Dynamic Tyrosine Phosphorylation Modulates Cycling
of the HSP90-P50CDC37-AHA1 Chaperone Machine
Wanping Xu,1Mehdi Mollapour,1Chrisostomos Prodromou,3Suiquan Wang,1Bradley T. Scroggins,1Zach Palchick,1
Kristin Beebe,1Marco Siderius,4Min-Jung Lee,2Anthony Couvillon,5Jane B. Trepel,2Yoshihiko Miyata,6Robert Matts,7
and Len Neckers1,*
1Urologic Oncology Branch
2Medical Oncology Branch
National Cancer Institute, Bethesda, MD 20892, USA
3Genome Damage and Stability Centre, University of Sussex, Brighton, UK
4Medicinal Chemistry, Faculty of Science, Vrije Universiteit, Amsterdam, The Netherlands
5Cell Biology, Growth and Viability Section, Cell Signaling Technology, Danvers, MA 01923, USA
6Department of Cell and Developmental Biology, Graduate School of Biostudies, Kyoto University, Kyoto, Japan
7Department of Biochemistry and Molecular Biology, Oklahoma State University, Stillwater, OK 74078, USA
Many critical protein kinases rely on the Hsp90 chap-
erone machinery for stability and function. After
initially forming a ternary complex with kinase client
and the cochaperone p50Cdc37, Hsp90 proceeds
through a cycle of conformational changes facilitated
by ATP binding and hydrolysis. Progression through
the chaperone cycle requires release of p50Cdc37
and recruitment of the ATPase activating cochaper-
one AHA1, but the molecular regulation of this
complex process at the cellular level is poorly under-
phorylation events, involving both p50Cdc37and
Hsp90, are minimally sufficient to provide direction-
on Y4 and Y298 disrupts client-p50Cdc37association,
while Hsp90 phosphorylation on Y197 dissociates
p50Cdc37from Hsp90. Hsp90 phosphorylation on
Hsp90 ATPase activity, furthering the chaperoning
process. Finally, at completion of the chaperone
tion of the client and remaining cochaperones.
The molecular chaperone Hsp90 regulates the stability and func-
tion of a host of cellular signaling proteins (‘‘clients’’), many of
which are protein kinases. The Hsp90-dependent chaperone
cycle requires sequential association and dissociation of various
to productively chaperone and release the client (Taipale et al.,
2010). In the absence of cochaperones, Hsp90 conformational
dynamics in solution are stochastically determined and thermally
driven and do not support a productive chaperone cycle (Ratzke
et al., 2012). How cochaperones regulate the directionality of
this complex molecular machine in the cellular milieu remains
and p50Cdc37. Phosphorylation of p50Cdc37on S13, mediated by
casein kinase 2 (CK2), has been proposed to promotethe forma-
tion of stable complexes with various client kinases including
Raf-1, Cdk4, and Src (Miyata and Nishida, 2004; Shao et al.,
2003b). However, dephosphorylation of S13 by the phosphatase
client kinase association with p50Cdc37(Vaughan et al., 2008),
suggesting that additional events likely participate in regulating
disassembly of the p50Cdc37-kinase complex in cells.
After initially mediating association of client kinase and Hsp90,
p50Cdc37must dissociate for the chaperone cycle to proceed,
because it inhibits Hsp90 ATPase activity (Siligardi et al.,
2002). In contrast, association of the cochaperone AHA1 stimu-
lates the inherently low ATPase activity of Hsp90, thus helping
to productively drive the chaperone cycle (Panaretou et al.,
2002). Although p50Cdc37and AHA1 likely do not bind simulta-
neously to Hsp90 (Harst et al., 2005), cellular regulation of their
sequential association and disassociation is poorly understood.
In higher eukaryotes, p50Cdc37and Hsp90 are phosphorylated
on tyrosine residues (Gilmore et al., 1982). However, the func-
tional significance of p50Cdc37tyrosine phosphorylation has re-
mained unexplored, while that of Hsp90 has only recently begun
to be elucidated (Mollapour and Neckers, 2012). Here, we
present data showing that human p50Cdc37tyrosine phosphory-
lation promotes p50Cdc37dissociation from client kinase pro-
teins. Further, phosphorylation of specific tyrosine residues in
Hsp90 is minimally sufficient to modulate association and disso-
ciation of p50Cdc37and AHA1 and thus significantly influences
Hsp90 ATPase activity and the directionality of the Hsp90
p50Cdc37Is Tyrosine Phosphorylated on Y4 and Y298
p50Cdc37isreported to bephosphorylated onY298, althoughthe
434 Molecular Cell 47, 434–443, August 10, 2012 ª2012 Elsevier Inc.
et al., 2005). Further, mutation of the conserved tyrosine residue
Y4 disrupts p50Cdc37interaction with heme-regulated eIF2a
kinase in reticulocyte lysate (Shao et al., 2003a). We examined
p50Cdc37tyrosine phosphorylation status in COS7 cells. Using
anti-phosphotyrosine antibody, we could not detect phosphory-
lation of p50Cdc37immunoprecipitated from untreated cells.
However, p50Cdc37isolated from cells treated briefly with the
cell permeable phosphatase inhibitor bpv(phen) was phosphor-
ylated on tyrosine to high levels (Figure 1A). Phosphorylated
p50Cdc37was resolved into two bands by electrophoresis,
a weak upper band and a more intense lower band. Mutation
of the Y4 residue abolished the strong lower band, while muta-
tion of Y298 abolished the weak upper band. Importantly,
p50Cdc37harboring Y298E phosphomimetic mutation migrated
similarly to p50Cdc37phosphorylated on Y298, suggesting an
influence of negative charge at this location on p50Cdc37electro-
phoretic mobility. Tyrosine phosphorylation was not observed
when both Y4 and Y298 were mutated, confirming that these
are the only p50Cdc37phosphotyrosine residues.
