Vol.
176,
May
15, 1991
No.
3, 1991
BIOCHEMICAL
AND
RESEARCH
COMMUNICATIONS
1338-1344
Pages
ASSOCIATION
OF HEAT SHOCK PROTEIN 90 WITH THE 16 kDa STEROID CORE FRAGMENT OF RAT GLUCOCORTICOID RECEPTORS Pradip K. Chakraborti
Steroid Hormones
Received
BIOPHYSICAL
March
BINDING
and S. Stoney Simons, Jr.’
Section, NIDDK, Bldg. 8, Rm B2A-07, National Institutes of Health, Bethesda, MD 20892
22,
1991
Summary: We have recently described a 16 kDa steroid binding core (Thrs,-/Args73) of the rat glucocorticoid receptor [Simons et al. (1989) J. Biol. Chem. 264, 14493-l 44971. Sedimentation analysis and size exclusion and anion exchange chromatography now suggest that other proteins are associated with the 16 kDa receptor, just as has been seen for the intact 98 kDa receptor. The 16 kDa fragment was also immunoprecipitable with anti-heat shock protein 90 (hsp90) antibody. These results argue that hsp90 binds to the 16 kDa core fragment and directly position the site of hsp90 association between Thr5s7 and Args7s of the rat glucocorticoid receptor. Q 1991Academic Press, Inc.
High affinity steroid binding to glucocorticoid and unactivated
complexes,
receptor molecules
(l-4).
which appear to be tightly associated with several nonHeat shock protein 90 (hsp90) is one such molecule
Activation of the receptor-steroid and DNA is accompanied to a monomeric and endogenous
receptors occurs only with steroid-free
complex to a species with increased affinity for nuclei
by dissociation
form (6, 7). Molybdate
of this heterooligomeric
form of the receptor
and other group VIA transition
metal oxyanions,
factors (7, S), are thought to stabilize both the association
with the receptor and the steroid binding activity. steroid receptors (9), which has heightened association.
(2, 5).
of hsp90
Hsp90 also binds to all of the other
the interest in defining the site(s) of hsp90
From indirect studies with deletion mutant receptors, it has been
concluded that hsp90 binds within the steroid binding domain of the rat glucocorticoid receptor between amino acids 568 to 671 (10-12). As most non-receptor less sensitive to proteolysis limited proteolysis
(13, 14), glucocorticoid
have been examined
0006-291X/91 Copyright All rights
should be addressed.
$1.50
0 1991 bv Academic Press, Inc. of reproduction in any form reserved.
receptor fragments obtained
by
and found to be associated with hsp90 (1, 5).
However, there has been no direct demonstration * To whom correspondence
proteins are
1338
that hsp90 can associate with any
Vol.
176,
No.
BIOCHEMICAL
3, 1991
AND
BIOPHYSICAL
RESEARCH
receptor fragment smaller than the 27-30 kDa mero-receptors
COMMUNICATIONS
(2, 5). We have recently
described a small 16 kDa steroid binding core fragment (Thr5s7-Args7.J glucocorticoid
of the rat
receptor (15), which includes all of the suspected hsp90 binding site(s).
