..............................................................
A high-affinity conformation
of Hsp90 confers tumour
selectivity on Hsp90 inhibitors
Adeela Kamal, Lia Thao, John Sensintaffar, Lin Zhang, Marcus F. Boehm,
Lawrence C. Fritz & Francis J. Burrows
Conforma Therapeutics Corporation, 9393 Towne Centre Drive, Suite 240,
San Diego, California 92121, USA
.............................................................................................................................................................................
Heat shock protein 90 (Hsp90) is a molecular chaperone that
plays a key role in the conformational maturation of oncogenic
signalling proteins, including HER-2/ErbB2, Akt, Raf-1, Bcr-Abl
and mutated p531–7
. Hsp90 inhibitors bind to Hsp90, and induce
the proteasomal degradation of Hsp90 client proteins6,8–11
.
Although Hsp90 is highly expressed in most cells, Hsp90 inhibitors
selectively kill cancer cells compared to normal cells, and
the Hsp90 inhibitor 17-allylaminogeldanamycin (17-AAG) is
currently in phase I clinical trials12,13
. However, the molecular
basis of the tumour selectivity of Hsp90 inhibitors is unknown.
Here we report that Hsp90 derived from tumour cells has a
100-fold higher binding affinity for 17-AAG than does Hsp90
from normal cells. Tumour Hsp90 is present entirely in multichaperone
complexes with high ATPase activity, whereas Hsp90
from normal tissues is in a latent, uncomplexed state. In vitro
reconstitution of chaperone complexes with Hsp90 resulted in
increased binding affinity to 17-AAG, and increased ATPase
activity. These results suggest that tumour cells contain Hsp90
complexes in an activated, high-affinity conformation that facilitates
malignant progression, and that may represent a unique
target for cancer therapeutics.
The naturally occurring ansamycin antibiotic geldanamycin
(GM) binds to a conserved binding pocket in the amino-terminal
domain of Hsp9014,15
, inhibiting ATP binding and ATP-dependent
Hsp90 chaperone activity16–18
. The GM derivative 17-AAG exerts
anti-tumour activity in preclinical models19
and is currently in
clinical trials12,20
. However, two paradoxical observations have
confounded understanding of the preferential effects of Hsp90
inhibitors on tumour cells. First, it is unclear why 17-AAG binds
purified Hsp90 protein with micromolar affinity in vitro21
but has
nanomolar activity in tissue culture1
. Second, although Hsp90 is
abundantly expressed in most cells, ansamycin drugs selectively
accumulate in human tumour xenografts in vivo12,22
(our unpublished
observations). It has been speculated that the physicochemical
properties of 17-AAG may contribute to cellular
accumulation23
, but the mechanism is unknown12
. Previous work
showing that the binding affinity of GM to Hsp90 from cell lysates1
was greater than that for the isolated protein21
suggested to us that
ansamycin binding to Hsp90 could be influenced by additional
intracellular factors. Furthermore, if the unidentified factors were
present to greater extent in tumour cells that normal cells, differential
binding affinity might also explain the tumour selectivity of
Hsp90 inhibitors.
To investigate if Hsp90 from tumour cells had a higher binding
affinity to 17-AAG than that from normal cells, we performed
competitive binding assays using a biotinylated GM (biotin-GM)
probe (Fig. 1a). Incubation of Hsp90 from BT474 breast carcinoma
cells with increasing concentrations of 17-AAG inhibited the binding
of Hsp90 to biotin-GM with an IC50 of 6 nM, whereas Hsp90
Figure 1 17-AAG has a higher binding affinity to Hsp90 from tumour cells than normal
cells or purified Hsp90 protein. a, b, 17-AAG (a) or ATP (b) has a higher binding affinity
to Hsp90 from BT474 tumour cells than normal cells (NDF and RPTEC) and purified
Hsp90. c, Apparent binding affinity of 17-AAG to Hsp90 from a panel of tumour and
normal cells. Individual values of IC50 (circles) and the mean ^ standard error of the
mean (line with error bars) is shown. d, Correlation of the binding affinity of 17-AAG to
Hsp90 to the cytotoxicity in different cells.
