IUBMB
Life, 64(3): 266–273, March 2012
Research Communication
STAT3 Interacts Directly with Hsp90
Earl Prinsloo,* Adam H. Kramer,* Adrienne L. Edkins and Gregory L. Blatchy
Biomedical Biotechnology Research Unit, Department of Biochemistry, Microbiology and Biotechnology,
PO Box 94, Rhodes University, Grahamstown 6140, South Africa
Keywords
Summary
Heat shock protein 90 (Hsp90) functionally modulates signal
transduction. The signal transducer and activator of transcription 3 (STAT3) mediates interleukin-6 family cytokine signaling.
Aberrant activation and mutation of STAT3 is associated with
oncogenesis and immune disorders, respectively. Hsp90 and
STAT3 have previously been shown to colocalize and coimmunoprecipitate in common complexes. Surface plasmon resonance
spectroscopy revealed a direct, high affinity specific interaction
between recombinant Hsp90b and STAT3b in the presence and
absence of adenosine triphosphate (ATP) in molar excess. Furthermore, comparative analysis using a phosphomimetic mutation at tyrosine 705 showed that the direct interaction appeared
to favor neither unactivated nor activated STAT3. Destabilizing
mutation of STAT3 at arginine residues 414/417 to alanine in the
DNA-binding domain, previously shown to disrupt nuclear translocation in vivo, reduced interaction with a STAT3 DNA binding
site oligonucleotide and Hsp90b in vitro, indicating that STAT3
requires a functional DNA-binding domain for full direct interaction with Hsp90. Site-directed mutagenesis of a mammalian
STAT3–EGFP-N1 fusion construct at RR414/417 and subsequent
transfection into human MCF7 epithelial breast cancer cells
showed no impaired nuclear translocation when observed by confocal laser scanning microscopy. However, costaining for
Hsp90a/b isoforms and colocalization analysis revealed a defined
decrease in pixel-on-pixel colocalization compared with the wildtype confirming the requirement of the DNA-binding domain for
high-affinity interaction. Ó 2012 IUBMB
IUBMB Life, 64(3): 266–273, 2012
Additional Supporting Information may be found in the online version of this article.
Received 28 April 2011; accepted 28 November 2011
*E.P. and A.H.K. contributed equally to this work.
Address correspondence to: Earl Prinsloo, Biomedical Biotechnology
Research Unit, Department of Biochemistry, Microbiology and Biotechnology, PO Box 94, Rhodes University, Grahamstown 6140, South
Africa. Tel: 127-46-603-8082. Fax: 127-46-622-3984.
E-mail: e.prinsloo@ru.ac.za
y
Present address of Gregory L. Blatch: School of Biomedical and
Health Sciences, Faculty of Health, Engineering and Science, PO Box
14428, Victoria University, Melbourne, Victoria 8001, Australia.
ISSN 1521-6543 print/ISSN 1521-6551 online
DOI: 10.1002/iub.607
chaperone; signaling; STAT3; Hsp90.
Abbreviations
STAT3, signal transducer and activator of transcription 3; Hsp90, heat shock protein 90; NLS, nuclear
localization signal; FT-IR, Fourier transformed infrared; SPR, surface plasmon resonance; DMEM,
Dulbecco’s modified Eagles medium; BSA, bovine
serum albumin; TBS, Tris-buffered saline; SH2, src
homology; ICQ, intensity correlation quotient; LIF,
leukemia inhibitory factor; IL-6, interleukin-6; PBS,
phosphate buffered saline.
INTRODUCTION
The signal transducer and activator of transcription 3
(STAT3) is functionally required for maintenance and development of normal and disease phenotypes (1, 2). Dogma dictates
that stimulation of the gp130 receptor by members of the interleukin-6 (IL-6) cytokine family results in the recruitment of
Janus kinases for phosphorylation of monomeric (‘‘unactivated’’) STAT3 at tyrosine 705 (Y705) resulting in ‘‘activated’’ STAT3 which dimerizes and undergoes active nuclear
targeting (3).
Evidence for the involvement of the ATP-dependent molecular chaperone, heat shock protein 90 (Hsp90) in this signal
transduction pathway has been mounting, including: association
of Hsp90 in STAT3-caveolin containing lipid raft complexes,
requirement of STAT3 DNA-binding domain for association in
Hsp90 complexes and requirement of leukemia inhibitory factor
(LIF) stimulation of the canonical pathway for Hsp90-STAT3
association (4–6). These studies have highlighted the interaction
by immunoprecipitation and at best show association in common molecular complex but do not show direct interaction.
