Phosphorylation and oligomerization states of native pig brainHSP90 studied by mass spectrometry

Cyrille Garnier1,*, Daniel Lafitte1,*, Thomas J. D. Jorgensen2, Ole N. Jensen2, Claudette Briand1

and Vincent Peyrot1

1UMR CNRS 6032, Faculte de Pharmacie, Marseille, France; 2Protein Research group, Department of Biochemistry and

Molecular Biology, University of Southern Denmark, Odense University, Denmark

HSP90 is one of the most abundant proteins in the cytosol

of eukaryotic cells. HSP90 forms transient or stable

complexes with several key proteins involved in signal

transduction including protooncogenic protein kinases and

nuclear receptors, it interacts with cellular structural

elements such as actin-microfilament, tubulin-microtubule

and intermediate filaments, and also exhibits conventional

chaperone functions. This protein exists in two isoforms

a-HSP90 and b-HSP90, and it forms dimers which are

crucial species for its biological activity. PAGE, ESI-MS

and MALDI-MS were used to study HSP90 purified from

pig brain. The two protein isoforms were clearly distin-

guished by ESI-MS, the a isoform being < six times more

abundant than the b isoform. ESI-MS in combination with

l phosphatase treatment provided direct evidence of the

existence of four phosphorylated forms of native pig brain

a-HSP90, with the diphosphorylated form being the most

abundant. For the b isoform, the di-phosphorylated was

also the most abundant. MALDI mass spectra of HSP90

samples after chemical cross-linking showed a high percen-

tage of a±a homodimers. In addition, evidence for the

existence of higher HSP90 oligomers was obtained.

Keywords: HSP90; mass spectrometry; phosphorylation;

cross-linking.

Heat shock proteins (HSP) are abundant eukaryotic proteinsconstituting < 1 or 2% of the cellular proteins underphysiological conditions, with an increased expression inresponse to stress [1]. These ubiquitous proteins, classifiedaccording to their apparent molecular mass, i.e. HSP27, 60,70 or 90, work in close association with cytosolic cofactorsto help the correct folding of several cytosolic proteins(reviewed in [2]).

The target of this study is HSP90, a dimeric phospho-protein involved in many cell processes such as cell cyclecontrol, transcriptional regulation or signal transduction(reviewed in [3]). Mammalian HSP90 exists in two iso-forms a- and b-HSP90 [4], differing in mass by < 2 kDa.Although dimers appear to be the functional form ofHSP90, oligomeric forms exist in solution under physio-logical conditions [5] and their amount tends to increaseunder hyperthermal exposure [6±9]. HSP90 possesses an

ATP binding site [10,11] and undergoes conformationalchanges in presence of ATP [12,13]. This protein alsopossesses ATPase activity [14]. The functions of HSP90 arein part regulated by phosphorylation, e.g. phosphorylatedHSP90 stimulates eIF-2a kinase activity [15] and phos-phorylation regulates the interaction between HSP90 andpp60v-src [16]. HSP phosphorylation may be mediated byvarious kinases [17,18] but this protein can also autophos-phorylate and act as a kinase for other protein substrates[19].

ESI-MS and MALDI-MS are the most suitable massspectrometric techniques for the analysis of macromole-cules [20,21]. The increase in resolution achieved by thelatest generation of mass spectrometers allows a detailedinvestigation of post-translational modifications of largeproteins and protein complexes [22]. This is particularlyimportant in the case of multiple post-translationalmodifications where various species with a small differencein the nominal masses exist in solution.

In this paper we provide the first mass spectrometricanalysis of the monomeric and dimeric forms of HSP90,and determine the relative abundance of the two isoforms insolution. The presence of four phosphorylation sites on apig brain HSP90 and a preferential di-phosphorylated stateof this protein isoform is demonstrated. The b pig brainHSP90, was only observed in the di-phosphorylated form.

