Cns1 Is an Activator of the Ssa1 ATPase Activity*

Hsp90 is a key mediator in the folding process of a growing number of client proteins. The molecular chaperone cooperates with many co-chaperones and partner proteins to fulfill its task. In Saccharomyces cerevisiae , several co-chaperones of Hsp90 interact with Hsp90 via a tetratricopeptide repeat (TPR) domain. Here we show that one of these proteins, Cns1, binds both to Hsp90 and to the yeast Hsp70 protein Ssa1 with comparable affinities. This is reminiscent of Sti1, another TPR-contain-ing co-chaperone. Unlike Sti1, Cns1 exhibits no influence on the ATPase of Hsp90. However, it activates the ATPase of Ssa1 up to 30-fold by accelerating the rate-limiting ATP hydrolysis step. This stimulating effect is mediated by the N-terminal TPR-containing part of Cns1, whereas the C-terminal part showed no effect. Competition experiments allow the conclusion that Hsp90 and Ssa1 compete for binding to the single TPR domain of Cns1. Taken together, Cns1 is a potent co-chaperone of Ssa1. Our findings highlight the importance of the regulation of Hsp70 function in the context of the Hsp90 chaperone cycle. Hsp90 1 molecular in the cytosol nonstress conditions. It essential In contrast to

Hsp90 1 is an abundant molecular chaperone in the cytosol even under nonstress conditions. It has been shown to be essential in both higher eucaryotes and yeast (1)(2)(3). In contrast to Hsp70, which is involved in a wide variety of functions including folding of nascent polypeptide chains (4 -6), Hsp90 seems to be a more specialized folding factor (6 -10). It cooperates with a still increasing number of co-chaperones. Some of these partner proteins exhibit chaperone function on their own, such as Cdc37 (11), the large prolyl isomerases, and p23 (12,13). Another partner protein, Sti1 (stress-inducible protein 1), inhibits the ATPase of yeast Hsp90 as a noncompetitive inhibitor (14,15). Furthermore, Sti1 was recently identified as a potent activator of the Ssa members of the Hsp70 family (6,16). Hop (Hsp70/Hsp90 organizing protein), the mammalian homologue of Sti1 (17,18), seems to be of central importance for regulating the Hsp90 chaperone cycle, as it binds to the C termini of both Hsp90 and Hsp70 (19), thus providing a physical link between the Hsp70 and Hsp90 chaperone machineries. Binding involves the TPR domains of Sti1/Hop. TPR domains consist of TPR motifs, which are degenerated sequences of 34 amino acids (14,20,21) that form a structurally conserved tandem array of two anti-parallel ␣-helices. Multiple TPR motifs adopt a righthanded superhelix structure, which creates a groove with a relatively large surface area, which is suitable for proteinprotein interactions. In the case of Hsp90 and Hsp70/Ssa1, binding of their highly conserved C-terminal residues EEVD to TPR domains is based predominantly on a "carboxylate clamp," which anchors the aspartate via electrostatic interactions to a conserved set of side chains inside this groove. The TPR domain of Cns1 (cyclophilin seven suppressor) can be expected to form this clamp, since the residues Lys 87 , Asn 91 , Asn 126 , Lys 156 , and Arg 160 agree with the respective consensus sequence (22). Several Hsp90 co-chaperones, including the prolyl isomerases Cpr6 and Cpr7, Ppt1 (protein phosphatase 1), and the protein Cns1 contain TPR motifs. Cns1 was discovered as a suppressor of a severe growth defect caused by a deletion of Cpr7 (23). A physical interaction between Cns1 und Hsp90 was shown by immunoprecipitation (23). Cns1 is a protein of 385 amino acids, resulting in a molecular mass of 44.1 kDa (24). The threedimensional structure of Cns1 is still unknown. Sequence analysis indicates the presence of three TPR motifs in the N-terminal part of the protein (Fig. 1). Cns1 shares weak homology of 24% and 18% sequence identity with Sti1. Thus, Cns1 was assigned to be a homologue of Sti1 (23). However, regions of high homology to Sti1 can only be found in the TPR domain of Cns1. In contrast to other co-chaperones containing a TPR domain, Cns1 is essential in yeast (4,23,(25)(26)(27)(28). Overexpression of Sti1, or of other TPR-containing proteins cannot compensate for the lethality of ⌬Cns1 (23,29). Furthermore, unlike Sti1, Cns1 expression is not up-regulated upon heat shock (23). Under different stress conditions, Cns1 even seems to be regulated in an opposite manner compared with Sti1 (30). Also, the expression levels of the two proteins seem to be completely different, since Sti1 was found to be expressed with 6.77 ϫ 10 5 molecules/cell, whereas Cns1 is present at ϳ670 copies/cell and is therefore among the most poorly expressed proteins in yeast (31).
