Sti1 is a novel activator of the Ssa proteins.

The molecular chaperones Hsp70 and Hsp90 are involved in the folding and maturation of key regulatory proteins in eukaryotes. Of specific importance in this context is a ternary multichaperone complex in which Hsp70 and Hsp90 are connected by Hop. In Saccharomyces cerevisiae two components of the complex, yeast Hsp90 (yHsp90) and Sti1, the yeast homologue of Hop, had already been identified, but it remained to be shown which of the 14 different yeast Hsp70s are part of the Sti1 complex and what were the functional consequences resulting from this interaction. With a two-hybrid approach and co-immunoprecipitations, we show here that Sti1 specifically interacts with the Ssa group of the cytosolic yeast Hsp70 proteins. Using purified components, we reconstituted the dimeric Ssa1-Sti1 complex and the ternary Ssa1-Sti1-yHsp90 complex in vitro. The dissociation constant between Sti1 and Ssa1 was determined to be 2 orders of magnitude weaker than the affinity of Sti1 for yHsp90. Surprisingly, binding of Sti1 activates the ATPase of Ssa1 by a factor of about 200, which is in contrast to the behavior of Hop in the mammalian Hsp70 system. Analysis of the underlying activation mechanism revealed that ATP hydrolysis is rate-limiting in the Ssa1 ATPase cycle and that this step is accelerated by Sti1. Thus, Sti1 is a potent novel effector for the Hsp70 ATPase.

The molecular chaperones Hsp70 and Hsp90 are involved in the folding and maturation of key regulatory proteins in eukaryotes. Of specific importance in this context is a ternary multichaperone complex in which Hsp70 and Hsp90 are connected by Hop. In Saccharomyces cerevisiae two components of the complex, yeast Hsp90 (yHsp90) and Sti1, the yeast homologue of Hop, had already been identified, but it remained to be shown which of the 14 different yeast Hsp70s are part of the Sti1 complex and what were the functional consequences resulting from this interaction. With a two-hybrid approach and co-immunoprecipitations, we show here that Sti1 specifically interacts with the Ssa group of the cytosolic yeast Hsp70 proteins. Using purified components, we reconstituted the dimeric Ssa1-Sti1 complex and the ternary Ssa1-Sti1-yHsp90 complex in vitro. The dissociation constant between Sti1 and Ssa1 was determined to be 2 orders of magnitude weaker than the affinity of Sti1 for yHsp90. Surprisingly, binding of Sti1 activates the ATPase of Ssa1 by a factor of about 200, which is in contrast to the behavior of Hop in the mammalian Hsp70 system. Analysis of the underlying activation mechanism revealed that ATP hydrolysis is rate-limiting in the Ssa1 ATPase cycle and that this step is accelerated by Sti1. Thus, Sti1 is a potent novel effector for the Hsp70 ATPase.
Molecular chaperones are a diverse set of functionally related proteins that are involved in protein folding in the cell. An important feature in this context are dynamic interactions of various components of the chaperone machinery. Two of the most abundant chaperones found in the eukaryotic cell are Hsp70 and Hsp90. The Hsp70 and Hsp90 chaperones cooperate in the formation and maintenance of protein structures of a set of target proteins in vivo (1)(2)(3)(4)(5)(6)(7).
In contrast to Hsp70, which is involved in a wide variety of functions including folding of nascent polypeptide chains (8,9), Hsp90 seems to be a more specialized folding factor (10 -15). The in vivo substrate proteins of Hsp90 include key regulatory factors such as kinases and steroid hormone receptors among others (2,3,16,17). In eukaryotes, Hsp90 fulfills its chaperone function in complex with a large number of cofactors (18 -20). Two of them were identified as Hsp70 and Hop in higher eukaryotes (21,22). A complex consisting of Hsp70, Hop, and Hsp90 is of central importance for the Hsp90 chaperone cycle, because it is able to activate steroid hormone receptors for steroid hormone binding in vitro (23). Hop/Sti1 is a unique co-chaperone consisting mainly of three tetratricopeptide repeat (TPR) 1 domains. TPRs are helical modules that mediate interactions between proteins (24). In the case of Hop, its TPR domains can bind to the C-terminal ends of both Hsp70 and Hsp90 (25), thus providing a physical link between the Hsp70 and Hsp90 chaperone machinery (26,27). Despite its importance for the Hsp90 cycle, the composition and function of this complex are still ill-defined.
