Conformational Switching of the Molecular Chaperone Hsp90 via Regulated Phosphorylation

Molecular Cell

Article Conformational Switching of the Molecular Chaperone Hsp90 via Regulated Phosphorylation Joanna Soroka,1 Sebastian K. Wandinger,1,3 Nina Ma¨usbacher,2,4 Thiemo Schreiber,2 Klaus Richter,1 Henrik Daub,2,3 and Johannes Buchner1,* 1Center for Integrated Protein Science Munich (CIPSM) at the Department of Chemistry Technische Universita ¨ t Mu¨nchen, Lichtenbergstr. 4, 85747 Garching, Germany 2Department of Molecular Biology, Max Planck Institute of Biochemistry, Am Klopferspitz 18a, 82152 Martinsried, Germany 3Present address: Evotec AG, Am Klopferspitz 19a, 82152 Martinsried, Germany 4Present address: Department of Chemistry Technische Universita ¨ t Mu¨nchen, Lichtenbergstr. 4, 85747 Garching, Germany *Correspondence: [email protected] DOI 10.1016/j.molcel.2011.12.031

SUMMARY

Hsp90 is an essential molecular chaperone in the eukaryotic cytosol. Its function is modulated by cochaperones and posttranslational modifications. Importantly, the phosphatase Ppt1 is a dedicated regulator of the Hsp90 chaperone system. Little is known about Ppt1-dependent phosphorylation sites and how these affect Hsp90 activity. Here, we identified the major phosphorylation sites of yeast Hsp90 in its middle or the C-terminal domain and determined the subset regulated by Ppt1. In general, phosphorylation decelerates the Hsp90 machinery, reduces chaperone function in vivo, sensitizes yeast cells to Hsp90 inhibition and affects DNA repair processes. Modification of one particular site (S485) is lethal, whereas others modulate Hsp90 activity via distinct mechanisms affecting the ATPase activity, cochaperone binding and manipulating conformational transitions in Hsp90. Our mechanistic analysis reveals that phosphorylation of Hsp90 permits a regulation of the conformational cycle at distinct steps by targeting switch points for the communication of remote regions within Hsp90.

INTRODUCTION Hsp90 is a conserved ATP-dependent molecular chaperone involved in the conformational maturation of client proteins in eukaryotes (Picard, 2002; Pearl and Prodromou, 2006; Young et al., 2001; Geller et al., 2007; Pratt and Toft, 1997; Picard et al., 1990; Kleizen and Braakman, 2004). Because several Hsp90 clients participate as important mediators in essential cellular processes, such as gene expression, signal transduction, and cell-cycle progression, the cell requires a strict regulation of this system. In this context, it is not surprising that Hsp90 inhibitors act as effective cancer therapeutics that can block the proliferation of cancer cells (Neckers, 2007; Whitesell and Lindquist, 2005).

Hsp90 is a dimeric protein, and each protomer contains an N-terminal ATP-binding domain (N-domain), a middle domain (M-domain) with binding sites for client proteins and cochaperones (Meyer et al., 2004; Retzlaff et al., 2010) and a C-terminal dimerization domain (C-domain), whose C-terminal MEEVD motif mediates the docking of TPR domain-containing cochaperones (Scheufler et al., 2000; Ratzke et al., 2010). Crystal structures of Hsp90 (Ali et al., 2006; Shiau et al., 2006), together with an exhaustive analysis of the ATPase cycle of various Hsp90 homologs (Richter et al., 2001; Frey et al., 2007; Richter et al., 2008; Weikl et al., 2000), revealed the conservation of the overall structural organization and its enzymatic properties (Shiau et al., 2006). ATPase activity is essential for the Hsp90 chaperone function in vivo. ATP hydrolysis requires a series of conformational rearrangements that lead, via defined intermediates, to a closed state in which the N-domains are transiently dimerized and associated with the M-domains (Hessling et al., 2009; Ali et al., 2006). The activity of Hsp90 is regulated at several levels, including the ATPase cycle, with slow intrinsic conformational changes (Hessling et al., 2009; Richter et al., 2006; Graf et al., 2009), association with conformation-specific cochaperones (Ali et al., 2006; Johnson and Toft, 1995; Prodromou et al., 1999; Wandinger et al., 2008; Richter et al., 2003, 2004), and posttranslational modifications (Retzlaff et al., 2009; Scroggins et al., 2007). Several studies have identified Hsp90 as one of the major eukaryotic phosphoproteins (Iannotti et al., 1988; Lees-Miller and Anderson, 1989). Furthermore, changes in phosphorylation were demonstrated to impact its function (Mollapour et al., 2010, 2011; Lees-Miller and Anderson, 1989; Wandinger et al., 2006; Zhao et al., 2001). Interestingly, among Hsp90 cochaperones there is a Ser/Thr phosphatase Ppt1 that specifically dephosphorylates Hsp90 and the cochaperone Cdc37 (Vaughan et al., 2008; Wandinger et al., 2006). The absence of Ppt1 leads to decreased activation of client proteins in vivo (Wandinger et al., 2006). However, how this process is controlled on the mechanistic level is not understood. Our mass spectrometry (MS) analysis revealed that Hsp90 is phosphorylated at multiple sites in the M- and C-domains and showed that Ppt1 directly dephosphorylates Hsp90 at two serine residues. This affects ATPase activity, cochaperone binding, and Hsp90 function in vivo by distinct mechanisms.

Molecular Cell 45, 517–528, February 24, 2012 ª2012 Elsevier Inc. 517

Molecular Cell Phosphoregulation of the Hsp90 Chaperone Machinery

open structure ATP

A

S602

N

S604 S379

M

C

S282 S297 S334 S379 S450 S456 S485

S485

S282 S297 S334 S456

S602 S604

S450 S485

T533

S485

T533

S602 S604 closed structure

All-A

RESULTS Quantitative Analysis of Ppt1-Mediated Hsp90 Dephosphorylation PP5/Ppt1 is a conserved phosphatase, which harbors a TPR domain that mediates its interaction with Hsp90. This interaction activates the phosphatase activity of Ppt1 (Kang et al., 2001; Suhre et al., 2006). Deletion of Ppt1 negatively affected the maturation of Hsp90-dependent clients (Wandinger et al., 2006). To analyze the cellular regulation of yeast Hsp90 phosphorylation sites by Ppt1 in vivo, we applied a quantitative proteomics approach based on stable isotope labeling with amino acids in cell culture (SILAC) (Gruhler et al., 2005; Ong et al., 2002). These SILAC screens identified ten phosphorylation sites in Hsp90 (Table S1), which were distributed in the M- and C-terminal domains (Figure 1A) and largely overlapped with phosphosites reported in recent phosphoproteomic studies (Albuquerque et al., 2008; Holt et al., 2009). Phosphorylation in the N-domain was not detected; in particular, we did not identify the recently reported phosphorylation sites T22 and Y24 (Mollapour et al., 2010, 2011). Weak ionization and/or low abundance of the corresponding phosphopeptides might have precluded their MS-based identification in our and other proteomics studies (Albuquerque et al., 2008; Holt et al., 2009). A comparison of Hsp90 phosphorylation patterns in wild-type (WT) and ppt1-deletion yeast cells exhibited a highly reproducible increase in phosphorylation at S485 and S604 in Ppt1-deficient yeast cells (Tables 1 and S1 available online). Thus, the

quantitative phosphoproteomics data establishes Ppt1 as a site-specific in vivo Hsp90 phosphatase for S485 and S604. To evaluate whether we detected all major phosphosites of Hsp90 in yeast, α−Hsp90 α−P-Ser Hsp90 was investigated by 2D gel electrophoresis. All ten phosphorylated amino acids were substituted by alanines (All-A Hsp90) and this variant was expressed as the sole Hsp90 source in yeast (Figures 1B and 1C) (Nathan and Lindquist, 1995). All-A Hsp90 supported yeast viability, but its phosphorylation was profoundly diminished compared to WT Hsp90 (Figures 1D and S1A), for which several different phosphoisoforms could be detected (Figures S1A and S1B). These data indicate that the ten phosphorylation sites detected by the MS screen represent the major fraction of Hsp90 phosphorylated in yeast. In our study, except for one threonine residue, only serine residues were identified as phosphoacceptor sites. All sites are strongly conserved or conservatively replaced by negatively charged phosphorylation-mimetic residues in a representative set of ortholog Hsp90 sequences. The only exceptions are S334 and S604, which are substituted predominantly by asparagine and methionine, respectively (Figure S2). Hsp90 Phospho Mutants Reveal Differential Effects on Cell Viability To assess in detail the effects of the individual phosphorylation sites on Hsp90 function, we introduced phosphorylationmimicking Hsp90 variants as the only source of Hsp90 in yeast. We tested glutamate and alanine substitutions at all sites to identify sites whose phosphorylation causes an alteration of the cellular properties of Hsp90. Three variants (S379E, S485E, and S604E) influenced yeast growth, whereas the other mutations had no observable effects. Strikingly, the Ppt1-regulated phosphorylation-mimicking variant S485E was unable to support yeast viability (Figure 2A). The functional importance of