p50Cdc37is also phosphorylated on serine 13 (Shao et al.,
2003b). Neither Y4 nor Y298 mutation affected p50Cdc37S13
phosphorylation (Figure 1B). In contrast, tyrosine phosphoryla-
tion on both Y4 and Y298 was more robust for p50Cdc37-S13A,
suggesting that dephosphorylation of p50Cdc37S13 may be
a prerequisite for optimal tyrosine phosphorylation to occur
p50Cdc37Tyrosine Phosphorylation Is Mediated by Yes
To identify the tyrosine kinase(s) responsible for p50Cdc37phos-
phorylation, we treated 293H cells expressing p50Cdc37mutated
at either Y4 or Y298 with several kinase inhibitors. Phosphoryla-
tion of both p50Cdc37-Y4F (indicating Y298 phosphorylation) and
p50Cdc37-Y298F (indicating Y4 phosphorylation) was dramati-
cally decreased by the kinase inhibitors PP1, PP2, dasatinib,
and SKI-606, all of which inhibit Src family and Abl kinases (Fig-
ure 1D). However, p50Cdc37tyrosine phosphorylation was not
affected by imatinib, which preferentially inhibits Abl. These
results suggested that one or more Src family kinases mediate
We used siRNA silencing to identify the Src family kinase(s)
mediating p50Cdc37tyrosine phosphorylation. Src, Fyn, and
Yes are ubiquitously expressed in different cell types (Thomas
Figure 1. p50Cdc37in Cells Is Tyrosine Phos-
(A) Tyrosine phosphorylation of cellular p50Cdc37
occurs only on amino acids Y4 and Y298. COS7
cells were transfected with indicated FLAG-tag-
ged p50Cdc37constructs. Twenty-four hours after
transfection, cells were treated with 100 mM
bpv(phen) for 30 min (to inhibit tyrosine phospha-
tases), lysed with 1% SDS buffer, and boiled for
5 min. p50Cdc37proteins were precipitated with
anti-FLAG antibody, and tyrosine phosphoryla-
tion was detected by blotting with antibody
4G10. Protein loading was monitored with anti-
FLAG antibody. Note that p50Cdc37-Y298E and
p50Cdc37-phosphoY298 have a similar mobility.
(B) p50Cdc37tyrosine phosphorylation status does
not affect p50Cdc37serine phosphorylation. Indi-
cated p50Cdc37proteins were expressed in COS7
cells and immunoprecipitated as described above
and S13 phosphorylation was detected.
(C) p50Cdc37serine phosphorylation status affects
p50Cdc37proteins were expressed in COS7 cells
and assessed as described above. Phosphoryla-
tion of Y4 and Y298 was assessed using site-
specific anti-phosphotyrosine antibodies (see
Figure S1 for antibody validation).
(D) p50Cdc37tyrosine phosphorylation is reduced
by Src family kinase inhibition. 293H cells were
transfected with indicated p50Cdc37plasmids.
Twenty-four hours after transfection, cells were
pretreated with indicated kinase inhibitors for 2 hr
and then with bpv(phen) for 20 min. Sample prep-
aration and analysis were as described above.
(E) Yes kinase is responsible for p50Cdc37tyrosine
phosphorylation. 293H cells were cotransfected
with siRNA specific for the indicated Src family
kinases and p50Cdc37-Y4F or p50Cdc37-Y298F.
Two days after transfection, cells were lysed and
samples were prepared and analyzed as above
(see also Figure S1).
Tyrosine Phosphorylation Modulates Hsp90 Function
Molecular Cell 47, 434–443, August 10, 2012 ª2012 Elsevier Inc. 435
and Brugge, 1997) and are thus likely candidates for this activity.
While silencing of Src and Fyn were without effect, silencing of
Yes resulted in dramatically reduced p50Cdc37tyrosine phos-
phorylation (Figure 1D). Y4 and Y298 phosphorylation were
equally reduced upon Yes knockdown, confirming Yes as the
predominant tyrosine kinase responsible for phosphorylating
p50Cdc37Chaperone Function Is Compromised by
Mutation of Tyrosine Phosphorylation Sites
p50Cdc37is required for optimal kinase activity in eukaryotes. To
investigate the physiologic implications of p50Cdc37tyrosine
phosphorylation, we examined whether mutation of Y4 or Y298
affects its chaperone function. First, we coexpressed wild-type
or mutant p50Cdc37together with the FLAG-tagged catalytic
domain of Raf-1 kinase in COS7 cells, and we assessed Raf-1
kinase activity by monitoring phosphorylation of the Raf-1
residue S338 as an indicator of Raf-1 activation. Importantly,
stable expression of exogenous Raf-1 catalytic domain in
COS7 cells requires excess p50Cdc37and is thus only detectable
when exogenous p50Cdc37is also present (Figure S1A). We
found that, when coexpressed with the p50Cdc37phosphomi-
metic mutant Y4E or the p50Cdc37nonphosphomimetic mutants
that seen with coexpressed wild-type p50Cdc37(Figure 2A).
However, when coexpressed with the phosphomimetic mutant
p50Cdc37-Y298E, Raf-1 catalytic domain activity was signifi-
cantly reduced (Figure 2A), indicating Y298 phosphorylation
status (but not Y4 phosphorylation status) is a key determinant
regulating p50Cdc37-dependent Raf-1 activation. Importantly,
upon Yes knockdown, Raf-1 catalytic domain activity was also
significantly impaired, even in the presence of excess p50Cdc37
Unlike Y298, Y4 is conserved in yeast Cdc37 (Y5 in yeast
Cdc37 is equivalent to Y4 in human p50Cdc37). Thus, the func-
tional implications of Y4 phosphorylation were investigated by
examining the ability of Cdc37 Y5 mutants to productively
tyrosine phosphorylation of cellular proteins in yeast expressing
wild-type Cdc37 or Cdc37-Y5F, but markedly decreased tyro-
sine phosphorylation was seen in yeast expressing Cdc37-Y5E
(Figure 2B). v-Src protein level was also dramatically reduced
in these cells. These effects cannot be ascribed to reduced
expression of Cdc37-Y5E, as both mutant and wild-type
proteins were expressed at comparable levels. Thedata indicate
that both protein expression and kinase activity of v-Src were
compromised in the presence of the phosphomimetic mutant
Y5 mutation did not compromise the ability of Cdc37 to stabi-
lize the active form of the Raf-1 homolog Ste11, which is Cdc37-
dependent (Abbas-Terki et al., 2000) (DN-Ste11, Figure 2C).
These data are in agreement with our earlier observation that
p50Cdc37Y4 mutation had no effect on the kinase activity of
exogenously expressed Raf-1 catalytic fragment in COS7 cells.
Figure 2. Mutation of Y298 and Y4 Affects
(A) Phosphomimetic mutation of Y298, but not of
Y4, compromises p50Cdc37-dependent activation
of Raf-1. COS7 cells were cotransfected with
indicated p50Cdc37plasmids and a plasmid ex-
pressing FLAG-tagged Raf-1 catalytic domain.
Raf-1 catalytic domain activity was assessed by
detecting S338 phosphorylation. The signal ob-
tained with phospho-S338 antibody was normal-
ized to the corresponding Raf-1 catalytic fragment
protein expression level. Please see Experimental
Procedures for quantitation method.
(B) Y5E-mutated yeast Cdc37 cannot chaperone
v-Src (Y5 in yeast Cdc37 is equivalent to Y4
in human p50Cdc37). Appreciable levels of v-Src
and tyrosine phosphorylated yeast proteins are
detectable in both wild-type Cdc37-expressing
yeast and in yeast expressing Cdc37-Y5F, but not
in cells expressing Cdc37-Y5E.
(C) Y5 mutation in yeast Cdc37 permits normal
chaperoning of the Raf-1 homolog Ste11DN in
yeast. The expression of Ste11DN was induced by
galactose in yeast expressing wild-type Cdc37,
Cdc37-Y5F, or Cdc37-Y5E. Cdc37 and Ste11DN
protein levels were visualized by western blotting.