Sato et al. (5) has previously proposed that STAT3 associates
with the Hsp90 N-terminal domain, shown via domain deletion
and immunoprecipitation and specific Hsp90 inhibition by
geldanamycin. The Hsp90 N-terminal/ATPase domain is an
absolute requirement for in vivo chaperone activity (7). The
Hsp90 chaperone mechanism is complex. Beyond binding
partially/fully folded substrate and delivering it as mature active
STAT3 INTERACTS DIRECTLY WITH Hsp90
protein to the site of activity, the interaction is mediated by
combined factors including nucleotide hydrolysis and cochaperone binding (reviewed in ref. 8). The current consensus model
of the Hsp90 chaperone cycle states that client proteins are
delivered to the Hsp90 dimer (typically via Hsp70 and Hsp90
organizing protein [Hop]) and that client binding stimulates
ATP hydrolysis at the N-terminal ATPase domain driving a
conformational change and thereby clamping the substrate into
Hsp90 (8–10). Interestingly, Hop has been implicated in the
modulation of STAT3 phosphorylation and therefore its nuclear
targeting (11). Xu et al. (12) demonstrated that specific inhibition of Hsp90 also resulted in decreased targeting of STAT3 in
sequestering endosomes in IL-6-induced human Hep3B hepatocytes. STAT3 nuclear (and other) targeting signals and mechanism require elucidation. Ma et al. (13) have previously shown,
following bioinformatic alignment with STAT1, a putative bipartite nuclear localization signal (NLS) at arginine 214/215
(RR214/215) and arginine 414/417 (RR414/417) in the coiledcoil and DNA-binding domains of STAT3, respectively. In contrast, Liu et al. (14) have argued, against the findings of Ma
et al. (13), for an association of STAT3 with importin-a3 via
the sequence DVRKRQDLEQKM between amino acids 150–
162 of the STAT3 coiled-coil domain, thereby exploiting the
NLS mechanism of importin-a3 for nuclear import. Curious
though is the inability of RR414/417AA mutant to enter the nucleus, although it does undergo phosphorylation at Y705 and
presumably dimerizes (13). Combined with multiple neighboring residues in the heavily b-barreled structure of the DNAbinding domain, R417 is functionally required for STAT3 DNA
binding activity (15).
Cumulatively, previous evidence indicates a potential control
checkpoint in STAT3 signaling. Here, using surface plasmon
resonance spectroscopy, we investigated a potential direct interaction between Hsp90 and STAT3. Furthermore, using sitedirected mutagenesis, in vitro biophysical analysis and confocal
laser scanning microscopy, we show that STAT3 requires a
functional DNA-binding domain for interaction with Hsp90.
MATERIALS AND METHODS
Reagents
Adenosine triphosphate (sodium salt) was obtained from Roche
(Germany). Bradford’s reagent was purchased from Sigma-Aldrich
(USA). Oligonucleotides and primers were obtained from IDT
Technologies (USA) and Inqaba Biotech Ltd (South Africa). DpnI
and Pfu Turbo polymerase were purchased from Promega (USA).
QuikChange I mutagenesis kit was obtained from Stratagene
(USA). Full-length polyclonal rabbit anti-green fluorescent protein
(GFP) antibody (cat no. sc-8334), goat anti-Hsp90a/b [N-17] (cat
no. sc-1055) antibodies were purchased from Santa Cruz Biotechnology (USA). Mouse anti-human Hsp90a/b (F-8), mouse anti-rabbit STAT3 (F2) and rabbit anti-mouse STAT3 (C20) antibodies
were purchased from Santa Cruz Biotechnologies (USA). Alexa
267
Fluor-633 donkey anti-goat IgG (H 1 L; cat no.: A21082) were
purchased from Invitrogen. Superdex pg200 16/60 HR was purchased from Pharmacia Amersham (Sweden). Recombinant human
Hsp90b was purchased from Stressmarq (Canada). Amicon Ultra15 10K centrifugal filters (Millipore, USA). Sensor chip CM5 and
amine coupling Kit, ECL Advance Western Blotting Detection kit
and goat anti-mouse IgG horseradish peroxidase-linked antibody
were purchased from GE Healthcare (Sweden). The pET32bSTAT3b-tc bacterial expression construct was a kind donation
from Dr Christoph Müller (European Molecular Biology Laboratory Heidelberg). Mammalian expression constructs pEGFP-N1
and pEGFP-N1 STAT3 were kind donations from Prof. Pravin
Sehgal (New York Medical College). Western blot chemiluminescent detection was performed using the ECL advanced Western
Blotting Detection kit (GE Healthcare, UK). Unless stated
otherwise, all reagents were of highest grade and purity.