M A T E R I A L S A N D M E T H O D S

HSP90 Purification

The 90-kDa heat shock protein was purified from porcinebrain according to Yonezawa et al. [23] modified by

Eur. J. Biochem. 268, 2402±2407 (2001) q FEBS 2001

Correspondence to D. Lafitte, UMR CNRS 6032, 27 bd Jean

Moulin, Faculte de Pharmacie, 13385 Marseille, France.

Fax: 1 33 491782024, Tel.: 1 33 491835506,

E-mail: daniel.laffite@pharmacie.univ-mrs.fr

Abbreviations: HSP, heat shock protein; EDC,

N-(3-dimethylaminopropyl)-N 0-ethylcarbodiimide hydrochloride.

Note: the following proteins have been deposited in the Swiss-Prot

database: a-HSP90 from pig (entry name HS9A_PIG, accession no.

O02705); b-HSP90from human (entry name HS9B_HUMAN,

accession no. P08238); b-HSP90from mouse (entry name

HS9B_MOUSE, accession no. P11499); b-HSP90from rat (entry

name HS9B_RAT, accession no. P34058).

*Note: the first two authors contributed equally to this work.

(Received 18 December 2000, revised 19 February 2001, accepted

23 February 2001)

Garnier et al. [8,24]. Samples were stored at 280 8C.Protein concentration was determined by UV absorbancewith an extinction coefficient of :280nm � 124 000 ^6000 m21´cm21 in 10 mm Tris/HCl buffer, pH 7.0 con-sidering that HSP90 is a dimer [8,24]. The absorption wascorrected for light scattering using the Beckman DU7400spectrophotometer software.

Chemical cross-linking experiments

Chemical cross-linking of HSP90 was performed usingN-(3-dimethylaminopropyl)-N 0-ethylcarbodiimide hydro-chloride: EDC (Sigma-Aldrich Chemical). HSP90(1 mg´mL21) in 10 mm ammonium acetate, pH 7.0 wascross-linked by addition of 5% of sample volume ofEDC stock solution (32 mm, 10 mm ammonium acetate,pH 7.0) and the mixture was incubated during 30 minat room temperature. To stop the cross-linking process,EDC excess was removed by using a PD10 column(Amersham Pharmacia Biotech), equilibrated with 10 mmammonium acetate buffer, pH 7.0. Samples were thenanalysed by native and SDS/PAGE and by MALDI massspectrometry.

Native and SDS/PAGE

Non cross-linked and cross-linked HSP90 samples wereanalysed on 4±15% gradient gels (native and SDS/PAGE)on MiniProteanII (Bio-Rad). Molecular mass markers usedwere HMW electrophoresis calibration kit (669, 440, 232,140, and 67 kDa) for native PAGE and HMW-SDS electro-phoresis calibration kit (212, 170, 116, 76, and 53 kDa)for SDS/PAGE (Amersham Pharmacia biotech). To detectoligomeric species, HSP90 quantities ranging from 7±35 mg were analysed on native 4±15% gradient gels usingPhastSystem apparatus (Amersham Pharmacia biotech). Allgels were stained with Coomassie brilliant blue.

Phosphatase treatment

A solution containing 400 mg of HSP90 was supplementedwith the 10 � l phosphatase reaction buffer (New EnglandBiolabs) and with 2 mm MnCl2 to reach the final volume of400 mL. After addition of 4000 units of l protein phos-phatase (a serine/threonine phosphatase; New EnglandBiolabs), the protein solution was incubated at 30 8C for3 h. The solution was then equilibrated in ammoniumacetate 10 mm, pH 7.0 using a PD10 column (AmershamPharmacia biotech) and submitted to ESI-MS analysis.

Size-exclusion chromatography

Samples containing 20 mm HSP90 (200 mL) in 10 mm Tris/HCl buffer, pH 7.0 were loaded on a Sephacryl S-300 HighResolution (Amersham Pharmacia biotech) 1 � 38 cm at aflow rate of 20 mL´h21, experiments were carried out at25 8C. The elution was detected by absorbance at 280 nm,fractions were analysed on 4±15% native PAGE using thePhastSystem apparatus (Amersham Pharmacia biotech).Fractions containing only dimers or only oligomers werethen analysed by MALDI mass spectrometry.

