Giannopoulos, G., Mulay, S., and Solomon, S. (1973) J. Biol Chem. 248,. Chem. 247 ... Yahamoto, K. R., Gehring, V., Stampfer, M. R., and Sibley, C. H. (1976).
Vol. 259, No.20, Issue of October 25, pp. 12915-12924,1984 Printed in U.S.A.
THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1984 by The American Society of Biological Chemists. Inc
Characterization ofthe Rat Liver Glucocorticoid ReceptorPurified by DNA-cellulose and LigandAffinity Chromatography* (Received for publication, February 21,1984)
Manjapra V, GovindanS and Hinrich GronemeyerQ From the Laboratoire de GhnPtique Moliculaire des Eucaryotes, Centre National de la Recherche Scientifique, U. 184 de Bwbgie Moliculaire et de Genie Genetique del’lnstitut National de la Sante et de la Recherche Medicale, Facult6 de Medecine,Institut de Chimie Biologique, 67085 Strasbourg Cedex, France
The importance of steroid receptors in modulating horTwo rapid and high yield purification methods for the rat liver glucocorticoid receptor based on differ- monal responses in target tissues is under intense investigaential DNA affinity (method A) and ligand affinity tion in many laboratories. However, the low concentration (method B) chromatography are described. In method and inherent lability of steroid receptor proteins causes seriA, the amount of receptor in rat livercytosol that can ous problems duringconventional purification techniques be activated and subsequently eluted from a DNA- which require several days of manipulation. In addition, the cellulose column has been increased to 80%by intro- reversible binding of a steroidal ligand to its receptor limits ducing a second heatactivationstep. Using this the analytical methods which can be used. Because of this method, 1.5 nmol of 25% pureglucocorticoid receptor latter problem, the technique of affinity labeling, which cocan be routinelyobtainedperdayfrom 15-20 rat valently links receptor hormone complexes, has been devellivers. Method B yields about 2.2 nmol of 60%pure oped for the analysis of steroid hormone-binding proteins receptor with an overallyield of -60%. using photochemically (1-11) or chemically (12-17) reactive The qualityof these purifications has been controlled steroids. by affinity labeling. In each case, more than 95% of purified binding activity represented the intact 92,000Using a combination of very rapid and quantitativepurifi2 400-Da glucocorticoid receptor polypeptide as shown cation methods with affinity labeling analysis, it should be by sodium dodecyl sulfate-gel electrophoresisand fluo- possible (i) to determine the precise molecular weight of a receptor polypeptide, (ii) todetermine the purity of a preparography. No difference in the labeling pattern was ration with respect to theabundance of this protein, and (iii) observedusing either [3H]triamcinoloneacetonide (photoaffinity labeling) or [3H]dexamethasone 21-me- to explain some of the discrepancies in reported molecular sylate (electrophilic labeling). The electrophilic label- masses of steroid receptors. For example, a number of purifiing step was performed in thecytosol prior to purifi- cation methods have been described for the glucocorticoid labeled components receptor, reporting molecular masses ranging from 37 to 150 cation by method A to compare the thus purified with those obtained when the photoaffin- kDa (18-32). These discrepancies in molecular weights might ity labeling was performedafter thepurification. Us- result from proteolytic degradation (9, 24, 25, 30, 31). ing this approach, distinct breakdown products of the Our aim has been to develop rapid and quantitative purifiglucocorticoidreceptor wererevealed,co-purifying cation methods for the rat liver glucocorticoid receptor such during DNA affinity chromatography. that they can be used for preparative purposes and toanalyze Cross-linked receptor obtainedby method A has been the purified material by affinity labeling techniques. We defurther purified to homogeneity by preparative sodium scribe in detail two purification procedures based on DNA dodecyl sulfate-gelelectrophoresisandsuccessfully affinity chromatography (methodA) and ligand affinity chroused as immunogen to raise glucocorticoid receptor matography (method B). The purified receptor obtained by antibodies in rabbits. These antibodies raised against both methods has been analyzed by photoaffinity labeling. In glucocorticoid receptor, as well as those previously obtained using affinity chromatography-purified re- addition, for method A we have performed a comparative ceptor, react with the receptor molecules irrespective study involving electrophilic labeling of the glucocorticoid of their method of purification. Glucocorticoid recep- receptor prior to purification. Using preparative SDS-gel tors purified bymethods A and B have been analyzed electrophoresis, we have further purified to homogeneity the for specific DNA-bindingproperties by the nitrocellu- receptor enriched by method Aand used it toobtain antibodies. The cross-reactivity of these antibodies, as well as those lose filter binding assay. raised against affinitychromatography-purified receptor (26), with receptor purified by either method was assayed by Western blotting analysis. Finally, the specificity of DNA binding * This work was supported by grants from Centre National de la is analyzed with glucocorticoid receptor purified by method A Recherche Scientifique, Institut National delaSantB et de la Re- or B to cloned MMTV DNA.
cherche MBdicale, MIR (83 V 0626), the Fondation pour la Recherche Mhdicale, the Association pour le DBveloppement de la Recherche sur le Cancer, and the Fondation Simone and Cino del Duca. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisernent” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. To whom correspondence should be sent. I Present address, Ludwig Institut fin Krebsforschung, Inselspital, CH-3010 Bern, Switzerland.
+
The abbreviations and trivial names used are: SDS, sodium dodecyl sulfate; MMTV-LTR, mouse mammary tumor virus long terminal repeat; TA, triamcinolone acetonide, 9a-fluoro-llp,l6a, 17,21-tetrahydroxypregna-1,4-diene-3,20-dione cyclic 16,17-acetal with acetone; dexamethasone, 9a-fluoro-l1~,17,2l-trihydroxy-16amethylpregna-1,4-diene-3,20-dione; ELISA, enzyme-linked immunosorbent assay; PBS, phosphate-buffered saline (20 mM phosphate buffer, pH 7.4, containing 150 mM NaCl); bp, base pair.
12915
12916
Glucocorticoid Receptor Purification and Characterization MATERIALS ANDMETHODS
on a DNA-cellulose I1 column (40 ml in a 5-crn diameter column). The flow-through of this column was reactivated (15 min, 20 'C) and [3H]Triamcinolone acetonide (31.3 Ci/mmol) and [3H]dexameth- rechromatographed on a third DNA-cellulose column (30 ml in a 5asone (25 and 49 Ci/mmol) were purchased from New England cm diameter column). A11 three DNA-cellulose columns were washed Nuclear, Dreieich, Federal Republic of Germany. Nonradioactive with 3 volumes of buffer C, 3 volumes of 0 . 