Here we were interested in analyzing the function of Cns1 in the regulation of the Hsp70 and Hsp90 chaperones. We show that Cns1 binds both to Hsp90 and Ssa1. Cns1 does not influence the ATPase activity of Hsp90, but it is a strong activator of the Ssa1 ATPase.
Protein Expression and Purification-Cns1 constructs were expressed in Escherichia coli HB101 (pUBS) (32). Cells were grown at 37°C to an A 600 of 0.8. After a temperature shift to 25°C, recombinant protein expression was induced by the addition of 1 mM isopropyl-1thio-␤-D-galactopyranoside. After 6 h, cells were lysed using a cell disruption system (Constant Systems Ltd., Warwick, UK). Cell lysis was performed in 40 mM K 2 HPO 4 /KH 2 PO 4 , pH 8.5, 400 mM KCl, 6 mM imidazole. Cns1 was purified in a three-step approach, beginning with a chelating Sepharose column (Amersham Biosciences), which was preloaded with 100 mM NiSO 4 . After applying the cell lysate, the column was washed with 10 volumes of 40 mM K 2 HPO 4 /KH 2 PO 4 , pH 8.5, 400 mM KCl, 20 mM imidazole before elution was performed in a step gradient with buffer containing 300 mM imidazole. Further purification was carried out on a Resource Q column (Amersham Biosciences), on which the protein was loaded in 50 mM Tris, pH 8.0, 20 mM KCl, 5 mM glycine, 5 mM DTT and eluted with a gradient from 20 mM to 1 M KCl. As a final step, a Superdex 75 26/60 HiLoad (Amersham Biosciences) was run in 40 mM Hepes, pH 7.5, 300 mM KCl. Cns1 was stored in 40 mM Hepes, pH 7.5, 40 mM KCl, 5 mM DTT at concentrations of 3-9 mg/ml at Ϫ80°C.
Ssa1-SBD was expressed in E. coli BL21 (DE3) codon ϩ cells (Stratagene, La Jolla, CA) that were grown at 37°C to an A 600 of 0.8. After the addition of 1 mM isopropyl-1-thio-␤-D-galactopyranoside, the recombinant protein was expressed for 4 h. The Ssa1-SBD was purified as described for Cns1. Ssa1-SBD was stored in 40 mM Hepes, pH 7.5, 50 mM KCl, at concentrations of 1.4 mg/ml at Ϫ80°C.
Saccharomyces cerevisiae Ssa1 was expressed using the Pichia pastoris expression system and purified as described earlier (33). Sti1 was purified as described elsewhere (15).
To verify the identity of Cns1, mass spectrometry of a tryptic digest of Cns1 was performed in a Biflex III matrix-assisted laser desorption ionization time-of-flight spectrometer (Bruker, Bremen, Germany). The results from electrospray ionization time-of-flight mass spectrometry (Ettan LC-MS system; Amersham Biosciences) confirmed the calculated mass of Ssa1-SBD and of the Hsp90-⌬MEEVD and -⌬16 constructs.
