. After incubation for an additional 30 min on ice, the samples were microcentrifuged, and the pellets were washed three times with ice-cold TEM buffer containing 0.1% Triton X-100 and 20 mM p-nitro-phenylphosphate. The pellets were then resuspended in protein sample buffer, kept on ice for 30 min (nondissociating conditions) or boiled for 5 min (dissociating conditions) and analyzed by SDS-PAGE.

RESULTS
Translation of Full-length 1 Protein Triggers p86 Phosphorylation-We have previously shown that Hsp90 interacts with the C-terminal half of protein 1 and that it is required for 1 maturation (i.e. assembly of the C-terminal globular head) (21). Because Hsp90 is a known phosphoprotein, we decided to investigate the phosphorylation status of Hsp90 during 1 assembly. To this end, full-length 1 was translated (for 15 min) in rabbit reticulocyte lysate in the presence of [␥-32 P]ATP, and the reactions were immunoprecipitated with a monoclonal anti-Hsp90 antibody, 3G3. At the time point chosen for immunoprecipitation, prior studies have shown that the majority of the 1 protein has trimerized within the N-terminal coiled-coil domain and a subpopulation has progressed to the mature form that has also folded within the C-terminal globular domain (21). Analysis of the immune pellet by SDS-PAGE revealed that a protein migrating at the ϳ86-kDa position was phosphorylated (Fig. 1A, left lane). In contrast, immunoprecipitates from a translation reaction of a truncated 1 (mutant d294) representing the N-terminal one-third of the protein, which does not associate with Hsp90 (21), did not contain a 32 Plabeled 86-kDa protein (Fig. 1A, right lane). This demonstrates thatphosphorylationofthe86-kDaprotein(p86)is1Cterminusdependent and not a by-product of translation events. Comparison of the migration rate of phospho-p86 with that of rabbit phospho-Hsp90 and bovine phospho-Hsp90 (both precipitable with 3G3) reveals that the three proteins have identical apparent molecular masses (Fig. 1B). These experiments therefore suggest that p86 is likely Hsp90 and is referred to as Hsp90 henceforth.
Phosphorylation of Hsp90 Occurs during Posttranslational Processing of 1-We previously showed that assembly of the C-terminal globular head of 1 occurs posttranslationally (19) and that this process involves Hsp90 (21). It would therefore be of interest to see whether phosphorylation of Hsp90 is also a posttranslational event. To this end, 1 mRNA was translated only long enough to generate immature trimers. These molecules have been previously characterized as the earliest posttranslational folding intermediate of 1 and possess loosely wound trimeric N termini and unfolded C termini (previously designated the unstable hydra form) (20,21). At least three chaperones are associated with these structures: Hsp90, p23, and Hsp70 (21). The reactions were then centrifuged at high speed to pellet polysomes, thereby preventing further translation. The supernatants, now containing only posttranslational complexes of 1 immature trimers, were incubated further at 37°C with [␥-32 P]ATP to promote further folding of the 1 trimers (i.e. assembly of the C-terminal head). The lysates were then removed at various times, diluted with buffer, and immunoprecipitated using the 3G3 antibody against Hsp90.
The results (Fig. 2) show that phosphorylation of Hsp90 occurred to a significant degree during further folding of the full-length 1 but was not detectable for the C-terminal truncation mutant d294 (control), although comparable amounts of the two polypeptides were synthesized in these reactions (not shown). Similar results were obtained when cyclohexamide was used to halt translation reactions (data not shown). These results therefore suggest that phosphorylation of Hsp90 is a posttranslational event and is linked to 1 C-terminal assembly. However, they provide no information as to whether Hsp90 phosphorylation occurs only when the target protein is folded correctly or whether interaction of Hsp90 with target sites alone is enough to trigger phosphorylation, even in the absence of maturation folding of the target protein.