Using purified components, it had been shown previously that yHsp90 and Sti1 form a binary complex in which Sti1 inhibits the ATPase of yHsp90 (28). However, neither the question of whether in addition to the Ssa class (29) any of the nine cytosolic Hsp70 proteins in yeast is incorporated into the Sti1-yHsp90 complex nor the functional consequences of this interaction had been addressed.
These nine Hsp70s can be divided into the four subfamilies: Ssa, Ssb, Sse, and Ssz (Ss denotes stress seventy-related; a, b, e, and z indicate the subfamily) (8). The Ssa subfamily, which is the only essential Hsp70 family in the cytosol, comprises four members (Ssa1-Ssa4). The expression of at least one of the four Ssa proteins is essential for viability (30). The Ssa1 gene is expressed at high levels at physiological temperature; in addition, its expression is stimulated ϳ10-fold after a temperature shift to 37°C, whereas Ssa2 is expressed at the same level at all temperatures (30). Ssa3 and Ssa4 are expressed at extremely low levels at physiological temperature, but after a temperature upshift, the amounts of Ssa3 and Ssa4 mRNAs increase severalfold (8). In contrast to the Ssa proteins, which are soluble in the cytosol, Ssb1 and Ssb2 are associated with nascent polypeptides of translating ribosomes (31)(32)(33). Similarly, Ssz1 forms a ribosome-associated complex with the Hsp40 protein Zuo1. This Ssz1-Zuo1 complex stimulates the mitochondrial translocation of ribosome-nascent chain complexes (34 -36).
Finally, Sse1 and Sse2 are up-regulated severalfold under heat shock conditions. Under normal growth conditions, Sse1 is present at moderate levels, and Sse2 is present at very low levels (37).
To analyze the function of the Hsp70⅐Sti1⅐Hsp90 complex from yeast, we first set out to identify the Hsp70 components. We show that the Ssa proteins are the only Hsp70 component of the complex and that Sti1, unlike Hop in the mammalian Hsp90 multichaperone complex, is a novel regulator for their ATPase activity.
The lysates for immunoprecipitations were prepared as follows. The strains W303 and CN11a were grown in 100 ml of YPD medium to a A 600 of ϳ1.0, harvested, and resuspended in 5 ml of 40 mM Hepes, pH 7.5, 150 mM KCl, 5 mM MgCl 2 . The cells were lysed using a cell disruption system (Constant Systems, Warwick, UK), shock frozen with liquid nitrogen, and stored in 100 l portions at Ϫ80°C.
Two-hybrid Screening-The yeast strain Y190 was transformed with either Sti1 in pAS1-CYH2 or pACTII or the nine different yeast Hsp70s in pAS1-CYH2 or pACTII as described in the manufacturer's instructions (Clontech). Single transformants were confirmed as His Ϫ and LacZ Ϫ to ensure that the fusion proteins alone do not exhibit transcriptional activity in Y190. Double transformants were incubated for 3 days at 30°C before positive clones were picked, restreaked onto His Ϫ Leu Ϫ Trp Ϫ plates and assayed for the lacZ phenotype (39).
Construction of the Mutants his-Ssa1 (K69Q) and his-yHsp90 (E33A)-The design of the Ssa1 (K69Q) mutant is described elsewhere (40). The yHsp90 (E33A) point mutation was generated in vitro using overlap PCR technique according to Mikaelian and Sergeant (41). The identity of this construct was confirmed by DNA sequencing.
Immunoprecipitations-The immunoprecipitations were performed with lysates from the S. cerevisiae wild type strain W303 or the Sti1 knock-out strain CN11a and with the purified proteins Ssa1, Sti1, and yHsp90 in a volume of 100 l. The immunoprecipitations were incubated at 25°C for 1 h in standard buffer with or without 2 mM ATP, using the indicated antibodies. Immune complexes were recovered by binding to protein A/G-Sepharose (Sigma). After three washes with 5 volumes of 1ϫ phosphate-buffered saline (4 mM KH 2 PO 4 , 16 mM Na 2 HPO 4 , 0.12 M NaCl, pH 7.4), the immunocomplexes were eluted with 1ϫ Laemmli SDS loading buffer. For immunodetection, the proteins were separated by SDS-PAGE and transferred to Immobilonnitrocellulose (Millipore, Bedford, MA) or Immobilon-polyvinylidene difluoride (Millipore) membranes and incubated with polyclonal antisera specific for either Ssa (C-terminal 80 amino acids), Ssb (C-terminal 80 amino acids) (48), Sti1 (this study), or yHsp90 (this study). As a secondary antibody, anti-rabbit IgG conjugated with peroxidase for ECL detection (Amersham Biosciences) was used.