Electrospray mass spectrometry

Protein samples were analysed in positive ion mode with aquadrupole-orthogonal acceleration TOF mass spectro-meter Q-TOFTM (Micromass, Manchester, UK), using aZsprayTM interface. Prior to mass spectrometry analysis,the samples were dissolved in a 1 : 1 (v/v) mixture ofacetonitrile and H2O containing 1% (v/v) formic acid. Thespray solution was introduced into a normal electrosprayion source at a flow rate of 10 mL´min21. The parametersfor the normal electrospray ion source were: spray voltage3.0 kV, cone 100 V, nebuliser gas N2 (flow 15 L´h21),desolvation gas N2 (140 8C, flow 350 L´h21). The massspectra of dephosphorylated HSP90 was obtained with ananoelectrospray ion source. The nanoelectrospray ionsource parameters were: spray voltage 800 V, cone 60 V.Raw spectra were deconvoluted using the maximumentropy algorithm (maxentTM) present in the masslynxTM

software (Micromass, Manchester, UK).

MALDI mass spectrometry

Positive ion MALDI-TOF mass spectra were recorded inlinear or reflectron mode using a voyager MALDI massspectrometer equipped with delayed extraction (AppliedBiosystems). External mass calibration was performedusing horse heart myoglobin and bovine serum albumin(Calbiochem). The samples were mixed in a 1 : 1 (v/v)ratio with a saturated solution of sinapinic acid (Aldrichchemicals) in 0.1% aqueous trifluoroacetic acid.

R E S U LT S A N D D I S C U S S I O N

Polyacrylamide gel electrophoresis

Native and cross-linked HSP90 were analysed on 4±15%gradient gels (Fig. 1). Native PAGE analysis (Fig. 1A) ofnative HSP90 (lane 1) revealed the presence of two specieswith apparent molecular masses of 190 kDa (dimers) and90 kDa (monomers). Dimeric species were highly pre-dominant. For cross-linked HSP90, native PAGE showed abroad band between 170 and 180 kDa corresponding todimers (Fig. 1A, lane 2), indicating that EDC blocked themonomer±dimer equilibrium in favour of the dimeric state.Noteworthy the cross-linked dimer has an apparent lowermolecular mass than the non cross-linked one suggestingthat EDC slightly modifies HSP90 conformation. SDS/PAGE analysis of native HSP90 showed only one band withan apparent molecular mass of < 85±90 kDa (Fig. 1B,lane 1). SDS/PAGE of the cross-linked protein showed asmear with molecular mass of around 170±210 kDa(Fig. 1B, lane 2). Native PAGE analysis of HSP90 quan-tities ranging from 7 to 35 mg (Fig. 1C) revealed thepresence of oligomeric species.

Mass spectrometry of monomeric species

Figure 2 shows the MALDI mass spectrum of pig brainHSP90. Two peaks are observed corresponding to the doubly(M 1 2H)21 and singly (M 1 H)1 charged monomers.The base peak (M 1 H)1 is broad and clearly bimodal withthe most abundant ions centred on m/z 84 880 ^ 100 Daand a shoulder centred on 83 360 ^ 100 Da confirming

q FEBS 2001 Phosphorylation and oligomerization states of HSP90 (Eur. J. Biochem. 268) 2403

that HSP90 is a mixture of two isoforms [4], which are notresolved by PAGE.

If the molecules are very different they often exhibitdifferent ionization efficiencies and therefore it is not trivialto estimate the abundance of two unrelated compounds

present in the original sample by a simple comparison oftheir mass spectra signal intensities. This is not the case fortwo similar products such as protein isoforms and thereforewe believe that the a isoform of HSP90 is < six times moreabundant in the pig brain than the b isoform.

