Biochem. J. (1979) 181,201-213 Printed in Great Britain

201

Chemical Cross-Linking of Chick Oviduct ProgesteroneReceptor Subunits by Using a Reversible Bifunctional Cross-Linking Agent By Maria E. BIRNBAUMER, William T. SCHRADER and Bert W. O'MALLEY Department of Cell Biology, Baylor College of Medicine, Houston, TX 77030, U.S.A.

(Received 27 November 1978) Chick oviduct progesterone-receptor proteins were treated in cytosol with the reversible cross-linking reagent methyl 4-mercaptobutyrimidate. The product of the reaction was a 7S complex that could be detected and recovered after sucrose-density-gradient centrifugation in 0.3 M-KCl. The extent of the reaction was dependent on the concentration of methyl 4-mercaptobutyrimidate and independent of the presence of bound hormone, since unlabelled receptors could also be cross-linked. The cross-linking reaction required conditions in which the cytosol 6S complex was preserved. A Stokes radius of 7.3 nm was determined by gel filtration in Agarose A-1.5m in 0.3M-KCI. The sedimentation coefficient, which was also determined in 0.3M-KCI, allowed us to calculate a mol.wt. of 228 000. We were also able to cross-link partially purified receptor forms isolated by using an Agarose A-15m column. On reduction with 16-mercaptoethanol the complex broke down to 4S monomers that were identified by DEAE-cellulose and phosphocellulose chromatography, adsorption on DNA-cellulose and gel filtration in an Agarose A-i1.5 m column. In most cases, A and B receptor proteins were released in equivalent amounts, implying that the cross-linked form was an A-B complex. Since the initial description of steroid-hormone receptors, considerable progress has been made

towards understanding their mechanism of action (O'Malley & Means, 1974; Jensen et al., 1974; Buller & O'Malley, 1976). These receptors render hormone specificity to the target tissues and interaction with these proteins is a necessary first step for steroids to exert their effects at the nuclear level. Several laboratories have studied the characteristics of cytoplasmic progesterone receptor of the chick oviduct (Sherman et al., 1970, 1976; Schrader & O'Malley, 1972; Schrader et al., 1972; Smith et al., 1974; Spelsberg & Cox, 1976; Toft et al., 1977) and our laboratory reported their purification (Schrader et al., 1976, 1977; Coty et al., 1979). Experiments in vitro showed that purified progesterone-receptor complexes can influence directly transcription of chick oviduct chromatin. The functional form of the receptor in these experiments was shown to be a 6S molecule (Buller et al., 1976; Schwartz et al., 1976). Previously we had shown preliminary evidence indicating that this 6S species is a dimer of two dissimilar A and B subunits (Schrader et al., 1975; Buller et al., 1976). Because the two subunits have different affinities for nuclear constituents it is important to know the composition of the 6S material and the relationship between the A and B proteins. Although the A and B subunits have been purified, there is no direct proof that they exist as a dimer in the target-cell cytosol. One test of this hypothesis Vol. 181

would be to attempt to covalently cross-link the native receptor and demonstrate the presence of both subunits, in equivalent amounts, in this complex. Cross-linking would also provide information about the spatial relationship that exists between the two subunits. Bifunctional cross-linking agents have been used in several systems to explore different properties of oligomeric proteins, ribosomal subunits and other complex systems (Wold, 1967; Peters & Richards, 1977). These techniques have not been applied before to the study of steroid-hormone-receptor proteins. Traut et al. (1973) reported the synthesis of methyl 4-mercaptobutyrimidate and its application as a cross-linking agent for the study of protein arrangements in the Escherichia coli 30S ribosomal subunit. This reagent is a substituted imidoester that reacts with available amino groups in proteins to produce an amidine group. This first reaction incorporates a thiol group into the protein and the second reaction is an oxidative step that creates disulphide bridges between neighbouring thiol groups, resulting in the chemical cross-linking of the proteins involved. The short aliphatic chain makes this reagent very selective, since the reactive groups of the proteins must be very close to be joined by it. The reagent is mild and the cross-linking obtained is reversible, which makes it an ideal choice for our studies. In the present paper we report the successful stabilization of specific receptor forms larger than

202

M. E. BIRNBAUMER, W. T. SCHRADER AND B. W. O'MALLEY

4S by use of this covalent cross-linking agent. The cross-linked material exhibits characteristics similar to those of unreacted 6S receptor. Finally, reduction of the cross-linked complexes yields receptor A and B proteins in equivalent amounts. These data are consistent with our previous hypothesis of receptor structure in which proteins A and B are subunits of a larger complex (Schrader et al., 1975).

Materials and Methods Animals Female 10-day-old White Leghorn chicks were implanted subcutaneously with 20mg of diethylstilboestrol in Carbowax vehicle. They were reimplanted at 17 days and used between 5 and 7 days later. The oviducts weighed between 1.2 and 2g each. Chemicals All chemicals were reagent grade. Tris and sucrose were Ultrapure grade from Schwarz-Mann, Orangeburg, NJ, U.S.A. [1,2-3H]Progesterone (sp. radioactivity 50Ci/mmol) was obtained from New England Nuclear, Boston, MA, U.S.A. Methyl 4-mercaptobutyrimidate, dimethyl suberimidate, dimethyl adipimate, dimethyl 3,3'-dithiobispropionimidate and dithiobis(succinimidylpropionate) were obtained from Pierce Laboratories, Chicago, IL, U.S.A. Reagent-grade triethanolamine base was from Fisher, Pittsburgh, PA, U.S.A. Diethylstilboestrol in Carbowax was purchased from Mattox and Moore, Indianapolis, IN, U.S.A.

Radioactivity determination Aqueous samples (0.5ml) were counted for radioactivity by using 4ml of Amersham/Searle (Arlington Heights, IL, U.S.A.) ACS scintillation fluid. Counting efficiency for 3H was 30%.

Sucrose-density-gradient centrifugation Linear gradients of 5 to 20 % sucrose in the indicated buffers were prepared by using a Beckman Gradient Former. The volume of gradients was 5 ml and 0.2ml of sample was layered on top. Centrifugations were performed in an SW 50.1 rotor for 16h at 45000rev./min (average force 189 000g). With gradients containing 10% glycerol in the buffer the centrifugation lasted 20h. The gradients were fractionated (0.2ml per fraction) in an ISCO fractionator at room temperature (22°C). When further experiments were to be carried out with the fractions, they were collected in the cold room. Sedimentation coefficients were compared with those of bacterial alkaline phosphatase (6S; Garen & Levinthal, 1960) and ovalbumin (3.7S; Kegeles & Gutter, 1951), both labelled with [14C]formaldehyde (Stancel & Gorski, 1975), and of unlabelled human

y-globulin (Schwarz-Mann; 7.3S; Oncley et al., 1947), detected by A280.