Circular Dichroism Measurements-Far UV CD spectra were recorded using a J-715 spectropolarimeter equipped with a peltier unit (Jasco, Gross-Umstadt, Germany). The measurements were performed in 40 mM NaH 2 PO 4 , pH 8.5, at 20°C. Far UV spectra were measured with protein concentrations of 300 g/ml in quartz-cuvettes of 0.1 cm in a wavelength range of 195-260 nm. All spectra were recorded in 10 accumulations and are buffer-corrected. Temperature transitions were recorded at a wavelength of 220 nm from 20 to 80°C with a heating rate of 30°C/h. CD spectra analysis was performed with the programs CDNN (CD deconvolution based on neural networks), using a network trained with 33 spectra (37), and k2d, trained with 100 spectra (38).
Determination of the Oligomeric State of Cns1 by SEC-HPLC-To define the quaternary structure of Cns1, SEC-HPLC was performed on a PU-1580 system (Jasco). A Superdex 75 HR column (Amersham Biosciences) was run at 25°C with 0.5 ml/min, using 10 mM Na 2 HPO 4 / NaH 2 PO 4 , pH 7.5, 150 mM NaCl. Protein fluorescence was detected using an FP-920 fluorescence detector (Jasco) with excitation at 280 nm and emission at 320 nm. The proteins used to generate a calibration curve were bovine serum albumin (67.0 kDa), ovalbumin (47.9 kDa), chymotrypsinogen (25.0 kDa), and ribonuclease A (15.7 kDa) (Amersham Biosciences). The standard proteins eluted at 17.06, 18.77, 24.06, and 26.14 min, respectively. The apparent molecular mass of Cns1 was determined by logarithmic interpolation from the calibration curve.
Citrate Synthase Aggregation Assay-CS (Roche Applied Science) was thermally denatured by incubation at 40°C in 40 mM Hepes, pH 7.5, 150 mM KCl, 10 mM MgCl 2 . Below, this buffer will be referred to as "standard buffer." Aggregation of CS was measured by following the absorbance at 360 nm in a Cary50 UV-visible spectrophotometer (Varian, Darmstadt, Germany) with a thermostated cell holder. The CS assay was performed according to Buchner et al. (39).
Insulin Aggregation Assay-The insulin aggregation assay was performed in standard buffer. Turbidity was monitored at 360 nm in a Cary50 UV-visible spectrophotometer (Varian). Insulin (50 mM) (Sigma) was preincubated at 25°C with varying concentrations of Cns1. The aggregation reaction was started by the addition of DTT to a final concentration of 20 mM. The assay was performed according to Scheibel et al. (40).
Peptidyl-Prolyl cis/trans-Isomerase (PPIase) Assay: Refolding of RCM-T1-To measure the PPIase activity of Cns1, we assayed the refolding of a reduced and S-carboxymethylated S54G/P55N variant of RNase T1 (RCM-T1), as described previously (41). RCM-T1 was preincubated in an unfolded state in 100 mM Tris, pH 8.0, and allowed to refold spontaneously upon transfer into 100 mM Tris, pH 8.0, 2 M NaCl. The S54G/P55N mutation renders the cis/trans-isomerization of the Tyr 38 -Pro 39 peptide bond as the limiting factor in refolding. The addition of a PPIase accelerates this reaction. The assay was run with 0.7 M RNase T1, at 15°C in a Spex FluoroMax-2 fluorescence spectrometer (Jobin Yvon, Edison, NJ) at an excitation of 268 nm and an emission of 320 nm.
Surface Plasmon Resonance Measurements-Surface plasmon resonance (SPR) experiments were carried out on a Biacore X TM Instrument (Biacore, Uppsala, Sweden). For determination of binding constants, Ssa1, Ssa1-SBD, and Hsp90-⌬16 were covalently linked on the matrix of a CM5 chip via amine coupling according to the supplier's instructions. Measurements were performed using standard buffer at a flow rate of 10 l/min at 30°C. To determine the binding constant, we directly measured the increase in resonance units (RU), when different concentrations of Cns1 were injected. 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 is the maximum signal, at which all molecules on the chip are saturated with the analyte (Equation 1).