Although the theoretical mass of the larger isoform (a) isreported in databases, i.e. 84 643 Da without any post-translational modification, no information exists concerningthe pig b isoform, however, the mass of the b isoformof other mammalian species is already known: 83 163,83 194, and 83 185 Da for human, mouse, and rat, respec-tively. The masses determined by MALDI were higher thanthe theoretical masses indicating the presence of some post-translational modification. The accuracy of measurementand the possible overlapping between protein ion signalsand signals corresponding to the protein-sinapinic acidadducts did not permit a precise assessment of the natureand number of post-translational modifications.

As HSP90 is a protein with numerous basic amino acids,high charge states were produced by electrospray ionizationin positive ion mode (Fig. 3). The peaks observed in themass spectrum ranged from m/z 650 (z � 130) to m/z 1500(z � 55) with the most abundant ion at m/z 840.52(z � 101) (insert Fig. 3). Deconvolution of the massspectrum clearly demonstrated the presence of at leastfive different protein species (Fig. 3). The base peak(84792 ^ 15 Da) was part of a cluster of four peakswhere the mass difference between adjacent peaks was80 ^ 10 Da. A single phosphorylation of serine, threonineor tyrosine results in a mass increase of 80 Da. Thus, theobserved mass differences led us to suspect the presenceof the a isoforms with various phosphorylation states. Asmaller peak with a mass of 86 266 ^ 15 Da couldcorrespond to the phosphorylated b isoform. To investigatethe possibility of multiple phosphorylation, the protein was

Fig. 1. Polyacrylamide gel electrophoresis of HSP90. (A) Native

PAGE analysis of native (lane 1) and cross-linked HSP90 (lane 2). (B)

SDS/PAGE analysis of native (lane 1) and cross-linked HSP90 (lane 2).

(C) Native PAGE analysis of a range of HSP90 quantities. M, D and O

indicate monomeric, dimeric and oligomeric species, respectively.

Fig. 2. MALDI spectrum of HSP90 in insert enlargement of the

(M 1 H)1 peak.

Fig. 3. Deconvoluted electrospray spectrum of HSP90 before

phosphatase treatment. Insert: raw spectrum.

2404 C. Garnier et al. (Eur. J. Biochem. 268) q FEBS 2001

treated with lambda phosphatase and reanalysed by ESI-MS. Figure 4 shows the deconvoluted electrospray spec-trum of dephosphorylated HSP90. Two species weredetected, a major one with a mass of 84 646 ^ 15 Dacorresponding to the nonphosphorylated a isoform (theor-etical mass 84 643 Da), and a minor one with a mass of83 128 ^ 15 Da and corresponding to the non phosphory-lated b isoform. Mass differences confirmed that withoutphosphatase treatment we detected a isoforms with one tofour phosphates, the di-phosphorylated form being the mostabundant, and b isoforms with two phosphates. Wang et al.[25±27] have shown that one to three lysines could bemethylated on HSP70 and HSP90, in fact, our analysisshowed that in the main species zero or one amino acid wasmethylated, the experimental error of 15 Da led to thisuncertainty.

The ratio between the two isoforms was on average 6/1(a-HSP90/b-HSP90). Minami et al. [4] have shown thatHSP90 purified from mouse lymphoma cells consistedapproximately of 40% a isoform and 60% b isoform,our analysis showed that the ratio between the twoisoforms was tissue and species dependent and thata-HSP90 was more inducible than b-HSP90, moreconstitutively expressed.

Concerning the role of phosphorylation on HSP functionother studies have shown that HSP27 oligomerization andchaperone function are down regulated by phosphorylation[28] and HSP70 phosphorylation is significantly reduced byheat stress [29]. HSP90 phosphorylation plays an importantrole on the stimulation of eIF-2a kinase activity [15] andon the regulation of the interaction with pp60v-src [16].Therefore mass spectrometric analysis could be extended toHSP90 samples extracted from cells submitted to variousstresses to determine the phosphorylation state of HSP90 asa function of various stimuli.