Hsp90 Phosphorylation Is Induced Independent of Folding but Is Dependent upon Hsp90/Target Protein Interaction-If binding to the target protein alone was sufficient to trigger Hsp90 phosphorylation, then 1 mutants that contained a binding domain for Hsp90 but that could not properly fold should also trigger this phosphorylation. On the other hand, if phosphorylation was folding-dependent, then these mutants should be incapable of inducing Hsp90 phosphorylation. To resolve this issue, we used both N-terminal and C-terminal deletion mutants of 1. The N-terminal deletion mutants included dI and dI,II (see schematic in Fig. 3A), both of which lack the coiled-coil trimerization domain and therefore do not trimerize (18). Extensive studies have shown that N-terminal trimerization is a prerequisite for assembly of the C-terminal globular head (18,19). However, because both mutants contain the C-terminal half of 1, they are nonetheless capable of interacting with Hsp90 (21). For C-terminal deletion mutants, we examined d30, d90 (lacking the C-terminal 30 and 90 residues, respectively; however, both still contain Hsp90 binding sites (21)), and d294, which contains no Hsp90-binding sites as negative control. All three C-terminal deletion mutants can trimerize within their N-terminal coiled-coil domains but fail to fold within vestigial regions C-terminal to it.
We first tested for Hsp90 association with these two families of truncation mutants by immunoprecipitation of [ 35 S]methionine-and [␥-32 P]ATP-labeled reaction mixtures with the 3G3 antibody. (The reactions were also immunoprecipitated with an anti-1 antibody to ensure that all 1 constructs were expressed at comparable levels; Fig. 3B, left panel.) Immunoprecipitation with the 3G3 antibody (Fig. 3B, right panel) revealed that all nascent proteins except d294 were associated with Hsp90. We then examined the ability of each product to trigger phosphorylation of Hsp90 in these same reactions by analyzing 32 P-labeled Hsp90 precipitable with 3G3. The results (Fig. 3C) show that with the exception of d294, all the deletion mutants were able to trigger Hsp90 phosphorylation to at least the same degree as FL 1. Thus the ability of these mutants to induce phosphorylation directly correlates with their ability to bind Hsp90, demonstrating that phosphorylation of Hsp90 is dependent on Hsp90/target protein interaction alone and not on the maturation (proper folding) of the target protein.
Phospho-Hsp90 Is Not Associated with the Target Protein-It was of interest to determine whether the phosphorylated form of Hsp90 was in association with or free from the target pro- The results (Fig. 4A) reveal that anti-N precipitated 1 but did not coprecipitate phospho-Hsp90 (compare lane 1 with lane 2), suggesting no association between these two proteins. On the other hand, the 3G3 antibody precipitated both phospho-Hsp90 and 1 (Fig. 4A, lane 2). Thus 3G3 reacted with both phosphorylated Hsp90 that was not associated with 1 and unphosphorylated Hsp90 that was 1-associated. This was confirmed using another anti-Hsp90 antibody, H90-10, which precipitated 1, but not phospho-Hsp90 (Fig. 4B). Preclearing the reactions using the H90-10 antibody effectively removed the Hsp90⅐1 complex while leaving behind phospho-Hsp90 precipitable by 3G3 (Fig. 4C). Collectively, these results suggest that 1-associated Hsp90 is unphosphorylated and that phosphorylation of Hsp90 is coupled to its release from the target protein.
