Structure and subunit dissociation of the mouse glucocorticoid receptor: Rapid analysis using vertical tube rotor sucrose gradients

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 230, No. 1, April, pp. 274-234, 1934

Structure and Subunit Dissociation of the Mouse Glucocorticoid Receptor: Rapid Analysis Using Vertical Tube Rotor Sucrose Gradients’ SARAH Department

B. EASTMAN-REKS, of Biochentistry,

Louisiana

Received September

CHERYL State

E. REKER,

University

Medical

AND Centa;

WAYNE New

6, 1983, and in revised form November

V. VEDECKIS’

0rleans,

Louisiana

70112

15, 1933

The structure and subunit dissociation of the glucocorticoid receptor from the mouse AtT-20 pituitary tumor cell line was analyzed on sucrose gradients using a Beckman VTI 80 vertical tube rotor. This technique afforded a very rapid analysis (65 min) of the variously sedimenting forms compared to swinging-bucket rotor sucrose gradients, which take 16 h to run. Thus, it was possible to detect and study the molybdatestabilized, oligomeric, untransformed receptor (9.1 S) in the presence of 0.3 M KCl. Under similar conditions using the swinging-bucket rotor, only the monomeric, transformed species (3.8 S) was observed. That is, artifactual subunit dissociation was minimized using the vertical tube rotor, allowing the study of the receptor structure in a more native state. Further studies demonstrated that Sephadex LH-20 chromatography causes receptor transformation. Thus, dextran-charcoal adsorption is preferred for the removal of unbound hormone under certain circumstances. Finally, using vertical tube rotor sucrose gradients, it was determined that the transformation of the mouse AtT20 glucocorticoid receptor involves a conversion of the oligomeric, 9.1 S, untransformed species to a 5.2 S, transformed moiety. This suggests that the 5.2 S, intermediate transformed species may be the physiologically relevant form of this gene regulatory protein. that the untransformed receptor had a sedimentation coefficient of 9 S, while two transformed species were obtained, with values of 5 S and 3.2 S (3). It was proposed that mouse GC-R transformation involves the dissociation of a 9 S, homotetrameric, untransformed receptor to a 5 S, transformed homodimer and/or a 3.2 S, transformed monomer. Alternatively, non-hormone-binding proteins, receptor binding factors (4-ll), or RNA (12-14) could be components of the 5 S and/or 9 S complexes. Since sucrose gradient ultracentrifugation is the best technique for resolving these three receptor forms, it was the method of choice for further studies on receptor transformation, including studies on the reversal of transformation and reconstitution experiments. Ever since its initial application to the study of steroid hormone receptors (15), sucrose gradient ultracentrifugation has been the most popular method of physi-

Previous studies from our laboratory (l3) have focused on the structure of the mouse glucocorticoid receptor (GC-R),3 and the mechanism of receptor transformation (conversion from a nonnuclear and nonDNA-binding form to species which do bind)! During these studies it was found 1 This research was supported by Grant AM-27033 from the National Institutes of Health, Grant BC436 from the American Cancer Society, and a grant from Cancer Crusaders of New Orleans. W.V.V. is a recipient of an NIH Research Career Development Award. a To whom all correspondence should be addressed. a Abbreviations used: Dex, dexamethasone (go-fluoro - 16 -o-methyl - 11&17,21- trihydroxypregna1,4diene-3,20-dione); GC-R, glucocorticoid receptor. ‘As suggested by Pratt (64) we will use the term transformation to indicate the change in receptor form accompanying its conversion to a species which binds to nuclear constituents. The term activation will be reserved for the conversion of the receptor from a form which is incapable of binding ligand to one which can bind hormone. 0003-9861/34 33.00 Copyright All rights

0 1984 by Academic Press, Inc. of reproduction in any form resewed.

274

GLUCOCORTICOID

RECEPTOR

cochemical analysis of these proteins. It has been used to draw conclusions about the process of receptor transformation, receptor presence and levels in various cells and tissues, and neoplastic conditions (16). The major drawback of this technique has been the extended time required for analysis. Typically, 16-18 h of centrifugation time is required for the efficient separation of receptor forms using swinging-bucket rotors. With the advent of vertical tube rotors it was hoped that sucrose gradient ultracentrifugation would be a much more desirable technique, since run times could be reduced to between one-fifth and onethirteenth the time. Besides obtaining data more rapidly, it was thought that the physical form of the receptor after the run would more closely resemble that layered on the gradient. For example, steroid hormone receptors are transformed in a dilution- and time-dependent manner (reviewed in (17, 18)). Since receptors are diluted five- to eightfold during a centrifugation which requires 16 h, it was not clear if the form obtained at the end of the analysis was the same as when it was begun. Despite these advantages, vertical tube rotor sucrose gradient ultracentrifugation has been utilized much less frequently than swinging-buckets in the study of steroid hormone receptors (14,19-28). One possible reason for this is that it has been claimed that this technique results in poor resolution of various receptor forms (25). Described below are conditions which yield very good resolution of the multiple forms of the mouse glucocorticoid receptor. Using this rapid technique we have characterized the sedimentation properties of the mouse AtT-20 GC-R, and alterations in structure which occur during receptor transformation. MATERIALS

AND

METHODS

Chemicals. Sucrose was “Ultra-Pure, Density Gradient Grade” from Schwa&Mann, and Tris was “Ultra-pure” from the same company. [1,2,4,6,7-3HJDexamethasone (Dex), 82 Ci/mmol, was obtained from Amersham. Molybdic acid, sodium salt (Sigma) was prepared as a 1 M stock solution, pH 7.0, in distilled water. Cytochrome c (1.7 S), ribonuclease A (1.9 S),