Preparation of crude receptorfraction Oviducts were excised from the oestrogenized animals and treated as before (Sherman et al., 1970, 1976; Schrader & O'Malley, 1972; Schrader et al., 1972; Smith et al., 1974; Spelsberg & Cox, 1976; Toft et al., 1977). Cytosol was prepared by using a standard buffer containing 20mM-triethanolamine, pH 8.0, 1 mM-EGTA, and 12mM-1-thioglycerol. The pH of all buffers was determined at 23°C. Unless otherwise indicated, the cytosol was labelled with 20nM-[1,2-3H]progesterone for 2h at 0°C before the chemical modification.

Gel-filtration chromatography of receptor An Agarose A-1.5m (200-400 mesh; Bio-Rad, Richmond, CA, U.S.A.) column (82cm x 26cm) was used to characterize the Stokes radius of the cross-linked material. The column was eluted with I0 mM-Tris, pH 7.4, containing 0.3 M-KCI and 1 mMEGTA. The column was calibrated with Blue Dextran 2000 (VO), thyroglobulin (Stokes radius= 8.6nm; Edelhoch, 1960), receptor monomer B (Stokes radius = 6.4nm; Kuhn et al., 1977), receptor monomer A (Stokes radius= 4.6nm; Coty et al., 1979), and KCl (V,). A column (2.6cmx 37cm) of Agarose A-15m (200-4W mesh; Bio-Rad) was used to prepare partially purified receptor dimers from the cytosol. This gel can resolve high-molecular-weight receptor from the monomer forms and from the bulk of the contaminating protein. The column was calibrated with Blue Dextran 2000 (VO), and KCI (V,) in separate runs, with 20mM-triethanolamine, pH8.0, containing 1 mM-EGTA, as buffer. When amidinated cytosol was chromatographed to partially purify receptor complexes and at the same time eliminate excess reagent and thiol groups, the buffer was the same. Samples (3 ml) were chromatographed as indicated in the text. Ion-exchange chromatography of receptors DEAE-cellulose and phosphocellulose chromatography were done as described by Buller et al. (1976). Ion-exchange columns (1.5 ml bed volume) in 20mMtriethanolamine, pH 8.0, containing 1 mM-EGTA were loaded with receptors prepared as described in the text (70000c.p.m. of bound [3H]progesterone) and eluted with a KCl gradient (0 to 0.5 M; total volume 40ml). Portions (500,1l) were counted for radioactivity. When reduced cross-linked receptor complexes or unmodified receptors were run, the buffers also contained 12mM-1-thioglycerol. KCI molarities were determined by conductivity. 1979

203

PROGESTERONE-RECEPTOR CROSS-LINKING DNA-cellulose chromatography The DNA-cellulose chromatography of receptor (Alberts & Herrick, 1971) was done as described by Schrader (1975). The samples were applied to small columns (1.5 ml packed bed) of DNA-cellulose equilibrated in 20mM-triethanolamine, pH8.0, containing 1 mM-EGTA. The columns were washed with 5 ml of the same buffer and then a KCl gradient (0 to 0.5M; total volume 40ml) was applied to elute the radioactivity. When reduced cross-linked complex was chromatographed, the buffer also contained 12mM-1-thioglycerol.

Polyacrylamide-gel electrophoresis Polyacrylamide-gel electrophoresis in the presence of 1 % sodium dodecyl sulphate was done by the procedure of Laemmli (1970). Slab gels (14cm x 24cm) contained 7.5% (w/v) acrylamide/0.2 % (w/v) bisacrylamide. They were stained with 0.05 % (w/v) Coomassie Brilliant Blue in 5 % methanol (v/v)/7.5 % (v/v) acetic acid. They were destained with 5% methanol/7.5 % acetic acid (v/v). Protein molecular-weight standards used were ovalbumin (mol.wt. 43000; Castellino & Barker, 1968), conalbumen (mol.wt. 86000; Fuller & Briggs, 1956) and heavy myosin (mol.wt. 200000; Gazith et al., 1970). Results Conditions for the cross-linking reaction with methyl 4-mercaptobutyrimidate The first step of the cross-linking reaction with methyl 4-mercaptobutyrimidate is the amidination of the protein. The reaction is rapid; owing to hydrolysis of the reagent it has a limited lifetime in aqueous solutions. This step can be done at different temperatures (0-30'C), and it is favoured by high pH values (>e8). The first step takes place in the presence of a thiol-specific compound to ensure that the methyl 4-mercaptobutyrimidate will not undergo spontaneous cross-linking with itself or with protein thiol groups. Once this phase is completed it is necessary to eliminate the excess reagent and the reducing agent before the oxidation that generates the disulphide bridges. In the present experiments this was achieved by dialysing the sample against 20mM-triethanolamine, pH8, containing 1 mMEGTA. The second step consists of the addition of H202 to create an oxidative environment and the formation of the disulphide links. When 2.5 g of chick oviducts was homogenized in 10ml of 20mM-triethanolamine, pH 8.0, containing 1 mM-EGTA and 12mM-1-thioglycerol, labelled with [3H]progesterone and analysed by sucrose-densitygradient ultracentrifugation in the absence of KCI, the 6S form of the receptor was the predominant form present, as shown in Fig. 1. A shoulder in the Vol. 181

8S region was also seen in some preparations. The 6S progesterone receptor has been tentatively identified as the form active in transcription studies in vitro (Buller et al., 1976; Schwartz et al., 1976). This material can convert into higher aggregates sedimenting at 8-9S or can dissociate to yield the two 4S monomeric receptor subunits (Schrader et al., 1975; Schwartz et al., 1977), particularly when exposed to solutions of high ionic strength. This buffer thus provided a convenient starting material of 6S receptor for cross-linking. The first step of the reaction, amidination, was favoured by an increase in the concentration of triethanolamine. The protocol involved exposing the cytosol briefly to the higher concentration of buffer and to the reagent. Typically 5 ml of the crude receptor preparation was made to a concentration of 5OmM-triethanolamine, pH8.0, by adding 150,ul of lM-triethanolamine, pH8.0. Then 13,u1 of 500mMmagnesium acetate was added to bring the solution to 1 mM-magnesium acetate. A freshly-made solution of 10mg of methyl 4-mercaptobutyrimidate/ml (43-860pl) was added immediately, bringing the final concentration in the mixture to 0.5-10mMmethyl 4-mercaptobutyrimidate. After 20min at 0°C in ice the preparation was dialysed against 20mM-triethanolamine, pH 8.0, containing 1 mmEGTA. This step eliminated excess unreacted methyl

12 ' 0

0

0

10

2

Top

Bottom

Fraction number

Fig. 1. Sucrose-density-gradient ultracentrifugation analysis of crude cytosol progesterone receptor preparation The receptor was prepared as de'scribed in the Materials and Methods section. The gradients were run in the absence of KCI (o) or in the presence of 0.3 m-KCI (o). Sedimentation coefficients were determined by comparison with ovalbumin, bacterial alkaline phosphatase and human y-globulin as mentioned in the Materials and Methods section.