GA Inhibits Hsp90 Phosphorylation-If phosphorylation of Hsp90 is indeed linked to its release from 1, then blocking this release should abrogate Hsp90 phosphorylation. In a previous study we showed that treatment of 1 folding reaction with the Hsp90 binding drug GA blocked maturation of the 1 trimer at a late stage intermediate (21). We also demonstrated that in the presence of GA, 1 remains associated with Hsp90. Thus GA either prevents release of Hsp90 from the 1⅐Hsp90 complex or stabilizes Hsp90/1 interactions concomitant with abrogation of 1 folding. In view of the above observations regarding Hsp90 phosphorylation, one would expect that GA would block Hsp90 phosphorylation. The results (Fig. 5) show that indeed, when used at a concentration that blocked forma-tion of the mature compact form of 1 (Fig. 5A), GA effectively prevented Hsp90 phosphorylation (Fig. 5B). This experiment therefore demonstrates that Hsp90 phosphorylation is coupled to its release from the target protein, which may represent a major regulatory mechanism for the proper assembly and/or folding of the target protein. DISCUSSION Although it has been known for some time that Hsp90 is a phosphoprotein and that it can be phosphorylated under certain in vitro conditions, the significance of this phosphorylation event in Hsp90 function is entirely unknown. In the present study, we investigate the possibility that Hsp90 phosphorylation is linked to its chaperoning function, specifically in the folding and assembly of nascent polypeptides. We show that Hsp90 is phosphorylated during the in vitro biogenesis of the reovirus attachment protein 1 in rabbit reticulocyte lysate. This phosphorylation is not detectable when the N-terminal portion of 1 is translated and is strictly dependent on the presence of the C-terminal portion of the protein. This is consistent with our previous demonstration that the N-terminal fibrous tail of 1 trimerizes in a chaperone/ATP-independent manner, whereas formation of the C-terminal globular head requires both chaperones (including Hsp90) and ATP. Interestingly, the phosphorylated form of Hsp90 is not associated with 1, suggesting that Hsp90 phosphorylation is linked to release of the chaperone from the target protein. Support for this notion has come from the demonstration that GA, which inhibits the release of Hsp90 from 1, also blocks Hsp90 phosphorylation. These observations have led us to suggest that Hsp90 phosphorylation is part of the Hsp90 cycling mechanism and likely plays an important role in its chaperoning function.
A most interesting observation from our present study is that the C terminus of 1, when translated as a truncated protein and therefore unable to undergo assembly (i.e. trim- [␥-32 P]ATP was then added, and the reaction was further incubated for 7 min. Aliquots of the reaction mixture were then immunoprecipitated (IP) with either an anti-1 antibody directed against the N terminus of 1 (designated anti-N) or the 3G3 antibody to Hsp90, and analyzed by SDS-PAGE and autoradiography. B, same as A, except the reaction was precipitated with another anti-Hsp90 antibody (designated H90-10) and compared with immunoprecipitation with 3G3. C, same as A, except the reaction was first preadsorbed (precleared) with the H90-10 antibody and then immunoprecipitated with the 3G3 antibody, the H90-10 antibody (Ab), and normal rabbit serum (NRS, as control). erization), can also trigger Hsp90 phosphorylation. This suggests that Hsp90 phosphorylation is strictly dependent on the capacity of the target protein to interact with Hsp90 and not on proper folding of the target protein or its potential to do so. Based on the current notion that Hsp90 functions by a binding and release mechanism and in view of our present demonstration that Hsp90 phosphorylation is accompanied by release of the chaperone from the target protein, one could further deduce that even in the case of misfolded proteins, binding and release of Hsp90 occurs normally. The corollary of such a scenario would be that compared with full-length wild type 1, which can form a functional globular head, mutants that possess Hsp90 binding sites but that cannot properly assemble should generate more phosphorylated Hsp90 because the end point could never be achieved. Preliminary data (not shown) from our laboratory appear to support this notion. Additionally, our demonstration that Hsp90 bound to 1 is not phosphorylated suggests that only unphosphorylated Hsp90 is capable of making the initial contact with the target, although the possibility of phospho-Hsp90 being quickly dephosphorylated upon interaction with 1 cannot be ruled out at present. The target-bound Hsp90 would remain unphosphorylated, and its subsequent release from the target would be coupled to (or triggered by) its phosphorylation. A schematic of such a model is presented in Fig. 6. Although the universality of such a mechanism remains to be seen, it is interesting to note that the Hsp90-pp60 v-src chaperone complex in v-src-transfected cells is destabilized by the serine/threonine phosphatase inhibitor, okadaic acid, and that this is accompanied by a drastic increase in Hsp90 phosphorylation (15). Here again, binding and release from the substrate appears to be coupled to Hsp90 phosphorylation/dephosphorylation.