(D) Client kinase binding to p50Cdc37is uniformly
affected by the phosphorylation status of Y298.
COS7 cells were transfected with wild-type ErbB2
and indicated p50Cdc37plasmids. Association of
the exogenously expressed kinase ErbB2, as well
as endogenous kinases Raf-1 and Cdk4, with
p50Cdc37proteins, was assessed as described
(see also Figure S2).
Tyrosine Phosphorylation Modulates Hsp90 Function
436 Molecular Cell 47, 434–443, August 10, 2012 ª2012 Elsevier Inc.
Further, they emphasize that the chaperone activity of Cdc37-
Y5E is not generally compromised, and they suggest that phos-
phorylation of Y4 and Y298 may differentially affect the ability of
p50Cdc37to chaperone individual kinases.
To explore this possibility further, we examined the impact of
these mutations on p50Cdc37association with several client
kinases in mammalian cells. We found that neither Y4 nor Y298
mutation affected p50Cdc37association with Hsp90 in COS7
cells (Figure 2D). However, association of client kinases was
substantially affected. The Y298F mutant associated with client
kinases with efficiency similar to wild-type p50Cdc37, while the
Y298E mutant lost association with all kinases examined,
including ErbB2, Cdk4, and Raf-1 (Figure 2D). The impact of
Y4 mutation was more variable. While Y4 mutation did not affect
p50Cdc37association with Cdk4 or Raf-1, interaction with ErbB2
was abolished (Figure 2D).
mutants in mammalian cells was compromised. Using v-Src
transformed NIH 3T3 cells, we indeed observed a negative
impact of Y4 mutation on p50Cdc37association with v-Src (Fig-
ure S2). In contrast, but in agreement with the data obtained in
COS7 cells, Y4 mutation did not affect p50Cdc37association
with endogenous Raf-1 in these cells. Further, p50Cdc37Y298F
mutation affected neither v-Src nor Raf-1 association, while
p50Cdc37Y298E mutation strongly reduced interaction with
both kinases. Taken together, these data suggest that phospho-
mimetic mutation of Y298 in the C-terminal domain uniformly
disrupts p50Cdc37association with client kinases, while mutation
of the p50Cdc37N-terminal domain residue Y4 has a more limited
impact, in this study affecting only ErbB2 and v-Src interaction.
Of note, the association profile of mutant p50Cdc37correlates
with the aforementioned impact of these proteins on kinase
Tyrosine Phosphorylation of p50Cdc37Disrupts
the Kinase-Chaperone Complex
Since phosphomimetic mutation of Y4 and Y298 dissociated
p50Cdc37and kinase proteins, the p50Cdc37that is in complex
with kinase clients is predicted to be unphosphorylated or hypo-
phosphorylated compared to the total cellular pool of p50Cdc37.
To test this hypothesis, we used an antibody raised specifically
to phosphorylated Y4 to examine the Y4 phosphorylation status
of p50Cdc37that was coprecipitated with ErbB2 (Figure 3A). No
phosphorylation on Y4 was detected in FLAG-p50Cdc37isolated
from cells not treated with the phosphatase inhibitor bpv(phen),
which is consistent with our earlier observations. Brief treatment
with bpv(phen) prior to lysis significantly increased Y4 phosphor-
ylation of the total p50Cdc37pool precipitated with anti-FLAG
antibody, but not in p50Cdc37that was coimmunoprecipitated
with antibody to ErbB2.
Next, we examined whether in vitro Y4 phosphorylation of
p50Cdc37would disrupt a preformed complex with ErbB2. Since
ErbB2 itself is a tyrosine kinase, we used a kinase-deficient
mutant (ErbB2-K753A) to eliminate any potential interference
in vitro from intrinsic ErbB2 kinase activity. We immunoprecipi-
tated the p50Cdc37-ErbB2 complex with an ErbB2-specific anti-
body, and we incubated the immune complex with purified Yes
kinase. Yes significantly increased the release of p50Cdc37from
ErbB2 immune complexes into the soluble phase of the kinase
assay (Figure 3B, anti-FLAG western blot and graphical display).
We confirmed that the p50Cdc37released from ErbB2 immune
complexes was phosphorylated on Y4. Importantly, Y4 phos-
phorylation was not evident in the p50Cdc37fraction remaining
in ErbB2 immune complexes.
Mutation experiments suggested that p50Cdc37phosphoryla-
tion on Y298 would dissociate it from the client kinase Raf-1
(Figures 2D and S2). Using a second phospho-specific antibody,
we compared Y298 phosphorylation of p50Cdc37in complex with
Raf-1 to that of the total p50Cdc37fraction. Similar to Y4 phos-
phorylation, brief treatment with bpv(phen) prior to lysis
increased Y298 phosphorylation of the total p50Cdc37pool, but
failed to increase Y298 phosphorylation in the p50Cdc37fraction
in complex with Raf-1 (Figure 3C, arrow identifies pY298, lower
band is nonspecific). Taken together, these data indicate that
p50Cdc37is hypophosphorylated on both Y4 and Y298 when
associated with client kinases.
Finally, we confirmed in vitro the disruptive effects of Y298
phosphorylation on association of p50Cdc37with client kinases.
We expressed FLAG-tagged wild-type or Y4F mutant p50Cdc37
in COS7 cells and we immunoprecipitated p50Cdc37with anti-
FLAG antibody. Endogenous Raf-1 and Cdk4 proteins were
coprecipitated. These complexes were incubated with Yes as
above. Release of Raf-1 and Cdk4 into the soluble fraction was
detected by western blot (also displayed graphically as a ratio
with kinase protein remaining on beads). We found that release
of both kinases from immune complexes of either wild-type or
Y4F p50Cdc37was substantially increased in the presence of
Yes, while bead-bound p50Cdc37was phosphorylated in vitro
on Y4 and Y298 (Figure 3D, arrow identifies pY298, lower band
is nonspecific). Substituting p50Cdc37-Y298F for wild-type
p50Cdc37, eliminating ATP from the assay buffer, or including
the Src family kinase inhibitor dasatinib in the assay buffer all re-
sulted in comparable and significant reduction of Yes-mediated
release of both Raf-1 and Cdk4 from p50Cdc37immune
complexes (Figure S3). These data confirm that p50Cdc37phos-
phorylation on Y298 disrupts association with client kinases.