518 Molecular Cell 45, 517–528, February 24, 2012 ª2012 Elsevier Inc.

wt Hsp90 All-A

l-A

100 80 60 40 20 0

wt Hsp90 All-A

α−Hsp90

Levels (%)

wt Hsp90

Al

sp w

tH

30°C

(A) Schematic domain organization of Hsp90 and structure of full-length yeast Hsp90 in the closed conformation (PDB 2CG9) with the ten identified phosphorylation sites depicted in red. In the two boxes on the right, a segment of Hsp90 is shown in the open (Hsp90 M-C domain crystal structure [PDB 2CGE]) and in the closed conformation (fulllength Hsp90 crystal structure [PDB code: 2CG9]), with the phosphorylation sites marked in red. (B) Growth of WT and All-A Hsp90-expressing yeast cells at 30
C. (C) Immunoblot analysis of Hsp90 levels in yeast cells expressing only WT or All-A Hsp90. (D) Hsp90 protein levels and Hsp90 phosphorylation levels in WT and All-A Hsp90-expressing yeast cells. The proteins were separated on 2D gels and detected by Hsp90- and phosphoserine-specific antibodies. The signal intensity of the spots was quantified using ImageJ and the data are displayed as a percentage of the WT cell.

D

C 90

B

Figure 1. Identification of Yeast Hsp90 Phosphorylation Sites

Molecular Cell Phosphoregulation of the Hsp90 Chaperone Machinery

Table 1. Quantitative Analysis of Yeast Hsp90 Phosphorylation Sites Hsp82 Phosphorylation Site

Exp. 1 Ratio Ppt1 ko/wt Yeast

Exp. 2 Ratio Ppt1 ko/wt Yeast

Exp. 3 Ratio Ppt1 ko/wt Yeast

S282

1.25

1.23

S297

1.50

1.09

S334

1.07

1.77

S379

1.24

1.18

1.03

S450

1.53

1.29

0.98

1.26 ± 0.27

S456

1.13

1.49

0.96

1.20 ± 0.27

S485

2.81

2.95

4.71

3.95 ± 1.27

9.98

6.39 ± 4.50

1.33

Average Ratio Ppt1 ko/wt Yeast

1.27

1.27 ± 0.05

1.26

1.28 ± 0.21 1.42 ± 0.49

5.34

T533

1.15 ± 0.11

ND

S602 S604

Exp. 4 Ratio Ppt1 ko/wt Yeast

ND 1.90

3.16

10.54

MS identification of Hsp90 phosphosites and SILAC-based quantification in WT versus Ppt1-deficient yeast cells was performed as described in Experimental Procedures. In experiments 1 and 3, WT cells were labeled with Arg0/Lys0 and Dppt1 cells were encoded with Arg10/Lys8, whereas a reciprocal SILAC scheme was used in experiments 2 and 4. Ratios shown represent the respective phosphosite ratios normalized for Hsp90 protein abundance. All MS data is provided in Table S1. SDs were determined from at least two individual experiments. ND, not determined.

S485 was already observed in previous unrelated mutagenesis studies, where its substitution with tyrosine had an impact on the cell growth and chaperoning of Hsp90-dependent clients (Fang et al., 1998; Hawle et al., 2006). Thus, to investigate the importance of this site, we substituted S485 with tyrosine. In conjunction with prior work, the tyrosine variant grew slower at physiological and elevated temperatures (Figure 2B). The second Ppt1-regulated phosphorylation-mimicking variant, S604E, conferred a temperature-sensitive growth phenotype. It is noteworthy that S602E located in close proximity to S604 exhibited no visible effects. Finally, yeast cells expressing S379E displayed slower growth and a decreased thermotolerance relative to WT Hsp90-expressing cells. All variations in the in vivo effects were not due to changes in protein expression because equal protein levels could be assured (Figure S1C). Phosphorylation-Mimicking Variants of Hsp90 Sensitize Yeast Cells to Hsp90 Inhibition To understand the biological role of Hsp90 phosphorylation in yeast, the sensitivity of phosphorylation-mimicking Hsp90 variants was probed by the Hsp90 inhibitor radicicol (RD) in vivo. Cells expressing S379E, S485Y, and S604E were hypersensitive to RD (Figure 2C). Interestingly, the sensitivity to the Hsp90specific inhibitor was unaffected in the S602E mutant. The effects seem to be restricted to the phosphorylated forms of Hsp90 because the corresponding alanine substitutions exhibited effects equivalent to WT protein (data not shown). These data identify S379, S485, and S604 as important determinants of Hsp90 sensitivity to RD. Notably, although the effect of the glutamate substitution mimics phosphorylation for S379 and S604, we cannot directly conclude that the effect of the tyrosine variant at position S485 is identical to that of phosphorylation because the glutamate mutant did not support viability. Endogenous Yeast DNA Repair Processes Depend on Hsp90 Phosphorylation Recent studies observed a role of Hsp90 in transcription, DNA repair, and telomere maintenance pathways in yeast (Toogun et al., 2008; Echtenkamp et al., 2011). To investigate the functional relevance of Hsp90 phosphorylation on DNA repair func-

tion, nucleotide excision repair was induced by exposing yeast cells to UV irradiation. This induces formation of pyrimidine dimers, thus activating the nucleotide excision repair pathway. A marked increase in susceptibility to the DNA mutagen was observed for S379E, S485Y, and, to lesser extent, the S604E variant (Figure 2D), whereas no differences were found between WT and alanine variants (data not shown). Hence, modification of Hsp90 at S379, S485, and S604 influences endogenous DNA repair processes. Client Protein Processing Is Affected in Hsp90 Phosphorylation-Mimicking Variants In addition, we examined the impact of phosphorylationmimicking variants on specific Hsp90 client proteins in vivo. To this end, we utilized the well-established Hsp90 model clients firefly luciferase (FFL) (Nathan et al., 1997), glucocorticoid receptor (GR) (Pratt and Toft, 1997), and the oncogenic tyrosine viral sarcoma (v-Src) kinase (Nathan and Lindquist, 1995). The S604E exchange had only minor effects on FFL and v-Src activity and the S602E mutant behaved like WT Hsp90 (Figure 3). However, with the S379E variant, the activity of GR and FFL was reduced to approximately one-half compared to WT Hsp90 and we observed an even more drastic decrease in v-Src activation. The S485Y mutant decreased the cellular activity of all three client proteins tested. In contrast, the respective alanine variants exhibited only minor influence on the maturation of Hsp90-specific clients (data not shown). Thus, it seems reasonable to conclude that phosphorylation sites S485 and S379 are important determinants for modulating the maturation of various substrate proteins, including kinases and steroid receptors, with the caveat that the phosphomimicking glutamate substitution of S485 could not be tested. Phosphorylation at S485 Alters Conformational Changes during ATP Hydrolysis In a further step, the effect of the all single phosphorylationmimicking variants on the ATPase activity of Hsp90 was determined (Table S2). We found that four phosphorylation-mimicking variants (S379E, S485E, S602E, and S604E) exhibited decreased rates of ATP hydrolysis (Figure 4A). Also, tyrosine