5 TBE ~ buffer (0.09 M steroids were from Sigma. [3H]Dexamethasone 21-methanesulfonate Tris-base, 0.09 M H3B03, and0.0025 M EDTA) followed by elution (dexamethasone 21-mesylate) was synthesized (49 Ci/mmol) on a with 0.5 M NaCl in buffer C. The DNA-cellulose eluates were pooled microscale from 0.5 mCi of [3H]dexamethasone as described (15,29). and immediately cross-linked or precipitated with 50% ammonium [LX-~SIATP (1000 Ci/nmol) was purchased from Amersham Buchler, sulfate asdescribed below. Braunschweig, Federal Republic of Germany. DNA restriction enCross-linked and noncross-linked receptor preparations were prezymes EcoRI and Sphl and T4 DNA polymerase were from Bethesda cipitated in 50% ammonium sulfate overnight in ice and collected by Research Laboratories, Neu Isenburg, Federal Republic of Germany. centrifuging for 30 min at 25,000 rpm in SW 27 rotors (Beckman). The MMTV-LTR deletion mutants p311, pdell, pdel2, and pdel3 After carefully drying the walls of the tubes, the cross-linked prepawere kindly provided by J. Majors (33). DNA restriction fragments rations were stored as wet pellets at -80 "C. The ammonium sulfate were end labeled and made flush with T4 polymerase. precipitates of noncross-linked preparations were redissolved in Protein was determined according to Bradford (34) using the Bio- buffer C and stored as aliquots of 100 g1 containing 1 g~ TA at Rad protein assay kit. SDS-polyacrylamide gel electrophoresis was -20 "C after the addition of glycerol to 60% (v/v). The receptor performed as described by Laemmli (35) using a Hoefer electropho- concentration in this stock solution was M as determined by the resis apparatus (Hoefer Scientific Instruments, San Francisco); silver amount of bound radioactive hormone. staining was according to Ansorge (36) and fluorography was carried Purification Method B-For ligand affinity chromatography, [3H] out with EN3HANCE (New England Nuclear) following the manu- dexamethasone 17~-carboxylicacid (29) was coupled to Affi-Gel-102 facturer's instruction. The acrylamide stock solution for electropho- (Bio-Rad) as described previously (29,39). The coupling efficiency of resis contained 30 g of acrylamide and 0.8 g of methylene bisacrylam- the matrix was 5 pmol of hormone bound/ml of packed gel. ide/lOO ml of water. For molecular weight calibration, the low molecRat liver cytosol was prepared and filtered through phosphocelluular weight kit from Pharmacia (Uppsala, Sweden) or the 14C-labeled lose as described above, without labeling with [3H]TA. The flowhigh molecular weight markers (fluorography) from Bethesda Re- through was applied onto the affinity matrix diluted 5-fold with search Laboratories were used. We did not observe any difference in Sepharose CL-4B (20 ml, 2.5-cm diameter column; flow rate, 1.5 ml/ the mobility between the labeled and unlabeled phosphorylase b; min) equilibrated in buffer A. The column was washed extensively therefore, we used the molecular mass of 94,000 Da for calibration. with 2 litersof buffer A followed by 0.5 liter of 0.25 M NaCl in buffer Sepharose CL-4B, Sephadex G-50 and protein A-Sepharose were A. The quantity of receptor retained on the column was measured as from Pharmacia. Radioactivity was determined using an Lntertech- the difference between specific binding with [3HJTAin the phosphonique (France) scintillation spectrophotometer with 30% counting cellulose flow-through and affinity column flow-through and washes. efficiency for tritium. DNA cellulose was prepared according to Al- The receptor was eluted from the affinity matrix by incubating 5 h berts and Herrick (37) and contained 0.8 mg of DNA/ml of settled at 0 "C with 1 g M [3H]TA(250 mCi/mmol) in 50 mM NaSCN, 10 mM gel. For phosphocellulose chromatography, P-11 (Whatman) was @-mercaptoethanol,20 mM Tris-HC1, pH 7.4, and 10% glycerol (20 washed before the first purification as described (38). Between sub- ml) (45). The column was washed with an additional 30 ml of elution sequent purifications, the phosphocellulose and the DNA-cellulose buffer without hormone. The eluate and the washings were pooled were washed with 1 M salt in buffer (see below). and 0.5 ml was precipitated with 10% trichloroacetic acid for SDSMale Wistar rats (-200 g) were adrenalectomized and maintained gel electrophoresis. For photoaffinity labeling, a 5-ml aliquot of this on a normal diet supplemented with 0.9% NaCl at constant temper- poolwas gel-filtered through Sephadex (2-50 and equilibrated in ature (27 "C) for 7-11 days and killed by cervical dislocation. buffer A, and theexcluded peak was irradiated. The remaining eluate Buffers were: lox buffer A, 10 mM EDTA, 200 mM sodium phos- was diluted 3-fold with buffer A, adjusted to pH 7.8, activated for 30 phate, pH 7.0, and 500 mM NaCI; buffer B, 10 mM EDTA, 200 mM min at 20 "C, and chromatographed on a 30-ml DNA-cellulose colsodium phosphate, pH 7.8, and 500 mM NaCI; 1OX buffer C, 10 mM umn. This column was washed and eluted, and theeluates were crossEDTA, 200 mM Tris-HC1, pH 7.8, and 900 mM NaCI; 1OX buffers Alinked or precipitated with 50% ammonium sulfate as described in C were diluted 10-fold by adding 10% (v/v) glycerol and ice-cold method A. For regeneration, after elution, the affinity column was double distilled water and adjusting the pHdescribed above, prior to washed with 500 mlof 1 M NaCl followed by 500 mlof warm 1%SDS (60"C), and finally with several liters of analytically pure ethanol the addition of 2 mmol/liter 0-mercaptoethanol. Purification Method A-This method is modified from the purifi- until 1 ml of the ethanol wash contained no detectable radioactivity. cations described by Eisen and Glinsman (27) and Wrange et al. (28). Then the affinity column was equilibrated in buffer A. Cross-linked Buffers A-C were essentially the same as described by Wrange et al. and noncross-linked preparations were precipitated with ammonium For one receptor purification, 16 rats were killed and the livers sulfate, centrifuged, and stored as described above after adjusting the were perfused in situ with 30 ml of ice-cold buffer A. After removal receptor stock solution to 1 X lo-' M. Photoaffinity and Electrophilic Labeling-Photoaffinity labeling of the livers, all the following procedures were performed at 0 "C unless otherwise stated. The livers were combined, minced, washed was performed using a 1-kilowatt mercury-xenon lamp and a 2-mmwith buffer A, and homogenized in 3volumes of buffer A by 10 strokes thick WG 320 long pass filter in ajacketed cylindrical quartz cuvette with a motor-driven Teflon-glass Potter-Elvehjem homogenizer. The (15 ml) at -10 "C as described (9).Cross-linking efficiencies, usually homogenate (360 ml) was centrifuged at 35,000 rpm for 70 min in a approximately 5%, were determined by precipitation in 10% trichloTi-35 rotor (Beckman) and thefloating fat layer was aspirated. After roacetic acid. Electrophilic labeling with [3H]dexamethasone 21-mesylate was filtration through two layers of Kleenex, the cytosol was labeled immediately by incubating with 100 nM [3H]TA (9 Ci/mmol) for 60 performed in the cytosol prior to thepurification according to method min at 0 "C. In order to determine the receptor content, aliquots of A. Because of the limited amount of the ligand available, cytosol of six rat livers was prepared in 20 mM Tris-HC1, pH 8.0, 50 mMKC1, the cytosol were incubated with 50 nM 13H]dexamethasone in the presence or absence of a 200-fold excess of cold hormone for 60 min 1 mM EDTA, and 10% glycerol and incubated with 100 nM [3H] at 0 "C. Unbound steroid was removedby the dextran-coated charcoal dexamethasone 21-mesylate (9 Ci/mmol) for 1 h at 0 "C. In parallel, the cytosol of 10 rat livers was prepared and incubated with 100 nM method (40). cold trichloroacetic acid as described above. Both cytosol fractions The [3H]TA-labeledcytosol (250 ml) was filtered through 400 ml of phosphocellulose in a 10-cm diameter column previously equili- were combined and the pH was adjusted to 7.0 with 1 N HCl. All brated in buffer A (flow rate, 25 ml/min). The column was washed subsequent purification steps were as described in method A with with buffer A and the combined flow-through (400 ml) was adjusted omission of the irradiation step. Preparative SDS-Gel Electrophoresis of Receptor Purified by to pH 7.8 with 1 N NaOH. For regeneration, the phosphocellulose was washed immediately with 1 liter of 1 M NaCl in buffer A and Method A-The ammonium sulfate-precipitated cross-linked receptor preparation from each DNA-cellulosecolumn eluate was dissolved reequilibrated in buffer A. The phosphocellulose flow-through was applied onto a DNA-ceI- separately in 1 ml of SDS sample buffer containing bromphenol blue lulose column (100 ml in a 5-cm diameter column) with a flow rate as marker, by heating 2-5 min at 80 "C. The preparative gel electroof 4 ml/min. Details of the DNA-cellulose chromatographic steps phoresis was performed in a Bethesda Research Laboratories preparincluding a time schedule are depicted in Fig. 1. The DNA-cellulose ative gel electrophoresis system. The separation was performed on a I flow-through was activated (30 min, 20 'C) and chromatographed 4-ml 7.5% SDS-polyacrylamide gel with a 2-ml 4% stacking gel.