Hsp90 Tyrosine Phosphorylation Disrupts the Hsp90-
p50Cdc37Complex and Promotes AHA1 Recruitment
p50Cdc37tyrosine phosphorylation disrupts its association with
kinase clients but does not affect its association with Hsp90
(Figures 2D and S2). Since p50Cdc37S13 phosphorylation status
likewise has minimal impact on association with Hsp90 (Miyata
and Nishida, 2004), we used a series of Hsp90 point mutants
and several site-specific Hsp90 phospho-antibodies to query
the importance of Hsp90 tyrosine phosphorylation in these
events. First, we compared the phosphorylation status of 3
previously identified Hsp90 phosphotyrosine residues (www.
phosphosite.org) in the (endogenous) Hsp90 fraction, either
remaining in p50Cdc37immune complexes or released into the
supernatant fraction following in vitro incubation of the immune
complexes with active Yes kinase (Figure 4A). In the absence
on either Y197, Y313, or Y627, and very little Hsp90 was sponta-
neously released from p50Cdc37immune complexes during the
course of the experiment. In the presence of Yes, however,
Tyrosine Phosphorylation Modulates Hsp90 Function
Molecular Cell 47, 434–443, August 10, 2012 ª2012 Elsevier Inc. 437
moreHsp90wasreleased intothesupernatant, andtheHsp90in
this fraction was strongly phosphorylated on these residues.
These results raised the possibility that Hsp90 phosphorylation
on one or more of these tyrosine residues may promote dissoci-
ation of p50Cdc37.
To explore this possibility, we made point mutations of each
residue. We found that the phosphomimetic mutation Y197E
dramatically diminished Hsp90 association with p50Cdc37(Fig-
ure 4B). These results, together with the data from the in vitro
kinase assay, suggest that Y197 phosphorylation promotes
dissociation of Hsp90 from p50Cdc37. A corollary of this hypoth-
esis suggests that constitutive phosphorylation of Y197 would
block initiation of the kinase-chaperone cycle by interfering
with p50Cdc37-Hsp90 association. To investigate this hypothesis
in an intact cell system, we studied Y197 phosphorylation and
Hsp90-p50Cdc37association in resting and activated human
peripheral blood mononuclear cells (PBMCs). Resting PBMCs
(G0 phase) activated by phobol myristate acetate and ionomycin
enter into G1 and proceed to DNA synthesis and cell division,
a process that depends on Hsp90 function (Schnaider et al.,
2000). We immunoprecipitated Hsp90 from resting PBMCs
and found that it was strongly phosphorylated on Y197 and con-
tained little coprecipitated p50Cdc37or (client kinase) Cdk4 (Fig-
no longer detectably phosphorylated on Y197 while significantly
morep50Cdc37andCdk4were coprecipitatedwith Hsp90.These
data support the possibility that Hsp90 Y197 phosphorylation
status may regulate both initiation and progression of the
After client loading, recruitment of the ATPase stimulatory co-
chaperone AHA1 is necessary to efficiently drive the chaperone
cycle to completion (Taipale et al., 2010). The phosphomimetic
mutation Y313E significantly increased Hsp90 association
with AHA1 (Figure 4C), suggesting that Y313 phosphoryla-
tion serves to recruit AHA1 to the Hsp90 complex. Consistent
with this hypothesis, the affinity of purified AHA1 protein for
purified Hsp90-Y313E was increased 3.5-fold compared to its
affinity for wild-type Hsp90 (3.7 ± 0.5 mM versus 13 ± 2.2 mM).
p50Cdc37In Vitro Disrupts the Kinase-Chap-
(A) ErbB2-associated p50Cdc37
were immunoprecipitated from transfected COS7
cells treated with or w/o bpv(phen) for 30 min.
p50Cdc37Y4 phosphorylation was detected with
(B) Yes kinase-mediated tyrosine phosphorylation
of p50Cdc37promotes its dissociation from ErbB2.
Kinase-deficient ErbB2-K753 was coexpressed
with FLAG-p50Cdc37in COS7 cells. ErbB2 was
immunoprecipitated with antibody-bound beads.
ErbB2 beads were extensively washed and sub-
jected to in vitro phosphorylation using purified
Yes kinase. Supernatant was collected, beads
were washed again, and remaining bound pro-
teins were eluted with SDS sample buffer.
p50Cdc37protein was detected in both superna-
tant and ErbB2 bead fractions by western blotting.
The ratio of p50Cdc37in the supernatant to that
remaining associated with ErbB2 beads, in the
presence and absence of Yes kinase, was quan-
tified and graphed. Yes-mediated phosphoryla-
tion of p50Cdc37-Y4 was visualized using anti-pY4
antibody. Y4 phosphorylation was only observed
on the fraction of p50Cdc37released into the
supernatant and was not detected on the fraction
of p50Cdc37that remained associated with ErbB2
(C) Raf-1-associated p50Cdc37is hypophosphory-
lated on Y298. FLAG-tagged Raf-1 or FLAG-
tagged p50Cdc37were immunoprecipitated from
transfected COS7 cells treated with or without
bpv(phen) for 30 min. p50Cdc37phosphorylation
on Y298 was detected with phospho-specific
3. Tyrosine Phosphorylationof
(D) Yes kinase promotes dissociation of Raf-1 and Cdk4 from p50Cdc37. FLAG-p50Cdc37was expressed in COS7 cells and immunoprecipitated as above. In vitro
phosphorylation was performed with Yes kinase. Raf-1 and Cdk4 proteins released into the supernatant or retained on the anti-FLAG (p50Cdc37) beads were
detected by western blotting. Phosphorylation of p50Cdc37on Y4 and Y298 was detected with specific antibodies. The ratios of Raf-1 and Cdk4 released into the
supernatant and remaining on the anti-FLAG (p50Cdc37) beads in the presence and absence of Yes kinase were quantified and graphed (see also Figure S3).
Please see Experimental Procedures for quantitation method.
Tyrosine Phosphorylation Modulates Hsp90 Function
438 Molecular Cell 47, 434–443, August 10, 2012 ª2012 Elsevier Inc.
Consistent with these findings, purified AHA1 more potently
stimulated the ATPase activity of Hsp90-Y313E compared to
wild-type Hsp90 (Figure 4D). Finally, we compared the Y313
phosphorylation level of the total cellular pool of Hsp90
(immunoprecipitated with the pan Hsp90-specific antibody
AC88) to that of the Hsp90 fraction found in FLAG-AHA1
immune complexes (Figure 4E). Hsp90 in complex with AHA1
was phosphorylated on Y313 to a level nearly 7 times that of
the total cellular Hsp90 fraction. Since the cellular concentration
of AHA1 in mammalian cells is less than 1% that of Hsp90
(Koulov et al., 2010), a regulatable mechanism to efficiently
promote association of these two proteins would be advanta-
geous. Our data are consistent with the hypothesis that
Hsp90 Y313 phosphorylation provides a cellular mechanism
to stimulate formation of Hsp90-AHA1 ATPase-active chap-
For the chaperone cycle to reach completion, the client and
any remaining cochaperones must dissociate from Hsp90. We
explored whetherHsp90 tyrosine
contribute to this process. We noted that Hsp90 Y627 was
strongly phosphorylated by Yes in vitro, but only in the fraction
of Hsp90 not associated with p50Cdc37(Figure 4A). This residue
has been suggested to be part of a flexible loop in the Hsp90
C-domain that forms a hydrophobic motif mediating client
protein binding (Fang et al., 2006; Sgobba et al., 2008; Shiau
et al., 2006). We determined whether Y627 phosphorylation
favored release of client and cochaperones. We found that inter-
action of AHA1, the client Cdk4, and PP5 with Hsp90-Y627E
Figure 4. Hsp90 Tyrosine Phosphorylation
Regulates the Association and Dissociation
(A) FLAG-tagged p50Cdc37was expressed in
COS7 cells and immunoprecipitated. In vitro
kinase assay with Yes kinase was as described.