Mass spectrometry of oligomeric species

Except in cases where the association constant is extremelyhigh [30], MALDI mass spectrometry is not suitable forobservation of noncovalently bound species. In order toobserve the oligomeric forms of HSP90, EDC was used tocross-link interacting species prior to MALDI massspectrometric analysis. Figure 5 shows a MALDI spectrumof HSP90 obtained after chemical cross-linking. We

observed a high percentage of dimeric form [peak(2M 1 H)11] but also a peak which could correspond tomonomeric species (main peak). One can argue that even ifthe cross linking was not complete and that noncovalentlybound dimer dissociated during the mass spectrometricanalysis, the main peak couldn't reflect the real abundanceof monomeric species in solution. There is two reasonsto this: (a) the higher is the mass, the less efficient isthe detection, (b) we believe that the main peak corre-sponded also to doubly charged dimeric species(2M 1 2H)21. This hypothesis was confirmed by the factthat we unambiguously detected abundant triply chargeddimeric species (2M 1 3H)31. Therefore due to the chargedistribution of species, abundant doubly charged dimericspecies (2M 1 2H)21 were also detected and overlappedwith monomeric species.

The peak corresponding to the dimer (2M 1 H)1 wascentred on 169 900 Da. The accuracy of measurement forsuch high masses was <^ 500 Da, taking that intoaccount, we believe that the peaks corresponded tophophorylated a±a homodimers (theoretical mass of thea±a homodimer 169 286 Da, 169 606 with four phos-phates), we did not see any evidence for the existence ofa±b heterodimers (considering the values obtained by elec-trospray for the b isoform, the mass of the a±b heterodimershould be around 167 771 Da without phosphate and168 091 Da with four phosphates), or b±b homodimers.This confirms previously reported results [5,31].

We also observed species with higher masses corre-sponding to trimers (3M 1 H)1 and tetramers (4M 1 H)1.To determine whether such oligomeric species reflectedbiological macromolecular complexes, or were instead

Fig. 4. Deconvoluted electrospray spectrum of HSP90 after treat-

ment with the l phosphatase.

Fig. 5. MALDI spectrum of HSP90 after crosslinking with EDC.

Species from the left to the right of the spectrum: Doubly charged

monomer (M 1 2H)21, triply charged dimer (2M 1 H)31, monomer

(M 1 H)1 and doubly charged dimer (2M 1 2H)21, dimer

(2M 1 H)1, trimer (3M 1 H)1, and tetramer (4M 1 H)1. In insert,

enlargement of the peak corresponding to the monomer and the doubly

charged dimer.

q FEBS 2001 Phosphorylation and oligomerization states of HSP90 (Eur. J. Biochem. 268) 2405

created in the gas phase during the MALDI process, thedimeric and higher order oligomers were separated by sizeexclusion chromatography. Elution fractions were tested bynative PAGE and checked by MALDI. Fractions containingdimers showed the existence of dimeric (2M 1 H)1 andmonomeric (M 1 H)1 forms and we did not detect thepresence of higher oligomers, in contrast, the fractionscontaining higher oligomers revealed the presence ofmonomeric and dimeric forms but also species whichcould correspond to trimers (measured mass 254 750 Da)and tetramers (342 650 Da). This analysis confirmed thatoligomers were present in our sample and were notartificially created in the gas phase. We also believe thatwe reached the limit of detection of our MALDI instrumentand therefore we were not able to detect higher oligomers.

In conclusion, we have shown that our mass spectro-metric analysis of HSP90 provides useful informationregarding post-translational modifications of HSP90. Bycombining ESI and MALDI, we have: (a) for the first time,determined the mass of the b-HSP90 isoform isolatedfrom pig brain, (b) shown that the two isoforms are mainlydi-phosphorylated, and (c) shown that the ratio of the a andb isoforms in the pig brain is on average 6 : 1. We alsoconfirmed the existence of a±a dimers the presence ofmonomers and higher oligomers. Further studies of thein vivo phoshorylation in response to various stimuli areenvisioned.

A C K N O W L E D G E M E N T S

We thank Drs P Barbier and Jean Michel Verdier for stimulating

discussions and critical reading of the manuscript. Dr J. Bonissel

(CNRS, Marseille) is acknowledged for MALDI analysis. We also

thank Ms S. Douillard and Mr G. Capriolo for technical assistance.