and Characterization
Glucocorticoid Receptor Purification Electrophoresis was at 3 mA and 70 V, and fractions of 1 m1/5 min were collected. Fraction number 1 contained the bromphenol blue. Radioactivity was determined in 10 pl of each fraction and aliquots of 2.5-10 pl were applied on a 7.5% SDS-polyacrylamide analytical gel, The radioactive fractions containing the glucocorticoid receptor band were pooled and precipitated with cold 20% trichloroacetic acid, and thepurity of this pool wasanalyzed again by SDS-polyacrylamide gel electrophoresis and fluorography. Immunization and Detection of Antibodies-Receptor purified by preparative gel electrophoresis was used as immunogen. Immunization was performed as described before (26). Antibodies were assayed by ELISA (46), Western blotting (47), and protein A immunoadsorbent assay (48). For ELISA, 12 wells in a Costar plate were coated with 500, 50, and 5 ng of purified receptor in atotal volume of 200 pl by incubating overnight at 4 "C in 50 mM sodium carbonate-sodium bicarbonate buffer, pH 9.6. The wells were washed twice with PBS, twice with PBS containing 0.5 ml of Tween-20/liter followed by two washings with PBS, for a total of 30 min a t room temperature with gentle shaking. Incubation with 1:lOO and 1:lOOO immune and preimmune serum diluted in PBS containing 2% bovine serum albumin (dilution buffer) was carried outa t 37 "C for 2 h. After the sequence of washing as described above, the wells were incubated for 2 h at 37 "C with 1:lOOO diluted peroxidase-conjugated anti-rabbit antibodies from sheep (anti-rabbit IgG, code 75011,Institut Pasteur, Paris) in dilution buffer. Unbound anti-rabbit IgGs wereremoved by the washing procedure described above, and theperoxidase reaction was performed with o-phenylenediamine (46). Western Blotting-After electrophoresis on a 10%polyacrylamide gel, the protein bands were transferred onto nitrocellulose using a Trans Blot Cell (Bio-Rad) according to the manufacturer's instructions. Blotting was performed in 25 mM Tris-base, 192 mM glycine, and 20% methanol (v/v) overnight (47). The washing and staining of the blots were according to the procedure described by Towbin et al. (49).A 1500 diluted antiserum was used for immunoreaction. Protein A Zmmunoadsorbent Assay-5 ml of rat liver cytosol was incubated for 2 h a t 4 "C with 100 nM [3H]dexamethasone21-mesylate (specific activity, 9 Ci/mmol) in the presence or absence of lOOx excess cold TA. After 40% ammonium sulfate precipitation (2 h), the precipitate was collected by centrifugation, dissolved in 2 ml of PBS, and filtered through a Sephadex G-50 column (20-ml bed volume) equilibrated in PBS. The flow-through (3.6 ml) containing the total binding activity was incubated for 30 min at 4 "C with 1.0 ml of protein A-Sepharose suspension (1:l in PBS). Theprotein A-Sepharose was separated by centrifugation. Aliquots of0.4ml from the supernatant were diluted with 0.6 ml of PBS andincubated overnight with 5-10 pl of immune or preimmune serum. 500 pl of protein ASepharose (1:l in PBS) was added to the incubation mixture and incubation was continued for 30 min at 4 "C with occasional shaking. The slurry was packed in a siliconized Pasteur pipette and washed with 10 mlof PBS at 4 "C. Bound activities were eluted with 0.1 M acetic acid and collected in 500-pl fractions. The radioactivity was determined in each fraction, andthe radioactive fractions were pooled, precipitated with trichloroacetic acid, and analyzed by SDSgel electrophoresis followed by fluorography (see above). DNA-binding Assay-Sequences within the MMTV-LTR required for the transcriptional regulation of MMTV by glucocorticoid in vivo have been characterized using a series of deletion mutants (33). We used the MMTV-LTR deletion mutants p311, pdell, pdel2, and pdel3 (see Ref. 33 for detailed construction) to study the specific interaction between purified glucocorticoid receptor by method A or B. The plasmids p311, pdell, pdel2, and pdel3were restricted with EcoRI and SpHI, which released a common 3.4-kilobasepair pBR322-herpes simplex virus thymidine kinase hybrid fragment and a second hybrid fragment of 1140, 730, 690, and 630 bp, respectively. These second DNA fragments contained pBR, MMTV-LTR, and herpes simplex thymidine kinase sequences. 1 pg of restricted DNAs was made flush and end labeled with [ ( U - ~ S I ~ Aand T P TTP using T4 polymerase to a specific activity of 1 x lo6 cpm/pg of DNA (50). Receptor-DNA interactions with MMTV-LTR deletion mutants were carried out at a final receptor concentration of -2 x 10-~ M as described before (51). The stock solutions of receptor purified by either method A or B were diluted 1:lO with ice-cold buffer C and incubated 5 min at 37 "C. After cooling on ice, 20 ng of receptor in 20 p1 were mixed with 50 ng of DNA (-5000 cpm) in 180 pl of buffer C. Following incubation for 10 min a t 37 'C, cooling in ice, the complexes were adsorbed onto nitrocellulose membranes by filtration and washed with 5 ml of buffer C. Bound material was eluted with
12917
0.2% SDS in 0.3 M sodium acetate, precipitated with ethanol, and subjected to electrophoresis on a 5% polyacrylamide gel. Electrophoresis was performed using bromphenol blue as marker at 200 V in TBE. After completion of electrophoresis, the gels were dried and autoradiographed using Kodak X-Omat films. RESULTS
Purification of Rat Liver Glucocorticoid Receptor by DNA Affinity Column (Method A)-A schematic description of this
method is presented in Fig. 1, which gives inadditiona detailed time schedule which is designed for the purification of glucocorticoid receptor from 15 to 20 rat livers/day. The principle of the method is based on the differential binding of nonactivated and thermally activated receptor to DNA-cellulose (27-29,42-44). After incubation with [3H]triamcinolone acetonide, the cytosol from 16 rat livers was filtered over a 400-ml phosphocellulose column to remove hormone-binding serum proteins and a major portion of DNA-binding proteins. The phosphocellulose flow-through adjusted to pH 7.8 was applied onto the first DNA-cellulose column. As soon as one-half of the flow-through was obtained, it was immediately thermally activated, cooled to 0 "C, and applied onto a second DNAcellulose column. During this time, the second batch of the first DNA-cellulose column flow-through including the washing was collected, thermally activated, cooled to 0 "C, and applied onto thesecond DNA-cellulose column. The firstDNA-cellulose column was washed with 3 volumes of buffer C, 3 volumes of 0.5X TBE, and eluted with 0.5 M NaCl in buffer C . As shown in Table I, approximately 5% of the purified hormone-binding activity was eluted. This eluate was immediately irradiated as described under "Materials and Methods." In a manner similar to the first DNA-cellulose flow-through, the flow-through obtained from the second DNA-cellulose column was collected in three fractions which 8.30- 9.W 9.05- 9.25 9.35-10.50 10.55-11.55
k i l l and Wrfuse hospcniration centrifugation incubation
I .#Just
- I
14.15 s t a r t washing
1
12.00-12.40 PHOSPWKELLULOSE fla through to pH 7.8 (1Y W.OH)
I
12.45 s t a r t DIU CELLULOSE I 13.30 first p a r t FT 14.15 second p a r t FT
I
15.30 e l u t i o n
I
cross1 ink
a c t i v a t e 30 .In M'C cool i n l c e l w a t c r f o r 10 n i n
I
14.55 DNA CELLULOSE 11-
18.30 s t a r t washing
1
activate 30 n i n zo'c cool i ni c c l w a t e r 10 min
I
14.10
15.50 first p a r t FT
17.20 second p a r t FT
I
I I
I
16.25
17.45 DIU CELLULOSE 111
I
1 18.30 t h i r dI part FT
20.35
1
21.30 e l u t i o n
crosslink
FIG. 1. Flow chart for purification by method A. Purification of rat liver glucocorticoid receptor according to method A, including a time schedule (see "Materials and Methods"). The startingtime for a particular step is indicated at the left; chromatographic column steps are presented incapitals. FT, flow-through.