Endogenous Hsp90 phosphorylation on tyrosine
residues was detected with site-specific anti-
bodies (see Figure S4 for antibody validation). ‘‘s’’
indicates Hsp90 released into the kinase assay
buffer; ‘‘b’’ indicates Hsp90 remaining bound to
(B) Phosphomimetic mutation of Hsp90-Y197
was expressed in 293A cells together with indi-
cated FLAG-tagged Hsp90 proteins, and p50Cdc37
was immunoprecipitated with anti-HA beads.
(C) Phosphomimetic mutation of Hsp90-Y313
promotes AHA1 association. FLAG-tagged Hsp90
was precipitated from transfected COS7 cells.
Association of AHA1 was detected by western
(D) AHA1 protein stimulates ATPase activity of
Hsp90-Y313E to a greater extent than that of wild-
type Hsp90. The ATPase activity of purified Hsp90
proteins (20 mM) was determined in the presence
or absence of AHA1 (5 mM).
(E) AHA1-associated Hsp90
phorylated on Y313. FLAG-tagged AHA1 was
expressed in 293A cells and immunoprecipitated
with anti-FLAG antibody. Associated Hsp90 was
examined for phosphorylation on Y313 using
antibody recognizing pY313, and the signal was
compared with Y313 phosphorylation of total
Hsp90 immunoprecipitated from 293A cells. The
pY313 signal intensity was normalized to the
total Hsp90 signal intensity and graphically dis-
(F) Phosphomimetic mutation of Hsp90-Y627
from Hsp90. Indicated
proteins were transiently expressed in 293A cells and immunoprecipitated with anti-FLAG. Association of endogenous AHA1 and Cdk4 was examined by
(G) Phosphomimetic mutation of Hsp90-Y627 disrupts PP5 association. Indicated FLAG-tagged Hsp90 proteins were expressed in 293A cells and immuno-
precipitated with anti-FLAG beads. Association of the cochaperone phosphatase PP5 was visualized with specific antibody. Loading equivalence was
confirmed by Coomassie stain of immunoprecipitated Hsp90 proteins (see also Figure S4). Please see Experimental Procedures for quantitation method.
Tyrosine Phosphorylation Modulates Hsp90 Function
Molecular Cell 47, 434–443, August 10, 2012 ª2012 Elsevier Inc. 439
was markedly reduced compared to wild-type Hsp90 (Figures
4F and 4G).
Can These Tyrosine Phosphorylation Events Be Placed
in an Ordered Sequence?
The Hsp90 chaperone cycle requires an ordered sequence of
transient association and dissociation of both client proteins
and cochaperones. After determining the impact on this process
of distinct p50Cdc37and Hsp90 phosphorylation events, we
investigated whether these phosphorylations can occur in an
appropriate sequential order. This possibility was initially sug-
gested when we noticed markedly reduced Y313 phosphoryla-
tion of the Hsp90-Y627E mutant (Figure S4B), consistent with
the lack of AHA1 binding to Hsp90-Y627E. We examined
whether certain cochaperones can coexist with Hsp90 in the
same complex. We found that p50Cdc37and AHA1 were mutually
excluded from complexes containing the other cochaperone
ure 5D). Since p50Cdc37forms the initial complex with client
kinases and Hsp90 prior to AHA1 binding and activation of
Hsp90 ATPase activity, these data support a model in which
p50Cdc37dissociates from Hsp90 prior to association of AHA1
occur preferentially after Y197 phosphorylation, based on the
impact of thesephosphorylation events on cochaperone binding
(see Figure 4). In fact, Y313 phosphorylation intensity of Hsp90-
Y197E is 2-fold that of wild-type Hsp90 (Figure S4C).
Both p50Cdc37and AHA1, as well as Hsp90 and client
protein, were present in PP5 immune complexes and PP5 was
present in both AHA1 and p50Cdc37immune complexes (Figures
5A–5D), indicating that PP5 can remain associated with Hsp90
throughout the chaperone cycle, until it dissociates upon
Hsp90 Y627 phosphorylation (Figure 4G). The generally disrup-
tive effects of Y627 phosphorylation on Hsp90 association with
Figure 5. AHA1 and p50Cdc37Bind to Hsp90 Simul-
taneously with PP5, but They Each Exist in Sepa-
rate Complexes with Hsp90
(A) p50Cdc37coexists with PP5 but not with AHA1 in Hsp90
complexes. FLAG-tagged AHA1 and PP5 were expressed
in 293A cells and immunoprecipitated with anti-FLAG
beads. Association of p50Cdc37was examined with anti-
(B)AHA1 coexists withPP5butnot withp50Cdc37inHsp90
complexes. FLAG-tagged p50Cdc37and PP5 were ex-
pressed in 293A cells and immunoprecipitated with anti-
FLAG beads. Association of AHA1 was examined with
(C) PP5 coexists with AHA1 and p50Cdc37on Hsp90.
cells and immunoprecipitated with anti-FLAG beads.
Association of PP5 was examined with anti-PP5 antibody.
(D) Association of Hsp90 and Raf-1 with p50Cdc37, AHA1
and PP5. FLAG-tagged p50Cdc37, PP5, and AHA1 were
expressed in 293A cells and immunoprecipitated with
anti-FLAG beads. Association of Hsp90 and Raf-1 were
examined with specific antibodies.
(E) AHA1-Hsp90-client kinase complexes are enriched
with Hsp90 mutants that cannot hydrolyze ATP. After
silencing endogenous Hsp90 (a and b) with siRNA, we
expressed either HA-tagged wild-type Hsp90a or the
ATPase-deficient mutant Hsp90a-E47A and FLAG-tag-
ged AHA1. Transfected Hsp90a is not affected by the
siRNA used to knock down endogenous Hsp90 because it
is targeted to a 30-UTR in Hsp90 mRNA and the Hsp90a
cDNA used for transfection lacks the 30-UTR. After anti-
FLAG immunoprecipitation, we blotted for endogenous
Raf-1 and Hsp90-pY313. AHA1-associated Hsp90 was
detected by Coomassie staining. Total Hsp90a (trans-
835 (Enzo Life Sciences), while endogenous Hsp90a was
detected with SPS-771 (Enzo Life Sciences) which
recognizes an epitope in the Hsp90 N terminus and shows
substantially reduced recognition of N-terminally tagged
Hsp90 (data not shown).
(F) Summary of the model proposing the sequential tyro-
sine phosphorylation of p50Cdc37and Hsp90.