Site-Directed Mutagenesis of STAT3b
Site-directed mutagenesis of the STAT3b DNA-binding domain
residues arginine 414 and 417 (to alanine/RR414/417AA) and src
homology (SH2) domain tyrosine 705 (to glutamine, phosphomimetic Y705D) were performed by linear (non-polymerase chain
reaction) whole plasmid amplification according to the Stratagene
Quikchange kit. Mutation primer DNA sets were as follows:
RR414/417AA forward, 50 -caagcacctgacccttgcggagcaggcatgtgggaa
tggagg-30 ; RR414/417AA reverse, 50 - cctccattcccacatgcctgctccgc
aagggtcaggtgcttg-30 ; Y705D forward, 50 - gtagtgctgccccggacctgaag
accaag-30 ; Y705D reverse, 50 - cttggtcttcaggtccggggcagcactac-30 .
All mutants were confirmed by DNA sequencing.
Expression and Purification of STAT3b
Wild-Type and Mutants
Mouse STAT3b encoded by the pET32b-STAT3b-tc was
expressed and purified as described previously (16). With the
following changes: FPLC was performed on an ÄKTA Basic
FPLC (Sweden) using 20 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid-hydrochloride (HEPES-HCl), pH 7.0, 200
mM NaCl and a Superdex pg200 16/60 HR size exclusion column at a flow rate of 1 mL min21. Fractions of 1 mL were
selected following 12% sodium dodecyl sulfate-polyacrylamide
gel electrophoresis (SDS-PAGE and western blot analysis), concentrated and buffer exchanged by Amicon Ultra-15 10K centrifugal filters into 40 mM HEPES, pH 7.4, 150 mM KCl and 5
mM MgCl2. Protein was quantified by Bradford’s method (17).
Fourier Transformed Infrared Spectroscopy
Recombinant STAT3b wild-type and mutant structural integrity was determined by Fourier transformed infrared (FT-IR)
spectroscopy. Briefly, scans were performed on a Perkin Elmer
Spectrum 100 FT-IR Spectrophotometer with attenuated total
reflectance attachment using Perkin Elmer Spectrum
(v6.35.0176) software at a path length of 2 cm21. Averages of
20 scans were collected in triplicate over a range of 650–4000
268
PRINSLOO ET AL.
cm21. Spectral scans of native and thermally denatured (1008C,
1 H) STAT3b wild-type and mutants were analyzed and compared for presence of Amide I and II bands indicating secondary and tertiary structural elements. Data were analyzed and
edited in Spekwin32 1.71.5 (available at http://www.effemm2.
de/spekwin) and Prism 4 (Graphpad Software, USA).
Surface Plasmon Resonance Spectroscopy
Surface plasmon resonsance (SPR) spectroscopy was performed on a BIAcoreX instrument (GE Healthcare, Sweden) at
258C using 40 mM HEPES-HCl, pH 7.4, 150 mM KCl and 5
mM MgCl2 as running buffer. Briefly, flow cell-1 of a CM5
sensor chip was preconditioned at 100 lL min21 with successive 20 lL pulses of 50 mM NaOH, 100 mM HCl, 0.1 mM
SDS and 0.085% (v/v) H3PO4. Recombinant Hsp90b (extensively dialyzed against 40 mM HEPES-HCl, pH 7.4, 150 mM
KCl, 5mM MgCl2) was immobilized at a level equivalent to
230 RU using amine coupling in 10 mM sodium acetate, pH
4.5 (determined by preconcentration pH scouting). Free amines
on flow cell-2 were blocked by ethanolamine and used as an
inline reference. All SPR measurements were performed in triplicate in the presence and absence of 1 mM ATP. Sensorgrams
were collected at 30 lL min21 by injection of 30 lL purified
STAT3b (wild-type and mutants, Y705D and RR414/417AA)
followed by 180-sec delay to monitor dissociation. Buffer
blanks with and without 1 mM ATP were performed for double
reference subtraction. DNA binding ability of the STAT3b
wild-type and RR414/417AA mutant was assessed by injection
of 1 lM STAT3b protein in the presence and absence of 1 lM
of a consensus STAT3 DNA binding site oligomer (50 TGCATTTCCCGTAAATCT-30 ; hybridized to its complementary strand) as a competitive analyte. Optimal regeneration conditions were obtained by 10-sec pulse injections of 50 mM
Tris-HCl, pH 8.0, 3 M guanidine-HCl. Triplicate injections
were performed for each concentration (31.25–2000 nM) to
account for statistical variability. Kinetic evaluation of the data
was performed based on the 1:1 association for the determination of the observed rate constant, kobs and Req, the steady state
binding level, respectively. The affinity constant (KD) was calculated from the ratio of the dissociation (kd) and the association rate constants (ka), (i.e., KD 5 kd/ka). Association rate constants were calculated following linear regression fitting of kobs
versus concentration of analyte plots according to the equation,
kobs 5 ka. Concentrationanalyte 1 kd. Dissociation rate constants
were calculated using an exponential decay model fit to the
dissociation phase data. Data and statistical analysis were performed in BIAevaluation 4.1.1 (GE Healthcare, Sweden) and
Prism 4 (Graphpad Software, USA).