letters to nature
NATURE | VOL 425 | 25 SEPTEMBER 2003 | www.nature.com/nature 407© 2003 NaturePublishing Group
from primary human renal epithelial cells (RPTEC) and normal
dermal fibroblasts (NDF) had an IC50 of 600 nM and 400 nM,
respectively. Purified Hsp90 protein binding was inhibited with
an IC50 of 600 nM, as previously reported21
. The apparent binding
affinity of ATP, the physiological ligand of Hsp90, was also higher in
the tumour cell lines (IC50 ¼ 100 mM in BT474) than the normal
cells (IC50 ¼ 1,000 mM in NDF and 2,900 mM in RPTEC) (Fig. 1b),
whereas purified Hsp90 had an IC50 of 3,000 mM. In addition, by
testing a large panel of cell lysates (see Methods), we observed that
Hsp90 from HER-2-overexpressing cancer cells (BT474, N87,
SKOV3 and SKBR3) had IC50 values of 5 ^ 1 nM, that from other
tumour cells (MCF7, A549, HT29, MDA468, SKMG3, U87, HT1080
and Hs578t) was 39 ^ 14 nM, whereas Hsp90 from normal human
primary cells (NDF, RPTEC, HMVEC, HMEC, HUVEC, Hs578Bst
and PBMC) had IC50 values of 943 ^ 290 nM (Fig. 1c). Total Hsp90
protein levels did not vary greatly between tumour and normal cells
(Supplementary Fig. 1a). Taken together, these results suggest that
the Hsp90 in tumour cells, particularly those overexpressing the
Hsp90 client protein HER-2, has a significantly higher binding
affinity for 17-AAG than does Hsp90 from normal cells. Furthermore,
the binding affinity of 17-AAG to Hsp90 from the different
cells directly correlates with the cytotoxic/cytostatic activity of
17-AAG in those cells (Fig. 1d).
Hsp90 interacts with many co-chaperone proteins that assemble
into multi-chaperone complexes that promote protein folding24
.
Tumour cells overexpress Hsp90 client proteins, suggesting that a
greater amount of Hsp90 in tumour cells might be engaged in active
chaperoning and present in multi-chaperone complexes that could
modulate the binding affinity of ligands to Hsp90. To assess whether
Hsp90 in tumour cells was present in multi-chaperone complexes,
we determined the levels of the co-chaperones p23 and Hop—
essential components of the two known multi-chaperone Hsp90
complexes13
—associated with Hsp90 from normal and tumour cells
(Fig. 2a). Co-immunoprecipitation in the normal and tumour cell
lysates with antibodies to Hsp90 revealed that more Hsp90 in
tumour cells was present in complexes with p23 and Hop compared
to normal cells (Fig. 2a). Control immunoprecipitations without
antibodies did not immunoprecipitate any Hsp90. Strikingly,
immunoprecipitation with antibodies to both p23 and Hop
revealed that the entire tumour cell pool of Hsp90 was present in
complexes unlike normal cells. Conversely, immunoprecipitation
with an antibody25
specific for the uncomplexed form of Hsp90
pulled down far more Hsp90 from normal cells than from tumour
cells. Neither the total levels of p23 and Hop nor the growth rate of
the tumour cells differed significantly from the normal cells (Supplementary
Fig. 1b). These results suggest that all Hsp90 in tumour
cells is in the form of multi-chaperone complexes that are actively
engaged in chaperoning client oncoproteins, and that the heightened
complex formation is not due to increased expression of
Hsp90 or co-chaperones. As the chaperone function of Hsp90 is
dependent upon its ATPase activity16
, we immunoprecipitated
Hsp90 from cell lysates and performed ATPase assays. Tumour
Hsp90 had markedly higher ATPase activity compared to Hsp90
from normal cells (Fig. 2b) and was inhibited by 17-AAG, radicicol
(another structurally unrelated Hsp90 inhibitor), and by 0.5 M KCl
(which is known to break apart chaperone complexes). Control
immunoprecipitation in the absence of Hsp90 antibody did not
have any significant ATPase activity (data not shown). Combined,
these results suggest that essentially all soluble Hsp90 in tumour
Figure 2 Hsp90 from tumour cells is present in chaperone complexes with high ATPase
activity. a, Normal (RPTEC, NDF and Hs578Bst) and tumour (BT474, MCF7 and Hs578t)
cell lysates were immunoprecipitated with indicated antibodies (IP Ab) and immunoblotted
with indicated western blot (WB) Ab. The bound and unbound fractions are shown. Hsp90*
antibody preferentially recognizes uncomplexed Hsp90. b, ATPase activity of Hsp90 from
normal (RPTEC, NDF and Hs578Bst) and tumour (BT474, MCF7 and Hs578t) cell lysates
that were untreated or treated with 10 mM 17-AAG, 10 mM radicicol and 0.5 M KCl.