M. E. BIRNBAUMER, W. T. SCHRADER AND B. W. O'MALLEY

204

4-mercaptobutyrimidate and 1-thioglycerol, both of which can interfere with the oxidation step. After dialysis for 1.5h with three changes of buffer, the amidinated cytosol was oxidized by addition of 7.5#i1 of 13M- (30°/0) H202 to reach a final concentration of 20mM-H202. After 20min (or longer periods) the cytosol was analysed for the presence of cross-linked material by sucrose-density-gradient ultracentrifugation in the presence of 0.3 M-KCI. Excellent stability of hormone-receptor complexes was obtained, since gel-filtration chromatography in Sephadex G-75 of the oxidized material showed almost no free progesterone present after the treatment with H202. During the complete reaction sequence only about 7 % of the receptor-bound radioactive progesterone was lost. Thus, the method is gentle and causes little denaturation of receptors. Concentration dependence of the reaction In the absence of methyl 4-mercaptobutyrimidate the dialysis and oxidation steps did not yield any receptor form resistant to dissociation into 4S monomers. Sucrose-density-gradient-ultracentrifugation analysis showed the typical 4S profile (Fig. 2a). Low concentrations of the reagent (1 mM) yielded a salt-resistant form that sedimented at a position corresponding to fraction 17 (Fig. 2a) with an apparent sedimentation coefficient of 7S. The total amount of hormone-receptor complex, as measured by total 3H entering into the gradient, did not decrease when the concentrations of methyl 4-mercaptobutyrimidate were increased, indicating

that higher concentrations of reagent did not denature or dissociate the receptors. When the concentration of methyl 4-mercaptobutyrimidate was increased, KCl-resistant receptor aggregates appeared in fraction 17 concomitant with the disappearance of 4S monomers from fraction 10 (Fig. 2b). In addition, high methyl 4-mercaptobutyrimidate concentrations also caused the appearance of higher-molecular-weight aggregates that sedimented almost to the bottom of the centrifuge tube. The 7S form was the predominant cross-linked entity seen at the lowest concentrations of reagent, so we decided to examine its properties and composition. To do this we characterized the product obtained with 2.5mM-methyl 4-mercaptobutyrimidate, when about 30% of the complexes sedimented as the salt-resistant 7S species. Since the change in sedimentation coefficient of the receptors could be due to non-specific coupling of receptor monomers to the bulk of the cytosol proteins, we examined the A280 of sucrose gradients run with control and cross-linked cytosol, as shown in Fig. 3. The profile was very similar in both cases. Polyacrylamide-gel electrophoresis in sodium dodecyl sulphate in the absence of fi-mercaptoethanol showed similar results. No gross change was detected in the protein pattern, even after reacting the cytosol proteins with large concentrations of cross-linking reagents (results not shown). Thus the reaction did not cross-link the major cell proteins to each other.

0.8

0.8

(b)

(a)

6

0.6

la

S. t-°

_.

_

.*

C

0.4

0.4-

O .°

12

Cd

C.) Cd

.O

0.6

CIO

E 16

*~

8

0.2

0.2-

.

0

x

4

Top

0 ,.D

0

10

20

4,S 10

0

.0

0

2

4

6

8

10

Top Bottom [Methyl 4-mercaptobutyFraction number rimidatel (mM) Fig. 2. Dependence of the cross-linking reaction on methyl 4-mercaptobutyrimidate concentration (a) Sucrose-density-gradient ultracentrifugation in 0.3 M-KCI of control cytosol (e) and cytosol after treatment with 2.5 mM-methyl 4-mercaptobutyrimidate (o) as described in the text. (b) Change in the amount of 7S material as a function of methyl 4mercaptobutyrimidate concentration. The quantification was done in sucrose gradients run in 0.3 M-KCI.

6S

4,S 6S

85S

20 Bottom

0

10

Togp

8S 20 Bottom

Fraction number Fig. 3. Effect of cross-linking reaction on cytosolproteins (a) Sucrose-density-gradient ultracentrifugation in the absence of KCI. A280 was determined for each 0.2ml fraction. Symbols: *, control cytosol; A, cytosol cross-linked as described in the Materials and Methods section with 2.5 mM-methyl 4-mercaptobutyrimidate. (b) Sucrose-density-gradient ultracentrifugation in the presence of 0.3M-KCI. Control cytosol (e) and cytosol cross-linked with 2.5mMmethyl 4-mercaptobutyrimidate (A) were prepared as described in the text. In both cases the gradients were run in 20mM-triethanolamine buffer, pH8.0, containing 1 mm-EGTA.

1979

205

PROGESTERONE-RECEPTOR CROSS-LINKING

This was consistent with, but does not prove that the receptor subunits were cross-linked to each other and not to other cell proteins. Dependence of the reaction on the integrity of the 6S receptor form To test whether formation of the 7S complex depends on the presence of 6S form, we converted the naturally occurring 6S receptors into 4S monomers by several methods and then reacted them with methyl 4-mercaptobutyrimidate. Exposure of cytosol to high ionic strength (>0.3 M-KCI) or brief warming of the material causes dissociation of receptor aggregates to monomers. When the cross-linking reaction was done in the presence of 0.3 M-KCI, salt-resistant complexes were almost completely obliterated (Fig. 4a). Also, dissociation of the 6S form by warming the cytosol at 30°C for 30min prevented the cross-linking reaction after the material was cooled at 0°C. In the case of the warming experiment there was no change in protein concentration or ionic strength of the buffers, only dissociation of the 6S receptor complex with loss of reaction (Fig. 4b). Since in both cases contaminating cytosol proteins were present, we conclude that intimate subunit-subunit contact was required for crosslinking and that the reaction did not tend to couple receptor monomers to other proteins in the mixture.

Reversibility of the reaction One criterion that successful cross-linking had occurred, rather than mere receptor aggregation or denaturation, was that the cross-linking obtained with methyl 4-mercaptobutyrimidate was reversible under reducing conditions. We performed next the experiment shown in Fig. 5. When cross-linked cytosol was reduced with 120mM-fi-mercaptoethanol, analysis in sucrose gradients containing 0.3 M-KCI showed exclusively 4S monomers and no salt-resistant forms. This experiment confirmed the reversibility of the reaction and also showed that the monomers retained their typical sedimentation characteristics after treatment with methyl 4-mercaptobutyrimidate, i.e. amidination by itself did not generate the saltresistant complexes. Thus it appeared that the methyl 4-mercaptobutyrimidate reaction was in fact crosslinking the receptor subunits as desired. Further characterization of the reaction product was therefore warranted and was carried out as described below. Characteristics of the cross-linked complex Cross-linked cytosol (3 ml) was applied to an Agarose A-1.5 m column equilibrated with 0.3 M-KCI,

20 18 4

W.