The identity of the kinase involved in 1-mediated Hsp90 phosphorylation is unknown at present; it can be an endoge-nous kinase or even Hsp90 itself (i.e. autophosphorylation). So far we have not been able to demonstrate the involvement of an endogenous kinase using a variety of kinase inhibitors including a specific inhibitor for casein kinase II. None of the inhibitors had any negative effect either on folding of the target protein or phosphorylation of Hsp90 (data not shown). Because autophosphorylation activity of cytosolic Hsp90 and its ER counterpart has been reported to be insensitive to a wide spectrum of kinase inhibitors, it is likely that Hsp90 autophosphorylation plays a role here. The exact site(s) on Hsp90 that is phosphorylated in this process is presently unknown and is under investigation. Neither is it known whether Hsp90 phosphorylation is linked to its ATP binding capacity and/or its ATPase activity. Recent studies have shown that ATP binding, but not hydrolysis, reduces the affinity of Hsp90 for hydrophobic residues (11). It would be of interest to see whether phosphorylated Hsp90 also contains bound ATP or whether any Hsp90-bound ATP is used for the phosphorylation event. Although the possible involvement of other chaperones in 1induced Hsp90 phosphorylation was not probed in the present study, it is likely that they actively participate in this process. We have previously demonstrated that assembly of the Cterminal globular head of 1 requires at least two other chaperones: Hsp70 and p23 (21). Recent evidence (data not shown) from our laboratory suggests that Hsp40 and p60/Hop/Sti1 are also involved. Antibodies against each of these chaperones can precipitate immature forms of 1. However, only the anti-p23 antibody can also precipitate phosphorylated Hsp90, suggesting that p23 remains associated with the released Hsp90. This association is not observed in the presence of GA, previously shown to prevent the release of Hsp90 from 1 while promoting the release of p23 from the 1-chaperone complex. This has provided yet another piece of supportive evidence for the re- The reaction was then centrifuged to pellet the ribosomes, and the supernatant was incubated further at 37°C in the presence and the absence of 7 M GA. At the various times indicated, the aliquots were taken and analyzed by SDS-PAGE under nondissociating conditions that allowed for the identification of all three 1 forms previously characterized (20,21). B, FL s1 transcripts were translated in vitro at 37°C for 8 min. [␥-32 P]ATP and GA (or 20% Me 2 SO as control) were then added to the reaction mixtures that were further incubated for 8 min. Aliquots of the mixtures were then immunoprecipitated (IP) with the 3G3 anti-Hsp90 antibody and analyzed by SDS-PAGE.
FIG. 6. Model for Hsp90 phosphorylation/dephosphorylation in target protein folding. Only unphosphorylated or hypophosphorylated Hsp90 (complexed with p23) can interact with the unfolded or partially folded target protein. Subsequent dissociation of the Hsp90target protein complex is coupled to Hsp90 phosphorylation and enhanced maturation of the target protein, which likely also involve other chaperones such as Hsp70, Hop, and Hsp40 (not shown). Reiterative cycles of binding and release concomitant with Hsp90 dephoshorylation and phosphorylation lead to the generation of the mature (properly folded) target protein.
lease of Hsp90 from the substrate being associated with Hsp90 phosphorylation.
Overall, the data presented in this study are in agreement with a model wherein Hsp90 phosphorylation is intimately linked to its chaperoning function. It also appears that this phosphorylation is not necessarily accompanied by correct folding of the substrate because binding and release of Hsp90 occurs normally for mutant targets and perhaps even more efficiently than for the wild type target because the end point (generation of a properly folded final product) can never be achieved. It remains to be seen whether Hsp90 phosphorylation is coupled to the ATP binding and ATPase activity of Hsp90. However, considering the highly coordinated nature of chaperone-mediated protein folding, this is most likely the case. It also seems safe to assume that other chaperones are also actively involved in this process. Clearly other systems (e.g. steroid hormone receptors and protein kinases) will need to be probed; the demonstration that Hsp90 phosphorylation also occurs in these systems would provide further support for the notion that phosphorylation/dephosphorylation represents a key regulatory mechanism for chaperone function.