ATPase Activity (Radioactive Assay)-ATPase assays were performed according to Kornberg et al. (49). The indicated amounts of Ssa1 Ϯ Sti1 Ϯ yHsp90 Ϯ Ydj1 Ϯ Hop, Hsp70 Ϯ Hop Ϯ Sti1, and Hsp90 Ϯ Hop Ϯ Sti1 (Ϯ, with or without) were incubated in standard buffer at 37°C. The ATPase reaction was initiated by the addition of the respective ATP concentrations containing [␣-32 P]ATP (Hartmann Analytic, Braunschweig, Germany). A sample contained 0.1 Ci of [␣-32 P]ATP. As a control for unspecific protein effects, Sti1 was added after incubation at 70°C for 30 min. For steady state hydrolysis measurements, a final ATP concentration of 15 mM was used. In the case of single turnover experiments, the ATP to protein ratio was kept constant at 0.8:1. The ATP to ADP ratio was quantified with a Typhoon 9200 PhosphorImager (Amersham Biosciences). The hydrolysis rates were corrected for uncatalyzed, spontaneous ATP hydrolysis.
ATPase Activity of Ssa1 (HPLC)-The formation of ADP in steady state and single turnover experiments was measured at 37°C by incubating Ssa1 (Ssa1 K69Q) Ϯ Sti1 and/or yHsp90 (yHsp90 E33A) with ATP for various times and subsequent nucleotide analysis by HPLC as described (50).
Biacore Measurements-Surface plasmon resonance (51, 52) data were collected on a Biacore X Instrument (Biacore, Uppsala, Sweden). For experiments in the yeast system, either Ssa1 or yHsp90 were covalently linked to the CM5 chip via amine residues according to the supplier's instructions; for determination of binding constants in the human system, Hop was linked the same way to the CM5 chip. The measurements were performed at a flow rate of 5 l/min at 25°C. To determine the binding constant, we directly measured the change in response units, when different amounts of Sti1 were injected onto a Ssa1-or yHsp90-coated chip in standard buffer with or without 2 mM ATP or different amounts of Hsp70 and Hsp90 were injected onto the Hop-coated chip in standard buffer with or without 2 mM ATP.
Data analysis for direct binding curves used the linear relationship between the resonance signal (RU) and the amount of protein (c) bound to the chip. RU max (Equation 1) is the maximum signal at which all of the Ssa1 molecules on the chip are saturated with Sti1.

RESULTS
The Hsp70 Component of the Sti1 Complex Is the Ssa Proteins-An important step in characterizing the Hsp90 multichaperone complex in S. cerevisiae is to elucidate which of the 14 Hsp70 proteins present in yeast interacts with cytosolic Sti1.
Nine of the 14 different yeast Hsp70 proteins are cytosolic and therefore potential partners of the co-chaperone Sti1. It had been shown previously that at least one member of the Ssa class interacts with Sti1 in yeast (29), but it was not clear whether other yeast Hsp70s also interact with Sti1 in the Hsp90 system.
To determine which of these Hsp70 proteins interacts with Sti1, we tested Ssa1-4, Ssb1-2, Sse1-2, and Ssz1 in a yeast two-hybrid reporter system for interaction with Sti1 ( Fig. 1 A). Only for the Ssa class (Ssa1-4) positive results were obtained suggesting that the Ssa proteins are the only Hsp70 component in the yHsp90 complex.