STRUCTURE

275

chymotrypsinogen A (2.6 S), ovalbumin (3.5 S), bovine serum albumin (4.4 S), aldolase (7.9 S), glucose oxidase (7.9 S), sweet potato @-amylase (9.4 S), and catalase (11.3 S) were all from Sigma Chemical Company. Ironfree human transferrin (4.9 S) was from Behringwerke AG (Marburg, Federal Republic of Germany). Standard proteins (1 mg/ml in 10 mM Tris-HCl, pH 7.4, at 25°C) were reductively methylated with r”C)formaldehyde, essentially as described by Jentoft and Dearborn (29). Briefly, sodium cyanoborohydride (Sigma Chemical Co.; 1 M in distilled water) was added to the protein solutions to a final concentration of 20 mM. The stock r%lformaldehyde (53 mCi/mmol; New England Nuclear) was diluted with 0.1 M radioinert formaldehyde to a final concentration of 0.112 M (final specific activity = 8.1 mCi/mmol). Ten microliters of this was then added to each milliliter of protein solution, followed by incubation overnight at room temperature. The samples were then dialyzed at 0-4°C versus 160 vol of TETg buffer, with two buffer changes, over 24 h. All other chemicals were reagent grade from J. T. Baker. CeZZculture The mouse AtT-20 pituitary tumor cell line was maintained in suspension cultures in l-liter spinner flasks as described previously (1). Cgtosolprepmtion Cell suspensions were harvested and washed with Tris-saline (10 mM Tris-HCl, pH 7.4, at 25”C, 0.148 M NaCl) as described previously (1). Two volumes of TETg buffer (20 mM Tris-HCl, pH 7.4, at 25°C 1 mhl NaaEDTA, 12 mM 1-thioglycerol added fresh daily) were added to the cellular pellet. Cells were swollen for 30 min at 0-4°C and then homogenized with five strokes in a stainless-steel Dounce homogenizer (Kontes). The homogenates were centrifuged at 190,0008,, at 2°C for 60 min using an SW 50.1 rotor. The lipid layer was removed by aspiration and the remaining supernatant was labeled with 2.4 X lo-* M PHjdexamethasone for 2 h. Vertical tube rotor w,crose gradient ultracent+gation Linear sucrose gradients (5.2 ml, 5-20s) were prepared in TETg buffer with 0.3 M KCI, 20 mM NarMoOd, or 0.3 M KC1 plus 20 mM NaaMoO, (KCl/ MOO:-). Gradients were prepared at room temperature in Beckman Quick-Seal tubes, using a Buchler gradient maker (Model 2-5104) and chilled at 0-4°C for 2 h. A 200-~1 volume of sample was applied to each tube. The gradients were centrifuged in a Beckman VTi 80 rotor and Beckman L8-80 preparative ultracentrifuge at 80,000 rpm, 2”C, to a preset w*t = 2.40 X lo”, or at 65,009 rpm, 2°C to the same w*t. At both centrifugation speeds the time of acceleration of speed was approximately 12 min. The time required to obtain an & of 2.40 X 10” was 65 and 93 min for the 80,000 and 65,000 rpm runs, respectively. Not included in this time is the deceleration time (7 min), which, when included, gave an accumulated & of 2.44 X 10”. All runs were performed using the “slow

276

EASTMAN-REKS,

REKER,

acceleration” mode with the brake on. Gradients were tapped using a Buchler fractionating device (Model 2-5010) and analyzed by liquid scintillation counting. A 4-ml volume of Beckman Ready-Solv EP cocktail was added to each l&drop fraction (0.2 ml), and the whole was counted in a Beckman LS 7500 liquid scintillation spectrophotometer at an efficiency of about 35%. swinging-bucm sucrose gradient cenh$b&ion, Sucrose gradients (520%) were prepared as described above. Gradients (5.0 ml) were layered with 200 pl of sample. Centrifugation was performed in an SW 50.1 rotor at 45,000 rpm 2°C for 16 h. Sucrose gradients were tapped, and 11-drop fractions were collected and analyzed as described above. LH-20 chromatography. Sephadex LH-20 (Pharmacia) columns (4 ml packed volume) were equilibrated in TETg or TETg-20 mr.i NazMoO,. Cytosol was applied and 0.5-ml fractions were collected. Aliquots (20 al) were analyzed by liquid scintillation counting to elucidate the location of the eluted glucocorticoid receptor. Receptor peak fractions, which are excluded from this column, were pooled for subsequent analysis. Llextmn-charcoal assay. Individual tubes containing 300 pl of charcoal suspension (2% Norit-A, 0.5% Dextran T-70, 10 mM Tris-HCl, pH 7.5, at 25°C 1 rnM EDTA) were centrifuged at 2000 rpm, 2°C for 5 min. The supernatant was then carefully removed. A 300pl volume of cytosol was added to the charcoal pellets and gently shaken for 5 min. Samples were centrifuged again at 2000 rpm, 2°C for 5 min. A 200-~1 volume of the supernatant was then applied to the sucrose gradients. Cytosol treutment. Following 60 min of incubation with Dex, aliquots of cytosol were treated with 20 rnM NazMoO,. Sixty additional minutes of incubation followed. This procedure stabilizes the untransformed species of the receptor (l-3,30-33). To achieve receptor transformation, aliquots of cytosol were either heated at 25°C for 1 h or pretreated with 0.3 M KC1 at O4°C for 1 h. RESULTS