Mean ^ standard deviation are shown from triplicate experiments.
Figure 3 In vitro reconstitution of purified Hsp90 with co-chaperones increased binding
affinity to 17-AAG and the ATPase activity. a, Hsp90 reconstituted with the indicated
co-chaperones has a higher apparent binding affinity to 17-AAG. b, Hsp90 reconstituted
with the indicated co-chaperones has increased ATPase activity, which can be inhibited
by 10 mM 17-AAG.
letters to nature
NATURE | VOL 425 | 25 SEPTEMBER 2003 | www.nature.com/nature408 © 2003 NaturePublishing Group
cells is present in fully active multi-chaperone complexes, whereas
Hsp90 in normal cells is in an uncomplexed inactive form.
As our data demonstrated that the Hsp90 from tumour cells had a
higher binding affinity to 17-AAG, and this correlates with presence
of increased multi-chaperone complexes and increased ATPase
activity, we tested whether reconstitution of purified Hsp90 with
co-chaperones would result in increased binding affinity and
ATPase activity. For in vitro reconstitution, we used five proteins
(Hsp90, Hsp70, Hsp40, Hop and p23) that have been shown to be
required for the in vitro chaperoning activity of Hsp9026
. Addition
of all four co-chaperones to purified Hsp90 increased the apparent
affinity of 17-AAG from 600 nM for Hsp90 alone to 12 nM in the
presence of the added proteins, whereas partial complexes did not
(Fig. 3a). Therefore, Hsp90 when reconstituted with co-chaperones
has nanomolar binding activity, in concordance with that observed
in tumour cell lysates (Fig. 1). The ATPase activity of Hsp90 was also
significantly enhanced by all four co-chaperones and was inhibited
by the addition of 17-AAG (Fig. 3b). Hsp70 had some minimal
ATPase activity, but this was not inhibited by 17-AAG. Thus, these
results suggest that Hsp90 present in multi-chaperone complexes
not only had a higher binding affinity for 17-AAG but was also more
biochemically active.
To determine if our observations in vitro also applied to mice and
to clinical cancer, we examined the binding affinity of 17-AAG to
Hsp90 in normal and malignant mouse and human tissue samples
(Fig. 4a, b). The apparent binding affinity of 17-AAG to Hsp90 from
mouse tumours (3T3-src, B16 and CT26) was 8–35 nM compared
to 200–600 nM for the mouse normal tissues (brain, kidney, liver,
lung and heart), even though Hsp90 was more abundantly
expressed in several of the normal tissues (Fig. 4a). For the
human tissues, using four samples of each tissue type, the IC50
was 6,170 ^ 1,060 nM for normal breast versus 29 ^ 4 nM for
breast carcinoma, and 2,725 ^ 736 nM for normal colon versus
43 ^ 4 nM for the colon carcinomas (Fig. 4b). To determine the
fraction of Hsp90 present as multi-chaperone complexes, we
performed co-immunoprecipitations and observed that there
were increased amounts of p23 with Hsp90 from the human
tumours compared to the normal tissues (Fig. 4c). Conversely,
there was significantly more uncomplexed Hsp90 co-immunoprecipitated
with the Hsp90* antibody from the normal tissues
compared to tumour tissues (Fig. 4c). Furthermore, the Hsp90
ATPase activity in both clinical tumour samples was significantly
higher relative to the normal tissues and was inhibited by 17-AAG,
radicicol and 0.5M KCl (Fig. 4d). These results indicate that the
Hsp90 in human clinical tumour tissues is in a high-affinity, multichaperone
complex with increased ATPase activity.