Molecular Cell 45, 517–528, February 24, 2012 ª2012 Elsevier Inc. 519

Molecular Cell Phosphoregulation of the Hsp90 Chaperone Machinery

A

- FOA

+ FOA S485A S485E

vector

S485Y

B

wt

25°C

30°C

37°C

30°C

RD 80 µM

RD 120 µM

30°C

UV 40 J/m2

UV 80 J/m2

wt Hsp90 S379E S485Y S602E S604E

C wt Hsp90 S379E S485Y S602E S604E

D wt Hsp90 S379E S485Y S602E S604E Figure 2. Effects of Hsp90 Phosphomutations on Yeast Growth, Resistance to Hsp90 Inhibition, and DNA Repair Processes (A) Viability of Hsp90 mutants at position S485. The S485E mutation is lethal to yeast cells, whereas S485A and S485Y support yeast growth. (B) Yeast cells with both genomic copies of hsp90 deleted but carrying either WT or Hsp90 phospho mutants were tested for temperature sensitivity. Cultures were adjusted to the same cell density, spotted in 5-fold serial dilutions, and incubated at the indicated temperatures. (C) Resistance of Hsp90 phospho mutants to RD treatment. Yeast cells expressing Hsp90 phospho mutants were incubated in media lacking or supplemented with RD at the indicated concentrations overnight at 30
C, spotted on a dropout plates, and incubated at 30
C. (D) DNA repair activity of Hsp90 phospho mutants. Yeast cells expressing Hsp90 phospho mutants spotted on droput plates were exposed to UV light (40 and 80 J/m2) and then incubated at 30
C.

substitution at S485 negatively influenced the ATP hydrolysis rate. Importantly, the two phosphorylation sites modified by Ppt1 were among the mutants with reduced enzymatic activity. Except for position S485, the respective alanine substitutions, which mimicked the dephosphorylated state, behaved like WT Hsp90. Here, the serine-to-alanine substitution was not able to restore WT-like ATPase activity, demonstrating the high sensitivity of this position to manipulation. The apparent KM values for ATP (Table S2) were equivalent for WT and both alanine and glutamate substitutions of all ten identified sites, indicating that binding of the nucleotide is not influenced by the phosphorylation status. The strongest in vivo consequences were observed for site S485. Mutants of S485 showed reduced ATPase activity,

although this site is located in a distant region from the ATPbinding pocket. We anticipated that this phosphorylation site exerts a long-range effect by inducing changes in the conformational dynamics. To analyze this hypothesis, we performed analytical ultracentrifugation (aUC) experiments, where the structural compaction of WT Hsp90 that occurs in response to the nonhydrolysable ATP analog, adenosine 50 -(b,g imidodiphosphate) (AMP-PNP), can be visualized by a shift in the sedimentation coefficient (Hessling et al., 2009). First, we tested the sedimentation profiles of S485E and Y phosphorylation-site mutants in the absence of nucleotide and found that S485 mutants sedimented slower than WT Hsp90 (Table S3). Furthermore, upon addition of AMP-PNP, S485 mutants exhibited a markedly reduced change in s values compared to WT protein.

520 Molecular Cell 45, 517–528, February 24, 2012 ª2012 Elsevier Inc.

Molecular Cell Phosphoregulation of the Hsp90 Chaperone Machinery

A

tH s S3 p90 79 S4 E 8 S6 5 Y 02 E

100 80

w

FFL Activity (%)

Figure 3. In Vivo Chaperone Function of Hsp90 Phosphovariants

C

120

60

α-P-Tyr

40 20

E 04

E S6

02

Y S6

85

E S4

79

sp tH w

S3

90

0

(A and B) Influence of Hsp90 phospho mutants on the activation of the client proteins FFL and GR. Data represent a percentage of the activity observed in WT Hsp90expressing cells. Error bars represent SD. (C) Effect of Hsp90 phospho mutants on the activity of v-Src. v-Src activity was analyzed by western blotting with a phospho-tyrosine antibody (4G10). PGK was used as a control. Yeast cells expressing v-Src under control of galactose promoter were spotted on the glucose and galactose-containing dropout plates.

B

PGK GR Activity (%)

120 GLC

100 80

GAL

60 40 20

E 04 S6

S6

02

E

Y 85

E 79

S4

S3

w

tH

sp

90

0

Thus, AMP-PNP failed to induce conformational rearrangements leading to a fully closed state. Although nucleotide-induced movements were impaired in the S485 variants, it was still unclear why S485 variants sedimented slower even in the absence of the nucleotide (Table S3). Lower s values could result either from changes in molecular weight caused by dissociation of Hsp90 dimers or by a different, more elongated conformation of the open form. To resolve this conundrum, we compared the sedimentation coefficients of S485 variants at different protein concentrations. If the mutations influenced the primary dimerization interface within the C-domains, sample dilution would induce further dissociation and, consequently, result in slower sedimentation. Similarly, applying a higher protein concentration would shift the equilibrium toward formation of the dimeric species, causing an increased sedimentation velocity. In the titration experiment at the concentrations tested, S485 mutants always sedimented with the same s values, indicating that the dimerization properties were not altered (data not shown). To further prove this observation, we performed a crosslinking experiment that confirmed equal ability to form functional dimers by WT and S485 mutants (Figure S3). Knowing that C-terminal contacts are not altered in S485 mutants, we further focused on potential effects on N-terminal dimerization during the ATPase cycle. As a reporter for the N-terminal contacts, we used lidless-Hsp90, a variant that is itself ATPase inactive but, when present in an Hsp90 dimer together with a WT subunit, stimulates the ATPase in the neighboring WT protomer (Figure 4B) (Richter et al., 2006). In our experiments, S485E showed about one-half of the lidless stimulation seen with WT Hsp90 (Figure 4C). S485Y partly restored activation by the lidless variant. These results support the notion that ATP induces an altered pattern of conformational changes in

Hsp90 with phosphorylated S485, leading to reduced N-domain dimerization. Different Hsp90 conformations can be recognized by conformation-specific cochaperones. Thus, we took advantage of p23/Sba1 as a biological sensor for the N-terminal dimerization because it requires this particular conformation for its association. We used surface plasmon resonance (SPR) and immobilized p23/Sba1. In agreement with the conformational studies above, S485 variants abrogated the association with p23/Sba1 in the presence of AMP-PNP (Figure 4D). p23/ Sba1 targets the ATP-bound closed conformation and reduces the ATP hydrolysis rate at a late stage of the cycle. Although the ATPase activity of WT Hsp90 was inhibited to approximately 60% of its basal value by p23/Sba1 (Figure 4E), its ability to inhibit ATPase activity was abrogated with the S485 mutants. This implies that phosphorylation at S485 severely reduces the conformational flexibility of the protein, which precludes the formation of the N-terminal dimers. Phosphorylation at S379 Disturbs Formation of the Active Conformation of Hsp90 We performed similar aUC experiments with the phosphorylation-mimicking S379E variant. These studies revealed that the S379E was not able to adopt the N-terminally dimerized conformation upon addition of AMP-PNP (Figure 5A and Table S3). However, in the absence of nucleotide, S379E sedimented with the same s value as the WT protein. This demonstrates that both proteins adopt comparable open conformations in their apo states. Because the phosphorylation-mimicking mutations at positions S379 and S485 affected conformational movements, we wished to determine whether the underlying mechanisms were similar. S379 is part of the catalytic loop, which is involved in the formation of the active site. Thus, local, rather than long distance, perturbations induced by phosphorylation at this serine residue can be envisioned. In consequence, S379E Hsp90 may be sensitive to different types of nucleotides. To test this, we took advantage of the slowly hydrolyzing ATP analog ATPgS. Interestingly, in contrast to AMP-PNP, ATPgS was able to induce nucleotide-dependent rearrangements in S379E, showing that this variant strongly differentiates between the two nucleotides (Table S3). This was not observed for S485 mutants. These results suggest that S379E is compromised in forming the hydrolysis-competent state. The reduced

Molecular Cell 45, 517–528, February 24, 2012 ª2012 Elsevier Inc. 521

Molecular Cell Phosphoregulation of the Hsp90 Chaperone Machinery

Figure 4. Regulation of the Hsp90 ATPase Activity and Association with Cochaperones