We therefore investigated
whether hsp90 was associated with this 16 kDa core. We
now report direct evidence for the association
of macromolecules,
and hsp90 in
particular, with the native form of this 16 kDa receptor fragment. MATERIALS
AND METHODS
Unless mentioned otherwise, all operations were carried out at 0°C. Chemicals and buffers: [3H]Dexamethasone (Dex; 40, 46 or 47 CVmmol) and [3H]dexamethasone 21 -mesylate (Dex-Mes; 49.9 Ci/mmol) were from Amersham Corp. and Du Pont-New England Nuclear respectively. Non-radioactive Dex and trypsin (TPCK-treated) (Sigma), Pansorbin, cytochrome C and HEPES (Calbiochem), DEAE-cellulose (Whatman), Sephadex G-100 (40-120~ mesh), TAPS (Ultrol grade, Behring Diagnostics), bacterial alkaline phosphatase (Worthington), molecular weight markers (Pharmacia-LKB Biotechnology), reagents for polyacrylamide gel electrophoresis (BioRad), and Ult-Emit autoradiography marker (Du Pont-New England Nuclear) were commercially available. All [3H]labeled samples were counted in Hydrofluor (National Diagnostics) at 40-55% counting efficiency in a Beckman 5801 liquid scintillation counter with automatic cpm to dpm conversion. TAPS buffer contained 25 mM TAPS, 1 mM EDTA and 10% glycerol (pH 8.8 or 9.5 at OOC). Preparation, labeling, and activation of receptors: The growth of HTC cells, the preparation of cytosol, and [3H]Dex binding to and [3H]Dex-Mes labeling of receptors have been described (16). 16 kDa fragments were obtained by treating steroid-free receptor solutions (30 or 60% v/v) + 20 mM sodium molybdate or tungstate (17) with trypsin (14-20 or 40pg/ml) for 1 hr followed by a lo-fold excess (wVwt) of soybean trypsin inhibitor to prevent further digestion (15). Steroid-bound or affinity labeled 16 kDa fragments were activated in the absence of sodium molybdate (or sodium tungstate) by heating at 20°C for 10 min, which is optimal for the activation of 98 kDa complexes (data not shown). The authenticity of 16 kDa fragments was ascertained by fluorography of Dex-Mes labeled receptors analyzed by SDS-PAGE (15, 18). RESULTS When unactivated, graphed on Sephadex molecular
AND DISCUSSION
[3H]Dex-bound
16 kDa receptor fragments
G-100 columns,
weight of 2150,000,
were chromato-
which exclude material with an apparent
52 & 18% (mean f SD, n = 3) of the added complexes
eluted in the void volume (fractions 7-l 2 of Fig. 1). Activation causes a dramatic decrease in apparent unactivated
molecular weight and sedimentation
98 kDa receptor-steroid
complexes
as a result of dissociation
receptor proteins such as hsp90 to give a monomeric being subjected to activating conditions,
coefficient of most of the of non-
complex (7, 19). Similarly,
the amount of 16 kDa complexes
after
eluting in
the void volume decreased to 18 f 7% (mean f SD, n = 3) (Fig. 1). In the absence of a functional assay for the “activation”
of 16 kDa fragments, it is not known whether this 1339
Vol.
176,
No.
BIOCHEMICAL
3, 1991
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
0 5
0 Fraction
15
10
Top Fraction
number
number
Fig. 1. Sephadex G-100 elution profile of 16 kDa core complexes. Aliquots (lOOpI) of unactivated (0) or “activated” (0) 16 kDa fragments bound with rH]Dex f excess nonradioactive Dex were chromatographed on Sephadex G-100 columns (10cm x lcm, bed volume = 7ml) and eluted with pH 8.8 TAPS buffer. The specifically bound dpm in five drop (280~1) fractions were determined by subtracting the values of competed from uncompeted samples and plotted as percentage of applied specifically bound dpm (determined by dextran-coated charcoal assay) in order to correct for different amounts of added complexes (“activated” complexes -29% of unactivated complexes). In the peak fraction (#9), the dpm for the non-competed and competed solutions were 3315 and 238 respectively for unactivated complexes (total added dpm = 24,346) and 328 and 87 for “activated complexes (total added dpm = 6535). The void volume (V,) was determined by exclusion of blue dextran and encompassed fractions 7-12; phenol red appeared in fractions 16-25. Fig. 2. Sucrose gradient sedimentation profile of 16 kDa core complexes. Aliquots (200~1) of unactivated (0, containing 20 mM sodium tungstate) or “activated”(o) 16 kDa fragments bound with i3H]Dex f excess nonradioactive Dex were centrifuged (2hrl OW334,OO xg/Beckman VTi 80 rotor) on linear gradients (4.6 ml) of 5-20% sucrose in pH 7.4 (r.t) phosphate buffer (+20 mM tungstate for unactivated complexes). Fractions (266~1) were collected from the bottom. The specifically bound dpm in 1OOul of each fraction was determined as in Fig. 1. In the peak fraction (#9). the dpm for the noncompeted and competed solutions were 1553 and 67 respectively for unactivated complexes (total added dpm = 32,937) and 250 and 146 for “activated” complexes (total added dpm = 10,043). The position of internal markers (AP = bacterial alkaline phosphatase, C = cytochrome C) was determined as described (20). represents
maximal “activation”.