2

16

C; 14

-

c)

x

> 12

C._

0

0

10

20

0

10

20

Bottom Top Fraction number Fig. 4. Effect of disruption of the 6S receptor form on the yield of cross-linked 7S material In both cases the cross-linking was done with 2.5 mMmethyl 4-mercaptobutyrimidate as described in the text and the sucrose gradients run in the presence of 0.3 M-KCI. (a) A portion (2ml) ofcytosol was brought to 0.3M-KCI by addition of 425,pl of 2M-KCI in 20mM-triethanolamine, pH 8.0, containing 1 mMEGTA and 12mM-l-thioglycerol and left at 0°C for 30min. The material was then treated with methyl 4-mercaptobutyrimidate. The control was diluted with the same volume of buffer without KCI. (b) The cytosol sample was amidinated and dialysed as described in the text. The sample was then divided into two portions. One half was warmed at 30°C for 30min, the other one kept at 0°C. The warmed material was then cooled at 0°C for 30min and then H202 was added to both samples to complete the

cross-linking reaction. Vol. 181

n 10 Cd pY. 8

x

aO

6

0

10 Top

20 Bottom

Fraction number Fig. 5. Reversibility of the cross-linking reaction wvith reducing agents All sucrose gradients were run in the presence of 0.3 M-KCI. Symbols: o, unmodified cytosol receptors; *, cytosol receptors cross-linked with 2.5 mM-methyl 4-mercaptobutyrimidate; A, cross-linked cytosol receptors reduced with fi-mercaptoethanol for 30min at 0°C.

M. E. BIRNBAUMER, W. T. SCHRADER AND B. W. O'MALLEY

206

20mM-triethanolamine, pH 8.0, and 1 mM-EGTA. The profile obtained is shown in Fig. 6. A peak of radioactive receptor-hormone complex was observed in fraction 62 (bold arrow), which eluted before the B and A monomer peaks. This cross-linked receptor had an apparent Stokes radius of 7.3 nm determined in the Agarose 1.5 m column described in the Materials and Methods section. This material in a sucrose gradient with the same buffer showed a sedimentation coefficient of 7.2S and co-sedimented with human y-globulin (results not shown). From these two measurements, the apparent molecular weight of the intact cross-linked complex was estimated by using the Svedberg formula (Siegel & Monty, 1966); a value for mol.wt. of 228000 was calculated. The protein-elution profile of the column is also shown in Fig. 6. The protein profile and the peaks of A and B monomers did not alter their elution on reaction with methyl 4-mercaptobutyrimidate, and good separation was achieved of the cross-linked material from the unreacted monomers and total protein. A prominent corticosteroid-binding globulin peak eluted after the A monomer, as described originally by Sherman et al. (1970). We next tested the ability of the methyl 4-mercaptobutyrimidate reagent to cross-link the aporeceptors. Cytosol was prepared as described in the Materials and Methods section, but the labelling step was omitted. The cross-linking reaction was then carried out with 2.5 mM-methyl 4-mercaptobutyrimidate by the protocol described above for labelled cytosol. After the oxidation step, the excess H202 was eliminated by gel-filtration chromatography through a Sephadex G-75 column (6cmx2cm) in 20mMtriethanolamine buffer, pH 8.0, containing 1 mm1-:

£

1.6

a

Vo0

ci

*--

A

Vt 1.2 0

c

,

4

0.8

042

'

0.4

x

"20

40

60

80

100

120

140

160

Fraction number

Fig. 6. Agarose A-I.5m chromatography of cross-linked cytosol receptors The column was run as described in the Materials and Methods section in lOmM-Tris (pH7.4)/0.3M-KCI/ lmM-EDTA. Symbols: *, radioactivity profile; o, A280. Calibration of the column is shown by the upper arrows. Vo, exclusion volume; B, elution of B monomer; A, elution of A monomer; V,, total volume determined by KCI elution.

12 10

0~~~~~ 0

%~~~~~~~

la

0

10

20

Top

Bottom

Fraction number

Fig. 7. Sucrose-density-gradient profile of cytosol crosslinked before (a) and after (0) labelling with radioactive progesterone at 20nM as described in the text The gradients were run in the presence of 0.3 M-KCI.

EGTA. The material in the void volume was labelled with 3H-labelled 20nm-progesterone for 2h and analysed by sucrose-density-gradient ultracentrifugation. As shown in Fig. 7, the 7S product was obtained in control and unlabelled material. This confirms our previous results that the presence of the receptor aggregate is independent of the presence of the hormone, and also demonstrates that any conformation changes caused by the steroid do not affect the groups involved in the methyl 4-mercaptobutyrimidate reaction. We have previously described (Schrader et al., 1975) the chromatographic properties of the native 6S receptor found in chicken oviduct cytosol: it binds to DEAE-cellulose and elutes as a single peak at 0.2M-KCI; it fails to bind to phosphocellulose and DNA-cellulose; and it dissociates in the presence of high salt concentration to yield equivalent amounts of A and B 4S monomers that can be identified by column chromatography. The 7S receptor form was predicted to have these properties as well if it were indeed a cross-linked A-B dimer. To test this, the 7S cross-linked receptor produced by 2.5mM-methyl 4-mercaptobutyrimidate was isolated from sucrose gradients run in 0.3 M-KCI. The gradients were pierced at 4°C and fractions from six gradients were pooled. Portions (20p1) of the pooled fractions were counted for 3H radioactivity to identify the radioactive receptor peaks. The material in fractions 16, 17 and 18 was combined for analysis as described in the following sections. It was determined by gelfiltration chromatography in Sephadex G-75 that approx. 30% of the counts in the 7S pool was free progesterone. The same was found when the 6S material from unmodified cytosol was tested. 1979

207

PROGESTERONE-RECEPTOR CROSS-LINKING Recoveries from columns have been corrected by this value. The pooled 7S material always contained 0.3 MKCI. Thus before ion-exchange or adsorption chromatography it was diluted with 20 mM-triethanolamine, containing 1 mM-EGTA to decrease the KCI concentration to 50mM. DEAE-cellulose chromatography of 7S complex The cross-linked material bound to DEAEcellulose columns, as shown in Fig. 8(a). Of the radioactivity 70% was retained and eluted in 0.2MKCI, similar to the elution of the native 6S form from such a column. After reduction of the complex in high salt concentrations, dilution and chromatography in DEAE-cellulose, the profile changes to show the presence of A-protein eluting in a plateau at 0.1 M-KCl and B-protein eluting at 0.2M-KCI (Fig. 8b).