Cell Culture, Transfection and Confocal
Laser Scanning Microscopy
The human breast cancer cell line MCF7 (HTB-22) was cultivated in a maintenance media comprising Dulbecco’s modified
Eagles medium (DMEM), containing 5% (v/v) heat inactivated
fetal calf serum and 50/50 penicillin/streptomycin (100 U
mL21). The cells were routinely incubated in a humidified incubator at 378C, in a 9% CO2 atmosphere. Endotoxin-free plasmids (pEGFP-N1, pEGFP-N1 STAT3 wild-type and pEGFP-N1
STAT3 RR414/417AA) were isolated using the Zyppy Plasmid
Midiprep kit (Zymo Research, USA). Trypsinized cells were
enumerated and resuspended (1 3 106) in phosphate buffered
saline, pH 7.4 (PBS; 16 mM Na2HPO4, 2 mM KH2PO4, 137
mM NaCl and 3 mM KCl) before electroporation with 12.5 lg
of plasmid (or PBS control) at 300 V (capacitance of 500 lF)
in a BioRad Gene Pulser II electroporation system. Following
electroporation, cells were plated into six-well dishes containing
sterile cover slips and 3 mL DMEM media. Samples were incubated for 3 days with frequent media changes. Transfected
MCF7 cells grown on cover slips were fixed in ice-cold methanol, air-dried and blocked with 1% (w/v) bovine serum albumin
in Tris-buffered saline (TBS; 0.8% (w/v) NaCl, 0.24% (w/v)
Tris and pH 7.6; BSA/TBS) for 30 min at room temperature,
followed by incubation with primary antibodies in 1% (w/v)
BSA/TBS at 48C overnight. Cover slips were washed twice
with 0.1% (w/v) BSA/TBS and incubated with the appropriate
secondary antibody at room temperature for 1 H in the dark.
Cells were counterstained with Hoechst-33342 (1 lg mL21 in
sterile water) for nuclear material and mounted using fluorescent mounting media (Dako, Denmark). Immunofluorescence
was visualized using a Zeiss LSM 510 Meta confocal laser
scanning microscope with a 633 oil objective. Colocalization
analysis was performed using ImageJ V1.42I (Macbiophotonics,
Canada).
RESULTS
STAT3 and Hsp90: A Specific and Direct Interaction
Recombinant STAT3b was expressed and purified as
described previously. Dot blot association assay (Supporting Information Fig. S1) initially revealed the STAT3b–Hsp90b interaction to be specific, as judged by noninteraction of thermally
denatured (1008C, 1 H) Hsp90 with STAT3b and inability of
STAT3b to interact with immobilized bovine serum albumin
(BSA). More noteworthy was the apparent ATP dependency of
the equilibrium interaction (Supporting Information Fig. S1).
Based on positive results obtained by dot blot association assay,
it was decided to investigate the interaction in real-time to further test for direct interaction in the presence and absence of
ATP using SPR spectroscopy. Recombinant Hsp90b was immobilized as ligand and STATb was used as analyte. High–low
followed by low–high analyte concentration injections were performed to test surface/ligand homogeneity; little variance was
observed, and we therefore concluded that the immobilized
Hsp90b exhibited little surface heterogeneity allowing for subsequent kinetic experiments (data not shown). Direct physical
binding of STAT3b to Hsp90b was found to be specific, com-
STAT3 INTERACTS DIRECTLY WITH Hsp90
269
were observed for 1 mM ATP; also 1 mM ATP was deemed
saturating based on previous literature reports (18). The interaction between Hsp90b and thermally denatured (1008C, 1 H)
STAT3b wild-type was minimal (Fig. 1A). Similarly, the interaction between Hsp90b and BSA corresponded to the baseline
(data not shown). Thermal denaturation resulted in loss of
STAT3b tertiary structure as observed by disappearance of the
amide II peak in FT-IR spectroscopy (Fig. 1B). The interaction
between Hsp90b and STAT3b wild-type showed distinct concentration dependency in the presence and absence of 1 mM
ATP (Supporting Information Fig. S2). Although conformational
change is what drives the Hsp90 chaperone activity, we have
not actively fit our data to a conformational change model due
to the rapid rate at which the biological conformational changes
likely occur (10). Therefore, we investigated the 1:1 binding of
STAT3b to nucleotide (ATP) free and bound forms of Hsp90,
thereby providing snapshots of the interaction events, that is,
isolated binding events. Kinetic evaluation of the association
phase by linear data fitting using a 1:1 model for a simple
bimolecular interaction (Fig. 1C) and exponential decay (data
not shown) of the dissociation phase data yielded rate constants
in the presence (ka 5 3.68 3 104 6 8.67 3 103 M21 sec21; kd
5 1.43 3 1023 6 0.14 3 1023 sec21) and absence (ka 5 3.47
3 104 6 1.80 3 104 M21 sec21; kd 5 1.54 3 1023 6 0.41 3
1023 sec21) of 1 mM ATP. By calculation, STAT3b wild-type
exhibited a high specific affinity for Hsp90b in the presence
(KD 5 40.3 6 10.1 nM) and absence (KD 5 59.9 6 44.9 nM)
of ATP.