Sucrose Graoknt Aru&sis of Protein Standards 14C-Methylated protein standards were centrifuged on 5-20s sucrose gradients in the VTi 80 or SW 50.1 rotor under the conditions described under Materials and Methods for cytosol samples. With the vertical tube rotor the standards migrated in a linear fashion in relation to their sedimentation values (data not shown). No significant differences in sedimentation

AND

VEDECKIS

position of the standards were obtained on MoOi- or KCl/MoOi- gradients. Sedimentation Properties of the Glucocorticoid Receptor Passed over Sephadex LH-20 Initial experiments utilized Sephadex LH-20 chromatography to separate the receptor-bound radioactivity from the free hormone. This had proven valuable in previous studies on receptor labeled with radioactive triamcinolone acetonide (l-3,39). Identical AtT-20 cytosol samples were then analyzed by sucrose gradient ultracentrifugation using both the standard SW 50.1 rotor (16 h; 45,000 rpm) and the VTi 80 vertical tube rotor (65 min; 80,000 rpm). These results are shown in Fig. 1. As had been found previously (3), centrifugation using the SW 50.1 swingingbucket rotor revealed a 3.2 S receptor peak on high salt (0.3 M KCI) gradients, and a 5 S peak on low salt gradients (Fig. 1D). Both of these receptor species have been designated as transformed (DNA-binding) forms of the receptor (1, 3). Molybdatetreated cytosol was passed over an LH-20 column which was equilibrated in 20 mM NazMo04. A portion of this sample was centrifuged on a sucrose gradient containing 20 mM NazMo04 (Fig. 1F). Again, as had been shown before for the AtT-20 cell GC-R (3), as well as for other steroid hormone receptors (34,37,38,40-45), a large, approximately 9 S form was obtained. Finally, the Moo:--stabilized receptor was either pretreated with 0.3 M KC1 for 1 h or centrifuged immediately on KC11 MOO;- gradients. In both cases, a 3.2 S transformed GC-R form was obtained (Fig. 1E). This is in contrast to previous experiments (3) in which the 5 S receptor form was obtained on KCl/Mo0,2- gradients. It has been found that the form of the GCR obtained under these conditions varies considerably from experiment to experiment. This may be due to slight variations in cytosol preparations and the long time required for centrifugation (see below). When the same samples were centrifuged using the vertical tube rotor, the high salt and low salt gradients gave results

GLUCOCORTICOID

RECEPTOR

VTI 80 3

TOP

10

20

TOP

10

20

FRACTION

FIG. 1. Sucrose gradient ultracentrifugation of the mouse AtT-20 pituitary tumor cell line glucocorticoid receptor after removal of free hormone by Sephadex LH-20 chromatography. Identical samples were analyzed by sucrose gradient ultracentrifugation using the VTi 30 vertical tube rotor (65 min run time), (A, B, and C) or the SW 50.1 swinging-bucket rotor (16 h run time) (D, E, and F). Samples (A, D) were chromatographed over an LH-20 column equilibrated in TETg buffer, the receptor-containing peak fractions were pooled, and the receptor was then centrifuged on low salt gradients (O), or gradients containing 0.3 M KC1 (0). Cytosol which was made 20 mM in NazMoOb was chromatographed on an LH-20 column equilibrated in TETg-20 mM NazMoO,. The pooled receptor peak fractions were then either left untreated (M, A), or were made 0.3 M in KC1 (using a TETg-1 M KC1 stock solution) and then incubated at 0-4°C for 1 h (0 = KCl-pretreated). Samples (m, 0) were then centrifuged on sucrose gradients made up in TETg-0.3 M KCl-20 mrd NazMoO, (B, E), or on gradients made up in TETg-20 mM NazMoO, (A). The arrow indicates the sedimentation position of human transferrin (4.9 S). The remainder of the details are given under Materials and Methods.

STRUCTURE

277

similar to those obtained with the swinging-bucket rotor (Fig. 1A). However, careful determinations of the sedimentation coefficients revealed slight differences. Whereas the values for the three GC-R forms using the swinging-bucket rotor were 3.2 S, 5 S, and 9 S (3), those obtained with the vertical tube rotor for the same receptor species were 3.8 S, 5.2 S, and 9.1 S. Perhaps this is due to hormone dissociation occurring during the lengthy swinging-bucket centrifugation; a mixture of free and receptor-bound hormone would tend to lower the sedimentation position of the observed radioactive peak. Thus, an immediate advantage of using vertical tube rotor sucrose gradients is evident; that is, this technique probably yields a more accurate estimate of the sedimentation coefficients of various receptor forms. The high salt vertical tube rotor gradient yielded a 3.8 S form, while a broad peak, centered in the 5.2 S region, was observed in low salt. In addition, a 9.1 S form was obtained when Moo:--treated cytosol was centrifuged on a Moo:--containing gradient (Fig. 1C). However, samples centrifuged on KCl/MoOz- gradients in the vertical tube rotor gave dramatically different results from those obtained with the swinging-bucket rotor. The predominant form obtained with the vertical tube rotor was the 9.1 S untransformed, oligomeric receptor (Fig. lB), whereas the 3.2 S transformed species was obtained with the swinging-bucket rotor (Fig. 1E). Furthermore, pretreatment of the receptor with 0.3 M KC1 for an hour promoted some dissociation of the 9.1 S receptor into 3.8 S subunits, when subsequently centrifuged on the KCl/MoOf gradient (Fig. 1B). It thus appears, at least for the non-pretreated receptor, that the form applied to the gradient is 9.1 S, and that MoOi- can counteract the dissociative effects of KC1 for short periods of time. However, when the receptor is in the presence of high salt and MOO:- for extended time periods (as with the swinging-bucket rotor), the protective effect of MOO!- is overridden by the dissociative effects of salt. Thus, the vertical tube rotor affords a rapid analytical technique for the study of receptor struc-