R E F E R E N C E S

1. Welch, W.J. & Brown, C.R. (1996) Influence of molecular and

chemical chaperones on protein folding. Cell Stress Chaperones 1,

109±115.

2. Hartl, F.U. (1996) Molecular chaperones in cellular protein

folding. Nature 381, 571±580.

3. Csermely, P., Schnaider, T., Soti, C., Prohazka, Z. & Nardai, G.

(1998) The 90-kDa molecular chaperone family: structure,

function, and clinical applications. Pharmacol. Ther. 79, 129±168.

4. Minami, Y., Kawasaki, H., Suzuki, K. & Yahara, I. (1991)

Analysis of native forms and isoform compositions of the

mouse 90-kDa heat shock protein, HSP90. J. Biol. Chem. 266,

10099±10103.

5. Nemoto, T. & Sato, N. (1998) Oligomeric forms of the 90-kDa

heat shock protein. Biochem. J. 330, 989±995.

6. Caplan, A.J. (1999) Hsp90s secrets unfold: new insights from

structural and functional studies. Trends Cell Biol. 9, 262±268.

7. Yonehara, M., Minami, Y., Kawata, Y., Nagai, J. & Yahara, I.

(1996) Heat-induced chaperone activity of HSP90. J. Biol. Chem.

271, 2641±2645.

8. Garnier, C., Protasevitch, I., Gilli, R., Tsvetkov, P., Lobatchov, V.,

Peyrot, V. & Briand, C. (1998) The two-state process of the heat

shock protein 90 thermal denaturation: effect of calcium and

magnesium. Biochem. Biophys. Res. Commun. 249, 197±201.

9. Chadli, A., Ladjimi, M.M., Baulieu, E.E. & Catelli, M.G. (1999)

Heat-induced oligomerization of the molecular chaperone Hsp90.

Inhibition by ATP and geldanamycin and activation by transition

metal oxyanions. J. Biol. Chem. 274, 4133±4139.

10. Prodromou, C., Roe, S.M., O'Brien, R., Ladbury, J.E., Piper, P.W.

& Pearl, L.H. (1997) Identification and structural characterization

of the ATP/ADP-binding site in the Hsp90 molecular chaperone.

Cell 90, 65±75.

11. Prodromou, C., Roe, S.M., Piper, P.W. & Pearl, L.H. (1997) A

molecular clamp in the crystal structure of the N-terminal domain

of the yeast Hsp90 chaperone. Nat. Struct. Biol. 4, 477±482.

12. Csermely, P. & Kahn, C.R. (1991) The 90-kDa heat shock protein

(hsp-90) possesses an ATP binding site and autophosphorylating

activity. J. Biol. Chem. 266, 4943±4950.

13. Csermely, P., Kajtar, J., Hollosi, M., Jalsovszky, G., Holly, S.,

Kahn, C.R., Gergely, P. Jr, Soti, C., Mihaly, K. & Somogyi, J.

(1993) ATP induces a conformational change of the 90-kDa heat

shock protein (hsp90). J. Biol. Chem. 268, 1901±1907.

14. Panaretou, B., Prodomou, C., Roe, S.M., O'Brien, R., Ladbury, J.,

Peper, P.W. & Pearl, L.H. (1998) ATP binding and hydrolysis are

essential to the function of the Hsp90 molecular chaperone in vivo.

EMBO J. 17, 4829±4836.

15. Szyszka, R., Kramer, G. & Hardesty, B. (1989) The phosphory-

lation state of the reticulocyte 90-kDa heat shock protein affects its

ability to increase phosphorylation of peptide initiation factor 2

alpha subunit by the heme-sensitive kinase. Biochemistry 28,

1435±1438.

16. Mimnaugh, E.G., Worland, P.J., Whitesell, L. & Neckers, L.M.

(1995) Possible role for serine/threonine phosphorylation in the

regulation of the heteroprotein complex between the hsp90 stress

protein and the pp. 60v-src tyrosine kinase. J. Biol. Chem. 270,

28654±28659.