As an independent proof for this interaction, we immunoprecipitated Sti1 from a wild type yeast lysate (strain W303) and tested the precipitate by Western blotting for the presence of Ssa proteins. Fig. 1B (lane 1) shows that the Ssa proteins were precipitated together with Sti1. When the cell lysates were supplemented with purified Sti1, the amount of Ssa1 co-precipitated increased with increasing amounts of purified Sti1 added (Fig. 1B, lanes 2 and 3). As a control for the specificity of the Sti1 antibody, we repeated the experiment using lysate from the Sti1 knock-out strain CN11a (Fig. 1B, lane 4). No Ssa1 could be detected in this case. To test whether the Sti1 complex contained only Ssa proteins and not Ssb proteins, we probed the blots with an antibody specific for Ssb proteins. This experiment showed that Ssb proteins could not be detected in the Sti1 complex (Fig. 1B, lane 5). Ssb proteins were, however, present in the lysate as shown in Fig. 1B (lane 6). This result is in agreement with previous studies in the mammalian system (25), because only the Ssa class of Hsp70s share the C-terminal sequence motif (I/L/M)EEVD-COOH.

Sti1 Accelerates the Ssa1 ATPase Activity
Because Ssa3 and Ssa4 are only expressed under heat shock, these proteins are not relevant for the function of the yHsp90 multichaperone complex under normal growth conditions. Therefore, we focused in the following on the Ssa1 protein, because Ssa1 and Ssa2 are 97% identical and expressed in equal amounts under these conditions (8).
Sti1 Binds to Ssa1 with Low Affinity-To further characterize the interaction between Ssa1 and Sti1, we purified both proteins and studied complex formation under physiological conditions concerning salt concentration and temperature.
First, we incubated the proteins and performed immunoprecipitations with anti-Sti1 or anti-Ssa antibodies and separated the precipitate by SDS-PAGE. Fig. 2A indicates that the Sti1-Ssa1 complex was formed efficiently under these conditions. We found that the amount of Ssa1 in the complex was not significantly affected by ATP ( Fig. 2A, lanes 2 and 4).
To gain quantitative data on the interaction between Ssa1 and Sti1, surface plasmon resonance spectroscopy was employed. Binding of different Sti1 concentrations with or without ATP to immobilized Ssa1 was analyzed. The data from the resulting sensorgrams were used to determine the dissociation constant of the Ssa1-Sti1 interaction (Fig. 2B). We found that in the absence of ATP, Sti1 binds to Ssa1 with an affinity of 7.5 Ϯ 2 M, whereas in the presence of ATP, the affinity is 3 Ϯ 1 M (Fig. 2B and Table I). Compared with the affinity of Sti1 for yHsp90 (28,43) (Table I), Sti1 binding to Ssa1 is about 2 orders of magnitude weaker.
Sti1 Is a Potent Activator of the ATPase Activity of Ssa1-It has been described previously that Sti1 inhibits the ATPase activity of yHsp90 (28). We were interested in determining whether Sti1 influences the ATPase activity of Ssa1 as well. Sti1 itself does not exhibit ATPase activity (data not shown). The ATPase activity of Ssa1 is similar to that of other Hsp70s studied, with a K m for ATP of 0.2 M and a k cat for ATP of 0.04 min Ϫ1 at physiological potassium concentrations (48). When Sti1 was added to Ssa1, we found that, unexpectedly, the ATPase activity of Ssa1 increased with increasing amounts of Sti1 (Fig. 3A). A stimulation by a factor of 200 was reached at a Sti1:Ssa1 ratio of 1:1 (Fig. 3B). Half-maximal activation was reached at a ratio of Ssa1:Sti1 ϭ 1:0.5. When thermally unfolded Sti1 was added in a control experiment, no significant The pAS1-CYH2 plasmids containing nine different Hsp70 genes were transformed together with pACTII-Sti1 in the yeast strain Y190. The strains were grown for 2 days at 30°C on His Ϫ Leu Ϫ Trp Ϫ plates. After transfer to a nitrocellulose filter, the interactions were identified by ␤-galactosidase activity (blue colonies). Only for the Ssa group, a positive result could be obtained. The reciprocal experiment yielded the same result (not shown). As a negative control, the plasmids were transformed separately in Y190. B, co-immunoprecipitations (IP). These experiments were performed at 25°C with yeast lysate from the wild type strain W303 (lanes 1-3 and 5) supplemented with the indicated amounts of purified Sti1 or the Sti1 knock-out strain CN11a (lane 4). Sti1 was precipitated with a polyclonal Sti1 antibody, and the precipitate was analyzed using SDS-PAGE. The proteins were then transferred to a polyvinylidene difluoride membrane, and the Ssa or Ssb proteins were detected by Western blot analysis using anti-Ssa-IgG (lanes 1-4) or anti-Ssb-IgG (lanes 5 and 6). As a positive control for the presence of Ssb proteins in the strain W303, we performed a Western blot analysis of the lysate with anti-Ssb-IgG (lane 6). wt, wild type.