278

EASTMAN-REKS,

REKER,

AND

VEDECKIS

ture by sucrose gradient ultracentrifugation, without artifactually generating different receptor forms. On the other hand, swinging-bucket rotor centrifugation, due to the long time required for analysis, yields artifactually low sedimentation coefficients, probably as a result of hormone dissociation and receptor subunit dissociation (or proteolysis) during the spin. Vertical Tube Rotor Sucrose Gradients of Receptor Stored Overnight Due to the length of time required to prepare cytosol, label the receptor with hormone, and perform various experimental manipulations, it would be advantageous to have a rapid assay for receptor structure. Ideally, it would be best to incubate the receptor overnight with hormone, since this allows saturation of the receptor and steady-state binding (W. V. Vedeckis, unpublished). Because some tissue sources may contain endogenous proteases which could alter receptor structure (2,46), long-term assays, such as swingingbucket rotor sucrose gradients or gel filtration, might be inappropriate. That is, up to 48 h would elapse from the time of cytosol preparation to obtaining the final results. Alternatively, the very rapid analysis afforded by the vertical tube rotor might be ideal for such studies. To determine if this approach was feasible, we analyzed the AtT-20 cell GC-R on sucrose gradients after an overnight labeling with rH]Dex. These results are shown in Fig. 2. The GC-R sedimented at 3.8 S in 0.3 M KCl, while a form resembling the 5.2 S receptor was seen on low salt sucrose gradients (Fig. 2A). In addition, the 9.1 S untransformed species was observed on gradients containing 20 mM NaaMoOd (Fig. 2B). Finally, MoOi--stabilized receptor centrifuged on sucrose gradients containing 0.3 M KC1 plus 20 mM NazMoOl sedimented with an value which was greater than 5.2 S, that is, similar to that for the untransformed, 9.1 S receptor (Fig. 2B). Thus, the GC-R which had been labeled overnight with hormone behaved similarly to that which was labeled for

-KCI

1

TOP

10

a

TOP

10

a

FRACTION

FIG. 2. Vertical tube rotor sucrose gradient ultracentrifugation of AtT-20 cell GC-R incubated with [‘H]Dex overnight. Aliquots of the same cytosol sample as was used in Fig. 1 were allowed to label overnight with the hormone. Cytosol was chromatographed over LH-20 columns equilibrated in TETg buffer (0, a), and Moo:--treated cytosol was chromatographed over LH-20 columns equilibrated in TETg-20 mM NaaMoOd (Cl, A). These samples were then centrifuged on sucrose gradients in low salt buffer (O), 0.3 M KC1 (0), 0.3 M KC1 plus 20 mM NkMoO, (II), or 20 mM NarMoOl (A). The arrow indicates the sedimentation position of human transferrin (4.9 S).

2 h. It should be noted that in all experiments performed, and as seen clearly in Fig. 2B, the Moo:--stabilized receptor centrifuged on KCl/MoOi- gradients sedimented at a slightly slower rate (fractions 15-16; approximately 8.6 S) than that observed for the MoOi--stabilized receptor at low ionic strength (fractions 17-19; approximately 9.5 S). Whether this 0.7-1.0 S difference represents a conformational change of the receptor at high ionic strength, or a dissociation of various components present in the low salt form, is unknown. These studies show that merely incubating receptors for 16-18 h does not cause the subunit dissociation seen when using swinging-bucket rotors. Rather, a combination of receptor dilution, the removal of low-molecular-weight cytosolic components, and the length of time result in the

GLUCOCORTICOID

RECEPTOR

279

STRUCTURE

smaller receptor forms. These effects are greatly minimized using vertical tube rotor sucrose gradients. Although MOO;-, hormone, and protease inhibitors can be used to stabilize the receptor species when using swinging-bucket rotors, this makes the study of important phenomena, such as receptor proteolysis and transformation, impossible.

SW 50.1

VTi 80

Sedimentation Properties of the CharcoalTreated Ileceptor As mentioned previously, a considerable variation in the form of the receptor was obtained on KCl/MoOi- gradients using the swinging-bucket rotor. It was also noted that the amounts of 3.8 S and 9.1 S receptor varied on KCl/MoOz- gradients with the vertical tube rotor. Additionally, when untreated cytosol was centrifuged on low salt gradients using the vertical tube rotor, not only did the amount of 5.2 S and 9.1 S receptor vary, but the peaks obtained were quite broad. Finally, since the 5.2 S receptor is apparently transformed ((3); see below), it was not immediately clear why it was obtained when the cytosol was centrifuged on low salt gradients. One possibility for these results was that the method used to remove excess free hormone (LH-20 chromatography) might produce receptor transformation. That is, since LH-20 is derived from Sephadex G25, gel filtration of the sample occurs. This could result in the removal of a low-molecular-weight inhibitor of receptor transformation, as has been shown previously for Sephadex G-25 (47-50). Thus, we utilized an alternative method to remove free hormone, namely charcoal-dextran adsorption. As was true for the GC-R chromatographed on LH-20 in TETg-20 mM NazMoOl, Moo:--treated cytosol (after dextran-charcoal adsorption to remove free hormone) gave untransformed receptor peaks when centrifuged in MOO;- gradients, using either the swinging-bucket (9 S) or the vertical tube (9.1 S) rotor (Figs. 3C and F). Thus, the untransformed receptor (9.1 S) was stabilized if the LH-20 chromatography buffer contained MOO:-