12918
and Characterization
Glucocorticoid Receptor Purification
were independently incubated for 15 min at 20 "C. This mode are not due to artifactual cross-linking, e.g. occurring during of fractionation was convenient because relatively low flow irradiation, as they were present with identical mobilities in rates are necessary for the quantitative binding of activated both affinity labeling reactions. Most probably they result receptor to DNA-cellulose (27,28).After cooling to 0 "C, each from receptor breakdown products with intact hormone and fraction was directly applied onto a third DNA-cellulose col- DNA-binding domain which are co-purified with the glucoumn. According to the time schedule given in Fig. 1, DNA- corticoid receptor. cellulose columns I1 and I11 were washed, and theeluates were The radioactivity eluted from the first DNA-cellulose colirradiated as described for the first DNA-cellulose column. umn represents bindingto the intactreceptor (Fig. 2d, lane I; Using the method of photoaffinity and electrophilic cross- 2e, lane fi and Table I), which may have been activated either linking, we investigated the nature of the hormone-binding endogenously or during theinitial stages of purification. The activities eluted from the three columns. As shown in Fig. 2, majority of the binding activity is recovered from the second the major site of photoaffinity ([3H]TA; Fig. 2d, lanes I-ZZZ) DNA-cellulose column (Fig. 2d, lane ZI; 2e, lane ZI; and Table as well as electrophilic labeling ( [3H]dexamethasone 21-me- I) and represents all the receptor molecules which have been sylate; Fig. 2e, lanes I-HZ) in all three DNA-cellulose eluates activated during the first thermal treatment (47.5% of the occurs in a protein with an apparent M , = 92,000 & 400 first DNA-cellulose flow-through). This is in agreement with (calculated from 18 7.5% SDS-polyacrylamide gels). previously published results (14, 28). Attempts to bind the Scanning of the fluorograph obtained after photoaffinity activity of the second DNA-cellulose flow-through fractions labeling (Fig. 2f, lanes Z, ZZ, and ZZZ) shows that more than on a further DNA-cellulose column or longer initial heat 95% of the bindingactivity present in thethree eluates treatments were unsuccessful. However, when this fraction represents the intactglucocorticoid receptor polypeptide. Sev- was again heated (15 min at 20 "C) and applied onto a third eral distinct minor labeled bands can also be detected onthe DNA-cellulose column, no binding activity could be detected fluorographs (Fig. 2d, lane ZZ, and 2e, lane ZZ). These bands in the flow-through. Analysis of the eluate obtained from this TABLE I Purification of rat liver glucocorticoid receptor (nethod A) Purity is calculated by assuming that thereceptor (M, = 92,000)is univalent; homogeneous receptor would have a specific activity of 10,870pmol of bound steroid/mg of protein. Total
step
w
PMl
6095Cytosol 3846 Phosphocellulose 0.973 flow-through 5840 3356 DNA-cellulose I 0.692 1.704 DNA-cellulose I1 16.3910.546 DNA-cellulose I11 2.784 2.381 DNA-cellulose I1 DEAE-purified a Specific activity 9 Ci/mmol. sI
II 111
s
s
II 111
111
0.09
II
s
I I
I
6.56 X 6500 10'
II
111
-fold
%
1 1 110
X 1 0' 115 25X 10'2700 2730 X 1 0'460 464
38
596
YieldPurity
Pmollmg
cpmlmg
3906 3266 196 1491 1105
P I
z c z if
Specific activity
receptor
6622
s
1
s
s
%
1
100 84 5
4
28
60
15
2
1
.-
'
94000-
-94000
67000-
-67000
43000-
ja
b
C
d
e
-
h
-43000
111
Jf L
7
n
2
9
h
i
FIG.2. Analysis of purified glucocorticoid receptor fractions by SDS-polyacrylamide gel electrophoresis. Purified fractions of receptor obtained according to methods A (a-e) and B (94) were run on 7.5% (a, b, d, e, g-i) or 10% (c) SDS-polyacrylamide gels (see "Materials and Methods"). a, silver staining of 3.5 pg of DNAcellulose eluates I, 11, and 111; b, Coomassie blue staining of 35 and 100 pg of DNA-cellulose eluates I1 and 111, respectively; c, Coomassie blue staining of -7.5 pg of DEAE-purified DNA-cellulose I1 eluate; d, fluorography of [3H]TA photoaffinity-labeled DNA-cellulose eluates I, 11, and I11 (approximately 30,000 cpm); e, fluorograph of DNA-cellulose eluates I, 11, and 111 obtained according to method A, but labeled with [3H]dexamethasone 21mesylate prior to purification. Lanes el-ell1 contained approximately 2,500, 15,000, and 10,000cpm, respectively. f, scan of d. g, lane 1, silver staining of 0.5 ml of affinity eluate after trichloroacetic acid precipitation. h, affinitypurified receptor further purified on a DNA-cellulose column, stained with silver: left lane, 2.5 pg; right lane, 5 pg. i, fluorograph of [3H]TA photoaffinity-labeled receptor purified by method B: lane I , affinity chromatography eluate (-2,000 cpm); lane 2, DNA-cellulose column eluate (-10,OOO cpm). S, molecular weight standard protein markers. Arrows, glucocorticoid receptor band.
Glucocorticoid Receptor Purificationand Characterization column by affinity labeling demonstrated that not only the activity present in the flow-through of the second DNA ~ 0 1 umn could be reactivated, but also that it could be quantitatively isolated as intactreceptor (Fig. 2d, lane Ilr; 2e, lane III; and Table I). With the inclusion of this step, the overall recovery of activated receptor from DNA-cellulose columns 11 and 111represented 80%of the phosphocellulose flow-through. 38 and 28% of the receptor initially present in thecytosol are removed in DNA-cellulose eluates I1 and 111with purifications of approximately 2700- and 460-fold, respectively. From the molecular weight determined as described above, the pure rat liver glucocorticoid receptor is 10,870 pmol/mg of protein, assuming one steroid-binding site/92-kDa receptor polypeptide. Therefore, the calculated purity of receptor in DNA column I1 and I11 eluates is 25 and 4%, respectively. Fig. 2a (lanes I , IZ, and ZII) shows asilver-stainedSDSpolyacrylamide gel of ammonium sulfate precipitates (3.5 pg of protein/slot) from DNA-cellulose eluates I, 11, and 111. Surprisingly, from this gel it would appear that the purity of the receptor protein (arrow) is actually lower than that calculated from the specific activity. However, when 35pgof protein of DNA-cellulose I1 and 100 pg protein of DNAcellulose I11 were stained with Coomassie blue after SDS-gel electrophoresis, the relative intensities of several protein bands were markedly different (Fig. 2b, lanes IZ and IZI). In particular, the receptor band (arrow) is much more prominent relative to other bands (e.g. in Fig. 2b, lanes I I and III) whereas some other bandsrevealed by silver staining aremuch weaker in the Coomassie-stained gels. In addition, the Coomassie staining appears in good agreement withthe purity calculated from the purification data. Similar differences in staining behaviors have been observed for the chick oviduct progesterone receptor.’ Two main contaminating proteins are present in the eluate of the DNA-cellulose I1 column (Fig. 26, lane I n . When necessary, the majority can be easily removed by DEAE chromatography, after removal of ammonium sulfate by extensive dialysis, to give approximately 60% pure receptor protein with an overall 15% recovery (28) (Fig. 2c, 10% SDSpolyacrylamide gel, and Table I). Purification of Rat Liver Glucocorticoid Receptor by Dexamethasone Affinity Chromatography (Method B)-For affinity chromatographic purification, we used a matrix containing dexamethasone 176-carboxylic acid bound by an amide bond to Aff-Gel-102. The non-[3H]TA-labeledcytosol was filtered through a phosphocellulose column as described for method A and the flow-through was applied onto the affinity column at a flow rate of 1.5 ml/min. This flow rate was optimum to enable the maximum interaction of the binding components with the stationary matrix. We chose a column technique because the purity of the material was greater than for a batchwise procedure; it also avoided losses due to sticking of the matrix to the walls of the incubation container and was more effective for removing unbound material. After washing the column extensively overnight with buffer A andthen 250 mM NaCl in buffer A, the elution was performed by incubating the matrix for 5 h at 0 “C in 20 mM Tris-HC1, pH 7.4, 10 mM mercaptoethanol, and 10% glycerol containing 50 m M NaSCN (45) and 1 p~ triamcinolone acetonide (20 ml). After elution, the column was washed with an additional 30mlof elution buffer and the eluate and the washing were combined. A 5-ml aliquot of this fraction was immediately filtered throughSephadex G-50 and theexcluded protein peak was used to determine the amount of [3H]TA-
* H. Gronemeyer and M. V. Govindan, manuscript in preparation.