ATPase Activity (Regenerative System)-ATPase activities were measured using a regenerative ATPase assay as described previously (42). In short, the decrease in NADH concentration was detected by the decrease of absorbance at 340 nm, using a Cary50 UV-visible spectrophotometer (Varian). Assays were performed at 30°C in standard buffer with 2 mM ATP added. Proteins were preincubated for 5 min at 30°C. Typical protein concentrations were 2 M for Hsp90, 5 M for Ssa1, and up to 30 M for the additional proteins. For full inactivation of the ATPase of Hsp90-⌬MEEVD, radicicol (Sigma) was used in a molar ratio of 2:1 to Hsp90-⌬MEEVD. Each protein was tested for contaminating ATPase activities. Data analysis was performed using Equation 2.
ATPase Activity (Radioactive Assay)-ATPase assays were performed according to Kornberg et al. (43). The proteins were incubated in standard buffer at 30°C. The ATPase reaction was initiated by the addition of the respective ATP concentrations including [␣-32 P]ATP (Hartmann Analytic, Braunschweig, Germany). A sample contained 0.1 Ci of [␣-32 P]ATP. As a control for unspecific protein effects, Cns1 was added after incubation at 70°C for 30 min. For steady state hydrolysis measurements, a final ATP concentration of 2 mM was used. In the case of single turnover experiments, the ATP/protein ratio was kept constant at 0.8:1. After thin layer chromatography, the ATP/ADP ratio was quantified with a Typhoon 9200 PhosphorImager (Amersham Biosciences). Hydrolysis rates were corrected for uncatalyzed, spontaneous ATP hydrolysis.

RESULTS
Structure and Stability of Cns1-The hydropathy plot (44) of Cns1 suggested two major structural units (Fig. 1). The first potential domain ranges from the N terminus to approximately amino acid 200, and the second is approximately between amino acid 215 and the C-terminal end of the protein. The Cns1 Is an Activator of the Ssa1 ATPase Activity respective constructs for each domain were produced with a view to analyze their contribution to the function of Cns1. Cns1 and Cns1 constructs were expressed in E. coli and purified to homogeneity.
CD spectroscopy confirmed that Cns1 and the respective constructs are folded proteins, with a high ␣ helix content in the full-length protein ( Fig. 2A and Table I) and the N-terminal construct (data not shown). CD data analysis performed with the CD deconvolution program CDNN (37) or k2d (38) resulted in 38.6 or 43% ␣-helix and 26.4 or 23% ␤-structure content, respectively (Table I).
Analytical size exclusion chromatography was performed to determine the quaternary structure of Cns1. Cns1 eluted at a elution time corresponding to a monomer (Fig. 2B). The fragments of Cns1 were also monomeric (data not shown).
To analyze the functional integrity of Cns1 under assay conditions, we determined the stability of Cns1 against thermal unfolding. We observed a slight decrease in the far UV CD signal intensity at temperatures above 38°C, followed by a defined transition in the range of 40 to 45°C. The melting point (T m ) was determined to be at 42.8 Ϯ 0.1°C. Due to aggregation of Cns1 at high temperatures, the transitions were not reversible (data not shown).
Chaperone/PPIase Activity of Cns1-Since several co-chaperones of Hsp90 exhibit chaperone activity on their own (11)(12)(13), we were interested whether Cns1 has positive effects on the denaturation and aggregation of model substrate proteins. To test the influence of Cns1 on the thermal denaturation of a protein, we used the CS aggregation assay (39). Here, even at a 10-fold molecular excess over CS, no effect of Cns1 was detectable (data not shown). The influence of Cns1 on chemical denaturation of a protein was assayed with insulin, which, upon reduction of its interchain disulfide bonds by DTT, is prone to aggregation (45). Cns1 was not able to prevent or delay insulin aggregation at any of the conditions tested (data not shown).