Figure 1. Direct interaction between STAT3b and Hsp90b in
the presence and absence of ATP. A: Specific interaction, compared with thermally denatured STAT3b, between 1 lM
STAT3b and immobilized Hsp90b in the presence and absence
of 1 mM ATP. B: Averaged FT-IR spectra for structural integrity of STAT3b wild-type showing native (open triangle) and
thermally denatured (solid black line) protein. Spectra represent
Amide II (1500–1600 cm21) and Amide I (1600–1700 cm21).
C: Linear plot of kobs versus [STAT3b] in the presence (r2 5
0.9381) and absence (r2 5 0.9303) of 1 mM ATP (see text for
details).
pared with heat denatured (1008C, 1 H) STAT3b, in the presence and absence of 1 mM ATP (Fig. 1A; 1 lM protein concentration shown). ATP concentration was selected after tests
for bulk shift effects (data not shown). No large bulk shifts
STAT3-Hsp90 Direct Interaction Requires a Fully
Functional STAT3 DNA-Binding Domain
Site-directed mutagenesis of key residues in the SH2 and
DNA-binding domains were performed to identify regions mediating the observed direct interaction between Hsp90b and
STAT3b. Mutant proteins were heterologously expressed in
Escherichia coli and purified to homogeneity. It should be noted
that STAT3b wild-type, Y705D and RR414/417AA were capable of forming dimers as revealed by size exclusion chromatography, SDS-PAGE and western blot analysis (Figs. 2A and 2B).
Size exclusion analysis of dimer fractions (data not shown)
resulted in separation into typical monomeric and dimeric peaks
indicating that the dimers observed by size exclusion chromatography most likely existed in equilibrium with a monomeric
pool. This is consistent with previous data that monomeric
STAT3b formed heterodimers in the presence of Mg21 (19).
In silico thermodynamic stability analysis based on homology
models revealed that the introduction of the negative charge,
that is, the Y705D mutation is extremely favorable (DDG 5
221.58 kcal mol21) when compared with the wild-type. Comparative FT-IR spectral analysis to STAT3b wild-type revealed
no definable changes in the amide I banding patterns (Fig. 3).
In silico thermodynamic analysis revealed thermodynamic instability (DDG 5 0.98 kcal mol21) in the RR414/417AA mutant,
270
PRINSLOO ET AL.
Figure 2. Representative chromatograms of STAT3b proteins showing monomeric and dimeric fractions verified by western blot
analysis. A: Chromatogram overlap showing typical elution profile of STAT3b Y705D (black solid line), blue dextran trace (grey
dashed line) shows column void volume, and comparative bovine serum albumin elution profile (grey dots) showing typical dimeric
(132 kDa indicated by ^ and monomeric (66 kDa fractions indicated by ^^. INSET shows similar traces for the STAT3b
wild-type and RR414/417AA purifications (arrows indicate monomer and dimer peaks). B: Western blot analysis of STAT3b
Y705D elution profile reveals dimeric (*) and momomeric (**) fractions indicated in (A). Elution volume fractions shown in milliliters.
but this instability was most likely localized to the DNA-binding domain as no distinct changes in the amide I banding patterns of STAT3b RR414/417AA were observed, compared with
the wild-type (Fig. 3). STAT3b Y705D showed decreased
response in its interaction with immobilized Hsp90b but shifts
observed were comparable with STAT3b wild-type in the presence and absence of ATP (Fig. 4). These data tentatively sug-
gested that Hsp90b favored neither ‘‘unactivated’’ nor ‘‘activated’’ STAT3. This is consistent with previous findings that
proposed that regions other than Y705 mediated STAT3 interaction with Hsp90, that is, the DNA-binding domain (5). The
DNA-binding mutant STAT3b RR414/417AA showed a substantial decrease in response to Hsp90b binding in the presence
and absence of 1mM ATP when compared with STAT3b wild-
Figure 3. Comparative FT-IR spectra overlay of Amide I band
(1600–1700 cm21) of STAT3b wild-type, Y705D and RR414/
417AA mutants.