Tyrosine Phosphorylation Modulates Hsp90 Function
440 Molecular Cell 47, 434–443, August 10, 2012 ª2012 Elsevier Inc.
cochaperones and client are consistent with the hypothesis that
this modification occurs at completion of the chaperone cycle.
Abundance of the client kinase Raf-1 in p50Cdc37immune
complexes is likely a reflection of the fact that this cochaperone
binds to Raf-1 by itself and in complex with Hsp90 (Figure 5D).
Raf-1 was present to a lower extent in AHA1 immune complexes
consistent with the fact that AHA1 stimulates Hsp90 ATPase
activity and rapidly drives the chaperone cycle to its completion,
with concomitant maturation and dissociation of the client
protein (Figure 5D). Indeed, Raf-1 (and Hsp90) association with
the AHA1-containing chaperone complex was noticeably
increased when endogenous Hsp90 was knocked down and re-
placed by exogenously delivered (siRNA-insensitive, see figure
legend) Hsp90 harboring a point mutation making it incapable
of hydrolyzing ATP (E47A mutation, Figure 5E). In this case,
Raf-1 is predicted to remain trapped in AHA1-bound chaperone
complexes that are unable to proceed to completion. Taken
together, these data support a sequential ordering of p50Cdc37
and Hsp90 tyrosine phosphorylation events that are minimally
By recruiting kinase clients to the Hsp90 chaperone machine,
p50Cdc37plays an important role in the maturation of numerous
protein kinases. Productive chaperoning of kinase clients
requires the regulated association and dissociation of p50Cdc37
from both the client kinase and Hsp90. Association of client
kinase has been suggested to require p50Cdc37S13 phosphory-
lation, mediated by CK2 (Miyata and Nishida, 2004; Vaughan
et al., 2008). S13 phosphorylation is also detected on uncom-
plexed p50Cdc37, and its dephosphorylation did not disrupt
preassembled chaperone-kinase complexes (Vaughan et al.,
2008), suggesting that additional regulatory events participate
in disassembly of p50Cdc37-kinase and p50Cdc37-Hsp90 com-
plexes in metazoans. Our current data support a model in which
p50Cdc37S13 phosphorylation and tyrosine phosphorylation
together modulate this process. In this model, p50Cdc37phos-
phorylation on S13 favors formation of an initial ternary complex
with kinase client and Hsp90. Binding of PP5 to this complex
promotes p50Cdc37S13 dephosphorylation and favors tyrosine
phosphorylation, which triggers release of kinase client from
p50Cdc37(Figure 6). It should be noted that p50Cdc37is not tyro-
sine phosphorylated in yeast (data not shown). Thus, this post-
translational modification is likely an adaptation of metazoans
that permits additional fine tuning of p50Cdc37activity in order
to cope with an expanded kinase repertoire.
Human p50Cdc37consists of three domains, with the con-
served N-terminal domain involved in interaction with kinase
client, and the middle domain involved in interaction with
Hsp90. No function has been ascribed to the C-terminal domain
(Shao et al., 2003a). In reticulocyte lysate, p50Cdc37consisting of
only N-terminal and middle domains formed a stable complex
with Hsp90 and the client heme-regulated eIF2a kinase (Shao
et al., 2003a). However, we found that transfected p50Cdc37
protein lacking the C-terminal domain failed to form a stable
complex with client kinases, even though truncated protein con-
taining the middle domain bound Hsp90 as well as full-length
C-terminal domain is necessary to stably associate with client
kinases in intact cells. These discrepant findings are not likely
due differences in p50Cdc37S13 phosphorylation, since compa-
rable levels of S13 phosphorylation are detected in vivo and
in vitro (Figure S5B). However, p50Cdc37tyrosine phosphoryla-
tion is not seen in reticulocyte lysate (Figure S5B).
The dynamic nature of p50Cdc37tyrosine phosphorylation
is consistent with a regulatory function for this posttransla-
tional modification. Using site-specific phospho-antibodies, we
characterized two tyrosine phosphorylation sites in human
p50Cdc37, one in the N-terminal domain (Y4) and the other in
the C-terminal domain (Y298). Y4 phosphorylation only affected
p50Cdc37interaction with the two relatively unstable kinases
examined in this study (ErbB2 and v-Src), and the equivalent
Y5 mutation in yeast Cdc37 affected productive chaperoning
of v-Src in yeast (but not that of the Raf-1 homolog Ste11). Since
(Figure S5A). This indicates that the p50Cdc37
Figure 6. A Series of Phosphorylation Events Drive the Hsp90 and
p50Cdc37-Mediated Chaperone Cycle for Client Kinases
Hsp90 and p50Cdc37, which is phosphorylated on S13 by CK2, bind to a client
kinase (‘‘k’’) (A). PP5 dephosphorylates p50Cdc37on S13 (B), priming the
complex for access by Yes kinase (C). Yes phosphorylates p50Cdc37on Y298,
weakening the interaction between p50Cdc37and the client kinase (some client
kinases are also released upon Y4 phosphorylation). Phosphorylation of
Hsp90 on Y197 promotes dissociation of p50Cdc37from the chaperone
complex, while Hsp90 phosphorylation on Y313 induces conformational
change favoring recruitment of AHA1 (D). AHA1 stimulates Hsp90 ATPase
activity, driving maturation of the client kinase. Finally, Hsp90 phosphorylation
on Y627 leads to dissociation of AHA1, PP5, and client kinase (E). The Hsp90
tyrosine phosphorylation events shown can be mediated by Yes (as shown by
our in vitro experiments), but it is not the sole kinase capable of doing so, since
Yes knockdown fails to completely abolish phosphorylation of these Hsp90
residues (data not shown). Tyrosine phosphorylated Hsp90 and p50Cdc37are
dephosphorylated by as yet unidentified phosphatase(s) allowing them to re-
enter the chaperone cycle (see also Figure S5).
Tyrosine Phosphorylation Modulates Hsp90 Function
Molecular Cell 47, 434–443, August 10, 2012 ª2012 Elsevier Inc. 441
Y4F (as well as Y4A and W7A, data not shown) did not support
p50Cdc37complex formation with either v-Src or ErbB2 as effi-
ciently as did wild-type, it is likely that this N-terminal motif
participates in chaperoning a subset of unstable kinase clients.
In contrast, Y298 phosphorylation uniformly disrupted p50Cdc37
association with all four kinases that we evaluated (ErbB2,
v-Src, Cdk4, and Raf-1), suggesting a general role for C-terminal
domain tyrosine phosphorylation in regulating p50Cdc37function.