seen in the included dissociation
No peak of dissociated,
volume for unactivated
from 16 kDa complexes
monomeric
or “activated”
16 kDa complexes.
during the 2 hrs of these experiments
major factor since the t,,,, oefore and after being subjected was 5.7 hr and -4.3 hr respectively
(data not shown).
the included volume may be due to dispersion, complexes
during chromatography. prior to chromatography
increase
in the amount of radioactivity
kOa complexes
of “activated”
Nevertheless,
has an apparent
to activating
Thus the absence
or sticking
In one experiment
removed
visible (data not shown).
complexes
to the column,
where
[3H]Dex-bound
was Steroid
was not a conditions, of a peak in of the
the free steroid was 16 kDa fragments,
in the included volume (fractions
13-18) was
it is obvious that the majority of unactivated
molecular
weight
1340
an
of greater than 150,000 daltons.
16
Vol.
176,
No.
BIOCHEMICAL
3, 1991
The peak of [3H]Dex-bound gradients to -4s
shifted from -7.8s
to -2.4s
after “activation”
at -7.8s.
RESEARCH
The specifically
complexes
bound radioactivity
centrifugation.
to bovine albumin, which sediments
be calculated
by the equation
on Sephadex
16 kDa fragment
the “unactivated”
16 kDa receptor
G-100 columns
is >118,000.
Collectively,
complexes
charged
complexes
of 16 kDa fragments
protein like hsp90 determines
antibody
sera.
respectively
The same antibody
stabilized
columns charged
molybdate-stabilized
of that
is thought (24).
The
16 kDa and of a
the elution of the macromolecular
at high salt on DEAE-cellulose
immunoadsorbed
it can
species.
hsp90 (22, 23), which is negatively
columns.
3.6 f 0.3% (mean f SD, n = 3) and 4.7 +
1.7% (mean f SD, n = 4) of [3H]Dex-bound, complexes
weight
was almost identical (Fig. 3). This argues that the association
negatively
Anti-hsp90
(21) and is
these data establish
on ion exchange
effect of salt on the DEAE binding of [3H]Dex-bound, 98 kDa complexes
at 4.4s
during
(data not shown),
does not exist as a monomeric
The elution of 98 kDa unactivated by the associated
from 16 kDa fragments
of Siegel and Monty (21) that the molecular
the “unactivated”
to be dictated
in Fig. 2
at the top of the gradient is
of associated
eluted after the 16 kDa fragment
(7, 19; data not
complexes
most likely due to dissociation By comparison
density
(Fig. 2). A similar shift from -9s
bound, unactivated
proteins
COMMUNICATIONS
on 5-20% sucrose
of intact 98 kDa receptor
About 15-20% of the specifically
sedimented
BIOPHYSICAL
16 kDa core fragments
is seen after the activation
shown).
AND
after subtraction
molybdate-stabilized of the background
also selectively
98 kDa and 16 kDa receptors
adsorbed
16 kDa and 98 kDa
obtained
with pre-immune
[3H]Dex-Mes-labeled,
(lanes l-6 and 7-12 respectively
KCI
concentration
molybdatein Fig. 4).
(mM)
Fi . 3. Effect of KCI on binding of 98 kDa and 16 kDa complexesto DEAE-cellulose. -P[ HJDexk excess nonradioactive Dex-bound samplesof 98 kDa (A) or 16 kDa (0) receptors were applied onto minicolumns(3OOr.tlof DEAE-cellulose),washed with pH 6.8 TAPS containing the indicated amount of KCI, and counted (25). The dpm specifically bound to DEAE-cellulosewere determinedas in Fig. 1 and plotted as percentage of applied specifically bound dpm. 1341
Vol.
176,
No.
BIOCHEMICAL
3, 1991
AND
BIOPHYSICAL
RESEARCH
16 kDa receptor
96 kDa receptor Antbdy.