Figure 4. Comparative interaction of STAT3b wild-type and
mutants (Y705D and RR414/417AA) with immobilized Hsp90b
using SPR. Representative sensorgrams of observed binding
responses of STAT3b wild-type (WT), Y705D and RR414/
417AA in the presence and absence of 1 mM ATP.
STAT3 INTERACTS DIRECTLY WITH Hsp90
Figure 5. DNA binding ability of STAT3b wild-type and
STAT3b RR414/417AA. Representative sensorgrams of the
interaction of (A) 1lM STAT3b wild-type and (B) 1lM
STAT3b RR414/417AA mutant with immobilized Hsp90b in
the presence (1lM) and absence of a consensus STAT3 18mer
DNA binding site (50 -TGCATTTCCCGTAAATCT-30 ).
type (Fig. 4). When normalized, the interaction observed for
1 lM STAT3b RR414/417AA, compared with STAT3b wildtype (Fig. 4), equates to a decrease of greater than 50%. Furthermore, SPR interaction analysis of STAT3b wild-type and
RR414/417AA in the presence and absence of the STAT3 18
base pair DNA binding site (50 -TGCATTTCCCGTAAATCT-30 )
confirmed the impaired DNA binding ability of the RR414/
417AA mutant by showing no change in response to interaction
with immobilized Hsp90b as opposed to that observed for the
STAT3b wild-type (Fig. 5).
In vivo interaction was confirmed by confocal laser scanning
microscopy following transfection with pEGFP-N1 STAT3
wild-type and RR414/417AA mutant (or pEGFP-N1, as negative control; Supporting Information Fig. S3) and costaining for
Hsp90a/b (Fig. 6). Briefly, STAT3-EGFP-N1 wild-type detection showed characteristic diffuse cytoplasmic and distinct nuclear staining in all transfected cells observed. STAT3-EGFPN1 RR414/417AA showed a similar staining pattern to the
wild-type. Surprisingly, the presence of the RR414/417AA
271
mutation did not appear to affect the nuclear localization.
Hsp90a/b immunofluorescent detection exhibited similar cellular
distribution patterns with decreased detection in the nuclear
region when compared with the Hoechst 33342 detection.
Colocalization analyses of these data were performed by calculation of the intensity correlation quotient (ICQ) (20). The ICQ
relates to the synchrony in variation of intensities between
images. If pixels are dependant (i.e., colocalize), they will vary
around the respective mean image intensities together. ICQ values range between 20.5 and 0.5 with random staining patterns
having values close to 0; segregated staining exhibit values
between 20.5 and 0 while dependant staining show values
between 0 and 0.5. The EGFP-N1/Hsp90a/b (Supporting Information Fig. S3) control exhibited a low ICQ (0.027 6 0.0017;
n 5 3) indicative of random staining and no interaction.
STAT3-EGFP-N1 wild-type/Hsp90a/b (Fig. 6A) showed a high
ICQ (0.485 6 0.009; n 5 7). Compared with the wild-type, the
mutant STAT3-EGFP-N1 RR414/417AA/ Hsp90a/b (Fig. 6B)
displayed a decreased ICQ (0.355 6 0.089). Taken together,
these data are suggestive of a decreased in vivo interaction of
STAT3 with Hsp90a/b upon mutation of the DNA-binding domain, thus confirming the in vitro biophysical data. Although
previously reported that the RR414/417AA mutation resulted in
impaired nuclear localization (13), it is clear from our data that
this is not the case and that STAT3 nuclear localization may
either be via the mechanism described by Liu et al. (14) or via
another undescribed pathway that may not require Hsp90 for
nuclear translocation.
DISCUSSION AND CONCLUSIONS
Here, we show using SPR spectroscopy that the high degree
of colocalization observed between STAT3b and Hsp90b (Fig.
6) is as a result of a direct interaction between the transcription
factor and Hsp90. Also, the in vitro and in vivo evidence
pointed to the interaction requiring a fully functional STAT3
DNA-binding domain for optimal association with Hsp90b.