D 500

0.5

400 RU

0.4 0.3

300 200

0.2

100

0.1

E

E

04

E

S6

02 S6

Y

85 S4

w

S4

85

90 tH

79

sp

E

E

04 S6

E

02 S6

Y

S4

85

E

85 S4

90

79

sp

S3

w

tH

E

0

0.0

S3

ATPase activity (min-1)

A

B

12 9 6 3

E 04

E

S6

02

85

S6

Y

20 15 10 5

E

E

04

S6

E

02

S6

85

Y

522 Molecular Cell 45, 517–528, February 24, 2012 ª2012 Elsevier Inc.

S4

85

E

activation potential for lidless-Hsp90 (Figure 4C) is in agreement with this notion. To further test the different response to nucleotides, we monitored the subunit exchange in the Hsp90 dimer by fluorescence resonance energy transfer (FRET) (Hessling et al., 2009). If Hsp90 is in an N-terminally closed state, the exchange will be much slower than in the open state. In agreement with previous studies, we observed a fast decrease in FRET efficiency in the absence of nucleotide upon addition of unlabeled Hsp90 (Figure 5B). Addition of AMP-PNP or ATPgS induced kinetically stable N-terminal dimers, precluding or decelerating the exchange of protomers with unlabeled Hsp90, respectively. Addition of ATPgS to the S379E FRET complex resulted in a chase with unlabeled Hsp90 (Figure 5C). Notably, the kinetics of S379E displacement differed significantly from those recorded for the WT protein. In particular, the slow kinetics observed after addition of ATPgS imply, first, that this nucleotide leads to the formation of N-terminal dimers and, second, that the hydrolysis occurs, albeit slowly. AMP-PNP did not inhibit dissociation of the S379E complex, supporting the view that this analog is not able to induce a kinetically stable closed state in S379E.

The analysis of interaction with p23/ Sba1 confirms the different conformational states induced by the two nucleotides in S379E. In the presence of AMP-PNP, S379E associated only weakly with this cochaperone; however, ATPgS restored WTlike binding properties (Figure 4D and data not shown). In addition, our ATPase studies revealed that p23/Sba1 could suppress weak basal ATPase of S379E, albeit its inhibitory potential was partly compromised (Figure 4E). Based on these results, it is reasonable to assume that phosphorylation at S379 affects the communication between the N- and the M-domains at the stage of forming the hydrolysis-competent state. Hsp90 Phosphorylation in the C-Domain at S602 and S604 Affects Intersubunit Communication In contrast to S379 and S485 phosphosite mutants, S602E and S604E had no influence on the nucleotide-induced N-domain association (Table S3). Further tests for the N-domain contacts revealed that neither interaction with p23/Sba1 nor activation mediated by the lidless-Hsp90 variant was affected by S602E and S604E mutations (Figures 4C and 4D). Also, the cochaperone Hop/Sti1, an inhibitor of the Hsp90 ATPase stabilizing the open conformation of Hsp90, inhibited ATPase activity of the S602 and S604 phosphomimetic mutants similar to the WT protein (Prodromou et al., 1999; Richter et al., 2003) (data not shown). Thus, it remained unclear how these phosphosites,

S4

79

sp tH w

S3

90

E

E

04

02 S6

S6

E

Y

85 S4

85 S4

E 79

S3

sp tH

25

0

0

w

S3

sp tH

Aha1 stimulation factor

15

90

Lidless Hsp90 Stimulation factor

w

F

C

E

0 90

lid

pe

le

ty

ss

ild

20

S4

w

40

E

P

P

P P P

80 60

85

P

P

P P P

P

P

P

P

P P P

P

ATP

100

79

Lid

120

S4

Normalized ATPase (%)

E

(A) ATPase activity of Hsp90 phospho mutants. The ATPase activity was determined as described in detail in Supplemental Information. Error bars represent SD. (B) Model for the ATPase activity of heterodimers consisting of WT Hsp90 and the lidless variant. (C) ATPase activity of heterodimers with lidlessHsp90. The N-terminal association properties of Hsp90 variants (1 mM) were investigated after formation of heterodimers with 10 mM lidlessHsp90 by their influence on the ATPase activity of the WT subunit. Error bars represent SD. (D) Binding of Hsp90 phospho mutants to p23/ Sba1. Interaction of Hsp90 phospho mutants with p23/Sba1 was determined by SPR analysis. AMPPNP concentration was 2 mM and Hsp90 concentration was 1 mM. (E) Effects of cochaperone p23/Sba1 on the ATPase activity of Hsp90 phospho mutants. Hsp90 phosphovariants (2.5 mM) were incubated with 35 mM p23/Sba1. Data are expressed as the percentage of the ATPase activity in the presence of p23/Sba1 versus the ATPase activity without p23/Sba1. Error bars represent SD. (F) Stimulation of the ATPase activity by Aha1. Data represent the ‘‘fold’’ increase of the Hsp90 ATPase activity (1 mM) in the presence of Aha1 (10 mM) versus to activity of the same Hsp90 variant in the absence of this cochaperone. Error bars represent SD.

Molecular Cell Phosphoregulation of the Hsp90 Chaperone Machinery

A

3.0

AMP-PNP

wt/Aha1

dc/dt

w/o

Least square fitted dc/dt

Figure 5. Phosphomicking Mutations Influence Conformational Dynamics

D

wt Hsp90

2.5

S379E/Aha1

2.0

S485Y/Aha1 S485E/Aha1

1.5 1.0

S602E/Aha1

Aha1

S604E/Aha1

0.5 0.0 S379E

0

0

2

4

6

8

10

2

4 6 8 Svedberg (S)

10

12

Svedberg (S)

B

E C

nucleotide

N

wt Hsp90 AMP-PNP wt Hsp90_Aha1 ATPγ S

S602E_Aha1 S602E

w/o nucleotide

ATP

wt Hsp90

12

(A) Conformational changes induced by AMP-PNP were investigated using aUC. (B and C) FRET analysis of the N-terminal dimerization. Donor- and acceptor-labeled WT Hsp90 or S379E were preincubated to enable formation of heterodimers. ATPgS (green), AMP-PNP (yellow), or ATP (red) (2 mM) were added or the heterodimers were preincubated in the absence of nucleotide (black). The chase reaction was initiated by addition of an excess of unlabeled WT Hsp90. (D) Analysis of the interaction of Aha1 with Hsp90 phospho variants. Complex formation between Aha1 and Hsp90 phopho mutants was monitored by fluorescence using aUC in the presence of 1 mM Alexa Fluor488-labeled Aha1 and 3 mM Hsp90 phospho mutants. (E) Subunit exchange of 100 nM Hsp90 heterodimers in the absence or presence of Aha1 upon addition of 2 mM unlabeled WT or S602E Hsp90. Normalized decrease of Hsp90 acceptor fluorescence over time is shown for WT Hsp90 (green), S602E (yellow), and WT Hsp90 in complex with Aha1 (black) and for S602E in complex with Aha1 (red). (F) The apparent half-life of subunit exchange (t1/2) is estimated using Equation 3 (see Supplemental Information).

F

C S379E

28

(Figure 5D). Additionally, the fraction of Hsp90 associated with Aha1 was diminATP ished. Titration experiments of S602 and 4.6 AMP-PNP S604 phosphorylation-mimicking muw/o nucleotide tants demonstrated that, even at high concentrations, the sedimentation coeffi0.26 0.27 cient of the Hsp90-Aha1 complex fraction did not shift to higher s values (data not shown). This can be explained by alterations of the complex shape, which seems to result in a different stimulation of the Hsp90 ATPase. Taken together, the Aha1-binding studies revealed that located within the C-domain, reduced ATPase activity and acti- Hsp90 phosphorylation at S602 and S604 impairs the cochapervation by the cochaperone Aha1, which accelerates rate-limiting one-induced structural rearrangements in Hsp90, which are conformational changes leading to the closed conformation of required for the conformational cycle. Of note, S485E and Y mutants that were weakly susceptible to Hsp90 (Hessling et al., 2009; Koulov et al., 2010; Retzlaff et al., 2010; Harst et al., 2005) (Figure 4F and Table S2). Both phos- Aha1 stimulation also exhibited slightly reduced binding to Aha1 phorylation sites are located in the region responsible for consti- and the complex sedimented slower than the corresponding tutive dimerization. One possible explanation for impaired Hsp90 complex with WT Hsp90 (Table S2 and Figures 4F and 5D). activity could lie in dissociation of the Hsp90 dimers. However, in This observation concurs with the altered conformations of these our studies, the dimerization properties were not affected by the mutants at early stages of the conformational cycle. In contrast mutations (data not shown), suggesting that S602E and S604E to S602/S604/S485, S379E interacted with Aha1 comparable to the WT protein, although Aha1-mediated stimulation was alter interdomain communication processes. To address this issue, we examined the ability of cochaperone significantly reduced in this phosphosite mutant. This confirms Aha1 to bind to Hsp90 mutants. We assessed complex forma- that later steps of the ATPase cycle (e.g., the formation of the tion between Aha1 and Hsp90 variants using aUC (Li et al., active site) are impaired. Knowing that S602E and S604E retain Aha1-binding and that 2011a). For the interaction of Aha1 with S602E and S604E, reduced s values of Hsp90 in complex with Aha1 were observed the activation mechanism requires the interaction of Aha1 in ATPγ S