For Cpr6 and Cpr7, PPIase activity has been demonstrated previously (26,35,46). Since Cns1 is a suppressor of the cellular defects observed upon deletion of the TPR-containing peptidyl-prolyl isomerase Cpr7, the formal possibility existed that Cns1 also may exhibit prolyl isomerase activity. Using the well established RNase T1 assay (47), we could confirm that Cpr6 and Cpr7 clearly accelerated the refolding of the T1 RNase. Cns1, however, had no influence on the reaction under all conditions tested and therefore cannot be considered as a PPIase (data not shown).
Cns1 Binds to Hsp90 -The presence of a TPR domain suggested that Cns1 may bind to the C-terminal end of Hsp90. To obtain quantitative data on the interaction between Cns1 and Hsp90-⌬16, SPR spectroscopy was applied. The analysis of sensorgrams of the binding of Cns1 at different concentrations to immobilized Hsp90-⌬16 allowed us to determine the dissociation constant of the Cns1-Hsp90 interaction to be 5.1 Ϯ 0.4 M (Fig. 3A). The TPR-containing fragment Cns1-N204 was also able to bind to Hsp90-⌬16 with comparable affinity of 5.3 Ϯ 0.8 M (data not shown). Binding of Cns1 to Hsp90-⌬16 required the TPR domain, since there was no detectable interaction between Hsp90-⌬16 and the C-terminal domain of Cns1, Cns1-218C (data not shown). In a different approach, using the C-terminally truncated Hsp90-⌬MEEVD construct immobilized on the SPR chip, we could not detect binding of Cns1 or its fragments Cns1-N204 and Cns1-218C (data not shown). In another approach, when Hsp90-⌬MEEVD was added to Cns1 prior to the injection on a Hsp90-⌬16 coated SPR sensor chip, it was not able to decrease the binding signal of Cns1 (Table III). A decrease would be expected for a protein that is able to bind to Cns1. The lack of the C-terminal residues of the Hsp90-⌬MEEVD construct abolishes the binding to Cns1. This demonstrates that the TPR domain of Cns1 interacts with the C-terminal end of Hsp90-⌬16.
Influence of Cns1 on the ATPase Activity of Hsp90 -Since it is known that binding of Sti1 to Hsp90 nearly abolishes the ATPase activity of Hsp90 (14,15), we wanted to explore whether Cns1 exhibits a similar effect on Hsp90. First, we determined the ATPase activity of Hsp90 under steady state conditions to be 0.3 min Ϫ1 . This is in the range of previously published results (14 -16). When increasing amounts of Cns1 were added to Hsp90, no apparent change in the ATPase activity of Hsp90 was observed, even at a 15-fold molar excess of Cns1 over Hsp90 (Fig. 3B). Different from this, Sti1, added at a 5-fold excess over Hsp90, inhibited the ATPase of Hsp90 severely (Table II, Fig. 5). Together with the results from the SPR measurements (Fig. 3A), this suggests that Cns1 binds to Hsp90 without altering its ATPase activity.
Cns1 Binds to Ssa1 and Activates Its ATPase-For Sti1, it had been shown recently that, in addition to binding to Hsp90, it also interacts with the Ssa proteins (16). Therefore, we tested whether Cns1 is also able to bind to Ssa1. Interestingly, we found that Cns1 binds to Ssa1 in a concentration-dependent manner. The affinity was determined to be 11.5 Ϯ 0.4 M (Fig.  4A). The finding that Sti1 is a potent activator of the Ssa1 ATPase (16) raised the question of whether Cns1 could act in a similar way and stimulate Ssa1. To test this hypothesis, we assayed the ATPase activity of Ssa1 in the presence of increasing amounts of Cns1 (Fig. 4B). We found a strong activation compared with the basal activity of Ssa1. The addition of Cns1 in a 15-fold molar excess over Ssa1 resulted in ATPase stimulation by a factor of 30. Half-maximum activation was reached at a ratio of Ssa1 to Cns1 of 1:4. From these data, we derived a binding constant of 16.5 Ϯ 1.1 M, which is in agreement with the results obtained in the SPR studies. In a single turnover experiment, where a substoichiometric amount of ATP is used, we analyzed which step of the Ssa1 ATPase reaction was accelerated by Cns1 (Fig. 4C). Cns1 was able to accelerate the ATPase activity of Ssa1 from 0.023 min Ϫ1 to a observed rate constant of 0.76 min Ϫ1 . This result suggests that Cns1 accelerates the ATP hydrolysis step or a conformational step preceding it in the ATPase cycle of Ssa1 similar to the mechanism of Sti1 (16).