Mutation of Y705 to D705 to mimic the negative charge conferred by phosphorylation showed potentially that the in vitro
interaction has no dependency on the activation state of
STAT3. However, previously we have described (6) in mouse
embryonic stem cells that Hsp90-STAT3 association, shown by
immunoprecipitation, occurs upon stimulation of the gp130 receptor complex with LIF. Thus, a separate undescribed signal
likely exists that recruits Hsp90 to STAT3 following phosphorylation.
It may be suggested that the observed decrease in STAT3b
RR414/417AA association with Hsp90b (both in vitro and in
vivo) could be caused by the removal of available positive surface charge (Supporting Information Fig. S4). This may be akin
to the observed interaction between Hsp90 and client kinases
that appear to share a common interaction region based on surface electrostatics (21). Csermely et al. (22) have successfully
shown using SPR that positive charges and hydrophobic surfa-
272
PRINSLOO ET AL.
Figure 6. Colocalization of endogenous Hsp90a/b with STAT3-EGFP-N1 wild-type and RR414/417AA mutant. Confocal laser
scanning microscopy images of (A) STAT3-EGFP-N1 wild-type and (B) STAT3-EGFP-N1 RR414/417AA transfected MCF7 cells
counterstained for endogenous Hsp90a/b (A and B; HSP90a/b) and Hoechst 33342 (A and B; HOECHST). (A and B, MERGE)
indicates merged blue (HOECHST), green (EGFP-N1) and red (HSP90) channels. Primary anti-Hsp90a/b applied at 1:100 dilution
and secondary (Alexa Fluor-633 donkey anti-goat IgG) at 1:1000. Scale bars indicate 20 lm. [Color figure can be viewed in the
online issue, which is available at wileyonlinelibrary.com.]
ces on client proteins are required for successful interaction with
Hsp90. Hence, it may be that the Hsp90-STAT3 interaction is dependent upon surface charges within the DNA-binding domain. In
structural description of the STAT3b crystal structure (1bg1,
www.rcsb.org), R414 and R417 are located on a b strand (be)
and a loop (ef, residues 417–425), respectively. Becker et al. (15)
commented on the relative flexibility of loop ef. It may be argued
that the mutation may result in loss of flexibility of the loop and
this could be a major determinant in the lack of Hsp90 interaction. However, molecular dynamics (23) and normal mode analysis (Supporting Information Fig. S5) experiments revealed little
conformational changes within individual domain structures of
STAT3b wild-type and RR414/417AA mutant, respectively. Certainly, the impaired DNA binding observed here for the RR414/
417AA mutant was as a direct result of the removal of the available positive charge of R417.
In conclusion, STAT3b interacts in a direct and specific
manner with Hsp90b in the presence and absence of ATP. This
was confirmed in vivo by confocal laser scanning microscopy.
Combined with previous data, by ourselves (6) and others (4–5,
12), we speculate that the Hsp90 interaction with STAT3
requires a functional DNA-binding domain. The contribution of
surface electrostatics within the DNA-binding domain requires
further investigation (in vivo and in vitro) to elucidate the minimal surface on STAT3 required for association with Hsp90.
The potential involvement of this direct interaction stimulating
STAT3 nuclear or endosomal targeting requires further exploration in vivo.
ACKNOWLEDGEMENTS
This work was supported by National Research Foundation of
South Africa, the Claude Leon Foundation and the Joint
Research Committee of Rhodes University. The authors thank
Dr. Christoph Müller (EMBL) for the kind donation of the
pET32b-STAT3b-tc bacterial expression construct, Prof. Pravin
Sehgal (New York Medical College) for the kind donation of
the pEGFP-N1 and pEGFP-N1 STAT3 and Dr. Eva-Rachele
Pesce for critical evaluation of the manuscript.
REFERENCES
1. Bromberg, J. F., Wrzeszczynska, M. H., Devgan, G., Zhao, Y., Pestell,
R. G., et al. (1999) STAT3 as an oncogene. Cell 98, 295–303.
2. Grivennikov, S., Karin, E., Terzic, J., Mucida, D., Yu, G.-Y., et al.
(2009) IL-6 and Stat3 are required for survival of intestinal epithelial
cells and development of colitis-associated cancer. Cancer Cell 15,
103–113.
3. Heinrich, P. C., Behrmann, I., Müller-Newen, G., Schaper, F., and
Graeve, L. (1998) Interleukin-6-type cytokine signalling through the
gp130/Jak/STAT pathway. Biochem. J. 334, 297–314.
4. Shah, M., Patel, K., Fried, V. A., and Sehgal, P. B. (2002) Interactions
of STAT3 with caveolin-1 and heat shock protein 90 in plasma membrane raft and cytosolic complexes. J. Biol. Chem. 277, 45662–45669.