Molecular Cell 45, 517–528, February 24, 2012 ª2012 Elsevier Inc. 523

Molecular Cell Phosphoregulation of the Hsp90 Chaperone Machinery

a cooperative manner with both the N- and M-domains of Hsp90 (Retzlaff et al., 2009), we assumed that, in particular, the Aha1induced conformational transitions are affected by the phosphorylation in the C-domain. To test this hypothesis, we assessed the effects of Aha1 on the subunit exchange of the Hsp90 dimer. The stability of each of the Aha1-Hsp90 FRET complexes was analyzed by displacement with an excess of unlabeled Hsp90 (Figure 5E). In the absence of Aha1, the halflife for subunit exchange was identical in WT and S602E, demonstrating that the C-terminal dimerization of Hsp90 is not affected by the mutations. In the presence of Aha1, however, the half-life of the Hsp90 subunit exchange reaction for the S602E mutant was substantially decreased compared to WT Hsp90 (Figure 5F). Thus, in the case of the C-domain phosphorylation-site mutants, Aha1 failed to coordinate the domains of the Hsp90 dimers in a way that permitted the conformational cycling as in WT Hsp90. Taken together, our data demonstrate that phosphorylation allows regulation of the Hsp90 chaperone cycle at different points and by different mechanisms. This implies that spatially distant residues modified by phosphorylation contribute to the communication between remote regions in Hsp90. DISCUSSION Hsp90 is a molecular chaperone that modulates the activity of hundreds of clients, many of which are important effector proteins (McClellan et al., 2007). Thus Hsp90 can be seen as a hub for the conformational regulation of proteins in the eukaryotic cell (Whitesell and Lindquist, 2005). Although it has been known for long time that Hsp90 is a phosphoprotein (Lees-Miller and Anderson, 1989), it remained unclear how this affects Hsp90 function globally. Our MS analysis revealed ten sites in yeast Hsp90, which are phosphorylated in the M- and the C-domain in vivo. These sites apparently constitute a major fraction of Hsp90 phosphorylation in yeast under nonstress conditions. To investigate the effects of the identified phosphorylation sites on Hsp90 function, we utilized glutamate and alanine mutations mimicking phosphorylated and nonphosphorylated states, respectively, and discovered that four phosphosites (S379, S485, S602, and S604) had an influence on various aspects of the Hsp90 chaperone machinery. They sensitized cells to higher temperatures and Hsp90 inhibition and affected DNA repair pathways, highlighting the importance of the corresponding phosphorylation events. Furthermore, this indicates that phosphorylation-mediated regulation of Hsp90 chaperone function may also involve nuclear processes that involve Hsp90. Two Hsp90 phosphorylation sites (S485 and S604) are specifically dephosphorylated by the Hsp90 cochaperone Ppt1, confirming its important role in the Hsp90 chaperone cycle. Our assays could not attribute specific functions to the other six identified phosphosites. This finding is consistent with the notion that there are many silent phosphorylation events without functional consequence, except for a potentially enhanced solubility of the protein (Lienhard, 2008). On the other hand, it cannot be excluded that these silent sites may have a particular function that was not addressed in our in vivo and in vitro assays. Mutation of each of the four active phosphosites (S379, S485, S602, and S604) reduced the ATPase activity of Hsp90 by

around 40%–60%, albeit by different mechanisms. The ATPase activity of Hsp90 is an essential determinant of its chaperone function in vivo. Point mutants with completely abrogated or very low ATPase activity are not viable in yeast (Nathan and Lindquist, 1995; Hainzl et al., 2009; Mollapour et al., 2010); however, variations in the ATP hydrolysis rate from 2% to 400% are tolerated (Hainzl et al., 2009). What could be the benefit of decreasing the already low ATP hydrolysis by phosphorylation? It seems reasonable to assume that the dwell time of clients in an Hsp90 complex correlates to the conformational cycle (Koulov et al., 2010). In consequence, longer cycling times observed for phosphorylation-mimicking variants may result in an increased association with clients. We observed a decrease in the activation potential of the phosphorylationsite mutants that exhibited reduced ATPase activity (S379 and S485) for the client proteins examined in this study. Thus prolonged interaction with Hsp90 may not always be beneficial. It cannot be excluded, however, that decelerated cycling could promote maturation for other clients. Recently, the Neckers group analyzed two phosphorylation sites in the Hsp90 N-domain influencing client protein maturation, cochaperone binding and drug sensitivity (Mollapour et al., 2010, 2011). The phosphorylation of the N-domain exerts short-distance effects on the Hsp90 ATPase. In contrast, three of the phosphorylation sites characterized in this study (S485, S602, and S604) affect long-range communication pathways in Hsp90 protein. In addition, short-distance effects for S379 can be envisioned. S379 is part of the catalytic loop of the M-domain and adjacent to the catalytically important residue R380, which is directly involved in ATP hydrolysis (Ali et al., 2006). Thus, phosphorylation may perturb local conformations and the network of stabilizing interactions in the N- and the M-domains that define the hydrolysis-competent state. In particular, a phosphoryl group at this serine residue will likely interact electrostatically with the arginine residue, thereby modulating the positive charge of the arginine and, consequently, disturbing formation of the active site (Woods and Ferre´, 2005; Johnson and Lewis, 2001). The phosphorylation-mimicking mutant S379E supported cell growth but failed to productively chaperone Hsp90 clients expressed in yeast, with v-Src activation being most sensitive to this posttranslational modification. Furthermore, this variant hypersensitized cells to Hsp90 inhibition with RD and affected DNA repair pathways. Experiments with different ATP analogs suggest that phosphorylation of this site affects conformational rearrangements at the step of the formation of the active site (i.e., at a late stage of the functional cycle, after the N-terminal dimerization has taken place and before the association of the N- and the M-domains) (Figure 6). Our analysis also identified S602 and S604 as two active phosphorylation sites in the C-domain of Hsp90. These phosphorylation sites seem to serve as rather specific modulators of Hsp90 function, affecting certain aspects of its chaperone cycle. Both sites are located in a loop that is resolved in the crystal structure of Hsp90 M-C but not in the crystal structure of the closed conformation of full-length yeast Hsp90 (Ali et al., 2006). S602E and S604E were dispensable for the activation of Hsp90 clients. However, specifically, phosphorylation of the Ppt1-regulated residue S604 plays a role under nonpermissive conditions,

524 Molecular Cell 45, 517–528, February 24, 2012 ª2012 Elsevier Inc.

Molecular Cell

P P

P P P

P P

P

P

P

P

P P P

ADP, Pi

P P P

P P P

P P P

P

P

P P

P P P

P P P

P

P

P

P

P

P

P

P

P

P

P

P

P

P

P P

P P P

P

P P P

ATP

P

P

P

P P P

ADP, Pi

P

P

P

P

P

P

P

P

S379

P P

P

P P

P

P P P

P P

P

ADP, Pi

P

P P

P

P

P

P P

ADP, Pi

lid R380

P

P

P

P

P P P

ATP

S485

P P P

P

P

P

Ppt1

P

P P P

P P P

P P P P P P

P P P

ATP

P

P

P

P

P

P P P

P

P P P

S604 P

ATP

S602 P

P

P

P P P

P P P

Ppt1

Phosphoregulation of the Hsp90 Chaperone Machinery

Figure 6. Mechanism of the Hsp90 Phospho-Regulation The four phosphorylation sites acting on the conformational dynamics of Hsp90 employ different mechanisms to regulate Hsp90 function. Phosphorylation at S379, located in a catalytic loop involved in ATP hydrolysis, affects conformational rearrangements during the ATPase cycle by interfering with the formation of the active site. Mutation of S485 located in a contact region between the C-terminal and the M-domain modulates conformational flexibility that is essential in the communication within dimeric Hsp90. Modification of S602 and S604 in the C-terminal domain influences the intersubunit communication required for efficient ATP hydrolysis.