12919
binding activity released from the matrix aswell as for photoaffinity labeling studies. As shown in Table 11, more than 90% of the binding activity could be eluted from the affinity matrix, being nearly 5900-fold purified. Analysis of this activity by silver staining (Fig. Zg, lane 1 ) and photoaffinity labeling (Fig. 2i, lune 1 ) shows that the same band is cross-linked as in method A. For further purification, the affinityeluate was diluted, adjusted topH 7.8, activated, and applied onto a DNAcellulose column. The column was washed and eluted, and the eluate was immediately cross-linked as described above (Fig. 2h and 2i, lane 2). Table I1 shows that the final purification thus achieved is approximately 7800-fold with a recovery of about 63%. Analysis of the DNA-cellulose-purified affinity eluate by affinity labeling (Fig. 2i, lune 2) and silver staining (Fig. 2h) shows that intact glucocorticoid receptor had been purified. Preparative SDS-Polyacrylamide Gel Ekctrophoresis of GlucocorticoidReceptor Purified by Method A-DNA-cellulose column eluates I1 and I11 were further purified by preparative SDS-polyacrylamide gel electrophoresis using a continuous flow elution device as described under “Materials and Methods.” 35 pgof DNA-cellulose I1 and 100 pg of DNA-cellulose I11 eluates were analyzed by SDS-polyacrylamide gel electrophoresis and fluorography (Fig. 3, CII and CUI) before the preparative gel electrophoresis. The differences observed between the autoradiograph in Fig. 2d (lunes II and III) and Fig. 3C (lunes II and III) may be due to proteolysis. Whether this proteolysis takes place during the storage or thawing is not yet clear. Even though most of the degradation products (Fig. 3C, arrows I, 2, and 3) arealso present initially (Fig. 2d, lane I n , they have increased during storage (see Figs. 201, 2e, and 3C). Most of the degradation products are common to both DNA-cellulose I1 and DNA-cellulose I11 eluates. Elution and recovery were followed by measuring cross-linked radioactivity in 1O-wlaliquots of each fraction. The elution patterns obtained are shown in Fig. 3, AII and AIM Fractions of the major peak were analyzed by SDS-polyacrylamide gel electrophoresis and silver staining. The preparative gel fractions obtained from DNA-cellulose I1 (Fig. 3AII) and I11 (Fig. 3AIII) eluates exhibited only one silver-stained protein band at the height of the major radioactive peak (Fig. 3, BII and SUI). As can be seen in Fig. 3, BII and BIII, these peak fractions contained two different components (e.g. 32-34 and 35-37, Fig. 3BII), which may be due to limited proteolysis. After pooling the major (35-37) as well as theminor radioactive peak fractions (6-8, 10-13, and 20-25of the DNAcellulose I1 eluate) and 10%trichloroacetic acid precipitation, an aliquot of each pool was applied onto a SDS-polyacrylamide gel and analyzed by fluorography. As shown in Fig. 3 0 , lane 4 , the major radioactive pool contained only the crosslinked 92-kDa receptor protein. The minor peak fractions contained cross-linkedpeptides with molecular masses of 36, 45, and 64 kDa (Fig. 3 0 , lunes 1-3, arrows I, 2, and 3). Antibody Studies-Low amounts of glucocorticoid receptor purified by method A and furtherpurified by preparative gel electrophoresis were sufficient for the production of antibodies. A total of 10 pg/rabbit was used for five injections as described previously (26). Antibodies could be detected by ELISA after the third injection (Fig. 4A). Antigen (fraction 40; see Fig. 3BIII) ranging from 500 to 5 ng (determined from the cross-linking efficiency and specific activity of [3H]TA) was employed in theassay. As little as 5 ng (Fig. 4 A , well 3b) of glucocorticoid receptor could be detected with 1:1000 diluted antiserum. Further characterization of the antibodies was performed using the immunoblotting technique(Fig. 4, B, C, and D).250
12920
Glucocorticoid Receptor Purification
and Characterization
TABLEI1 SkP
Cytosol
Purification of rat liver glucocorticoid receptor (method B) Total activitySpecificProtein receptor 10-3 x Pml mg cpm/mg pm1/-
3491’
4025
Phosphocellulose flow-through 3818’ 3452 327’ 2810 Affinity column flow-through Affinity eluate 3158‘ 0.612d DNA-cellulose eluate 2185‘ 0.325d a See Table I. ‘Specific activity,25 Ci/mmol. e Specific activity, 250 mCi/mmol. Determined after ammonium sulfate precipitation.
15.9 20.3 944 1230
0.867 1.106 5160 6723
Purifi- Yield Purity’
cation -fold
%
1 1.275 5900 7800
%
100 109 47 62
90 63
pg of cytosolic proteins (lunes 1 and 9 ) ,1.5 pg ofprotein from and B-We investigated the specific interaction of purified ammonium sulfate-precipitated DNA-cellulose I (lunes 2 and glucocorticoid receptor with cloned MMTV sequences by a l o ) , I1 (lunes 3 and II), and 111 (lunes 4 and 12) eluates, and nitrocellulose retention assay. To reveal the specificity of this 150 (lunes 5 and 13), 125 (lunes 6 and 14), 100 (lane 7), and interaction, a series of deletion mutants containing varying 75 (lane 8 ) ng of preparative SDS-gel-purified receptor (Fig. lengths of DNA sequences 5’ to the MMTV transcription 3SZff, fraction 40) were electrophoresed in a 10% SDS-poly- initiation site was chosen. The plasmids p311, pdell, pdel2, acrylamide gel. After electrophoretic transfer onto nitrocel- and pdel3 contained 210-, 190-, 140-, and 80-bp upstream lulose paper, two identical blots were incubated with immune sequences, respectively, from the MMTV transcription initi(Fig. 4B, lunes 1-8) or preimmune serum (Fig. 4B, lunes 9- ation site(33). They contained also a uniform length of coding 14), followed by immunodetection using a peroxidase-labeled sequences of 100 bp. Digestion of the plasmids with EcoRI second antibody. Fig. 4B (lunes 1-8) demonstrates that only and Sphl is shown in Fig. 5A. Before performing the binding one protein band is immunoreactive. For the cytosol, only a experiments, aliquotsof the labeled DNA fragments (Fig. 5B, very faint band could be seen at the same position on the lanes a, c, e, and g) were tested by filtration through nitroceloriginal blot. No staining was observed on blots incubated lulose membranes to determine any nonspecific trapping. As with preimmune serum (Fig. 4B, lunes 9-14). The staining shown in Fig. 5B (lunes b, d , f , and h ) , the end-labeled was roughly proportional to the amount of the 92-kDa glu- fragments were notretainedon the filter,indicating the cocorticoid receptor protein. Additional blotting experiments absence of single-stranded regions. In Fig. 5C, lunes b, d , f , with 3 wg of DNA-cellulose 11 (Fig. 4CII) and I11 (Fig. 4CIZI) and g, the receptor-DNA interaction studied usingthe recepeluates and antibodies elicited against the hormone affinity tor purified by method A and in Fig. 5 0 , lunes b, d , f , and g, chromatography-purified glucocorticoid receptor (26) gave the DNA retained after incubation with the receptor purified identical results. When the same experiment was performed by method B are shown. As can be seen, the specificity of with 0.5 pg of affinity eluate (Fig. 2g, lune I) or with 0.5 pg of interaction of glucocorticoid receptor with MMTV DNA is affinity eluate further purified by DNA-cellulose chromatog- independent of the method of purification. This interaction raphy (Fig. 2h) and antibodies raised against glucocorticoid is specific to MMTV-LTR as the retention of the second receptor purified by method A, immunoreaction could be hybrid fragment not containingthe binding site is very weak. detected with the intact glucocorticoid receptor band (Fig. With glucocorticoid receptor purified by method A or B and 40, lanes 1 and 2). Western blotting analysis with the degra- stored as described, nitrocellulose-binding assays have been dation products purified by preparative SDS-gel electropho- successfully performed up to 6-8 weeks after purification. resis (36,45, and 64 kDa; Fig. 3D, lunes 2 and 3 ) demonstrated DISCUSSION that only 45- and 64-kDa receptor fragments reacted with the antibodies (not shown). In this paper, we describe two rapid and quantitative proTo analyze whether these antibodies will also react with cedures for partial purification of native rat liver glucocortinative nonpurified glucocorticoid receptor molecules, a pro- coid receptor. Method A is based on the differential DNAtein A-Sepharose immunoadsorbent assay was performed. In cellulose-binding properties of the activated and nonactivated order to decrease the unspecific binding of cytosolic proteins form of the receptor. In method B, the specific binding of the to protein A-Sepharose, a preincubation with this matrix was receptor to the dexamethasone affinity matrix was the major found to be necessary. Thereafter, 1 ml of cytosol containing purification step. 50,000 cpm of binding activity was incubated with immune or Method A is a modification of the purification procedures preimmune serum followed by adsorption to protein A-Seph- described by Eisen and Glinsman (27) and Wrange et al. (28). arose as described under “Materials and Methods.” Totals of 50-80% pure preparations with yields between 20 and 45% 13,500 and 17,000 cpm were eluted from protein A-Sepharose had been achieved previously by specific dissociation of the receptor-hormone complex from DNA-cellulose using either incubated with 1:200 and 1:lOO diluted immune serum, respectively. No radioactivity was absorbed on the protein A- pyridoxal 5“phosphate (28) or magnesium chloride (53). To Sepharose column when the cytosol was treated with radio- avoid quenching of the photoaffinity labeling reaction due to active hormone in the presence of a lOOX excess of cold pyridoxal 5’-phosphate, this reagent was omitted for the steroid. Similarly, no radioactivity was retained after incu- dissociation of hormone-receptor from DNA-cellulose. For bation of labeled cytosol with preimmune serum. Fluorogra- the isolation and furtheranalysis of all the hormone-binding phy of the eluates after trichloroacetic acid precipitation and components bound to DNA-cellulose, we found that quantiSDS-gel electrophoresis (Fig. 4E) shows that a single band of tative elution can be obtained with NaCl alone. As has been described by several groups (14, 28), the effiM , = 92,000 -+ 400 reacted with the antibodies. DNA-binding Assay with Receptor Purified by Methods A ciency for DNA-cellulose binding of rat liver glucocorticoid
Glucocorticoid ReceptorPurification and Characterization
-
Fracllon
'
w
k
I ,
.
i
A
*
S t 2 3 4
s II 111 8
-
cc
94000
-67000
/ -
h-.