Phosphocellulose chromatography of the 7S complex The 7S material isolated from sucrose gradients did not bind to phosphocellulose columns (results not shown), in agreement with the properties of the native 6S receptor. To test the subunit structure of this cross-linked species, the material was reduced with 120mM-fl-mercaptoethanol in the presence of 0.3M-KCI, diluted to 0.05M-KCI, and then applied to a phosphocellulose column equilibrated in 0.05 MKCI. About 80% of the bound counts were retained by the column, as shown in Fig. 9(b). A KCI gradient was then applied to the column and both the B and A forms were eluted from it in approximately equal amounts. Similar results were obtained when 6S material from unmodified cytosol was isolated from sucrose gradients, treated with 0.3M-KCI and then

;>

(a)

(b)

3 E 4

0

A B,6S

20

40

B

-0.4

A

60 40 20 0 Fractiola number

60

Fig. 8. DEAE-cellulose chromatography of cross-linked 7S receptor obtainedfrom sucrose gradients in 0.3 M-KCI Arrows indicate the elution position of authentic A and B monomers and unreacted 6S material in companion control experiments. (a) The 7S cross-linked complex was chromatographed as described in the Materials and Methods section. The buffer was 20mM-triethanolamine, pH 8.0, containing I mMEGTA. (b) After reduction with 120mM-,8-mercaptoethanol the fraction was diluted and chromatographed as in (a). Fractions (500,1) were assayed for 3H (o) and for conductivity (o).

Vol. 181

E ci 6

(a)

;

20

(b)

B A 4

> 1.5

>I.

It

u

co o

1.0

la x

0.5

0

0

20

t 40

0

20

40

Fraction number Fig. 9. Subuttit composition of unmodified 6S and crosslinked 7S receptor complexes anialysed by phosphocellildose

chromlatography The material was isolated by sucrose-density-gradient centrifugation. (a) Unreacted 6S material was treated with 0.3 M-KCI, diluted and chromatographed in a phosphocellulose column. (b) The cross-linked 7S material was reduced with 120mM,B-mercaptoethanol in the presence of 0.3 M-KCI, diluted and chromatographed. Both gradients were run in lOmM-Tris (pH7.4)/i mM-EDTA/12mM-lthioglycerol. Fractions (500pu) were collected and assayed for 3H (e) and conductivity (c).

chromatographed in a similar column (Fig. 9a). The KCI molarities at which the B and A proteins eluted in the two panels were identical, indicating that the reaction with methyl 4-mercaptobutyrimidate caused little change in electrostatic properties of the two subunits. DNA-cellulose chromatography of the 7S complex We wished to test possible functional properties of the cross-linked preparation and chose binding of the 7S complex to DNA-cellulose. As -shown in Table 1, the 6S form of the receptors does not bind significantly to DNA-cellulose columns when it is applied in 50mM-KCl (Schrader & O'Malley, 1977). However, on dissociation to 4S monomers by KCI treatment, approximately half (47 %) of the receptor binds to the DNA-cellulose column (Schrader, 1975) at 50mM-KCI. Table 1 also shows the results obtained with cross-linked 7S receptors. Similar to the control 6S form, the cross-linked receptors showed low binding to DNA-cellulose. On reduction and liberation of 4S subunits about 60 % of the receptors bound to DNA-cellulose, approximately half of the recovered counts. Thus the 7S cross-linked material has two additional similarities to the native dimer: (1) it does not bind to DNA-cellulose; and (2) a DNA-binding subunit is liberated on dissociation of the complex by reduction. The DNA-binding form

O

208

M. E. BIRNBAUMER, W. T. SCHRADER AND B. W. O'MALLEY Table 1. DNA-cellulose chromatography of receptor complexes

Preparation* 6S pool of unreacted cytosol

Total radioactivity

Non-retained

applied (c.p.m.)

radioactivity (c.p.m.)

Treatmentt None Dissociatedt

Radioactivity eluted (c.p.m.) 0

Amount of recovered complexes bound to DNA§ (Y.)

55820 44410 0 55820 22930 20390 47 7S pool of cross-linked cytosol None 39800 24800 5000 17 39800 10900 15900 60 Dissociated$ * Sucrose-gradient fractions were pooled from tubes containing either unreacted cytosol or cytosol treated with 2.5 mMmethyl 4-mercaptobutyrimidate as described in the text. t Unreacted cytosol 6S receptor pool (2ml) was treated with 3M-KCI (220,u1) to bring the KCI concentration to 0.3 M. After 4h, the sample was diluted 6-fold and applied to the DNA-cellulose column. t Cross-linked 7S pool (already in 0.3 M-KCI) was treated with 13 M-/1-mercaptoethanol as described in the text. § DNA-cellulose columns were run as described in the Materials and Methods section.

I--

d

600

Vo I

B

A

Cross-linking ofpartially purified receptor complexes

V

II

CL

;>

400

*._

co

-_o

200

20

40

60

80

100

120

140

160

Fraction number Fig. 10. Gel-filtration analysis of cross-linked 7S complex The 7S receptor complex isolated from sucrose gradients in 0.3M-KCI was reduced with 120mM,8-mercaptoethanol and applied to-an Agarose A1.5m column (2.6cmx82cm) equilibrated with 20mM-triethanolamine/ 1 mM-EGTA/ 12mM- -tthioglycerol/0.3 M-KCI. The buffer also contained 0.5mg of bovine serum albumin/ml. Fractions (3ml) were collected and I ml portions were counted for 3H radioactivity.

of the receptor eluted from the column at 0.18 M-KCI, a value identical with that seen for authentic receptor A subunit (Coty et al., 1979).

Gel-filtration chromatography Next the reduced 7S material was chromatographed on an Agarose A-1.5 m column equilibrated in 10mM-Tris, pH7.4, containing I mM-EDTA, 12mM-1-thioglycerol, 0.3 M-KCI and 0.5 mg of bovine serum albumin/ml (Fig. 10). The albumin helped to stabilize the receptors during prolonged gel filtration. This column also showed that after reduction the receptor yielded approximately equal amounts of A and B monomers. The column yielded 90% of the counts applied, and both peaks had the same Stokes radii as authentic A and B monomers from untreated cytosol (Fig. 10, arrows B and A).