Our data show that Hsp90 in tumour cells exists in a functionally
distinct molecular form, and clarify three fundamental aspects of
the role of Hsp90 in tumour biology. First, the nanomolar binding
of 17-AAG to high-affinity tumour Hsp90 is now consistent with
the nanomolar anti-tumour activity of this family of drugs. Second,
the markedly higher affinity of tumour Hsp90 in vivo explains the
remarkable ability of ansamycins to accumulate progressively at
tumour sites in animals. And third, the complete usage of tumour
Hsp90 may provide a selection pressure leading to further upregulation
of Hsp90 that is observed in many advanced tumours27,28
. We
propose a model of Hsp90-dependent malignant progression in
which, as tumour cells gradually accumulate mutant and overexpressed
signalling proteins, Hsp90 becomes engaged in active
chaperoning and stabilization of oncoproteins, and adopts a highaffinity
form induced by bound co-chaperone proteins. Thus,
Hsp90 may serve an analogous role in the somatic evolution of
tumours as in darwinian evolution29,30
—rescuing potentially misfolded
mutant proteins to prevent the mutation becoming lethal to
Figure 4 Hsp90 from clinical tumour samples is in a high-affinity complex with increased
ATPase activity. a, Hsp90 protein levels and apparent binding affinity of 17-AAG to Hsp90
from mouse tumour and normal tissues. b, Apparent binding affinity of 17-AAG to Hsp90
from human normal and tumour tissues. Mean ^ standard error is shown (N ¼ 4).
c, Human normal and tumour tissue lysates co-immunoprecipitated with Hsp90
antibodies and immunoblotted with indicated WB Ab. d, Hsp90 ATPase activity from
human normal and tumour tissues treated with 10 mM 17-AAG or radicicol and 0.5 M KCl.
Mean ^ standard error is shown (N ¼ 4).
letters to nature
NATURE | VOL 425 | 25 SEPTEMBER 2003 | www.nature.com/nature 409© 2003 NaturePublishing Group
the cell, and thus allowing these oncoproteins to support tumour
cell proliferation, survival and malignancy. Dependence on the
activated, high-affinity chaperone could make Hsp90 an ‘Achilles
heel’ of tumour cells, driving the selective accumulation and
bioactivity of pharmacological Hsp90 inhibitors, and making
tumour Hsp90 a unique cancer target. A
Methods
Cells and reagents
Normal and malignant cell lines were obtained from Clonetics and ATCC unless otherwise
indicated. Tumour cells used included high-HER-2 breast carcinoma (BT474, SKBR3),
stomach carcinoma (N87) and ovarian carcinoma (SKOV3), MCF7 breast carcinoma,
MDA468 breast carcinoma, Hs578T breast carcinoma and paired normal epithelium
Hs578Bst, A549 lung carcinoma, HT29 colon carcinoma, U87 glioblastoma, SKMG3
glioblastoma (a gift from C. Thomas) and HT1080 fibrosarcoma. Normal primary cells
used were renal proximal epithelial cells (RPTEC), normal dermal fibroblasts (NDF),
human microvascular endothelial cells (HMVEC), human mammary epithelial cells
(HMEC), human umbilical vascular endothelial cells (HUVEC) and peripheral blood
mononuclear cells (PBMC). Normal mouse tissue (brain, kidney, liver, lung and heart)
was obtained from nude mice and the mouse tumours were from nude mice injected with
3T3-src (a gift from L. Neckers), B16 and CT26 cells. The clinical samples (benign breast
tissue, metastatic ductal breast carcinoma, normal colonic mucosa and colon
adenocarcinoma) were obtained from Bainbridge Genomic Foundation and BioResearch
Support, Inc. Antibodies used were: Hsp90 (Stressgen, SPA-845; recognizes Hsp90a and
Hsp90b and immunoprecipitates free and complexed Hsp90), Hsp90* (Stressgen, SPA-
830; recognizes Hsp90a and Hsp90b and immunoprecipitates uncomplexed Hsp90), p23
(Alexis Biochemicals, 804-023-R100) and Hop (a gift from D. Toft). 17-AAG was
synthesized from GM, and the biotin-GM probe was prepared by displacing the
17-methoxy of GM with a biotinyl-linked amine.