A11
% C
D
FIG. 3. Further purification of glucocorticoidreceptor prepared according to methodA by preparative SDS-gel electrophoresis. Preparative SDS-gel electrophoresis of DNA-cellulose fractions I1 and 111 was carried out asdescribed under "Materialsand Methods." AI1 and A I I I are theradioactive elution profiles of DNA-
12921
receptor after thermalactivation is in the order of 30-50%. It has been speculated that this is due to receptor degradation or a limitation in the activation step. We show here that the hormone-binding activity present in the flow-through of the column could be nearly quantitatively bound to andrecovered from an additional DNA-cellulose chromatography if a further activation step was employed. Analysis of this material by affinity labeling showed that no degradation had occurred during these two activation steps and that intact receptor was isolated. The binding of receptor to the last DNA-cellulose column was entirely dependent on the second activation step andnot due to insufficient primaryactivation or limited capacity of the matrix. These possibilities were excluded by increasing the time for the initial activation step, decreasing the flow rate, ordoubling the column volume. Silver staining is a very important tool to control the complexity of aprotein mixture. However,in the case of glucocorticoid and chick progesterone receptor: the picture obtained by silver staining is a t variance with the calculated purification data in showing an apparently lower purification. These findings should be compared with results from Coomassie blue staining (e.g. Fig. 2, a and b) which we believe to be a more quantitative method for estimating the amountof proteins. Previous studies with dexamethasone 17/3-carboxylic acid bound to a disulfide-containing affinity matrix has demonstrated theisolation of highly purified receptor (29,51). This type of affinitymatrixcan be used only once, since the hormone-receptor complex is dissociated by cleavage of the disulfide bond between the hormone and the stationary matrix (5, 26, 29, 51). Previously described affinity matrices (26, 29, 51,52) for the purification of glucocorticoid receptor contain 12-14-CH2-groups as spacers between the ligand andthe stationary matrix. By affinitychromatography and subsequent DNA-cellulose chromatography (29) or DEAE chromatography (51), glucocorticoid receptor has been purified 10,548-fold with a 68% recovery. Failla et al. (52) were able to purify the glucocorticoid receptor 2,000-fold using a 11deoxycorticosterone derivative-bound affinity matrix. Using this matrix, theycould absorb 80%of the receptor from crude cytosol and elute 40-50% of the absorbed receptor with 5 X M [3H]TA. Lustenberger et al. (39) have described an affinity matrix derived from dexamethasone 17B-carboxylic acid (29, 39). Using a batchwise procedure, they purified the receptor 7,000-fold in two steps with an overall yield of 18%. We prepurified the crude cytosol by phosphocellulose exclusion chromatography and performed the affinity binding by chromatographyonadexamethasone 17/3-carboxylic acidbound matrix. After thorough washing of this column with buffers of different ionic strength, we eluted the hormonereceptor complex quantitatively by incubating the affinity matrix with abuffer containing sodium thiocyanate (45). We further purified the receptor after thermal activation of the affinity eluate by DNA-cellulose chromatography. The glucocorticoid receptor could be purified approximately 7800-fold by employing the hormone-bound affinity matrix cellulose I1 and 111, respectively. BII and BIII, SDS-polyacrylamide gel electrophoresis of selected fractions from A I I (2.5 pl) and A I I I (10 p l ) aliquotsafter silver staining. Numbers indicate the fractions. Fluorography (C) of 35 and 100 pg of DNA-cellulose eluates I1 and 111, respectively, before preparative SDS-gel electrophoresis. Arrows I , 2, and 3: see text; fluorography of an aliquot of the pooled fractions of preparative gel electrophoresis after preparative SDS-polyacrylamide gel electrophoresis (DNA-cellulose 11, A m . D l , fractions 6-8; 02, fractions 10-13, 03, fractions 20-25, 0 4 , fractions 35-37. S, standard proteins. Arrow, glucocorticoid receptor band.
Glucocorticoid Receptor Purificationand Characterization
12922 1
2132
3 4 5 6 7 8 ,. .
.
9 10 11 121314
r
11 111
1 2
C
D
s -
1 2 3
a
B
E
FIG. 4. Immunological studies. A, ELISA with purified glucocorticoid receptor after preparative gel electrophoresis (fraction 40; Fig. 3BIIr). Wells al, bl, e l , and d l contained 500 ng of receptor; wells a2, b2, c2, and d2 contained 50 ng and wells a3, b3, c3, and a7 contained 5 ng of purified receptor. Wells al-a3 and b1-M were incubated with 1:lOO and 1:lOOO diluted antireceptor antiserum,respectively. Wells cl-c3 and dl-& were incubated with 1:lOO and 1:1000 diluted preimmune serum, respectively. Wells between a, b, c, and d were not used for incubation. B, immunoblotting analysis with antireceptor antiserum. 250 pg of cytosol (lanes I and 9 ) , 1.5 pg of DNA-cellulose I (lanes2 and IO),I1 (lanes3 and I I ) , and I1 (lanes 4 and 12) eluates, 150 ng (lanes 5 and 13), 125 ng (lanes 6 and 14), 100 ng ( l a n e 7),and 75 ng ( l a n e 8) of purified receptors from fraction 40 (see Fig. 3BIIr) were electrophoresed in 10% SDS-polyacrylamide gels and electrophoretically transferred onto nitrocellulose paper. Lanes 1-8 show the incubation with 1:500 diluted antireceptor antiserum and lanes 9-14 were incubated with 1:500 diluted preimmune serum as described under “Materials and Methods.” Arrow, glucocorticoid receptor band. C, immunoblotting of 3 pg of DNA-cellulose I1 ( l o n e 1 ) and DNA-cellulose 111 ( l a n e 2 ) with receptor antibodies raised against affinity-purified receptor (26). D,immunoblotting of 0.5 pg of affinity eluate ( l a n e I ) and 0.5 pg of affinity eluate further purified on DNA-cellulose ( l a n e 2), with receptor antibodies raised against the 92,000-Da band purified by method A and preparative SDS-gel electrophoresis. E, protein A immunoadsorbent assay of [‘HI dexamethasone 21-mesylate-labeled cytosolic glucocorticoid receptor. Radioactivity retained by protein A-Sepharose from labeled cytosol incubated with 1:200 diluted immune serum ( l a n e I ) , with 1:100 diluted immune serum (lane 2), and 1:lOO diluted preimmune serum ( l o n e 3 ) . 7.5% SDS-polyacrylamide gel electrophoresis and fluorography after elution with 0.1 M acetic acid and trichloroacetic acid precipitation. Each slot contained one-half of the trichloroacetic acid-precipitated material. S, molecular weight standards.
with an overall yield of 63%. The affinity matrix absorbed 90% of the receptor present in the phosphocellulose flowthrough. The elution of the receptor was quantitative using sodium thiocyanate and 1 p~ triamcinolone acetonide. This low concentration of chaotropic salt hadno negative effect in thermal activation and furtherpurification by DNA-cellulose chromatography. Using one affinity column, this method allows only one purification per week due to time consuming and extensive regeneration of the matrix required before the next isolation. The majority of the receptor present in the affinity eluate (-70%) could be thermally activated, bound to DNA-cellulose, and eluted with0.5 M NaCl. The major silverstained protein band (Fig. 2h) of this preparation has identical mobility as thereceptor purified previously (26,29,51). The integrity of the receptor purified by method A, which is based on the DNA-cellulose-binding property of the activated form of the receptor, and receptor purified by ligand affinity chromatography (methodB) have been scrutinized by two independent affinity labeling techniques, i.e. electrophilic affinity labeling in crude cytosol and subsequent purification (method A), as well as photoaffinity labeing after the final purification step (methodsA and B).