5. Sato, N., Yamamoto, T., Sekine, Y., Yumioka, T., Junicho, A., et al.
(2003) Involvement of heat-shock protein 90 in the interleukin-6-mediated signaling pathway through STAT3. Biochem. Biophys. Res. Comm.
300, 847–852.
6. Setati, M. M., Prinsloo, E., Longshaw, V. M., Murray, P. A., Edgar D.
H., et al. (2010) Leukemia inhibitory factor promotes Hsp90 association
with STAT3 in mouse embryonic stem cells. IUBMB Life 62, 61–66.
STAT3 INTERACTS DIRECTLY WITH Hsp90
7. Obermann, W. M. J., Sondermann, H., Russo, A. A., Pavletich, N. P.,
and Hartl, F. U. (1998) In vivo function of Hsp90 is dependent on ATP
binding and ATP hydrolysis. J. Cell. Biol. 143, 901–910.
8. Wandinger, S. K., Richter, K., and Buchner, J. (2008) The Hsp90 chaperone machinery. J. Biol. Chem. 283, 18473–18477.
9. McLaughlin, S. H., Sobott F., Yao, Z., Zhang, W., Nielsen P. R., et al.
(2006) The co-chaperone p23 arrests the Hsp90 ATPase cycle to trap
client proteins. J. Mol. Biol. 356, 746–758.
10. Richter, K., Soroka, J., Skalniak, L., Leskovar, A., Hessling, M., et al.
(2008) Conserved conformational changes in the ATPase cycle of
human Hsp90. J. Biol. Chem. 283, 17757–17765.
11. Longshaw, V. M., Baxter, M., Prewitz, M., and Blatch, G. L. (2009)
Knockdown of the co-chaperone Hop promotes extranuclear accumulation of Stat3 in mouse embryonic stem cells. Eur. J. Cell. Biol. 88,
153–166.
12. Xu, F., Mukhopadhyay, S., and Sehgal, P. B. (2007) Live cell imaging
of interleukin-6-induced targeting of ‘‘transcription factor’’ STAT3 to
sequestering endosomes in the cytoplasm. Am. J. Physiol. Cell Physiol.
293, C1374 – C1382.
13. Ma, J., Zhang, T., Novotny-Diermayr, V., Tan, A. L. C., and Cao, X.
(2003) A novel sequence in the coiled-coil domain of Stat3 essential for
its nuclear translocation. J. Biol. Chem. 278, 29252–29260.
14. Liu, L., McBride, K. M., and Reich, N. C. (2005) STAT3 nuclear
import is independent of tyrosine phosphorylation and mediated by
importin-a3. Proc. Natl. Acad. Sci. USA 102, 8150–8155.
15. Becker, S., Groner, B., and Müller, C. W. (1998) Three-dimensional structure of the Stat3b homodimer bound to DNA. Nature 394, 145–151.
273
16. Becker, S., Corthals, G. L., Aebersold, R., Groner, B., and Müller, C.
W. (1998) Expression of a tyrosine phosphorylated, DNA binding Stat3beta dimer in bacteria. FEBS Lett. 441, 141–147.
17. Bradford, M. M. (1976) Rapid and sensitive method for the quantitation
of microgram quantities of protein utilizing the principle of protein-dye
binding. Anal. Biochem. 72, 248–254.
18. 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.
19. Novak, U., Ji, H., Kanagasundaram, V., Simpson, R. and Paradiso, L.
(1998) STAT3 forms stable homodimers in the presence of divalent
cations prior to activation. Biochem. Biophys. Res. Comm. 247, 558–
563.
20. Li, Q., Lau, A., Morris, T. J., Guo, L., Fordyce, C. B., et al. (2004) A
Syntaxin 1, Gao, and N-type calcium channel complex at a presynaptic
nerve terminal: analysis by quantitative immunocolocalization. J. Neurosci. 24, 4070–4081.
21. Citri, A., Harari, D., Shohat, G., Ramakrishnan, P., Gan, J., et al. (2006)
Hsp90 recognizes a common surface on client kinases. J. Biol. Chem.
281, 14361–14369.
22. Csermely, P., Miyata, Y., Söti, C., and Yahara, I. (1998) Binding affinity of proteins to Hsp90 correlates with both hydrophobicity and positive surface charges. Life Sciences 61, 411–418.
23. Lin, J., Buettner, R., Yuan, Y.-C., Yip, R., Horne, D., et al. (2009) Molecular dynamics simulations of the conformational changes in signal
transducers and activators of transcription, Stat1 and Stat3. J. Mol.
Graphics 28, 347–356.