Hsp90 binding assays
Purified native Hsp90 protein (Stressgen) or cell lysates in lysis buffer (20 mM HEPES, pH
7.3, 1 mM EDTA, 5 mM MgCl2, 100 mM KCl) were incubated with or without 17-AAG for
30 min at 4 8C, and then incubated with biotin-GM linked to BioMag streptavidin
magnetic beads (Qiagen) for 1 h at 4 8C. Tubes were placed on a magnetic rack, and the
unbound supernatant removed. The magnetic beads were washed three times in lysis buffer
and heated for 5 min at 958C in SDS–PAGE sample buffer. Samples were analysed on SDS
protein gels, and western blots done using indicated antibodies. Bands in the western blots
were quantified using the Bio-rad Fluor-S MultiImager, and the percentage inhibition of
binding of Hsp90 to the biotin-GM was calculated. The IC50 reported is the concentration of
17-AAG needed to cause half-maximal inhibition of binding. For in vitro reconstitution,
5 mM of purified Hsp90 (Stressgen) was combined with 1 mM each of Hsp70, Hsp40, p23
and Hop purified proteins (a gift from D. Toft) as previously described26
.
MTS assay
Cells were seeded in 96-well plates at 2,000 cells per well in a final culture volume of 100 ml
for 24 h before the addition of increasing concentrations of 17-AAG that was incubated for
5 days. Viable cell number was determined using the Celltiter 96 AQueous Nonradioactive
Cell Proliferation Assay (Promega). The value of the background absorbance
at 490 nm (A490) of wells not containing cells was subtracted. Percentage of viable
cells ¼ (A490 of 17-AAG treated sample/A490 untreated cells) £ 100. The IC50 was defined
as the concentration that gave rise to 50% viable cell number.
Co-immunoprecipitation
Cell lysates were prepared as described above. Protein-A Sepharose beads (Zymed) were
pre-blocked with 5% BSA. The cell lysates were pre-cleared by incubating with 50 ml of
protein-A Sepharose beads (50% slurry). To 100 ml of the pre-cleared cell lysate, either no
antibody or antibodies to Hsp90, p23 and Hop were added, and incubated by rotating for
1 h at 4 8C. 50 ml of pre-cleared beads (50% slurry) was then added and incubated by
rotating for 1 h at 4 8C. Bound beads were briefly centrifuged at 3,000g and unbound
samples collected. Beads were washed thrice in lysis buffer and once with 50 mM Tris, pH
6.8, and then SDS-sample buffer added for 5 min at 95 8C. Bound and unbound samples
were analysed by SDS–PAGE and western blots using indicated antibodies.