‘: 5! x E 0

TOP

10

10

TOP

10

20

FRACTION

FIG. 3. Sucrose gradient ultracentrifugation of the AtT-20 cell GC-R after removal of free hormone by dextran-charcoal adsorption. Cytosol was either left untreated (0,O) or was made 20 mM in NazMoOl (0, n , A) and incubated with [‘HjDex for 3 h. One sample was made 0.3 M in KC1 for the last hour of incubation (0 = KCl-pretreated). Free hormone was then removed by dextran-charcoal adsorption as described under Materials and Methods. These samples were then analyzed on 5-2076 sucrose gradients using either the VT1 30 vertical tube rotor (A, B, and C) or the SW 59.1 swinging-bucket rotor (D, E, and F). Untreated cytosol was centrifuged on either low salt gradients (0) or gradients containing 0.3 M KC1 (0). Moo:--treated samples were centrifuged on gradients containing 0.3 M KC1 and 20 rnM N&Mood (Cl, n ) or 20 mM NazMoOI alone (A). The arrow indicates the sedimentation position of human transferrin (4.9 S).

(Figs. 1C and F). However, the promotion of receptor dissociation (or proteolysis) by LH-20 chromatography is shown dramatically by comparing Figs. 1A and D with Figs. 3A and D. Whereas untransformed GC-R chromatographed on LH-20 and then

280

EASTMAN-REKS. REKER, AND VEDECKIS

centrifuged on a low salt sucrose gradient had given a broad peak centered in the 5 S, transformed receptor region (Figs. 1A and D), the charcoal-treated untransformed GC-R sedimented as an authentic, untransformed peak on both the swingingbucket (9 S) and vertical tube (9.1 S) rotors (Figs. 3A and D). Thus, receptor transformation was not promoted by charcoaltreatment, while it clearly was by LH-20 chromatography. In addition, since the 9.1 S peak was obtained in both the absence and presence of NazMo04, it is unlikely that the Moo:--stabilized, 9.1 S form observed for this and other steroid hormone receptors is merely an artifact. The charcoal-treated receptor sedimented as 3.2 S and 3.8 S peaks on high salt gradients in the swinging-bucket and vertical tube rotors, respectively (Figs. 3A and D) in a manner similar to that seen for the LH-20-chromatographed GC-R (Figs. 1A and D). However, the most dramatic results obtained were, again, seen with the KCl/MoO$- gradients (Figs. 3B and E). Whereas only the transformed, 3.2 S GC-R was obtained using the SW 50.1 rotor, vertical tube rotor centrifugation yielded, exclusively, the 9.1 S, untransformed moiety. Thus, as before, it appears that MOO:- can stabilize the receptor against salt-induced dissociation of subunits if the analysis is performed rapidly. However, subunit dissociation is promoted by dilution and extended time of exposure to high salt, as occurs using swingingbucket sucrose gradients. These results emphasize the value of vertical tube rotor sucrose gradient ultracentrifugation in identifying the receptor form originally layered on the gradients. In addition, care must be taken to remove unbound hormone from the sample by utilizing a technique which does not promote receptor transformation (dextran-charcoal adsorption). Finally, minimal dissociation of hormone from the receptor occurs during vertical tube rotor centrifugation (Figs. 3A, B, and C), while considerable amounts of free hormone are released during the long centrifugation time using the swinging-bucket rotor (Figs. 3D, E, and F). As mentioned previously, this could contribute to the

slightly lower S values obtained using the swinging-bucket rotor. Receptor Forms Obtained at DiIkrent Centrifugation Speeds to the Same w4 One advantage of currently available centrifuges is the ability to set the runs to a preset cumulative centrifugal effect. This is termed the &. mode for the Beckman L8-80 ultracentrifuge. By using this mode, it is possible to obtain highly reproducible run conditions, independent of the run speed and acceleration and deceleration times. Thus, we wished to determine if the AtT-20 GC-R would sediment identically at different centrifugation speeds. Initial studies had shown that satisfactory conditions for the resolution of the various GC-R forms at a speed of 80,000 rpm was to a preset &. of 2.40 X 10” rad2/ s (data not shown). Including the deceleration time, the cumulative centrifugal effect at the end of the run was 2.44 X 10” rad2/s. Identical samples of AtT-20 GC-R were centrifuged to a preset a2t of 2.40 X 10” at either 80,000 rpm (463,000ga,) or 65,000 rpm (305,OOOg,,).At 80,000 rpm the total run time was 65 min, while at 65,000 rpm it took 93 min to achieve the same W2t. The results of this experiment are shown in Fig. 4. An excellent correspondence of the receptor peak positions was obtained at both speeds under these conditions. Because of the relatively small difference in run times, there was no significant broadening of the peaks at the slower speed due to diffusion. Thus, because excellent reproducibility is obtained, the u2t mode is ideal for the study of the AtT-20 GC-R, especially since this protein undergoes alterations in sedimentation properties due, apparently, to subunit interactions (3). Receptor Transfomnatiwn Studies The rapidity of the analysis of receptor forms using vertical tube rotor sucrose gradient ultracentrifugation has allowed a detailed analysis of GC-R transformation (C. E. Reker, S. B. Eastman-Reks, and W. V. Vedeckis, submitted for publication). Some preliminary results are shown in Fig. 5. Glucocorticoid receptor from AtT-20