The 7S cross-linked material prepared from crude cytosol thus has properties of the native A-B dimer. As a final test of the method, we wished to prepare receptor aggregates in partially purified form and then cross-link them. To do this, receptors in cytosol were chromatographed on a gel-filtration column of Agarose A-I Sm, which allows isolation of intact AB complexes at low ionic strength. When unreacted chick oviduct cytosol was chromatographed in an Agarose A-15m column at low ionic stength, a good separation was achieved between the large receptor complexes, and the bulk of the cytosol proteins, as shown in Fig. 11(a). The receptor B and A monomers are poorly resolved from each other on this column, and co-elute with the major protein band. This chromatography purifies the 6S form approx. 15-fold. In our experiment, a companion cytosol was labelled as usual and amidinated with 2.5 mM-methyl 4-mercaptobutyrimidate. The amidinated cytosol was then applied to the Agarose A-ISm and chromatographed in 20mMtriethanolamine, pH8, containing I mM-EGTA. The radioactive profile of the column (Fig. I lb) was identical with unreacted cytosol and contained a high percentage of non-dissociated receptor aggregates. Fractions 45-53, belonging to the aggregate peak, were then oxidized with 20mM-H202 and analysed in sucrose gradients containing 10 % glycerol. A portion of fraction 48 was also analysed before oxidation. The results of the sucrose/glycerol gradients are shown in Fig. 12. Before the oxidation step, fraction 48 from the A-i Sm column (Fig. 1 b) contained predominantly 6S receptor complex when run in the absence of 0.3 M-KCI (Fig. 1 2a), consistent with the behaviour of unreacted cytosol chromatographed in this way. After the oxidation step, crosslinking had again occurred by this procedure since a 7S peak appeared in the gradient run in 0.3M-KCI 1979

209

PROGESTERONE-RECEPTOR CROSS-LINKING

between the liquid phase and the gel phase for a given protein) as unmodified cytosol receptors. This result, together with the sucrose-gradient data, suggest that the conformation ofcross-linked receptor dimer is very similar to that of the unreacted material.

300

c)

0.

c; 0.

0

x0

> 200

4._ c)

Cd 0

CU

x

100

0 0I 10

0

Top

20

0

Bottom Top

10

20

Bottom

Fraction number Fig. 12 Sucrose-density-gradient ultracentrifugation of fraction 48 from Fig. 11(b) (a) Sucrose-gradient pattern of fraction 48 run in the absence of 0.3 M-KCI. (b) Sucrose-gradient pattern in 0.3 M-KCI of fraction 48 after oxidation with 20mM-H202.

Fraction number Fig. 11. Isolation of receptor aggregates by gel filtration in low-ionic-strength buffer Three cytosol receptor samples were prepared, one (a) being untreated. The second sample (b) was amidinated, but not oxidized. The third sample (c) was amidinated and cross-linked by oxidation. Samples (3 ml) of each were chromatographed on an A-i.5m column (2.6cmx40cm) in 20mM-triethanolamine (pH 8.0)/1 mM-EGTA as described in the Materials and Methods section. Fractions (1.5ml) were analysed for 3H (e) and A280 determined to measure protein concentration (o). Arrows denote elution positions of void volume (V0) and total volume (V,) respectively. Fractions of each aggregate peak (horizontal bars) were pooled for further analysis.

(Fig. 12b). Thus partially purified receptor 6S dimers could also be cross-linked by methyl 4mercaptobutyrimidate. The reaction product had the same sedimentation coefficient as the nonoxidized material. As shown in Fig. 11(c), when the cross-linked receptors from whole cytosol after oxidation were chromatographed on the same A-15 column used to isolate the amidinated fraction, they eluted with the same Kav. (the partition coefficient Vol. 181

Analysis of partially purified receptor fractions by polyacrylamide-gel electrophoresis under denaturing conditions

The Agarose A-15m column purification described above provided fractions containing receptor dimer nearly devoid of ovalbumin. Thus we were able to analyse interactions between minor proteins by polyacrylamide-gel electrophoresis in the presence of sodium dodecyl sulphate. Peaks obtained from an Agarose A-15m column that contained receptor (Figs. 1 la and 1 ic) were analysed by this technique. Fig. 13 shows the distribution of protein bands in both samples by gel electrophoresis. Since the crosslinking is reversed by reduction, the usual gel electrophoresis protocol involving boiling of the proteins in lOOmM-,B-mercaptoethanol would separate the cross-linked proteins. Thus an attempt was made to run gels of both reduced and non-reduced samples in sodium dodecyl sulphate. Tracks (1) and (2) in Fig. 13 show the patterns obtained when the pooled peaks containing receptor from cross-linked (Fig. 1 ic) and unreacted (Fig. 1 la) Agarose A-iSm peaks were applied to the gel omitting the treatment with 8-mercaptoethanol. Large amounts of protein failed to enter the stacking and running gel in both cases, more in the case of the

210

M. E. BIRNBAUMER, W. T. SCHRADER AND B. W. O'MALLEY

cross-linked pool. Tracks (3) and (4) show the patterns obtained after reduction of both samples. Surprisingly, there were only slight differences in the distribution and quantity of protein bands present in both tracks, instead of a definite change in pattern in the cross-linked sample. This gel shows that the cross-linked peak was not significantly enriched in low-molecular-weight proteins; therefore, the crosslinking had not induced major aggregation of small peptides. Neither was the peak depleted of largemolecular-weight proteins, which could have been cross-linked out of the receptor pool. Thus the reaction is more specific than we had anticipated. In the absence of functional tests for a wide variety of cytosol proteins, we cannot rule out the probability that methyl 4-mercaptobutyrimidate cross-links other proteins, but the procedure seems to favour

1

(2) 13} (4)

r

__ 200

r

!1,$

Fot t F.N

h...X. .e 2

f..MS zw

CZ tl''.

-

i .*

....___

86H

b... 7:MU.

F

u

43K x

BPB L

_ L4.

-SH

-SH

Fig. 13. Polyacrylamide-gel electrophoresis in the presence of 0.1 % sodium dodecyl sulphate stained with Coomassie Brilliant Blue Fractions were pooled (as shown in Figs. 1 la and 11 c), dialysed against 5OmM-NH4HCO3, concentrated by freeze-drying and resuspended in buffer containing 63mM-Tris (pH6.8)/3% sodium dodecyl sulphate. Each sample was divided in two portions, and one half of it was boiled as such and the other one in the presence of 5 % f-mercaptoethanol. Tracks (1) and (2) contained non-reduced samples of cross-linked and untreated peaks (-SH), and tracks (3) and (4) contained the reduced samples (+SH). Abbreviation used: BPB, Bromophenol Blue.