Hsp90 ATPase assay
Hsp90 was co-immunoprecipitated from normal and tumour cell lysates as described
above using an N-terminal Hsp90 antibody. For in vitro reconstitution, 5 mM Hsp90 was
combined with 1 mM each of Hsp70, Hsp40, p23 and Hop purified proteins25
. The ATPase
activity of the immunoprecipitated or in vitro reconstituted Hsp90 was measured by
detection of free inorganic phosphate (Pi) using a PiPer Phosphate Assay kit (Molecular
Probes). The assay measures an increase in fluorescence absorption of an Amplex Red
reagent that is proportional to the amount of Pi in the sample. Standards were done using
known concentrations of Pi (100 mM with six half-log dilutions), and the amount of Pi
released from Hsp90 in the cell lysates or in the in vitro reconstituted proteins was
calculated. For the cell lysates, the amount of Hsp90 immunoprecipitated with the Hsp90
antibody was quantified to normalize the data to mmol Pi per mol Hsp90 per min. All
assays were done in triplicate.
Received 4 June; accepted 11 July 2003; doi:10.1038/nature01913.
1. Schulte, T. W. & Neckers, L. M. The benzoquinone ansamycin 17-allylamino-17demethoxygeldanamycin
binds to HSP90 and shares important biologic activities with geldanamycin.
Cancer Chemother. Pharmacol. 42, 273–279 (1998).
2. Xu, W. et al. Sensitivity of mature Erbb2 to geldanamycin is conferred by its kinase domain and is
mediated by the chaperone protein Hsp90. J. Biol. Chem. 276, 3702–3708 (2001).
3. Sato, S., Fujita, N. & Tsuruo, T. Modulation of Akt kinase activity by binding to Hsp90. Proc. Natl
Acad. Sci. USA 97, 10832–10837 (2000).
4. Basso, A. D., Solit, D. B., Munster, P. N. & Rosen, N. Ansamycin antibiotics inhibit Akt activation and
cyclin D expression in breast cancer cells that overexpress HER2. Oncogene 21, 1159–1166 (2002).
5. Schulte, T. W., Blagosklonny, M. V., Ingui, C. & Neckers, L. Disruption of the Raf-1-Hsp90 molecular
complex results in destabilization of Raf-1 and loss of Raf-1-Ras association. J. Biol. Chem. 270,
24585–24588 (1995).
6. An, W. G., Schulte, T. W. & Neckers, L. M. The heat shock protein 90 antagonist geldanamycin alters
chaperone association with p210bcr-abl and v-src proteins before their degradation by the
proteasome. Cell Growth Differ. 11, 355–360 (2000).
7. Blagosklonny, M. V., Toretsky, J. & Neckers, L. Geldanamycin selectively destabilizes and
conformationally alters mutated p53. Oncogene 11, 933–939 (1995).
8. Whitesell, L., Mimnaugh, E. G., De Costa, B., Myers, C. E. & Neckers, L. M. Inhibition of heat shock
protein HSP90-pp60v-src heteroprotein complex formation by benzoquinone ansamycins: Essential
role for stress proteins in oncogenic transformation. Proc. Natl Acad. Sci. USA 91, 8324–8328 (1994).
9. Miller, P. et al. Depletion of the erbB-2 gene product p185 by benzoquinoid ansamycins. Cancer Res.
54, 2724–2730 (1994).
10. Mimnaugh, E. G., Chavany, C. & Neckers, L. Polyubiquitination and proteasomal degradation of the
p185c-erbB-2 receptor protein-tyrosine kinase induced by geldanamycin. J. Biol. Chem. 271,
22796–22801 (1996).
11. Schulte, T. W., An, W. G. & Neckers, L. M. Geldanamycin-induced destabilization of Raf-1 involves the
proteasome. Biochem. Biophys. Res. Commun. 239, 655–659 (1997).
12. Workman, P. Auditing the pharmacological accounts for hsp90 molecular chaperone inhibitors:
Unfolding the relationship between pharmacokinetics and pharmacodynamics. Mol. Cancer Ther. 2,
131–138 (2003).
13. Isaacs, J. S., Xu, W. & Neckers, L. Heat shock protein 90 as a molecular target for cancer therapeutics.
Cancer Cell 3, 213–217 (2003).
14. Prodromou, C. et al. Identification and structural characterization of the ATP/ADP-binding site in the
Hsp90 molecular chaperone. Cell 90, 65–75 (1997).