GLUCOCORTICOID

RECEPTOR

11 TOP

10

al

TOP FRACTION

cryptic hormone-binding sites (receptor activation) which were either unfilled during the 2-h labeling period, or were destroyed during thermal transformation. To analyze these possibilities, cytosol was labeled with hormone for either 2 h or overnight, thermally transformed (25”C, 1 h), made 20 mM in N%Mo04, and then centrifuged on MoOf-containing sucrose gradients. As can be seen in Fig. 5B, the sample which was labeled overnight showed virtually all 5.2 S receptor after transformation, while distinct 5.2 S and 9.1 S peaks were obtained for the sample which was

FIG. 4. Vertical tube rotor sucrose gradient ultracentrifugation of the GC-R at different speeds to the same preset o’t. Identical samples of MoOi--treated AK-20 cytosol were centrifuged on 5-20% sucrose gradients made up in TETg-20 mM NarMoO, (A) or TETg-0.3 M KCl-20 mM NarMoOl (B). On both runs the preset cumulative centrifugal effect (&) was 2.40 X 10” rad’/s, and the run was performed in the preset & mode. Samples were centrifuged at either 65,000 rpm (305,OOOg,,), which took 93 min (Cl, A), or 80,006 rpm (463,OOOg,,), which took 65 min (m, A). The arrow indicates the sedimentation position of human transferrin (4.9 S).

cells was transformed by incubation of the cytosol at 25°C for 1 h. An aliquot was then made 20 mM in N%MoOl, while another received an equal volume of water and served as a control. After charcoal treatment, the Moo:--treated sample was centrifuged on a Moo:--containing sucrose gradient, while the control sample was spun on a low salt sucrose gradient in the absence of MOO:-. The control, thermally transformed, GC-R sedimented as a single 5.2 S peak (Fig. 5A), confirming previous studies which suggested that the 5.2 S receptor form is a transformed species (3). The MoOf-stabilized, transformed receptor also sedimented predominantly as a 5.2 S species, but also contained a distinct, small peak in the 9.1 S region (Fig. 5A). Three possibilities exist for the latter result with the Moo:--treated sample. First, a reversal of transformation (5.2 S 9.1 S) could be occurring. Second, molybdate may improve the resolution between a mixture of 5.2 S and 9.1 S species. Lastly, molybdate may be causing a recovery of

281

STRUCTURE

6

T s x % -

6

TOP

lo

al

FRACTION

FIG. 5. Vertical tube rotor sucrose gradient ultracentrifugation of the thermally transformed AtT-20 cell GC-R. (A) The GC-R from AtT-20 cells was transformed by warming cytosol to 25°C for 1 h. After charcoal adsorption of the free hormone, one aliquot of the cytosol was made 20 mM in NaaMoO, and centrifuged on a MO@--containing sucrose gradient (O), while another aliquot received an equal volume of water and was centrifuged on a low salt sucrose gradient (0). (B) Samples of cytosol were incubated with [*H]Dex for 2 h (0) or overnight (0). Both samples were made 20 mM in N&MOO, and centrifuged on sucrose gradients containing 20 rnrd NaxMoO,. The arrow in both panels represents the sedimentation position of human transferrin (4.9 S).

282

EASTMAN-REKS.

REKER,

thermally transformed after only a 2-h incubation with hormone. In addition, other studies have demonstrated a 1.5 to 2-fold increase in total receptor-bound radioactivity after an overnight incubation, as compared to a 2-h labeling period (data not shown). Since ligand binding protects the hormone-binding site of the glucocorticoid receptor against thermal denaturation, and molybdate can also protect the hormone-binding activity as well as enhance a recovery of this activity after heat treatment (31,32,51-53), the results seen in Fig. 5B are interpreted as follows. The overnight labeling of the receptor allowed steady-state binding to occur, with a concomitant saturation of receptor hormonebinding sites, prior to thermal transformation. Thus, all of the receptor contained bound hormone in the untransformed state, and these receptors were converted predominantly to the 5.2 S transformed species upon warming. Conversely, cytosol incubated with hormone for only 2 h contained a significant number of untransformed receptors with vacant hormonebinding sites. Upon warming, these untransformed receptors were probably converted to a form incapable of binding hormone (receptor inactivation). Subsequent treatment of this sample at low temperature with molybdate may have allowed these untransformed receptors to regain their hormone-binding activity, thus yielding the observed 9.1 S peak. Therefore, instead of the appearance of the 9.1 S peak in Fig. 5A being due to a reversal of transformation, we believe the addition of molybdate caused receptor activation. Lastly, we have also found that the addition of MOO:- to partially transformed, fully labeled receptors, followed by centrifugation on MoOi--containing sucrose gradients does, indeed, yield a superior resolution of a mixture of untransformed and transformed receptors (C. E. Reker, S. B. Eastman-Reks, and W. V. Vedeckis, unpublished observations). This could also contribute to the distinct 5.2 and 9.1 S peaks seen for the molybdate gradient in Fig. 5A, compared to the broader 5.2 S peak seen on the low salt gradient lacking molybdate.