intra- over inter-molecular cross-linking according to the present results. This suggests that very close proximity between peptide chains is required to join them with a disulphide bridge. Other cross-linking reagents Other types of bifunctional cross-linking reagents react with the receptor subunits and join adjacent amino groups in one step (see Table 2). Typical representatives of this group are dimethyl suberimidate (Davies & Stark, 1970) and dimethyl adipimate (Hartman & Wold, 1967). We explored the reactivity of the receptors with these reagents to gain further insight into the receptor conformation. All of these bifunctional reagents react very rapidly with neighbouring amino groups of the subunits, and the solutions are very unstable. A typical reaction involves first adjusting the buffer concentration to the desired value by adding 1 M-triethanolamine, pH 8.0, and then dissolving the appropriate amount of reagent in 100,ul of cytosol and rapidly mixing it with the rest of the cytosol. This method minimizes the time needed to add the reagent and exposes only a small amount of cytosol to a high concentration of the cross-linking agent. When the receptor complexes were exposed to these reagents at the buffer concentration at which methyl 4-mercaptobutyrimidate reacts, variable results were obtained. Dimethyl suberimidate yielded only material with a sedimentation coefficient of 8S or more, as analysed by sucrose-density-gradient ultracentrifugation. Dimethyl adipimate produced a 7-7.2S cross-linked complex. Both reagents introduced irreversible cross-links in the molecule and the composition of their products could not be determined. Their efficiency of reaction was higher with increasing buffer concentration. Two additional reversible reagents were tested: dimethyl dithiobispropionimidate and dithiobis(succinimidyl propionate). These differ from methyl 4-mercaptobutyrimidate both in the length of the expected cross-link and by the fact that reactive amino groups in both subunits are required. Thus, reversing the coupling with these two agents by using 2-mercaptoethanol reduction is expected to leave the imidoester fragments bound to the proteins. As shown in Table 2, both dithio compounds did cross-link, but their reaction yielded receptor complexes larger than those obtained with methyl 4-mercaptobutyrimidate. More study of the composition of these cross-linked receptor complexes is required. However, methyl 4-mercaptobutyrimidate has yielded the desired material with characteristics close to the original 6S dimer in a reaction that permits the identification of the components after reduction with thiol-specific reagents. 1979

211

PROGESTERONE-RECEPTOR CROSS-LINKING

4

000= o

WI_)OSm

ef)

4-R en

en

*_~~~~~~~~~~~~~~~~~~~~1

00

En

0

*

IN

CU

G

0 0

e

0~~~~~~~~~~~~~~~~~ - -

A8 G°.

£.-

-

oE

co

co

X~~~~~~~~~~~~~~~~~~~~~~t 0 0

°4

CbnC-

tn

-

0,,

0o

0

'O

o

sO e o0O

00

00

0 CU

-

E

4)~~~~~~~~~~ s

00 oto

o

4-

CU

w

0

E F Qi 0~~

1~~~~~~-

00

00'0

.4

U)~~~~~~~~~~~~~~~~~~~~~~ 1~~

'4)

~0

4~)

4)

CUO

= ;, - I) OIC j Q .4)~~~~~~~~~~ E 0~~~~~~~~~~~~~~~~~~~0c w = s Z 0 0

CU~~~~~~. 0 0

.0 '~ ~ ~ ~ ~ ~ Z~ ~ ~ ~ ~ 0W

CUU~~~

0

0

W

~

E Q 4)

'm~~~~~~~0CU

U~~

U 4

o

o~~~C

x

a 0-ua

4

EUUE~~~~~~~~~.U~U en

CUa

N

Un

.+4

a!~~~~~+.& E ==6 .2

0 0=

0 -

C)~~~~~4

C)

4)

.

Z=--u

Z=U

Vol. 181

4

4)

4)

212

M. E. BIRNBAUMER, W. T. SCHRADER AND B. W. O'MALLEY

Discussion To elucidate the molecular mechanism by which steroid hormones modify gene transcription, it is necessary to know the structure of the receptors in the target cell. The progesterone receptor of the chick oviduct consists of two dissimilar proteins, monomers A and B, able to bind the steroid with the same affinity and with identical rate constants. These proteins differ, however, in their binding to chromatin and DNA. We have postulated that the progesterone receptor in the chick oviduct exists as a 6S dimer of the A and B proteins. The recombination of the isolated subunits to generate the 6S dimer would be the definitive proof for this hypothesis, but our efforts to achieve this have been unsuccessful. An alternative way to explore receptor composition is via crosslinking experiments such as those presented above. Thus, with low concentrations of methyl 4mercaptobutyrimidate a 7S material was obtained. With higher concentrations of reagent the 7S material persisted, but products of larger S values were also produced. We therefore studied the characteristics of this 7S material, which can be isolated as a discrete entity. The 7S material failed to bind to phosphocellulose columns and bound to DEAE-cellulose columns, which is characteristic of the 6S dimer. When a KCI gradient was applied to the DEAEcellulose, the cross-linked material eluted at 0.2MKCl as does the 6S material (Schrader et al., 1975). When the 7S cross-linked material was reduced with 1 % i-mercaptoethanol and analysed by ion-exchange and gel-filtration chromatography, it was shown to contain both A and B proteins in approximately equivalent amounts. The estimated molecular weight was reasonably consistent with our estimate of the molecular weight of the 6S dimer, in that small errors in estimation for both could account for the discrepancy. The different sedimentation constants of the reacted complex could be due to changes in molecular shape that might result from cross-linking. The calculated mol.wt. of 228000 is larger than the value of 194000 predicted from the sum of the molecular weights for the purified B and A subunits (Schrader et al., 1976; Coty et al., 1979). These results support our assumption that the 7S material is the product of cross-linking the 6S receptor. The question now arises as to whether the 6S form is indeed an A-B dimer. Our results cannot be explained merely by the initial presence of A and B subunits in equal amounts, together with their artifactual cross-linking to non-receptor proteins in vitro. The reagent failed to cross-link dissociated A and B proteins to larger forms in the presence of cytosol proteins after warming or salt treatment. Furthermore, the reagent does not generally crosslink all proteins, as the gel electrophoresis experiments show.