Although p50Cdc37tyrosine phosphorylation promotes disso-
ciation of client kinases, this modification does not perturb
association with Hsp90. In order to effectively chaperone its
clients, Hsp90 must proceed through a series of conformational
changes coupled to ATP hydrolysis. This requires the carefully
orchestrated association/dissociation of a number of cochaper-
ones. Since p50Cdc37binding to Hsp90 restrains these confor-
mational changes at an early point in the chaperone cycle and
inhibits Hsp90 ATPase activity, p50Cdc37must dissociate from
Hsp90 for the chaperone process to continue. By mutating
a number of Hsp90 phosphotyrosine residues that were previ-
ously identified by mass spectrometry (www.phosphosite.org),
rupted association of Hsp90 with p50Cdc37. We also showed that
Hsp90 Y197 phosphorylation status and association with
p50Cdc37/client kinase was inversely related in PBMCs, and we
confirmed that in vitro phosphorylation of this residue dissoci-
ated Hsp90 and p50Cdc37. In the crystal structure of the Hsp90
N-terminal domain complexed with p50Cdc37, Y197 does not
directly interact with p50Cdc37. Consequently, phosphorylation
of this residue likely affects the Hsp90-p50Cdc37complex by
inducing a conformational change in Hsp90.
The cochaperone AHA1 facilitates the Hsp90 conformational
cycle and strongly activates Hsp90 ATPase activity (Panaretou
et al., 2002). We showed that Hsp90 in complex with AHA1
was phosphorylated to high levels on Y313 and that phos-
phomimetic mutation of Y313 resulted in a dramatic increase
in AHA1 association. AHA1 affinity for Hsp90-Y313E was
increased 3.5-fold compared to wild-type, and AHA1 stimulation
of Hsp90-Y313E ATPase activity was significantly greater com-
pared to wild-type Hsp90. We propose that Hsp90 Y313 phos-
phorylation provides an environmentally responsive regulatable
mechanism to recruit AHA1 to Hsp90, thereby stimulating
Hsp90 ATPase activity and driving the chaperone cycle forward.
In the crystal structure of the Hsp90 middle domain complexed
with AHA1, Y313 is not directly involved in the interaction of
the two proteins; rather, it is on the opposite side of the binding
interface (Meyer et al., 2004). Therefore, Y313 phosphorylation
likely promotes AHA1 association indirectly by affecting Hsp90
Recently, Hsp90 in solution was shown to fluctuate stochasti-
cally between thermally determined conformational states, while
ATP bound to both open and closed conformations and was
unable to direct a productive conformational cycle (Ratzke
et al., 2012). These data emphasize the importance of cochaper-
ones for maintaining proper Hsp90 function in cells. Although
additional Hsp90 phosphorylation sites and other Hsp90 phos-
phorylating kinases have been identified, the phosphorylation
events described herein provide a minimally sufficient biochem-
of the Hsp90-kinase chaperone cycle in higher eukaryotes (see
Figure 6), and they emphasize the importance and complexity
of posttranslational mechanisms for optimal function of the
Cells, Plasmids, and Antibodies
COS7 and v-Src-NIH 3T3 cells were purchased from American Type Culture
Collection (Manassas, VA). 293H and 293A cells were purchased from
Invitrogen (Carlsbad, CA). All cells were grown in DMEM plus 10% fetal bovine
(University of Tokyo). Plasmids expressing truncated p50Cdc37proteins have
been described previously (Shao et al., 2003a). Point mutations were made
using the QuikChange site-directed mutagenesis method. To construct
the yeast CDC37 expression plasmid, CDC37 promoter (855 bp upstream
of the start codon) and the terminator (498 bp downstream of the stop
codon) were clonedinto HindIII/XhoI-KpnI/EcoRI
CDC37His6 (tagged at the C-domain) and mutated forms CDC37His6-Y5F
and CDC37His6-Y5E were cloned into the XhoI/KpnI of the above plasmid.
These constructs were transformed into cdc37 deletion strain XX201 contain-
solid media containing 5-fluoroorotic acid (5-FOA), in order to lose the
YCplac33. DN-Ste11 was provided by J. Johnson (University of Idaho).
ErbB2, Raf-1, and Cdk4 antibodies were purchased from Santa Cruz
Biotechnology (Santa Cruz, CA). Fyn and Yes antibodies were from BD
Bioscience (San Jose, CA). Cdc37 antibody was from Thermo Scientific
(Fremont, CA). Anti-phosphotyrosine antibody 4G10 was from Upstate
Biotechnology (Lake Placid, NY). Anti-pS13-Cdc37 antibody was described
previously (Miyata and Nishida, 2007). Site-specific phosphotyrosine anti-
bodies to p50Cdc37and Hsp90 used in this study were developed in collabo-
ration with Cell Signaling Technology (Beverly, MA). These antibodies were
validated for specificity by western blotting of specific phosphosite mutant
proteins (see Figures S1 and S4).
Immunoprecipitation and Western Blotting
Cells were transfected with Lipofectamine 2000. Twenty-four hours after
transfection (48 hr for siRNA), cells were washed with PBS and lysed. For
coimmunoprecipitation, cells were lysed with HEPES buffer containing
10 mM Na2MoO4, 30 mM NaF, 2 mM b-glycerol phosphate, 2 mM sodium
vanadate, 100 mM bpv(phen), and Complete protease inhibitors (Roche
Applied Science, Indianapolis, IN). For p50Cdc37tyrosine phosphorylation,
cells were lysed with 1% SDS buffer, boiled for 5 min, and then diluted to
0.1% SDS. M2 anti-FLAG antibody-linked beads were used to immunoprecip-
itate p50Cdc37proteins. Proteins were resolved by SDS-PAGE and transferred
onto PVDF membrane. Membranes were probed with indicated antibodies.
For quantitation, films with appropriate exposure were scanned by densitom-
etry, band densities were obtained using NIH Image software, and standard
deviations were calculated in Excel. Where graphical quantification of band
densities is shown, experiments were repeated twice.
In Vitro Kinase Assay
ErbB2-K753A was expressed in COS7 cells and immunoprecipitated with
mouse anti-ErbB2 antibodies (Ab-3 and Ab-5, EMD Biosciences, San Diego,
CA) prebound to protein G-agarose beads (Invitrogen, Carlsbad, CA). The
beads were washed three times with lysis buffer (20 mM HEPES [pH7.2],
100 mM NaCl, 1 mM MgCl2, 0.1% Nonidet P40, 10 mM Na2MoO4, Complete
protease inhibitors and PhosSTOP phosphatase inhibitors [Roche Applied
Science, Indianapolis, IN]), and then once with MOPS buffer (8 mM MOPS
[pH7.0], 100 mM NaCl, 0.2 mM EDTA, 10 mM Na2MoO4). The beads were
resuspended in 50 ml buffer containing 10 ng purified Yes kinase protein
(Millipore, Billerica, MA) according to the manufacturer’s protocol, and incu-
bated at 30?C for 1 hr with stirring. After incubation, beads were spun
down, and supernatant was transferred to a new tube and mixed with 5x
SDS-sample buffer. Beads were washed twice with lysis buffer, resuspended
in SDS-sample buffer, and heated at 100?C for 5 min to elute remaining
Tyrosine Phosphorylation Modulates Hsp90 Function
442 Molecular Cell 47, 434–443, August 10, 2012 ª2012 Elsevier Inc.
proteins. Eluate and supernatant were examined by western blot as
Isolation of Peripheral Blood Mononuclear Cells and Lymphocyte
Buffy coats were provided anonymously as a byproduct of whole blood dona-
tionsfrom paid healthyvolunteer donors through aNational Institutes of Health
Institutional Review Board–approved protocol, and they were processed to
obtain the mononuclear cell fraction. Mononuclear cells were pelleted and
frozen immediately to obtain the resting cell population or activated by incu-
bating overnight at 37?C in medium containing phorbol myristate acetate
(PMA) (1 ng/ml) and ionomycin (1 mg/ml).