AND

VEDECKIS DISCUSSION

Suitable conditions for the study of the mouse GC-R using vertical tube rotor sucrose gradient ultracentrifugation have now been defined. A number of important advantages for using this technique are obvious. First, the Beckman VTi 80 vertical tube rotor holds eight samples, rather than having the six positions of conventional swinging-bucket rotors; this significantly increases the number of samples which can be simultaneously monitored. Second, by using the cumulative centrifuge effect (&%) feature of the modern centrifuges, very reproducible gradient fractionations of the various receptor forms can be obtained with this rotor. Third, the resolution obtained using the conditions described here is quite satisfactory for distinguishing the three differently sedimenting forms of the AtT-20 GC-R: the 9.1 S, untransformed species and the 5.2 S and 3.8 S, transformed moieties. For the present study, the most dramatic effect of the rapid analysis with the vertical tube rotor was observed with respect to subunit dissociation, which was greatly reduced compared to that obtained using a conventional swinging-bucket rotor. An interesting application for the use of vertical tube rotor sucrose gradient ultracentrifugation will be in the analysis of various receptor forms fractionated from gelfiltration, ion-exchange, and adsorption columns in the presence of high salt concentrations. By running these columns rapidly, adding molybdate to the fractions to inhibit further salt-induced transformation, and subsequently analyzing the sedimentation of the peak fractions with vertical tube rotor sucrose gradients, it should be possible to determine the physical form of the receptor for each of the column peaks. Thus, it would be interesting to determine if the transformed receptor which elutes from DEAE-cellulose at 0.08 M KC1 (l-3) is the 5.2 S form, the 3.8 S form, or both. Because of the long centrifugation time required using swingingbucket rotors, and the resultant possibility of further subunit dissociation, these experiments have been difficult to perform convincingly in the past.

GLUCOCORTICOID

RECEPTOR

The most exciting prospect for the use of this technique is for studies on receptor transformation. The preliminary experiments reported here strengthen our suspicion that the authentic, transformed AtT-20 GC-R is a 5.2 S species. Using the vertical tube rotor, we have also shown that other methods of causing receptor transformation (Sephadex G-25 gel filtration, high salt treatment, dialysis) also result in the formation of a 5.2 S receptor form, and that the in viva-transformed AtT-20 cell GC-R sediments at 5.2 S (C. E. Reker, S. B. Eastman-Reks, and W. V. Vedeckis, submitted for publication). Also, a kinetic analysis of receptor transformation has shown that the appearance of the 5.2 S species parallels the conversion of the 0.20 M KCl-eluting peak from DEAE-cellulose to the 0.08 M KCl-eluting form. These experiments were greatly facilitated by the rapid analysis provided by the vertical tube rotor. Future experiments where this technique will be of great value is in the study of the reversal of receptor transformation. Thus, it should be possible to analyze the 5.2 S - 9.1 S shift upon transformation reversal, by minimizing the amount of subsequent subunit dissociation during the sedimentation analysis. In addition, reconstitution studies of receptor subunits, including the possible role of receptorbinding factors (4-ll), will also be greatly facilitated by reducing the possibility of artifactual receptor forms being generated during the sucrose gradient ultracentrifugation assay. The application of vertical tube rotor sucrose gradient ultracentrifugation to the study of steroid structure should, thus, answer some important questions about the components of the oligomeric forms, and the molecular mechanism of receptor transformation. ACKNOWLEDGMENT We wish to thank Dr. Merry R. Sherman for valuable comments on this manuscript. REFERENCES 1. VEDECKIS, W. V. (1981) Biochemistry 7245.

20, 7237-

STRUCTURE

283

2. VEDECKIS, W. V. (1983) Biochemistry 22, 19751983. 3. VEDECKIS, W. V. (1983) Biochemistry 22, 19831989. 4. MURAYAMA, A., FUKAI, F., HAZATO, T., AND YAMAMOTO, T. (1980) J. Biochem (Tokyo) 38.955961. 5. MURAYAMA, A., FUKAI, F., HAZATO, T., AND YAMAMOTO, T. (1980) J. B&hem. (Tokyo) 88,963968. 6. MURAYAMA, A., FUKAI, F., AND YAMAYOTO, T. (1980) J. Biochem (Tokyo) 88.969-976. 7. MURAYAMA, A., FUKAI, F., AND YAMAMOTO, T. (1980) J. Biochem (Tokyo) 88, 1457-1466. 8. MURAYAMA, A., FUKAI, F., AND YAMAMOTO, T. (1980) J. Biochem (Tokyo) 88, 1305-1315. 9. FUKAI, F., AND MURAYAMA, A. (1981) J. B&hem (Tokyo) 89.1829-1837. 10. MURAYAMA, A., AND FUKAI, F. (1981) J. Biochem (Tokyo) 90, 823-832. 11. COLVARD, D. S., AND WILSON, E. M. (1981) Endocrinologg 109,496-504. 12. COSTELLO, M. A., AND SHERMAN, M. R. (1980) in Proceedings of the Endocrine Society, 62nd Annual Meeting, p 174. 13. HUTCHENS, T. W., MARKLAND, F. S., AND HAWKINS, E. F. (1982) Biochem. Biophys. Res. Commun 105,20-27.

14. TYMOCZKO, J. L., AND PHILLIPS, M. M. (1983) Endocrinology 112.142-149. 15. TOFT, D., AND GORSKI, J. (1966) Proc NatL Acad Sci. USA 55,1574-1581. 16. KUTE, T. E., HEIDEMANN, P., AND WITLIFF, J. L. (1978) Cancer Res. 38,4307-4313. 17. MILGROM, E. (1981) in Biochemical Actions of Hormones (Litwack, G., ed.), Vol. 13, pp. 465492, Academic Press, New York. 18. SCHMIDT, T. J., AND LITWACK, G. (1982) Physid Rev. 62,1131-1192. 19. HAWKINS, E. F., HORN, D., AND MARKLAND, F. S., JR. (1980) Cancer Res. 40,2174-2178. 20. HAWKINS, E. F., LIESKOVSKY, G., AND MARKLAND, F. S., JR. (1981) J. Clin Endocrinol. Metub. 53, 456-458. 21. FOEKENS, J. A., PEERBOLTE, R., MULDER, E., AND VAN DER MOLEN, H. J. (1981) Mel CeU. E& m-in01 23,173-186. 22. HUGHES, M. R., COMPTON, J. G., SCHRADER, W. T., AND O’MALLEY, B. W. (1981) Biochemistry 20, 2481-2491. 23. MAGGI, A., COMPTON, J. G., FAHNESTOCK, M., SCHRADER, W. T., AND O’MALLEY, B. W. (1981) J. Steroid Biochem 15,63-68. 24. MULDER, E., VRIJ, L., AND FOEKENS, J. A. (1981) Md Cell Endocrinol 23.283-296. 25. TRAISH, A. M., M~J.,LER, R. E., AND Wo~rz, H. H. (1981) J. Biol Chem 256, 12028-12033. 26. CHANG, C. H., ROWLEY, D. R., LOBL, T. J., AND

284

27. 28. 29. 30. 31.