FIG. 2. In vitro analysis of Ssa1⅐Sti1 complexes.
A, co-immunoprecipitations. The experiments were carried out at 25°C with the purified proteins Ssa1 (5 g) and Sti1 (5 g) using polyclonal antibodies against Sti1 (lanes 1, 2, and 6) or Ssa1 (lanes 3-5). If present, ATP was added to a final concentration of 2 mM. The precipitate was separated by 4 -12% SDS-PAGE. We found that the dimeric complex forms in the absence and presence of ATP. B, surface plasmon resonance analysis. These experiments were performed with Ssa1-coated chips as described in the experimental protocol. The plot shows the signal of binding (response units, RU) versus increasing Sti1 concentrations. The data points were fitted, and the dissociation constant K D was calculated as described in the experimental protocol. E, Ϫ ATP; q, ϩ 2 mM ATP. increase in the ATPase activity of Ssa1 was detected (Table II).
To determine which step in the Ssa1 ATPase cycle is accelerated by Sti1, we compared steady state measurements with single turnover experiments in the presence and absence of Sti1 (Fig. 3). It was not known whether ATP hydrolysis or nucleotide exchange represents the rate-limiting step in the ATPase cycle of Ssa1. To discriminate between the two possibilities, we analyzed the steady state k cat and the single turnover k cat . In steady state measurements, excess of ATP is used, and therefore Ssa1 hydrolyzes ATP with maximum velocity, whereas in single turnover assays, about 80% of Ssa1 is saturated with ATP. Thus, a single hydrolysis step could be monitored, which allows, in comparison with the steady state measurement, determination of the rate-limiting step of the ATPase cycle. We found that the k cat values (0.04 min Ϫ1 ) were identical in the absence of Sti1. Thus, we conclude that ATP hydrolysis, or a conformational change preceding it, is rate-limiting for the Ssa1 ATPase cycle. Increasing amounts of Sti1 resulted in an acceleration of the single turnover ATPase activity of Ssa1 (Fig.  3C). A stimulation factor of 200 could also be seen in the single turnover experiments at a Sti1 to Ssa1 ratio of 1:1. In the presence of Sti1, both the steady state k cat and the single turnover k cat were accelerated (Fig. 3). This demonstrates that Sti1 stimulates ATP hydrolysis and confirms that this step is rate-limiting in the Ssa1 ATPase cycle.
As previously described (53), Ydj1 is able to accelerate the ATPase of Ssa1 by a factor of 10. When Ydj1 was added to a preformed Ssa1-Sti1 complex at a 1:1 ratio or at high excess, the activation resulting from Sti1 was not influenced (data not shown). Further, if Sti1 was added to an Ssa1-Ydj1 complex, a 200-fold activation was observed.
Influence of Sti1 on Ssa1 and yHsp90 in the Ternary Multichaperone Complex-To study the ternary multichaperone complex consisting of Ssa1, Sti1, and yHsp90, we incubated the purified Ssa1, Sti1, and yHsp90, precipitated with anti-Ssa1, anti-Sti1 or anti-yHsp90 antibodies and separated the precipitate by SDS-PAGE (Fig. 4A). The three proteins could be detected in equal amounts no matter which antibody was used. This complex is also present in the absence of ATP (Fig. 4A,  lane 4).
To study the influence of Sti1 on the ATPase activities of yHsp90 and Ssa1 in the ternary complex, it was necessary to incorporate either a Ssa1 mutant (K69Q) (40) or a yHsp90 mutant (E33A) (54, 55), both of which are defective in ATP hydrolysis in the complex. Using these variants, complex formation with Sti1 was not affected (data not shown). When we analyzed the ATPase activity of Ssa1 in a Ssa1-Sti1-yHsp90(E33A) complex and the ATPase activity of yHsp90 in a Ssa1(K69Q)-Sti1-yHsp90 complex, we could not detect any difference compared with the respective binary wild type complexes (Fig. 4B). Thus, the effect of Sti1 on the Ssa1 and yHsp90 ATPase activities in the ternary complex seems to be the same as in the dimeric complexes.