Molecular Cell 45, 517–528, February 24, 2012 ª2012 Elsevier Inc. 525

Molecular Cell Phosphoregulation of the Hsp90 Chaperone Machinery

including elevated temperatures or RD treatment. In addition, a moderately reduced association with Aha1 was observed, although the S602 and S604 are neither located directly in the interaction interface with Aha1, nor do they influence the constitutive dimerization properties of the C-domains (data not shown). Phosphorylation at S602 and S604 impairs the conformational crosstalk within domains of Hsp90, specifically the intersubunit communication that is a prerequisite for efficient ATP hydrolysis. The most profound effects on the Hsp90-chaperone machinery were observed for the Ppt1-regulated site S485. Because its phosphorylation-mimicking variant was not viable, it was impossible to query the consequences of this modification on chaperone activity in vivo. S485E seemingly disrupts the essential function of Hsp90 in yeast. S485 was previously identified by mutational analysis as a residue affecting various aspects of Hsp90 function (Fang et al., 1998; Hawle et al., 2006), not knowing that it is phosphorylated in vivo. The previously described viable mutant S485Y failed in the study presented here to productively chaperone Hsp90 clients, sensitized cells to the Hsp90 inhibitor RD and influenced DNA repair processes, implying general effects on client processing in yeast. Due to inability of the glutamate mutant at position S485 to support yeast viability, it should be noted that the effects observed for the tyrosine substitution do not necessarily directly reflect the impact of phosphorylation. The S485E and S485Y mutants showed in vitro compromised association with p23/ Sba1 but binding of Aha1 was only moderately affected. However, the ability of Aha1 to stimulate the intrinsic ATPase activity significantly varied between both S485 mutants. Conformational studies revealed the mechanistic underpinnings of these observations: the S485 mutants are defective in a specific step of the conformational cycle, the formation of the N-terminally closed state required for efficient ATP hydrolysis and maturation of clients. Interestingly, the phosphorylation site S485 is located outside the contact region between the C-domain and the M-domain, in a loop that exhibits significant conformational plasticity, varying between a helical structure, when Hsp90 is in the open form, and an extended loop when Hsp90 adopts the closed conformation (Ali et al., 2006). Thus, S485 is a key switch point for the conformational communication within the Hsp90 protein (Figure 6). It affects Hsp90 ATPase activity, although it resides around 70 A˚ distant from the ATP binding site. The data presented here show that phosphorylation at sites in the M- and the C-domains, specifically, and by different molecular mechanisms modulates conformational rearrangements during the ATPase cycle of Hsp90. The modulation results in altered ATPase activity, association of cochaperones, and activation of client proteins. The finding that two of these sites are regulated by the dedicated cochaperone Ppt1 highlights their importance. It should be noted that all Hsp90 molecules are modified in yeast expressing the phosphomicking variant. This does not reflect the in vivo situation wherein WT cells representing only a fraction of Hsp90 seem to be phosphorylated at a respective given site. It seems plausible that differentially phosphorylated isoforms of Hsp90 are populated in response to various stimuli. Based on our study, several scenarios are possible. When Ppt1 is inac-

tive or absent, Hsp90 may be phosphorylated at all four active sites (S379, S485, S602, and S604). In this case, its chaperone function is attenuated. Since Ppt1 levels are strongly regulated in response to different environmental conditions (Gasch et al., 2000), this will couple Hsp90 activity to the physiologic state of the cell. In a second scenario, when both Ppt1-dependent sites (S485 and S604) are dephosphorylated and S379 and S602 are still phosphorylated, Hsp90 chaperone activity is partly restored. Finally, in the fully dephosphorylated form, Hsp90 reaches maximum activity. Thus, the control of the conformational cycle driven by phosphorylation makes Hsp90 a precisely tunable folding machine. The four active phosphorylation sites (S379, S485, S602, and S604) identified in this study affect distinct steps of the conformational cycle of Hsp90 and, especially, the formation of the closed conformation of Hsp90 (S379 and S485), which is the rate-liming step of the cycle. Because late stages of the chaperone cycle are thought to be important for activation of client proteins (Ratzke et al., 2010; Pratt and Toft, 1997; Li et al., 2011b), changes in the conformational dynamics of Hsp90 can be readily exploited to fine-tune chaperone activity toward client proteins and environmental conditions. Thus, the cell is able to maintain a high pool of Hsp90 as a protective reservoir and still restrict its activity in a specific manner by reversible inactivation. EXPERIMENTAL PROCEDURES Detailed experimental procedures can be found in Supplemental Information. SILAC and MS Analysis To enable quantitative phosphoproteomic studies, a SILAC approach was applied to compare phosphorylation changes in Hsp90 in Dppt1 versus WT yeast cells. Hsp90 was isolated from the cells by immunoprecipitation and digested by LysC and trypsin. Phosphorylated peptides were purified with TiO2 beads and analyzed by liquid chromatography (LC)-MS analysis on LTQ-Orbitrap and LTQ-Orbitrap-Velos instruments (Thermo Fisher Scientific, Schwerte, Germany). Protein Purification All proteins were expressed recombinantly and purified according to standard protocols using NiNTA affinity chromatography (GE Bioscience, Munich, Germany), followed by ResourceQ anion exchange (GE Bioscience, Munich, Germany), a hydroxyapatite mixed-mode ion exchange column and, finally, Superdex200 pre-grade size exclusion chromatography (GE Bioscience, Munich, Germany) (Richter et al., 2001). ATPase Measurements ATPase assays were performed using an ATP regenerating system (Richter et al., 2001). To determine the ATPase activity of Hsp90 phospho mutants, 2 mM of Hsp90 proteins were used in the presence of 0–4 mM ATP at 30
C. The reaction was inhibited by adding 50 mM RD (Sigma, St Louis, USA). To investigate the influence of cofactors, 1 or 2.5 mM Hsp90 was preincubated with different concentrations of Aha1 or Hop/Sti1 and p23/Sba1, respectively, before 2 mM ATP was added. For heterodimer formation, 1 mM Hsp90 and 10 mM lidless-Hsp90 were mixed and preincubated at 30
C for 20 min in standard assay buffer before 2 mM ATP was added (Richter et al., 2006). Surface Plasmon Resonance For all SPR measurements, p23/Sba1 was coupled to the CM5 sensor chip (GE Healthcare, Freiburg, Germany) (Hainzl et al., 2004). p23/Sba1 coupling gave a signal of 750 RU. The AMP-PNP concentration was 2 mM and Hsp90 concentration was 1 mM.