To determine whether the Ssa1 ATPase may be stimulated by TPR proteins in general, we also tested the influence of Cpr7 and Ppt1 on the ATPase activity of Ssa1. At a ratio of Ssa1 to TPR protein of 1:4, none of these proteins was able to increase the ATPase of Ssa1 substantially (Table II, Fig. 5). In contrast, the TPR domain-containing N-terminal fragment of Cns1 (Cns1-N204) activated the ATPase of Ssa1 at the same ratio as the full-length protein. The C-terminal Cns1 fragment, Cns1-218C, showed no influence on the ATPase of Ssa1, even at a 15-fold excess.
Influence of Ydj1 on the Stimulation of the Ssa1 ATPase by Cns1-Next we were interested to test whether Ydj1 influences the ability of Cns1 to stimulate the ATPase of Ssa1. Ydj1 is an established co-chaperone of Ssa1, which is capable of activating the ATPase of Ssa1 about 5-10-fold (16,36). In our studies, Ydj1 was able to accelerate the ATPase of Ssa1 by a factor of 5 at a 4-fold molar excess over Ssa1. Previously, it was shown that Ydj1 did not impair the Sti1-mediated activation of the Ssa1 ATPase (16). To test whether Ydj1 acted in a competitive, additive, or synergistic manner with Cns1, we performed ATPase assays in the presence of both proteins. When we added Ydj1 in excess to a preformed Ssa1⅐Cns1 complex, the ATPase rate of Ssa1 increased by an amount similar to the stimulation of Ydj1 on Ssa1 alone. However, at high concentrations of Cns1, Ydj1 did not further contribute to the stimulation of the Ssa1 ATPase. Still, Ydj1 did not exhibit an inhibitory effect on the stimulation by Cns1, assuming two different binding sites of Ydj1 and Cns1 on Ssa1. Although the ternary complex between Ssa1, Cns1, and Ydj1 could not be directly detected, we assume based on the ATPase stimulation that both proteins bind simultaneously to Ssa1.
Influence of Hsp90 and Ssa1-SBD on the Stimulation of the Ssa1 ATPase by Cns1-As shown before, no influence of Cns1 on the ATPase of Hsp90 was observed. This finding allowed experiments to further define the influence of Hsp90 on the Cns1-mediated activation of the Ssa1 ATPase. Since the basal ATPase activity of yeast Hsp90 exceeds that of Ssa1 by more than a factor of 10 (Table II) (16), we used the ATPase-defective Hsp90 deletion mutant Hsp90-⌬16 (14) in these experiments. When Hsp90-⌬16 was added in a 10-fold excess (40 M to 4 M) to Ssa1⅐Cns1, a significant decrease of 40% upon the stimulation of the ATPase of Ssa1 by Cns1 was observed (Table III). This decrease in activity suggests a competition between Ssa1 and Hsp90-⌬16 for the interaction with Cns1. Hsp90 exclusively binds to the TPR domain of Cns1; a diminished activation of the Ssa1 ATPase upon the addition of Hsp90-⌬16 suggests that Ssa1 and Hsp90 compete for binding to the TPR domain of Cns1. This model is supported by the finding that Hsp90-⌬MEEVD could not compete with Ssa1 for binding to Cns1 and therefore had no influence on the Ssa1 ATPase stimulated with Cns1, whereas Hsp90-⌬16 and Ssa1-SBD are also able to compete for binding using SPR analysis (Table III). In the ATPase assay, Ssa1-SBD was able to suppress the Cns1mediated stimulation of the Ssa1 ATPase by 72%. Competing with immobilized Hsp90-⌬16 or Ssa1-SBD for Cns1 binding, Hsp90-⌬16 was able to hold off 93 or 65% of the Cns1 molecules from binding to the respective chip surfaces. The respective values for Ssa1-SBD as a competitor were Ϫ62 and Ϫ54%.