15. Stebbins, C. E. et al. Crystal structure of an Hsp90-geldanamycin complex: Targeting of a protein
chaperone by an antitumor agent. Cell 89, 239–250 (1997).
16. Obermann, W. M., Sondermann, H., Russo, A. A., Pavletich, N. P. & Hartl, F. U. In vivo function of
Hsp90 is dependent on ATP binding and ATP hydrolysis. J. Cell Biol. 143, 901–910 (1998).
17. Panaretou, B. et al. ATP binding and hydrolysis are essential to the function of the Hsp90 molecular
chaperone in vivo. EMBO J. 17, 4829–4836 (1998).
18. Grenert, J. P., Johnson, B. D. & Toft, D. O. The importance of ATP binding and hydrolysis by
hsp90 in formation and function of protein heterocomplexes. J. Biol. Chem. 274, 17525–17533
(1999).
19. Solit, D. B. et al. 17-Allylamino-17-demethoxygeldanamycin induces the degradation of androgen
receptor and HER-2/neu and inhibits the growth of prostate cancer xenografts. Clin. Cancer Res. 8,
986–993 (2002).
20. Blagosklonny, M. V. Hsp-90-associated oncoproteins: Multiple targets of geldanamycin and its
analogs. Leukemia 16, 455–462 (2002).
21. Chiosis, G. et al. A small molecule designed to bind to the adenine nucleotide pocket of Hsp90 causes
Her2 degradation and the growth arrest and differentiation of breast cancer cells. Chem. Biol. 8,
289–299 (2001).
22. Egorin, M. J. et al. PC3 human prostate xenograft retention of, and oncoprotein modulation by 17(allylamino)-17-demothoxygeldanamycin(17AAG)
in vivo. Proc. Am. Assoc. Cancer Res. 40, 3409
(1999).
23. Chiosis, G. et al. 17AAG: Low target binding affinity and potent cell activity-finding an explanation.
Mol. Cancer Ther. 2, 123–129 (2003).
24. Richter, K. & Buchner, J. Hsp90: Chaperoning signal transduction. J. Cell. Physiol. 188, 281–290
(2001).
25. Schuh, S. et al. A 90,000-dalton binding protein common to both steroid receptors and the Rous
sarcoma virus transforming protein, pp60v-src. J. Biol. Chem. 260, 14292–14296 (1985).
26. Kosano, H., Stensgard, B., Charlesworth, M. C., McMahon, N. & Toft, D. The assembly of progesterone
receptor-hsp90 complexes using purified proteins. J. Biol. Chem. 273, 32973–32979 (1998).
27. Gress, T. M. et al. Differential expression of heat shock proteins in pancreatic carcinoma. Cancer Res.
54, 547–551 (1994).
28. Yano, M., Naito, Z., Tanaka, S. & Asano, G. Expression and roles of heat shock proteins in human
breast cancer. Jpn J. Cancer Res. 87, 908–915 (1996).
29. Rutherford, S. L. & Lindquist, S. Hsp90 as a capacitor for morphological evolution. Nature 396,
336–342 (1998).
30. Queitsch, C., Sangster, T. A. & Lindquist, S. Hsp90 as a capacitor of phenotypic variation. Nature 417,
618–624 (2002).
Supplementary Information accompanies the paper on www.nature.com/nature.
Acknowledgements We thank D. Toft for critical reading of the manuscript. We also thank
G. Timony for preparing the scatter plots, P. Karjian for quantitative western blotting of Hop and
p23, and other team members at Conforma Therapeutics Corporation for discussions and
suggestions.
Competing interests statement The authors declare competing financial interests: details
accompany the paper on www.nature.com/nature.
Correspondence and requests for materials should be addressed to F.J.B.
(fburrows@conformacorp.com).
letters to nature
NATURE | VOL 425 | 25 SEPTEMBER 2003 | www.nature.com/nature410 © 2003 NaturePublishing Group