When preparations obtained according to both DNA-cellulose chromatography(method A) and hormoneaffinity chromatography (method B) were photoaffinity labeled and electrophoresed on SDS-polyacrylamide gels, the corresponding fluorographs showed that a major polypeptide of M , = 92,000 f 400 is labeled (Fig. 2, d and i). Theapparent molecular weight of the glucocorticoid receptor is identical after purification by method A, using cytosol labeled with dexamethasone 21-mesylate (Fig. 2e). The second intensely labeled component of M , = 45,000 (Fig. 2, d l l and i, lane 2) reported earlier (26) is the major degradation product. Two additional components of 79,000 and 72,000 Da are reported by Payvar et al. (53), which are minor cross-linked components in our purification. This may be due to the decreased time required for DNA-cellulose chromatography as shown in Fig. 1. In quantitative respects, the efficiency of electrophilic labeling was 10-fold higher than the -5% cross-linking efficiency obtained by photoaffinity labeling. We (9) and others(54) have described a 42,000-Da rat liver glucocorticoid receptor obtained from frozen tissue. As previously suggested, it is very likely that this is a fragment of the receptor generated by proteolytic attack after lysosomal
Glucocorticoid Receptor
Purification.arzd Characterization
12923
polypeptide can be identified already in the cytosol (7,8,1214) and even “fingerprinted” by limited protease digestion (10). In the preparations of glucocorticoid receptor purified L b P by method A, 95% of the cross-linked radioactivity represents intact receptor as demonstrated by the scans in Fig. 2h. The B.3 residual 5% is comprisedof about seven polypeptides showing -5.1 distinct bands on the fluorograph (e.g. Fig. 2, dII and ell). These peptides are also co-purified with the intact receptor -3.2 when the labeling is performed inthe cytosol withdexameth-3.1 c asone 21-mesylate. We therefore conclude that they contain -2.a both the ligand-binding as well as theDNA-binding domain. -2.2 - 2.1 These bands are seen after photoaffinity labeling of receptor preparations using method A or B, whereas they disappear when an excess of nonlabeled hormone is included in the -13 incubations. It is likely that they are endogenous breakdown -.9f products present already in the cytosol, rather than a result of degradation during the purification. Similar results have been obtained previously usinga variety of methods to analyze crude and partially purified receptor (5, 8, 13, 14, 27, 53, 55). A Analysis of the photoaffinity labeled DNA-cellulose eluates stored at -70 “C before preparative SDS-gel electrophoresis (Fig. 3, CII and CIII) compared to samples analyzed immediately after the purification (Fig. 2, dII and dIII) show the a b c d e f g h a b c d e f g h enhancement of proteolytic degradation. There are at least four distinct intensely labeled regions inthe autoradiograph, labeled polypeptidesof M , = 92,000,45,000, and 36,000 being the most predominant bands. The region 64,000-67,000 Da is labeled to a lesser extent in the DNA-celluloseI1 eluate (Fig. 3CII) compared to the DNA-cellulose I11 (Fig. 3CIII) eluate. This enhancement of lower molecular weightfragments during storage or thawing can only be due to proteolytic degradation. The detection and analysis of the receptor during preparative SDS-gel electrophoresis could be monitored by followingthe irreversibly bound radioactive hormone (Fig. 3). The proteolytic activity co-purified inthe preparations seems to digest the receptor to fragments of defined sizes (Fig. 3BII). Additional characterization of the receptor molecules by raising antibodies against the purified component shows that these antibodies react not only with the purified or partially purified cross-linked and noncross-linked denatured receptor band (Fig. 4B)but also with native cytosolic receptor as shown by the protein A-Sepharose assay (Fig. 4E). The antibodies raised against the receptor obtained by affinity chromatogFIG.5. DNA-binding studies with glucocorticoid receptor raphy (26) reacted with the 92,000-Da polypeptideprepared purified by method Aor method B.Aliquots of the plasmid DNAs by method A in immunoblotting analysis (Fig. 4C) and the from p311, pdell, pdel2, and pdel3 after digestion with EcoRI and antibodies raised against the purified glucocorticoid receptor Sphl (A, lanes a-d, respectively) after electrophoresis on a1%agarose gel stained with ethidium bromide; s, adenovirus Hind111 fragments obtained by method A reacted with hormoneaffinity-purified as size markers. B, %%labeled restriction fragments from plasmids receptor equally well (Fig. 40). The M , = 92,000 +- 400 ofthe unactivated cytosolic receptor p311, pdell, pdel2, and pdel3 before filtration through nitrocellulose filters (lanes a, c, e, and g, respectively) and DNA retained on protein, observed after protein A immunoadsorbent assay nitrocellulose filters in the absence of receptor (lanes b, d, f, and h, (Fig. 4E) and photoaffinity labeling of the glucocorticoid respectively). C, nitrocellulose-retained DNA fragments from p311 (lane b), pdell (lane d ) , pdel2 (lane fl, and pdel3 (lane h) after receptor in the affinity eluate (Fig. 2g and 2i, lane I ) , reported agreement with the size determination incubation with glucocorticoid receptor purified by method A (DNA- in this paper, is in close cellulose 11). D, DNA fragments from plasmids p311 (lane b), pdell of “activated” glucocorticoid receptor from rat liver (4, 7, 8, (laned), pdel2 (lanefl,and pdel3 (lane h) adsorbed onto thenitrocel- 11-14,26-29,51,53). lulose filters after incubation with receptor purified by method B. The proteolytic products purified by preparative SDS-gel Lanes a, c, e, and g in C and D represent the labeled DNA fragments electrophoresis were analyzed by Western blotting. The refrom p311, pdell, pdel2, and pdel3, respectively, before filtration. ceptor fragments of 45,000 and 64,000 Da reacted with antireceptor antiserum, whereas the 36,000-Da fragment did not breakage (9,25,30). In thereceptor preparations described in react (not shown). Thus, the proteolytic fragments of glucothis paper, no such fragment was seen even in the group of corticoid receptor (45,000 and 64,000 Da) contain not only the hormone-binding and DNA-cellulose-bindingdomains, endogenous breakdown products (Fig. 2d). In fact, wenow routinely check the quality of our preparations for the absence but also the immunoreactive domain(56). Studies are in progress to characterize the receptor “degradation”products of this fragment. The assay for receptor purification described here is a which donot contain hormone-bindingactivity, but the DNAsignificant improvement on the existing analytical tools for cellulose binding and immunoreactivedomains. Using preparative SDS-gel electrophoresis, it is possibleto evaluating receptor purity and integrity. The intact receptor
a b c d
S
a b c d e f g h
12924
Glucocorticoid Receptor Purification and Characterization