Hsp90-⌬MEEVD showed no effect on the SPR binding signal of Cns1 on a Hsp90-⌬16 chip. Since the binding affinities of Ssa1 and Hsp90 toward Cns1 are comparable, the inhibition of the Ssa1 ATPase by the addition of Hsp90-⌬16 is less than  2. Structural characterization of Cns1. A, far UV CD spectrum of Cns1. To analyze the secondary structure of Cns1, a far UV CD spectrum of Cns1 was recorded from 190 to 250 nm, with a protein concentration of 300 g/ml. B, SEC-HPLC of Cns1. SEC-HPLC runs were performed as described under "Materials and Methods." The straight line represents the fit obtained from the elution time of the standard proteins (gray dots) (bovine serum albumin (BSA), ovalbumin, chymotrypsin, and ribonuclease A). Cns1 (black square) has a calculated molecular mass of 50 kDa. The inset shows an SEC-HPLC run of Cns1, with a peak maximum at 18.8 min.
expected. This may be explained by the presence of a second, yet uncharacterized and weak binding site between Cns1 and Ssa1. For example, a second, weaker N-terminal binding site for Sti1 on Hsp90 was identified recently (14). DISCUSSION In this study, we show that Cns1 from S. cerevisiae is a novel TPR-containing modulator of the yeast Hsp70 ATPase activity. Among the TPR-containing proteins that are able to associate with Hsp90, Cns1, Sti1, and Cpr7 share the highest degree of homology. However, the homology is limited to the TPR domain. Marsh et al. (48) suggested that Cns1 is acting as a ⌬Cpr7 suppressor via its TPR domain. Thus, the PPIase function of Cpr7 does not seem to be important. This finding is further supported by sequence alignments between the PPIase domains of human cyclophilin A, Cpr6, Cpr7, and the C-terminal domain of Cns1, where Cns1 shares only two of the 11 conserved catalytic residues of a cyclophilin PPIase domain (49). PPIase assays performed in this study demonstrate directly the absence of PPIase activity in Cns1.
Cns1 contains only one TPR domain, and so far it has not been clear whether it can bind to Hsp90 and/or Hsp70. We show that Cns1 binds to yeast Hsp90 with micromolar affinity but does not have any influence on the Hsp90 ATPase activity. Furthermore, it also binds to the yeast Hsp70 protein Ssa1, which is the yeast Hsp70 most closely related to Hsp70s from higher eukaryotes. Cns1 is able to stimulate the basal Ssa1 ATPase rate up to 30-fold. This is well above the stimulation rates known for Hsp70 substrate proteins and DnaJ-like cofactors (6). The most extensively studied cochaperones that modulate Hsp70 function are DnaJ and GrpE in the E. coli DnaK system. DnaJ stimulates the ATP hydrolysis step in the DnaK ATPase cycle (50), which yields the high affinity ADP state for substrate binding. To release the bound substrate and allow the cycle to begin anew, GrpE, which functions as a nucleotide exchange factor, is needed in a next step (51,52). In S. cerevisiae, the co-chaperones Ydj1 and Sis1, which belong to the Hsp40/DnaJ family, stimulate the ATPase activity of Ssa1 ϳ10-fold (36,53,54). Although Cns1 and Ydj1 both stimulate the rate-limiting hydrolysis step or a conformational step preceding it, our studies suggest that the binding sites for Ydj1 and Cns1 are not overlapping, since Ydj1 does not decrease the stimulation of a preformed Cns1⅐Ssa1 complex.