526 Molecular Cell 45, 517–528, February 24, 2012 ª2012 Elsevier Inc.

Molecular Cell Phosphoregulation of the Hsp90 Chaperone Machinery

FRET Measurements To allow the formation of heterodimers, 100 nM donor- and acceptor-labeled Hsp90 were mixed and the emergence of a FRET signal was monitored (Hessling et al., 2009). The signal was chased by the addition of an excess of unlabeled Hsp90 in the presence or absence of 2 mM nucleotide. In the case of conformational changes induced by Aha1, 15 mM of the cochaperone was preequillibrated with Hsp90 (Retzlaff et al., 2010). Analytical Ultracentrifugation To probe Hsp90 conformational changes, experiments were performed with a standard interferometer that was attached to a UV/VIS detection system (Hessling et al., 2009; Li et al., 2011a). Hsp90 concentration was 0.5 mg/ml and 2 mM nucleotide was added as indicated. To monitor cochaperone binding, a fluorescence detection system was used. 1mM fluorescent-labeled Aha1 was used in the presence of different concentrations of Hsp90 phospho mutants. Nuclear Hsp90 Chaperone Activity Assay Exponentially growing yeast cells expressing Hsp90 phospho mutants were adjusted to the same optical density, spotted in 1:5 dilution series into selective agar plates, exposed to UV light (40 and 80 J/m2) at room temperature, and incubated at 30
C, and the growth of yeast cells was compared. Hsp90 Inhibitor Assay Yeast cells expressing Hsp90 phospho mutants were incubated in media supplemented with the indicated concentrations of RD. Cells were adjusted to the same optical density, spotted in 1:5 dilution series into selective agar plates, and incubated at 30
C, and the growth of yeast cells was compared. GR, v-Src, and Firefly Luciferase Activity Assays GR, FFL, and v-Src activities were determined as described elsewhere (Nathan et al., 1997; Nathan and Lindquist, 1995). v-Src expression was under the control of the galactose-inducible promotor. Growth of WT and Hsp90 phosphovariants was examined on glucose- and galactose-containing selective agar plates. SUPPLEMENTAL INFORMATION Supplemental Information includes Supplemental Experimental Procedures, three figures, and three tables and can be found with this article online at doi:10.1016/j.molcel.2011.12.031. ACKNOWLEDGMENTS The authors gratefully acknowledge the support of the TUM’s Thematic Graduate Center/Faculty Graduate Center Chemistry at Technische Universita¨t Mu¨nchen. We thank Davide Roaschio, Witold Jaworek, and Jaroslaw Surkont for excellent practical assistance, Bettina Richter for performing 2D gels, and Felix Oppermann for MS data analysis. We thank Brian Freeman for information on the analysis of DNA repair. The work was supported by grants SFB594 A2 from the Deutsche Forschungsgemeinschaft and by the Fonds der Chemischen Industrie. Received: January 19, 2011 Revised: September 26, 2011 Accepted: December 5, 2011 Published: February 23, 2012

Echtenkamp, F.J., Zelin, E., Oxelmark, E., Woo, J.I., Andrews, B.J., Garabedian, M., and Freeman, B.C. (2011). Global functional map of the p23 molecular chaperone reveals an extensive cellular network. Mol. Cell 43, 229–241. Fang, Y., Fliss, A.E., Rao, J., and Caplan, A.J. (1998). SBA1 encodes a yeast hsp90 cochaperone that is homologous to vertebrate p23 proteins. Mol. Cell. Biol. 18, 3727–3734. Frey, S., Leskovar, A., Reinstein, J., and Buchner, J. (2007). The ATPase cycle of the endoplasmic chaperone Grp94. J. Biol. Chem. 282, 35612–35620. Gasch, A.P., Spellman, P.T., Kao, C.M., Carmel-Harel, O., Eisen, M.B., Storz, G., Botstein, D., and Brown, P.O. (2000). Genomic expression programs in the response of yeast cells to environmental changes. Mol. Biol. Cell 11, 4241– 4257. Geller, R., Vignuzzi, M., Andino, R., and Frydman, J. (2007). Evolutionary constraints on chaperone-mediated folding provide an antiviral approach refractory to development of drug resistance. Genes Dev. 21, 195–205. Graf, C., Stankiewicz, M., Kramer, G., and Mayer, M.P. (2009). Spatially and kinetically resolved changes in the conformational dynamics of the Hsp90 chaperone machine. EMBO J. 28, 602–613. Gruhler, A., Olsen, J.V., Mohammed, S., Mortensen, P., Faergeman, N.J., Mann, M., and Jensen, O.N. (2005). Quantitative phosphoproteomics applied to the yeast pheromone signaling pathway. Mol. Cell. Proteomics 4, 310–327. Hainzl, O., Wegele, H., Richter, K., and Buchner, J. (2004). Cns1 is an activator of the Ssa1 ATPase activity. J. Biol. Chem. 279, 23267–23273. Hainzl, O., Lapina, M.C., Buchner, J., and Richter, K. (2009). The charged linker region is an important regulator of Hsp90 function. J. Biol. Chem. 284, 22559– 22567. Harst, A., Lin, H., and Obermann, W.M. (2005). Aha1 competes with Hop, p50 and p23 for binding to the molecular chaperone Hsp90 and contributes to kinase and hormone receptor activation. Biochem. J. 387, 789–796. Hawle, P., Siepmann, M., Harst, A., Siderius, M., Reusch, H.P., and Obermann, W.M. (2006). The middle domain of Hsp90 acts as a discriminator between different types of client proteins. Mol. Cell. Biol. 26, 8385–8395. Hessling, M., Richter, K., and Buchner, J. (2009). Dissection of the ATPinduced conformational cycle of the molecular chaperone Hsp90. Nat. Struct. Mol. Biol. 16, 287–293. Holt, L.J., Tuch, B.B., Ville´n, J., Johnson, A.D., Gygi, S.P., and Morgan, D.O. (2009). Global analysis of Cdk1 substrate phosphorylation sites provides insights into evolution. Science 325, 1682–1686. Iannotti, A.M., Rabideau, D.A., and Dougherty, J.J. (1988). Characterization of purified avian 90,000-Da heat shock protein. Arch. Biochem. Biophys. 264, 54–60. Johnson, L.N., and Lewis, R.J. (2001). Structural basis for control by phosphorylation. Chem. Rev. 101, 2209–2242. Johnson, J.L., and Toft, D.O. (1995). Binding of p23 and hsp90 during assembly with the progesterone receptor. Mol. Endocrinol. 9, 670–678. Kang, H., Sayner, S.L., Gross, K.L., Russell, L.C., and Chinkers, M. (2001). Identification of amino acids in the tetratricopeptide repeat and C-terminal domains of protein phosphatase 5 involved in autoinhibition and lipid activation. Biochemistry 40, 10485–10490. Kleizen, B., and Braakman, I. (2004). Protein folding and quality control in the endoplasmic reticulum. Curr. Opin. Cell Biol. 16, 343–349.

REFERENCES Albuquerque, C.P., Smolka, M.B., Payne, S.H., Bafna, V., Eng, J., and Zhou, H. (2008). A multidimensional chromatography technology for in-depth phosphoproteome analysis. Mol. Cell. Proteomics 7, 1389–1396. Ali, M.M., Roe, S.M., Vaughan, C.K., Meyer, P., Panaretou, B., Piper, P.W., Prodromou, C., and Pearl, L.H. (2006). Crystal structure of an Hsp90-nucleotide-p23/Sba1 closed chaperone complex. Nature 440, 1013–1017.

Koulov, A.V., Lapointe, P., Lu, B., Razvi, A., Coppinger, J., Dong, M.Q., Matteson, J., Laister, R., Arrowsmith, C., Yates, J.R., III, and Balch, W.E. (2010). Biological and structural basis for Aha1 regulation of Hsp90 ATPase activity in maintaining proteostasis in the human disease cystic fibrosis. Mol. Biol. Cell 21, 871–884. Lees-Miller, S.P., and Anderson, C.W. (1989). The human double-stranded DNA-activated protein kinase phosphorylates the 90-kDa heat-shock protein, hsp90 alpha at two NH2-terminal threonine residues. J. Biol. Chem. 264, 17275–17280.

Molecular Cell 45, 517–528, February 24, 2012 ª2012 Elsevier Inc. 527

Molecular Cell Phosphoregulation of the Hsp90 Chaperone Machinery

Li, J., Richter, K., and Buchner, J. (2011a). Mixed Hsp90-cochaperone complexes are important for the progression of the reaction cycle. Nat. Struct. Mol. Biol. 18, 61–66.

Retzlaff, M., Hagn, F., Mitschke, L., Hessling, M., Gugel, F., Kessler, H., Richter, K., and Buchner, J. (2010). Asymmetric activation of the hsp90 dimer by its cochaperone aha1. Mol. Cell 37, 344–354.

Li, J., Soroka, J., and Buchner, J. (2011b). The Hsp90 chaperone machinery: conformational dynamic and reguation by co-chaperones. Biochim. Biophys. Acta, in press. Published online September 16, 2011. 10.1016/ j.bbamcr.2011.09.003.