obtain homogeneous receptor material for primary structural studies. We aim a t isolating thetrypticor chymotryptic peptides generated from the receptor purified by preparative SDS-gel electrophoresis by reverse phase liquid chromatography. Partial amino acid sequence data may enable the synthesis of oligonucleotide mixtures which could be used for screening a rat liver cDNA library. The purified receptor preparations have been tested for specific DNA interaction using the nitrocellulose-binding assay. This is to our knowledge the first comparison of the receptor purified by different approaches binding to specific DNA sequences in MMTV-LTR in vitro. Sequences defined by this mapping have been shown to be essential for the regulation of MMTV transcription by glucocorticoids in vivo (33, 41, 57-60). Extensive in vitro binding and footprinting studies with purified glucocorticoid receptor and cloned MMTV sequences have demonstrated the specific binding region to be between 100 and 210 nucleotides upstream of the transcription initiation site (51, 61-63). Elucidation of the role of the interaction of glucocorticoid receptor with specific sequences in eliciting hormonal response is awaiting. The binding of the receptor purified by either of the two methods described in thispaper appears to occur with equal efficiency to thepreviously described sequences in the MMTV-LTR. Acknowledgments-Wewouldlike to express our thanks to A. Alberga, J. P. Le Pennec, and P. Oudet for expert criticism and stimulating discussions. We thank C. Aron and B. Boulay for the preparation of the manuscript. We would like to thankG. Richards a n d T. Reudelhuber for many useful comments. We thank J. Majors for the MMTV-LTR deletion mutants. We are grateful to Prof. P. Chambon for support, encouragement and critical reading of the manuscript. REFERENCES 1. Benisek, W. F. (1977)Methods Enzymol. 46,469-479 2. Taylor, C. A,, Jr., Smith, H. E., and Danzo, B. J. (1980)Proc. Natl. Acad. Sci. U. S. A. 77,234-238 3. Dure, L. S., IV, Schrader, W. T., and OMalley, B. W. (1980)Nature ( L o n d . ) 283,784-786 4. Govindan, M. V., and Manz, B. (1980)AUergobgk 4,204-213 5. Gronemever. H.. and Ponm, 0.(1980)Proc. NatL Acad. Sci. U. S. A. 77, 2108-2112 6. WestDhal. H. M.. Fleischmann., G.., and Beato.. M. (1981) . . Eur. J. Biochem. 119,101-106' 7. Nordeen, S. K., Lan, N. C., Showers, M. 0..and Baxter, J. D. (1981)J. Biol. Chem. 256,10503-10508 8. Dellweg, H. G., Hotz, A., Mugele, K., and Gehring, U. (1982)EMBO J. 1, 285-289 9. Gronemeyer, H., Harry, P., and Chambon, P. (1983)FEBS Lett. 156,287292 10. Birnbaumer. M. E.. Schrader. W. T., and OMalley, B. W. (1983)J. Biol. Chem. 258,1637-1644 11. Gehring, U., and Hotz, A. (1983)Biochemistry 22,4013-4018 12. Simons, S. S., Jr., and Thompson, E. B. (1981)Prm. Natl. Acad. Sci. U. S. A. 78,3541-3545 13. Eisen, H. J., Schleenbaker, R. E., and Simons, S. S., Jr. (1981)J. Biol. Chem. 256,12920-12925 14. Simons. S. S.. Jr.. Schleenbaker. R.E., and Eisen, H. J. (1983)J. Biol. Chem. 258; 2229-2238 15. Weisz,A,, Buzard, R.L., Horn, D., Li, M. P., Dunkerton, L.V., and Markland, F. S., Jr. (1983)J. Steroid Biochem. 18,375-382 16. Katzenellenbogen,J. A,, Carlson, K. E., Heiman, D. F., Robertson, D., Wei, L. L., and Katzenellenbogen, B. S. (19813)J. B+l. Chem. 258.3487-3495 17. Homes, S. D., and Smrth, R. G. (1983)BmhemrPtry 22,1729-1734
18. Koblinsky, M., Beato, M., Kalimi, M., and Feigelson, P. (1972)J. Biol. Chem. 247,7897-7904 19. Giannopoulos, G., Mulay, S., and Solomon, S. (1973)J. Biol Chem. 248, n n l ~-_" nnm """
20. Litwack, G., Filler, R., Rosenfield, S. A., Lichtash, N., Wishman, C. A., and Ginger, S. (1973)J. Biol. Chem. 248,7481-7486 21. Litwack, G., Cake, M.H., Fillers, R., and Taylors, K. (1978)Biochem. J. 169,445-448 22. Climent, F., Doenecke, D., and Beato, M. (1977)Biochemistry 16, 46944703 23. Yahamoto, K. R., Gehring, V., Stampfer, M. R., and Sibley, C. H. (1976) Ree. Prog. Horm. Res. 32,332 24. Sherman, M. R., Pickening, L. A., Rollwagen, F. M., and Miller, L. K. K. (1978)Fed. Proc. 37,167-173 25. Wrange, 8.,and Gustafsson, J.-A. (1978)J. Biol. Chem. 253,856865 26. Govindan, M. V.,and Sekeris, C. E.(1978)Eur. J. Biochem. 89,95-104 27. Eisen, H. J., and Glinsman, W. H. (1978)Biochem. 171,177-183 28. Wrange, 8.,Carlstedt-Duke, J., and Gustafsson, J . - d (1979)J.Bwl. Chem. 264,9284-9290 29. Govindan, M. V., and Manz, B. (1980)Eur. J. Bioehem. 108,47-53 30. Carlstedt-Duke, J., Wran e 6 , Dahlerg, E., Gustafaaon, J.-A., and Hogberg! B. (1979)J. BioL 254,1537-1539 31. McBlrun, W. A., and Shyamala, G. (1980)J. Bml. Chem. 256,3884-3891 32. Sherman, M. R., Moran, M. C., Tuozon, F. B., and Stevens, Y. W. (1983) J. BioL Chem. 258, 10366-10377 33. Majors, J., andVarmus, H. (1983)Proc. Natl. Acad. Sci. U. S. A. 80,5866m7n
34. Br;;dford, M. M. (1976)Anal. Biochem. 72,26254 35. Laemmli, U.K. (1970)Nature ( L a n d . ) 227,680-685 36. Ansorge, W. (1982)in Proceedings of Electrophoresis (Stathakos, D.,ed) p. 235-245,Walter de Gruyter, Berlin 37. Algerts, B. M., and Herrick, G. (1971)Methods Enzymol. 21,198-217 38. Burgess, R. R. (1969)J. Bml. Chem. 244,6160-6167 39. Lustenberger, P., Formstecher, P., and Dautrevaux, M. (1981)J. Steroid Biochem. 14,697-703 40. Rousseau, G. G., Baxter, J. D., and Tomkins, G. M. (1972)J. Mol. BioL 67,99-115 41. Lee, F., Mulligan, R., Berg, P.,and Ringold, G. (1981)Nature ( L a n d . ) 294,
-"99R-329 "_
42. Rousseau, G. G., H i e n s , S. J., Baxter, J. D., Gelfand, D., and Tomkins, G. M. (1975)J..BmL Chem. 250.6015-6021 43. Atger, M., and MII om, E. (1976)J. BmL Chem. 251,4758-4762 44. Eisen. H. J.. and ginsman., W. .(1976) . Biochem. Bwzhvs. . - Res. Commun. 70;367-372 45. Green, G. L., NOhru, C., Engler, J. P., and Jensen,E. V. (1980)Proc. Natl. Acad. Sei. CJ. S. A. 77,5119-5119 46. Engvall, E. (1980)Methods Enzymol. 70,419-439 47. Burnette, N. W. (1981)A d . Biochem 112,195-203 48. 0 et, S., Carlsted-Duke, J., Wrange, 8.,Carlstrom, K., and Gustafsson, J. %.(1981)Biochrm. Bmphys. Acta 677,205-219 49. Towbin, H., Staehelin, T., and Gordon, J. (1979)Proc. NatL Acad. Sei. U. S. A. 76,4350-4354 50. Maniatis, T., Fritsch, E.F., and Sambrook, J. (1982)Molecular Cloni A Laboratorv Manual. uu. 117-119. Cold Surina ~- Harbor Laboratorr,?old Spring H & b r , NY' 51. Govindan, M. V., Spiess, E., and Majors, J. (1982)Pro~.Natl. Acad. Sci. U. 5'. A. 79,5157-5161 52. Failla, D., Tomkins, G. M., and Santi, D. (1975)Proc. Natl. Acad. Sei. U. S. A. 72,3849-3852 53. Pa ar, F., De Franco, D., Firestone, G. L., Edgar, B., Wrange, 8.,Okret, Gustafsson, J. A., and Yamamoto, K. R. (1983)CeU 36,381-392 54. Wes hal H M., and Beato, M. (1980)Eur. J. Bmhem. 106,395-403 55. Hou2ey, b. R . , and Pratt, W.B. (1983)J. BioL Chem. 258,4630-4635 56. Carlstedt-Duke, J., Okret, S., Wrange, 8., and Gustafsson, J. A. (1982) Proc. Natl. Acad. Scr. U. S. A. 79,4260-4264 57. Huang, A. L., Ostrowski, M. C., Berard, D., and Hager, G. L. (1981)Cell 27.245-256 58. Groner, B., Kennedy, N., Rahmdorf, U., Herrlich, P., Van Ooyen, A., and Hynes, N. E. (1982)in Hormones and CeU Regulation (Dumont, J. E., Nunez, J., and Schutz, G., e&) pp. 217-228, Elsevier/North-Holland Biomedical Press, Amsterdam 59. Chandler, V. L., Maler, B. A., and Yamamoto, K. R. (1983)Cell 33, 489499 60. Hynes, N., Van Ooyen A. J. J. Kennedy, N., Herrlich, P., Ponta, H., and Groner, B. (1983)P k . NatL'Acad. Sci. U. S. A. 80,3637-3641 61. Pa ar, F., De Fran 0,D., Firestone, G. L., Edgar, B., Wrange, 8.. Okret, Gustafsson, J.X., and Yamamoto, K. R. (1983)Cell 36,381-392 62. Scheidereit, C., Geisse, S., Westphal, M., and Beato, M. (1983)Nature ( L a n d . ) 304,749-752 63. Pfahl, M. (1982)CeU 31.475-482 ~~
~