In a recent study, we showed that Sti1 is a new member of the family of Hsp70 ATPase modulators, which is able to stimulate the ATPase activity of Ssa1 up to 200-fold (16). Sti1/Hop contains three TPR-domains (17). The N-terminal TPR domain of Sti1/Hop interacts with Ssa1/Hsp70, and the TPR domains in the C-terminal half seem to be important for the interaction with Hsp90 (22). Cns1 binds via its TPR domain to the EEVD motif of yeast Hsp90 and of yeast Hsp70. We found that Cns1-218C, even if added at a 15-fold excess over Ssa1, does not stimulate the ATPase activity of Ssa1, whereas Cns1-N204, the fragment containing the TPR domain, activates the Ssa1 ATPase similar to the full-length Cns1.
There are several interesting differences between the activators of the Hsp70 ATPase. First, Sti1 is an inhibitor of the Hsp90 ATPase, whereas Cns1 does not affect this reaction. Second, Sti1 forms ternary complexes with Hsp70 and Hsp90 (16,18,55), whereas in the case of Cns1, Hsp70 and Hsp90 seem to compete for binding to the single TPR domain, but we assume additional contacts of Cns1 with the N-terminal ATPase domain of Hsp70 and can therefore not exclude the possibility that the formation of a ternary complex between The k cat values were determined as described under "Materials and Methods." For the investigated ATPases, Ssa1 and Hsp90, a concentration of 5 M was used. Combinations of Ssa1 or Hsp90 with the indicated proteins were measured at 1:4 ratios. "Activation factor" denotes the relative activation that is caused by the addition of the respective proteins, compared with the basal activities of the Ssa1 and Hsp90 ATPases, which are set as 1 . 3. Interaction of Cns1 with Hsp90. A, Cns1 binding to Hsp90. SPR was used to measure the binding of Cns1 to Hsp90. Hsp90 was immobilized as described under "Materials and Methods." Cns1 concentrations were 125 nM to 20 M. The observed resonance signals were analyzed using least square data analysis and resulted in a binding constant of 5.1 Ϯ 0.4 M. B, influence of Cns1 on the ATPase activity of Hsp90. Kinetics of the ATPase activity of 2 M Hsp90 were measured using the regenerative ATPase assay as described under "Materials and Methods." Cns1 was added in concentrations ranging from 1 to 20 M. The best fit of the data is shown by the continuous line. Ssa1, Cns1, and Hsp90 is possible, even if a dimeric complex is more likely. Third, compared with Sti1, the levels of Cns1 are very low in the cell (56). Considering comparable binding affinities for Ssa1 and Hsp90, it is unlikely that Cns1 has the capability to uncouple the Hsp70/Hsp90 multichaperone cycle by competing with Sti1 for binding to Hsp70 and/or Hsp90. Indeed, it is more likely that Cns1 exhibits a specific function possibly in the context of the regulation of a subset of proteins. However, this function seems to be very important, since a Cns1 knock out is lethal (23). In this context, it seems to be necessary to elucidate the function of the C-terminal part of Cns1. So far, no known sequence motif could be assigned to this domain.
Taken together, our results define Cns1 as a potent co- The relative influences of Ssa1-SBD, Hsp90-⌬16, and Hsp90-⌬MEEVD as competitors for Cns1 binding were determined by assays with the Ssa1 ATPase or by SPR binding assays, as described. The stimulation/binding of Cns1 alone was set as 100%. The indicated proteins were added in a 10-fold excess (40 M) over Cns1 (4 M). ND, not determined. Single turnover ATPase assays were carried out at 30°C, using [␣-32 P]ATP. The ratio of Ssa1 to ATP was 1:0.8. The change in the ADP/ATP ratio was monitored over a time period of 60 min. No ATPase activity could be detected for Cns1 in the absence of Ssa1 and for the buffer alone. chaperone of Ssa1, which also binds to Hsp90 via TPR interactions without affecting the ATPase activity. This further highlights the importance of the regulation of Hsp70 function as part of the Hsp90 chaperone cycle.