Premmune serum
Buffer --
Nonradfoactrve
Dex: -
+
-
+
1
2
3
4
lane
COMMUNICATIONS
Ant!-hsp90 arm&y -
5
Prelmmune Serum
Buffer ---
+
-+-+-
6
7
Ant\-hspW antbody +
6
9
10
11
12
Fig. 4. Fluorographs of 13H]Dex-Mes labeled (A) 98 kDa and (B) 16 kDa receptors immunoprecipitated by anti-hsp90 antibody. Unactivated 98 kDa or 16 kDa complexes labeled with [3H]Dex-Mes f excess nonradioactive Dex were incubated with buffer, pre-immune sera, or anti-hsp90 antibody (diluted 1:6; prepared against a C-terminal peptide of human heat shock protein(s) (26) for 2 hr followed by precipitation using Pansorbin, analysis by SDS-PAGE, and fluorography as described (14, 27). The positions of the molecular weight standards (P = phosphotylase b, M, = 97.4 kDa; B = bovine serum albumin, M, = 66.3 kDa; 0 = ovalbumin, M, = 45 kDa; C = carbonic anhydrase, M, = 30.6 kDa; S = soybean trypsin inhibitor, M, = 21.5 kDa and L = a-lactalbumin, M, = 14.4 kDa) are indicated. The arrows designate the positions of the 98 and 16 kDa receptors. The weaker signal for 16 kDa fragments is due to the fact that the yield of 16 kDa fragments under these labeling conditions is only -20% of that for the intact 98 kDa receptors (15).
The yield of 16 kDa fragments,
as quantitated
by slice-and-counting
0.9% (mean f SD, n = 3; = 1.5% with pre-immune were
reported
receptors
for the immunoadsorption
with anti-hsp90
concentrations
antibody
of hsp90 (l-2%
sera).
of unactivated
(28), presumably
of cytosol
of gels, was 4.9 +
Similar low yields (-5%) mouse L-cell glucocorticoid
due to the high cytosolic
protein) (29).
Nevertheless,
these results
show that heat shock protein binds to the 16 kDa core. The antibody (hsp70)
used in these experiments
and hsp90 (26). However
hsp70 was associated 30). Furthermore,
antibody
hsp90 is associated
both heat shock
hsp90 is associated
with the 98 kDa glucocorticoid
Therefore,
facts that the 16 kDa fragment
the data strongly
This conclusion
is reinforced
(7, 9) by this
argue that by the
binds Dex with good affinity (15) and that the associa-
tion of hsp90 is required for steroid binding to both the 98 kDa receptor 27 kDa mero-receptor
receptor
were immunoadsorbed
collectively,
with the 16 kDa fragment.
that no
receptor from mouse L cells (12,
of 98 and 16 kDa complexes
(see above).
protein 70
Pratt et al. found, with this same antibody,
with the intact glucocorticoid
and the same percentage anti-hsp90
recognizes
(2). Thus the 16 kDa core is the smallest
which both steroid
binding and association
These conclusions
in turn further support the current hypothesis
(2, 31) and the
receptor sequence
for
with hsp90 have been directly observed.
for the high affinity binding of steroid to glucocorticoid 1342
receptors
that hsp90 is required (2, 7, 31).
Vol.
176,
No.
3, 1991
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
In addition to the hsp90 binding site(s), the 16 kDa core fragment is stabilized by molybdate
(15) and contains two “vicinal thiols” that are involved in steroid binding to
the glucocorticoid of the elements
receptor (14, 16, 32, 33). Thus the 16 kDa fragment contains many of the intact 98 kDa receptor that are involved in steroid binding.
Further studies with the 16 kDa core promise to yield valuable information the interactions
of steroid with the binding cavity of the glucocorticoid
concerning
receptor.
ACKNOWLEDGMENTS We thank Dr. Ettore. Appella (NCI, NIH) for the gift of anti-hsp90 antibody, Mr. Hung Luu for excellent technical assistance, Dr. William. 6. Pratt (Univ. of Michigan) for helpful discussions and for sharing his unpublished results, and Dr. Alice H. Cavanaugh for critical review of the paper. REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.