17. Lees-Miller, S.P. & Anderson, C.W. (1989) The human double-

stranded DNA-activated protein kinase phosphorylates the 90-kDa

heat-shock protein, hsp90 alpha at two NH2-terminal threonine

residues. J. Biol. Chem. 264, 17275±17280.

18. Lees-Miller, S.P. & Anderson, C.W. (1989) Two human 90-kDa

heat shock proteins are phosphorylated in vivo at conserved

serines that are phosphorylated in vitro by casein kinase II. J. Biol.

Chem. 264, 2431±2437.

19. Park, M., Yong Kang, C. & Krishna, P. (1998) Brassica napus

hsp90 can autophosphorylate and phosphorylate other protein

substrates. Mol. Cell. Biochem. 185, 33±38.

20. Fenn, J.B., Mann, M., Meng, C.K., Wong, S.F. & Withehouse,

C.M. (1989) Electrospray ionization for mass spectrometry of

large biomolecules. Science 246, 64±71.

21. Karas, M. & Hillemkamp, F. (1988) Laser desorption ionization of

proteins with molecular masses exceeding 10,000 daltons. Anal.

Chem. 60, 2299±2301.

22. Loo, J.A. (1997) Mass spectrometry in the combinatorial

chemistry revolution. Eur. Mass. Spectrom. 3, 93±104.

23. Yonezawa, N., Nishida, E., Sakai, H., Koyasu, S., Matsuzaki, F.,

Iida, K. & Yahara, I. (1988) Purification and characterization of

the 90-kDa heat-shock protein from mammalian tissues. Eur. J.

Biochem. 177, 1±7.

24. Garnier, C., Barbier, P., Gilli, R., Lopez, C., Peyrot, V. & Briand,

C. (1998) Heat-shock protein 90 (hsp90) binds in vitro to tubulin

dimer and inhibits microtubule formation. Biochem. Biophys. Res.

Commun. 250, 414±419.

25. Wang, C., Gomer, R.H. & Lazarides, E. (1981) Heat shock

proteins are methylated in avian and mammalian cells. Proc. Natl

Acad. Sci. USA 78, 3531±3535.

26. Wang, C. & Lazarides, E. (1984) Arsenite-induced changes in

methylation of the 70,000 dalton heat shock proteins in chicken

embryo fibroblasts. Biochem. Biophys. Res. Commun. 119,

735±743.

27. Wang, C., Lin, J.M. & Lazarides, E. (1992) Methylations of

70,000-Da heat shock proteins in 3T3 cells: alterations by arsenite

treatment, by different stages of growth and by virus transfor-

mation. Arch. Biochem. Biophys. 297, 169±175.

28. Martin, J.L., Hickey, E., Weber, L.A., Dillmann, W.H. & Mestril,

R. (1999) Influence of phosphorylation and oligomerization on the

2406 C. Garnier et al. (Eur. J. Biochem. 268) q FEBS 2001

protective role of the small heat shock protein 27 in rat adult

cardiomyocytes. Gene Expr. 7, 349±355.

29. Cvoro, A., Dundjerski, J., Trajkovic, D. & Matic, G. (1999) The

level and phosphorylation of Hsp70 in the rat liver cytosol after

adrenalectomy and hyperthermia. Cell Biol. Int. 23, 313±320.

30. Moniatte, M., van der Goot, F.G., Buckley, J.T., Pattus, F. & Van

Dorsselaer, A. (1996) Characterisation of the heptameric pore-

forming complex of the Aeromonas toxin aerolysin using MALDI-

TOF mass spectrometry. FEBS Lett. 384, 269±272.

31. Nemoto, T., Ohara-Nemoto, Y., Ota, M., Takagi, T. & Yokoyama,

K. (1995) Mechanism of dimer formation of the 90-kDa heat-

shock protein. Eur. J. Biochem. 233, 1±8.

q FEBS 2001 Phosphorylation and oligomerization states of HSP90 (Eur. J. Biochem. 268) 2407