v-Src Activity in Yeast
XX201 yeast strain-expressing Cdc37-His6 or various cdc37 mutant (Y5F or
Y5E) alleles were transformed with YpRS316v-SRC (Nathan and Lindquist,
1995). v-SRC is under control of the GAL1 promoter. Its activity was analyzed
as described previously (Panaretou et al., 2002), with the exception that yeast
cells were grown on 2% glucose in order to repress the GAL1 promoter. v-Src
protein levels were detected with EC10 mouse antibody (Millipore, Billerica,
MA), and v-Src activity was assessed by blotting yeast lysates with 4G10
mouse anti-phosphotyrosine antibody (Millipore). Cdc37-His6 was detected
with Tetra-His monoclonal antibody (QIAGEN, Valencia, CA).
Supplemental Information includes five figures, Supplemental Experimental
Procedures, and Supplemental References and can be found with this article
online at doi:10.1016/j.molcel.2012.05.015.
We are grateful to Susan Leitman and the Department of Transfusion Medi-
cine, Clinical Center, National Institutes of Health, for their help in obtaining
and processing peripheral blood leukocytes from healthy donors.
Received: February 2, 2012
Revised: April 13, 2012
Accepted: May 10, 2012
Published online: June 21, 2012
Abbas-Terki, T., Donze ´, O., and Picard, D. (2000). The molecular chaperone
Cdc37is required for Ste11 function and pheromone-inducedcell cycle arrest.
FEBS Lett. 467, 111–116.
Fang, L., Ricketson, D., Getubig, L., and Darimont, B. (2006). Unliganded and
hormone-bound glucocorticoid receptors interact with distinct hydrophobic
sites in the Hsp90 C-terminal domain. Proc. Natl. Acad. Sci. USA 103,
Gilmore, T.D., Radke, K., and Martin, G.S. (1982). Tyrosine phosphorylation of
a 50K cellular polypeptide associated with the Rous sarcoma virus transform-
ing protein pp60src. Mol. Cell. Biol. 2, 199–206.
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.
Koulov, A.V., Lapointe, P., Lu, B., Razvi, A., Coppinger, J., Dong, M.Q.,
Matteson, J., Laister, R., Arrowsmith, C., Yates, J.R., 3rd, 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.
Meyer, P., Prodromou, C., Liao, C., Hu, B., Mark Roe, S., 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, 511–519.
Miyata, Y., and Nishida, E. (2004). CK2 controls multiple protein kinases by
phosphorylating a kinase-targeting molecular chaperone, Cdc37. Mol. Cell.
Biol. 24, 4065–4074.
Miyata, Y., and Nishida, E. (2007). Analysis of the CK2-dependent phosphor-
ylation of serine 13 in Cdc37 using a phospho-specific antibody and phospho-
affinity gel electrophoresis. FEBS J. 274, 5690–5703.
Mollapour, M., and Neckers, L. (2012). Post-translational modifications of
Hsp90 and their contributions to chaperone regulation. Biochim. Biophys.
Acta 1823, 648–655. Published online August 10, 2011. 10.1016/j.bbamcr.
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,
Panaretou, B., Siligardi, G., Meyer, P., Maloney, A., Sullivan, J.K., Singh, S.,
Millson, S.H., Clarke, P.A., Naaby-Hansen, S., Stein, R., et al. (2002).
Activation of the ATPase activity of hsp90 by the stress-regulated cochaper-
one aha1. Mol. Cell 10, 1307–1318.
Ratzke, C., Berkemeier, F., and Hugel, T. (2012). Heat shock protein 90’s
mechanochemical cycle is dominated by thermal fluctuations. Proc. Natl.
Acad. Sci. USA 109, 161–166. Published online December 19, 2011. 10.
Rush, J., Moritz, A., Lee, K.A., Guo, A., Goss, V.L., Spek, E.J., Zhang, H., Zha,
X.M., Polakiewicz, R.D., and Comb, M.J. (2005). Immunoaffinity profiling of
tyrosine phosphorylation in cancer cells. Nat. Biotechnol. 23, 94–101.
Schnaider, T., Somogyi, J., Csermely, P., and Szamel, M. (2000). The Hsp90-
specific inhibitor geldanamycin selectively disrupts kinase-mediated signaling
events of T-lymphocyte activation. Cell Stress Chaperones 5, 52–61.
Sgobba, M., Degliesposti, G., Ferrari, A.M., and Rastelli, G. (2008). Structural
models and binding site prediction of the C-terminal domain of human Hsp90:
a new target for anticancer drugs. Chem. Biol. Drug Des. 71, 420–433.
Shao, J., Irwin, A., Hartson, S.D., and Matts, R.L. (2003a). Functional dissec-
tion of cdc37: characterization of domain structure and amino acid residues
critical for protein kinase binding. Biochemistry 42, 12577–12588.
Shao, J., Prince, T., Hartson, S.D., and Matts, R.L. (2003b). Phosphorylation of
serine 13 is required for the proper function of the Hsp90 co-chaperone,
Cdc37. J. Biol. Chem. 278, 38117–38120.
Shiau, A.K., Harris, S.F., Southworth, D.R., and Agard, D.A. (2006). Structural
Analysis of E. coli hsp90 reveals dramatic nucleotide-dependent conforma-
tional rearrangements. Cell 127, 329–340.
Siligardi, G., Panaretou, B., Meyer, P., Singh, S., Woolfson, D.N., Piper, P.W.,
Pearl, L.H., and Prodromou, C. (2002). Regulation of Hsp90 ATPase activity by
the co-chaperone Cdc37p/p50cdc37. J. Biol. Chem. 277, 20151–20159.
Taipale, M., Jarosz, D.F., and Lindquist, S. (2010). HSP90 at the hub of protein
homeostasis: emerging mechanistic insights. Nat. Rev. Mol. Cell Biol. 11,
Thomas, S.M., and Brugge, J.S. (1997). Cellular functions regulated by Src
family kinases. Annu. Rev. Cell Dev. Biol. 13, 513–609.
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). Hsp90-
dependent activation of protein kinases is regulated by chaperone-targeted
dephosphorylation of Cdc37. Mol. Cell 31, 886–895.
Tyrosine Phosphorylation Modulates Hsp90 Function
Molecular Cell 47, 434–443, August 10, 2012 ª2012 Elsevier Inc. 443