Comparison with the Mammalian System-Taken together, our results clearly show that Sti1 accelerates the ATPase activity of Ssa1. This is in contrast with previous studies on the function of Hop in the mammalian system. Here, Hop did not exhibit an effect on the ATPase activity of Hsp70 (56). To directly compare the yeast with the mammalian system, we purified the mammalian proteins. First, we determined the binding constants between Hsp70, Hop, and human Hsp90 The K D values were obtained by SPR spectroscopy as described under "Experimental Procedures." The experiments were performed in 40 mM Hepes, pH 7.5, 150 mM KCl, 5 mM MgCl 2 , and 1 mM dithiothreitol at 25°C. For experiments in the yeast system, Ssa1 or yHsp90 were covalently linked to the CM5 chip via amine residues according to the supplier's instructions; for determination of binding constants for the homologous human proteins, Hop was linked to the CM5 chip. ND, not determined. (hHsp90) by surface plasmon resonance spectroscopy. As can be seen in Table I, Hsp70 binds with low affinity (K D ϭ 1.5 M) to Hop. The affinity is increased in the presence of hHsp90 (K D ϭ 0.3 M). hHsp90 binds with higher affinity to Hop (K D ϭ 0.1 M), and the binding is not affected by Hsp70. The values are comparable with previously published data determined by other experimental procedures (57).
These data allowed us to perform ATPase experiments under conditions where the respective proteins are associated. The steady state ATPase activities of hHsp90 and Hsp70 were determined to be 0.04 and 0.4 min Ϫ1 , respectively. This is in the range of previously published results (58,59). In agreement with data from Johnson et al. (56), our experiments show that Hop does not influence the ATPase activity of Hsp70 or that of hHsp90 (Table II).
To test whether the effect of Sti1 is specific for Ssa1 and yHsp90, we analyzed the influence of Sti1 on the ATPase activities of Hsp70 and hHsp90 and the influence of Hop on the ATPase activities of Ssa1 and yHsp90 (Table II). We found that Sti1 influenced neither the ATPase activity of Hsp70 nor that of hHsp90, even if added in high excess. Furthermore, the experiments showed that Hop had no influence on the ATPase activity of Ssa1 or yHsp90. DISCUSSION In eukaryotes, multiple Hsp70s exist in one compartment. This highlights the functional diversity of these proteins. In yeast, the subgroups Ssa, Ssb, Sse, and Ssz are present in the cytosol. However, as we show here, only the Ssa class is involved in cooperation with the yHsp90 chaperone machinery. This allows specific regulation of one subset of Hsp70s.
Co-chaperones, which modulate Hsp70 function, have been extensively studied for E. coli DnaK. DnaK is regulated by DnaJ, which stimulates the ATP hydrolysis step in the DnaK ATPase cycle (60), yielding the high affinity ADP state for substrate binding and GrpE, which functions as a nucleotide exchange factor for release of substrate and allows the cycle to begin anew (61,62). The only yeast co-chaperones reported so far that have a stimulatory effect on the ATPase activity of Ssa1 are Ydj1 and Sis1, which belong to the DnaJ/Hsp40 family. They activate the ATPase activity of Ssa1 about 10-fold by catalyzing the ATP hydrolysis step in the Ssa1 ATPase cycle (53,63,64).
More divergent are GrpE homologues, which have so far only been identified in prokaryotes (65) and mitochondria (66). The eukaryotic protein Bag-1, which is discussed as a nucleotide release factor of Hsp70 (67), regulates the activity of Hsp70 proteins by binding to the N-terminal ATPase domain of Hsp70 (68). Snl1 from S. cerevisiae may be the yeast homologue of Bag-1 (69). In previous studies, a protein with similarity to Hop from rabbit reticulocyte lysate was shown to activate the ATPase of Hsp70 about 7-fold (70, 71) by stimulating nucleotide exchange, which implies that this protein activates Hsp70 through a GrpE-like mechanism. Hop itself does not have an influence on the ATPase activity of Hsp70 in the mammalian system (56).