Richter, K., Muschler, P., Hainzl, O., and Buchner, J. (2001). Coordinated ATP hydrolysis by the Hsp90 dimer. J. Biol. Chem. 276, 33689–33696.

Lienhard, G.E. (2008). Non-functional phosphorylations? Trends Biochem. Sci. 33, 351–352.

Richter, K., Muschler, P., Hainzl, O., Reinstein, J., and Buchner, J. (2003). Sti1 is a non-competitive inhibitor of the Hsp90 ATPase. Binding prevents the N-terminal dimerization reaction during the atpase cycle. J. Biol. Chem. 278, 10328–10333.

McClellan, A.J., Xia, Y., Deutschbauer, A.M., Davis, R.W., Gerstein, M., and Frydman, J. (2007). Diverse cellular functions of the Hsp90 molecular chaperone uncovered using systems approaches. Cell 131, 121–135.

Richter, K., Walter, S., and Buchner, J. (2004). The Co-chaperone Sba1 connects the ATPase reaction of Hsp90 to the progression of the chaperone cycle. J. Mol. Biol. 342, 1403–1413.

Meyer, P., Prodromou, C., Liao, C., Hu, B., Roe, S.M., Vaughan, C.K., Vlasic, I., Panaretou, B., Piper, P.W., and Pearl, L.H. (2004). Structural basis for recruitment of the ATPase activator Aha1 to the Hsp90 chaperone machinery. EMBO J. 23, 1402–1410.

Richter, K., Moser, S., Hagn, F., Friedrich, R., Hainzl, O., Heller, M., Schlee, S., Kessler, H., Reinstein, J., and Buchner, J. (2006). Intrinsic inhibition of the Hsp90 ATPase activity. J. Biol. Chem. 281, 11301–11311.

Mollapour, M., Tsutsumi, S., Donnelly, A.C., Beebe, K., Tokita, M.J., Lee, M.J., Lee, S., Morra, G., Bourboulia, D., Scroggins, B.T., et al. (2010). Swe1Wee1dependent tyrosine phosphorylation of Hsp90 regulates distinct facets of chaperone function. Mol. Cell 37, 333–343. Mollapour, M., Tsutsumi, S., Truman, A.W., Xu, W., Vaughan, C.K., Beebe, K., Konstantinova, A., Vourganti, S., Panaretou, B., Piper, P.W., et al. (2011). Threonine 22 phosphorylation attenuates Hsp90 interaction with cochaperones and affects its chaperone activity. Mol. Cell 41, 672–681. Nathan, D.F., and Lindquist, S. (1995). Mutational analysis of Hsp90 function: interactions with a steroid receptor and a protein kinase. Mol. Cell. Biol. 15, 3917–3925. Nathan, D.F., Vos, M.H., and Lindquist, S. (1997). In vivo functions of the Saccharomyces cerevisiae Hsp90 chaperone. Proc. Natl. Acad. Sci. USA 94, 12949–12956. Neckers, L. (2007). Heat shock protein 90: the cancer chaperone. J. Biosci. 32, 517–530. Ong, S.E., Blagoev, B., Kratchmarova, I., Kristensen, D.B., Steen, H., Pandey, A., and Mann, M. (2002). Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurate approach to expression proteomics. Mol. Cell. Proteomics 1, 376–386. Pearl, L.H., and Prodromou, C. (2006). Structure and mechanism of the Hsp90 molecular chaperone machinery. Annu. Rev. Biochem. 75, 271–294. Picard, D. (2002). Heat-shock protein 90, a chaperone for folding and regulation. Cell. Mol. Life Sci. 59, 1640–1648. Picard, D., Khursheed, B., Garabedian, M.J., Fortin, M.G., Lindquist, S., and Yamamoto, K.R. (1990). Reduced levels of hsp90 compromise steroid receptor action in vivo. Nature 348, 166–168. Pratt, W.B., and Toft, D.O. (1997). Steroid receptor interactions with heat shock protein and immunophilin chaperones. Endocr. Rev. 18, 306–360.

Richter, K., Soroka, J., Skalniak, L., Leskovar, A., Hessling, M., Reinstein, J., and Buchner, J. (2008). Conserved conformational changes in the ATPase cycle of human Hsp90. J. Biol. Chem. 283, 17757–17765. Scheufler, C., Brinker, A., Bourenkov, G., Pegoraro, S., Moroder, L., Bartunik, H., Hartl, F.U., and Moarefi, I. (2000). Structure of TPR domain-peptide complexes: critical elements in the assembly of the Hsp70-Hsp90 multichaperone machine. Cell 101, 199–210. Scroggins, B.T., Robzyk, K., Wang, D., Marcu, M.G., Tsutsumi, S., Beebe, K., Cotter, R.J., Felts, S., Toft, D., Karnitz, L., et al. (2007). An acetylation site in the middle domain of Hsp90 regulates chaperone function. Mol. Cell 25, 151–159. Shiau, A.K., Harris, S.F., Southworth, D.R., and Agard, D.A. (2006). Structural Analysis of E. coli hsp90 reveals dramatic nucleotide-dependent conformational rearrangements. Cell 127, 329–340. Suhre, M.H., Wegele, H., and Wandinger, S.K. (2006). Expression, purification and refolding of the phosphatase domain of protein phosphatase 1 (Ppt1) from Saccharomyces cerevisiae. Int. J. Biol. Macromol. 39, 23–28. Toogun, O.A., Dezwaan, D.C., and Freeman, B.C. (2008). The hsp90 molecular chaperone modulates multiple telomerase activities. Mol. Cell. Biol. 28, 457–467. Vaughan, C.K., Mollapour, M., Smith, J.R., Truman, A., Hu, B., Good, V.M., Panaretou, B., Neckers, L., Clarke, P.A., Workman, P., et al. (2008). Hsp90dependent activation of protein kinases is regulated by chaperone-targeted dephosphorylation of Cdc37. Mol. Cell 31, 886–895. Wandinger, S.K., Suhre, M.H., Wegele, H., and Buchner, J. (2006). The phosphatase Ppt1 is a dedicated regulator of the molecular chaperone Hsp90. EMBO J. 25, 367–376. Wandinger, S.K., Richter, K., and Buchner, J. (2008). The Hsp90 chaperone machinery. J. Biol. Chem. 283, 18473–18477. Weikl, T., Muschler, P., Richter, K., Veit, T., Reinstein, J., and Buchner, J. (2000). C-terminal regions of Hsp90 are important for trapping the nucleotide during the ATPase cycle. J. Mol. Biol. 303, 583–592.

Prodromou, C., Siligardi, G., O’Brien, R., Woolfson, D.N., Regan, L., Panaretou, B., Ladbury, J.E., Piper, P.W., and Pearl, L.H. (1999). Regulation of Hsp90 ATPase activity by tetratricopeptide repeat (TPR)-domain co-chaperones. EMBO J. 18, 754–762.

Whitesell, L., and Lindquist, S.L. (2005). HSP90 and the chaperoning of cancer. Nat. Rev. Cancer 5, 761–772.

Ratzke, C., Mickler, M., Hellenkamp, B., Buchner, J., and Hugel, T. (2010). Dynamics of heat shock protein 90 C-terminal dimerization is an important part of its conformational cycle. Proc. Natl. Acad. Sci. USA 107, 16101–16106.

Young, J.C., Moarefi, I., and Hartl, F.U. (2001). Hsp90: a specialized but essential protein-folding tool. J. Cell Biol. 154, 267–273.

Retzlaff, M., Stahl, M., Eberl, H.C., Lagleder, S., Beck, J., Kessler, H., and Buchner, J. (2009). Hsp90 is regulated by a switch point in the C-terminal domain. EMBO Rep. 10, 1147–1153.

Woods, A.S., and Ferre´, S. (2005). Amazing stability of the arginine-phosphate electrostatic interaction. J. Proteome Res. 4, 1397–1402.

Zhao, Y.G., Gilmore, R., Leone, G., Coffey, M.C., Weber, B., and Lee, P.W. (2001). Hsp90 phosphorylation is linked to its chaperoning function. Assembly of the reovirus cell attachment protein. J. Biol. Chem. 276, 32822– 32827.

528 Molecular Cell 45, 517–528, February 24, 2012 ª2012 Elsevier Inc.