Gehring, U, and Arndt, H (1985) FEBS Letters 179,138-142. Bresnick, EH, Dalman, FC, Sanchez, ER, and Pratt, WB (1989) J. Biol. Chem. 264, 4992-4997. Bresnick, EH, Dalman, FC, and Pratt, WB (1990) Biochemistry 29, 520-527. Nemoto, T, Ohara-Nemoto, Y, Denis, M, and Gustafsson, J-A (1990) Biochemistry 29, 1880-l 886. Denis, M, Gustafsson, JA, and Wilstrom, AC (1988) J. Biol. Chem. 263, 1852018523. Mendel, DB, Bodwell, JE, Gametchu, B, Harrison, RW, and Munk, A (1986) J. Biol. Chem. 261, 3758-3763. Pratt, WB (1987) J. Cell. Biochem. 35,51-68. Bodine, PV, and Litwack, G (1988) J. Biol. Chem. 263, 3501-3512. Catelli, MG, Binart, N, Jung-Testas, I, Renoir, JM, Baulieu, EE, Feramisco, JR, and Welch, WJ (1985) EMBO J. 4,3131-3135. Pratt, WB, Jolly, DJ, Pratt, DV, Hollenberg, SM, Giguere, V, Cadepond, FM, Schweizer-Groyer, G, Catelli, M-G, Evans, RM,and Baulieu, E-E (1988) J. Biol. Chem. 263, 267-273. Howard, KJ, Halley, SJ, Yamamoto, KR, and Distelhorst, CJ (1990) J. Biol. Chem. 265, 11928-l 1935. Dalman, FC, Scherrer, LC, Taylor, LP, Akil, H, and Pratt, WB (1991) J. Biol. Chem. 266, 3482-3490. Reichman, ME, Foster, CM, Eisen, LP, Eisen, HJ, Torain, BF, and Simons, SS Jr (1984) Biochemistry 23, 5376-5384. Chakraborti, PK, Hoeck, W, Groner, B, and Simons, SS Jr (1990) Endocrinology 127, 2530-2539. Simons, SS Jr, Sistare, FD, and Chakraborti, PK (1989) J. Biol. Chem. 264, 14493-l 4497. Miller, NR, and Simons, SS Jr (1988) J. Biol Chem. 263, 15217-15225. Rafestin-Oblin, ME, Lombes, M, Lustenberger, P, Blanchardie, P, Michaud, A, Cornu, G, and Claire, M (1986) J. Steroid Biochem. 25, 527-534. Simons, SS Jr (1987) J. Biol. Chem. 262, 9669-9675. Holbrook, NJ, Bodwell, JE, Jeffries, M, and Munck, A (1983) J. Biol. Chem. 258, 6477-6485. Yamamoto, KR (1974) J. Biol. Chem. 249, 7068-7075. Sherman, MR, Moran, MC, Tuazon, FB, and Stevens, Y-W (1983) J. Biol. Chem. 258, 10366-l 0377. 1343
Vol.
22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33.
176,
No.
3, 1991
BIOCHEMICAL
AND
BIOPHYSICAL
RESEARCH
COMMUNICATIONS
Sanchez, ER, Meshinchi, S, Tienrungroj, W, Schlesinger, MJ, Toft, DO, and Pratt, WB (1987) J. Biol. Chem. 262,6986-6991. Nemoto, T, Ohara-Nemoto, Y, Kurokawa, R, Sato, M, and Cta, M (1988) Biochem. Int. 16,973-981. Mendel, BD, and Orti, E (1988) J. Biol. Chem. 263, 6695-6702. Cavanaugh, AH, and Simons, SS, Jr (1990) Biochemistry 29,991-996. Ehrhart, JC, Duthu, A, Ullrich, S, Appella, E, and May, P (1988) Oncogene 3, 595-603. Urda, LA, Yen, PM, Simons, SS Jr, and Harmon, JM (1989) Mol. Endocrinol. 251, 251-260. Sanchez, ER, Toft, DO, Schlesinger, MJ, and Pratt, WB (1985) J. Biol. Chem. 260, 12398-12401. Riehl, RM, Sullivan, WP, Vroman, BT, Bauer, VJ, Pearson, GR, and Taft, DO (1985) Biochemistry 24, 6586-6591. Sanchez, ER, Hirst, M, Scherrer, LC, Tang, H-Y, Welsh, MJ, Harmon, JM, Simons, SS Jr, Ringold, GM, and Pratt, WB (1990) J. Biol. Chem. 265, 20123201 30. Housley, PR (1990) Biochemistry 29, 3578-3585. Simons, SS Jr, Chakraborti, PK, and Cavanaugh, AH (1990) J, Biol. Chem. 265, 1938-1945. Lopez, S, Miyashita, Y, and Simons, SS Jr (1990) J. Biol. Chem. 265, 1603916042.
1344