Sti1 is a new member of the family of Hsp70 ATPase modulators. Sti1 is composed of three TPR domains. TPR domains are loosely conserved repeats of roughly 34 amino acids known to mediate protein-protein interactions (24,25). In Hop and Sti1, they serve to link Hsp70 and Hsp90 (1,72). The Nterminal TPR domain of Hop interacts with Hsp70, and the TPR domains in the C-terminal half seem to be important for the interaction with Hsp90 (25). We show here that Sti1 is a  4. Analysis of complexes between Ssa1, Sti1, and yHsp90 in vitro. A, co-immunoprecipitations. These experiments were carried out at 25°C with the purified proteins Ssa1 (5 g), Sti (5 g), and yHsp90 (5 g) using polyclonal antibodies against Sti1 (lanes 1 and 5), Ssa1 (lanes  2 and 6), or yHsp90 (lanes 3, 4, and 7). If present, ATP was added to a final concentration of 2 mM. The precipitate was separated by 4 -12% SDS-PAGE. We found that the ternary complex forms in the absence and presence of ATP. B, ATPase effects in the ternary multichaperone complex. ATPase assays were carried out at 37°C using [␣- 32  To determine the effects in the ternary complex, Ssa1 and yHsp90 were replaced by the respective ATPase mutants Ssa1(K69Q) and yHsp90(E33A). The reactions were incubated for 120 min, and changes in the ADP/ATP ratios were monitored. The effect of Sti1 on the Ssa1 and yHsp90 ATPase activities in the ternary complex seems to be the same as in the dimeric complexes. potent activator of the ATPase activity of Ssa1. Sti1 does not exhibit significant sequence homology to the DnaJ or GrpE class. Interestingly, as described above, all of the stimulation factors known for Ssa1, including peptides, show no more than 10-fold activation, whereas Sti1 increases the overall ATPase activity of Ssa1 up to 200-fold. Comparing steady state and single turnover ATP hydrolysis, we show that the hydrolysis step or conformational rearrangements preceding it are ratelimiting for the Ssa1 ATPase cycle. Sti1 specifically accelerates this step. Thus, Sti1 is the first "non-DnaJ" activator of this step. Moreover, Sti1 is able to activate Ssa1, even if Ssa1 is already stimulated maximally by Ydj1, and Ydj1 is not able to replace Sti1 in a preformed Ssa1-Sti1 complex, even if added in high excess. Therefore, we conclude that Sti1 and Ydj1 have separate, unique binding sites on Ssa1. Further, we show that binding of Sti1 to yHsp90 and Ssa1 allows to activate Hsp70 and inhibit Hsp90 at the same time.
Our studies of the yeast and mammalian proteins show that Sti1 acts differently from Hop. Whereas Hop only seems to act as a passive linker protein between Hsp70 and hHsp90 (1), Sti1 is a regulatory linker protein for Ssa1 and yHsp90. It is able to activate the ATPase activity of Ssa1 and inhibit the ATPase activity of yHsp90 at the same time. Moreover the experiments with mixed complexes between yeast and human components show that the effect of Sti1 on Ssa1 is specific, because Hop has no influence on the ATPase activity of Ssa1, and Sti1 has no influence on the ATPase activity of Hsp70. Because the ATPase activity of Hsp70s can be stimulated by peptides and denatured protein substrates, the formal possibility existed that Sti1 binds unspecifically to the substrate-binding site of Ssa1 and stimulates ATP hydrolysis this way rather than being an allosteric effector. However, the analysis of the effect of unfolded Sti1 on Ssa1 and the finding that Sti1 does not influence human Hsp70 allow direct exclusion of this possibility. Therefore, we conclude that Sti1 acts as a specific activator of Ssa1.
This raises the question of how the regulation of Hsp70 and Hsp90 is accomplished in the context of the human Hsp90 system. Interestingly, the basic ATPase activity of hHsp90 is 10-fold weaker compared with yHsp90, and the ATPase activity of Hsp70 is about 10-fold higher compared with Ssa1. Thus, compared with the yeast system, Hsp70 is already activated, and hHsp90 is inhibited. Therefore, different modes of regulation may be required for the human system. For the yeast system, clearly, the role of Sti1 in the Hsp90 chaperone cycle is not only that of a scaffold protein for Hsp70 and Hsp90 but also that of a chimeric regulator for both proteins, which changes the state of activity of the two components several hundred-fold. In the Hsp90 chaperone cycle, Hsp70 is the first component of the Hsp90 complex that can be coprecipitated with steroid hormone receptors (20). Hsp90 and Sti1/Hop then join the complex. We suggest that the differential regulation of the ATPase activities of both yHsp90 and Ssa1 by Sti1 is the basis for the committing step of the yeast Hsp90 chaperone cycle.