32. 33.

34. 35. 36. 37. 38.

39.

40.

41.

EASTMAN-REKS,

REKER,

TINDALL, D. J. (1982) Biochemistry 21, 41024109. GIAMBIAGI, N., AND PASQUALINI, J. R. (1982) Endocrinology 110,106’7-1069. WOODY, C. J., AND JAFFE. R. C. (1982) Gen Comp. Erwkcrinol 47, 28-35. JENTOFT, N., AND DEARBORN, D. G. (1979) J. Bid Chem 254, 4359-4365. JOHN, J. K., AND MOUDGIL. V. K. (1979) Biochem Biophys Rea Commun 90,1242-1248. LEACH, K. L., DAHMER, M. K., HAMMOND, N. D., SANDO, J. J., AND PRATT, W. B. (1979) J. Bid Chem. 254,11884-11890. TOFT, D., AND NISHIGORI, H. (1979) J. Steroid Biochem 11,413-416. GRODY, W. W., COYPTON, J. G., SCHRADER, W. T., AND G’MALLEY, B. W. (1989) J. steroid Biochem 12, 115-120. NISHIGORI, H., AND TOFT, D. (1980) Biochemistry 19, 77-83. SCHMIDT, T. J., HARMON, J. M., AND THOMPSON, E. B. (1980) Nature (London) 286.507-510. BARNETT, C. A., SCHMIDT, T. J., AND LITWACK, G. (1980) Biochemistry 19,5446-5455. SHYAMALA, G., AND LEONARD, L. (1980) J. Bid Chem 255, 60286031. WOLFSON, A., MESTER, J., CHANG-REN, Y., AND BAULIEU, E. E. (1980) Biochem Biophys. Res. Cemmun 95, 1577-1584. SHERMAN, M. R., BARZILAI, D., PINE, P. R., AND TUAZON, F. B. (1979) in Steroid Hormone Receptor Systems (Leavitt, W. W., and Clark, J. H., eds.), pp. 357-375, Plenum, New York. HUTCHENS, T. W.. MARKLAND. F. S., AND HAWKINS, E. F. (1981) Biochxm Biophys. Ra Ccnnmun 103,60-67. MILLER, L. K., TLJAZON, F. B., NIU, E.-M., AND SHERMAN, M. R. (1981) Endocriw 108,13691378.

AND

VEDECKIS

42. NIU, E.-M., NEAL, R. M., PIERCE, V. K., AND SHERMAN, M. R. (1981) J. Steroid Biochem. 15,1-10. 43. REDEUILH, G., SECCO,C., BAULIEU, E.-E., AND RICHARD-FOY, H. (1981) J. Bid Chem, 256, 1149611502. 44. SHERMAN, M. R., T~AZON, F. B., AND SBMJEN, G. J. (1981) in Physiopathology of Endocrine Diseases and Mechanisms of Hormone Action (Soto, R. J., De Nicola, A., and Blaquier, J., eds.), pp. 321-337, Liss, New York. 45. SHERMAN, M. R., MORAN, M. C., NEAL, R. M., NIU, E.-M., AND TUAZON, F. B. (1982) in Progress in Research and Clinical Applications of Corticosteroids (Lee, H. J., and Fitzgerald, T. J., eds.), pp. 45-66, Heyden & Son, Philadelphia. 46. SHERMAN, M. R., MORAN, M. C., TUAZON, F. B., AND STEVENS, Y.-W. (1983) .J. Biol. Chem. 258. 10366-10377. 47. CAKE, M. H., GOIDL, J. A., PARCHMAN, L. G., AND LITWACK, G. (1976) Biochem Biophys. Ra. Commun 71, 45-52. 48. BAILLY, A., SALLAS, N., AND MILGROM, E. (1977) J. Biol Chem 252, 858-863. 49. GOIDL, J. A., CAKE, M. H., DOLAN, K. P., PARCHMAN, L. G., AND LITWACK, G. (1977) Biochemistry 16,2125-2130. 50. SATO, B., NOMA, K., NISHIZAWA, Y., NAKAO, K., MATSUMOTO, K., AND YAMAMURA, Y. (1980) EndocrinoZogy 106, 1142-1148. 51. NIELSEN, C. J., SANDO, J. J., VOGEL, W. M., AND PRA?T, W. B. (1977) J. Biol. Chem 252, 75687578. 52. LEACH, K. L., GRIPPO, J. F., HOUSLEY, P. R., DAHMER, M. K., SALIVE, M. E., AND PRATT, W. B. (1982) J. Biol Chem 257, 381-388. 53. MCBLAIN, W. A., AND SHYAMALA, G. (1980) J. Biol Chem 255, 3884-3891. 54. WHEELER, R. H., LEACH, K. L., LA FOREST, A. C., O’TOOLE, T. E., WAGNER, R., AND PRAY, W. B. (1981) J. Biol Chem 256, 434-441.