Another possibility is that the 6S form is an equimolar mixture of A-A and B-B dimers. The existence of an A-A dimer is inconsistent with our results on cross-linking, since the cross-linked 7S form exhibits a Stokes radius of 7.3 nm, significantly larger than the value expected from an A-A dimer of mol.wt. 160000. An A-A dimer, given a sedimentation coefficient of 7S, should have had a Stokes radius of only 5.3 nm, or alternatively, given a Stokes radius of 7.3 nm, should have had a sedimentation coefficient of only 5.1S. This argument does not rule out the existence of B-B dimers. The expected molecular weight of such a B-B dimer would be 230000, the same as our estimate of molecular weight for the 7S complex. If the cross-linked complex were indeed a B-B dimer, to explain our results we need to postulate that half of the B monomer produced on dissociation is transformed into A monomers. Cytosol preparations from oviducts have been prepared in our laboratory by using a wide range of conditions: different buffers, different salt concentrations, the addition of proteinase inhibitors and a variety of chemicals, and all the tissue extracts have yielded both monomers in the same proportions. Also repeated attempts of various kinds to convert B monomers into A monomers have always failed to show such conversion. Thus it seems unlikely that such a complex exists. Finally, it is possible to envisage a receptor structure containing another protein subunit that does not bind hormone and that consequently could only be detected in purified cross-linked receptor complexes. A mixture of A and B proteins each complexed separately with a specific subunit that does not bind progesterone could also yield these results. This latter possibility has been suggested by Yamamoto & Alberts (1972) and Notides & Nielsen (1974) to explain the increase in molecular weight due to transformation in the oestrogen receptor of the rat uterus. Such alternative explanations would require that all complexes involved exhibit: (i) identical susceptibility to disruption by salt and heat; (ii) identical size and sedimentation coefficients; (iii) identical surface-charge distribution to account for our results on ion-exchange chromatography of the 6S form of the receptor; and (iv) identical susceptibility to chemical reaction with methyl 4-mercaptobutyrimidate. It seems unlikely that A and B proteins complexed with other subunits would both have all these characteristics in common. The ratio of A and B proteins in receptor aggregates after various treatments is always close to unity. However, a rigorous test of this possibility can only be accomplished by purifying the cross-linked complexes to homogeneity and identifying the components directly by electrophoresis. Thus we consider our data reasonable, but 1979

PROGESTERONE-RECEPTOR CROSS-LINKING not yet conclusive, evidence that the cross-linked material is a dimer of A and B proteins. The availability of cross-linked receptor complexes will be very important to our future studies of receptor-subunit interactions in cytosol preparations from target cells. In addition, the cross-linked receptor should be more amenable to extreme conditions during receptor purification that tend to disrupt the 6S dimers. These purified complexes can be utilized intact or after reduction to examine the biochemical reactions involved in the hormonereceptor-induced alterations of eukaryotic gene expression. Purification of the cross-linked material is required toward these ends. We thank Ms. Marinan Freeman for her expert technical assistance. This work was supported by U.S.P.H.S. grants HD-07857 and HD-07495 and by a Lalor Foundation Fellowship Award to M. E. B. This is the sixteenth in a series of articles from our laboratory dealing with progesterone-binding proteins. Coty et al. (1979) is the preceding paper in the series.

References Alberts, B. M. & Herrick, G. (1971) Methods Enzymol. 21, 198-217 Buller, R. E. & O'Malley, B. W. (1976) Biochem. Pharmacol. 25, 1-12 Buller, E. R., Schwartz, R. J., Schrader, W. T. & O'Malley, B. W. (1976) J. Biol. Chem. 251, 5178-5186 Castellino, F. T. & Barker, R. (1968) Biochemnistry 7,22072217 Coty, W. A., Schrader, W. T. & O'Malley, B. W. (1979) J. Steroid Biochem. 10, 1-12 Davies, G. E. & Stark, G. R. (1970) Proc. Natl. Acad. Sci. U.S.A. 66, 651-656 Edelhoch, H. (1960) J. Biol. Chem. 235, 1326-1334 Fuller, R. A. & Briggs, D. R. (1956) J. Am. Chem. Soc. 78, 5253-5262 Garen, A. & Levinthal, C. (1960) Biochim. Biophys. Acta 38, 470-483 Gazith, J., Himmelfarb, S. & Harrington, W. F. (1970) J. Biol. Chem. 245, 15-22 Hartman, F. C. & Wold, F. (1967) Biochemistry 6, 24392448 Jensen, E. V., Mohla, S., Gorell, T. A. & De Sombre, E. R. (1974) Vitam. Horm. (N. Y.) 32, 89-127 Kegeles, G. & Gutter, F. J. (1951) J. Am. Chem. Soc. 73, 3770-3777

Vol. 181

213 Kuhn, R. W., Schrader, W. T., Coty, W. A., Conn, P. M. & O'Malley, B. W. (1977) J. Blol. Chein. 252, 308-317 Laemmli, U. K. (1970) Nature (London) 227, 680-685 Notides, A. C. & Nielsen, S. (1974) J. Biol. Chem. 249, 1866-1873 O'Malley, B. W. & Means, A. R. (1974) Science 183, 610620 Oncley, J. R., Scatchard, G. & Brown, J. (1947) J. Phys. Colloid Chem. 51, 184-185 Peters, K. & Richards, F. M. (1977) Annu. Rev. Biochem. 46, 523-551 Schrader, W. T. (1975) Methods Enzymol. 36, 187-210 Schrader, W. T. & O'Malley, B. W. (1972) J. Biol. Chem. 247, 51-59 Schrader, W. T. & O'Malley, B. W. (1977) in Receptors and Hormone Action (O'Malley, B. W. & Birnbaumer, L. eds.), pp. 189-224, Academic Press, New York and London Schrader, W. T., Toft, D. 0. & O'Malley, B. W. (1972) J. Biol. Chem. 247, 2401-2407 Schrader, W. T., Heuer, S. S. & O'Malley, B. W. (1975) Biol. Reprod. 12, 134-142 Schrader, W. T., Kuhn, R. W. & O'Malley, B. W. (1976) J. Biol. Chem. 252, 299-307 Schrader, W. T., Coty, W. T., Smith, R. G. & O'Malley, B. W. (1977) Ann. N. Y. Acad. Sci. 286, 64-80 Schwartz, R. J., Kuhn, R. W., Buller, R. E., Schrader, W. T. & O'Malley, B. W. (1976) J. Biol. Chem. 251, 5166-5177 Schwartz, R. J., Chang, C., Schrader, W. T. & O'Malley, B. W. (1977) Ann. N. Y. Acad. Sci. 286, 147-160 Sherman, M. R., Corvol, P. L. & O'Malley, B. W. (1970) J. Biol. Chem. 245, 6085-6096 Sherman, M. R., Tuazon, F. B., Diaz, S. C. & Miller, L. K. (1976) Biochemistry 15, 980-988 Siegel, L. M. & Monty, K. J. (1966) Biochim. Biophys. Acta 112, 346-362 Smith, H. E., Smith, R. G., Toft, D. O., Neergaard, J. R., Burrows, E. P. & O'Malley, B. W. (1974) J. Biol. Chem. 249, 5924-5932 Spelsberg, T. C. & Cox, R. F. (1976) Biochim. Biophys. Acta 435, 376-390 Stancel, G. M. & Gorski, J. (1975) Methods Enzymol. 36, 166-176 Toft, D., Moudgil, V., Lohmar, P. & Miller, J. (1977) Ann. N. Y. Acad. Sci. 286, Z9-41 Traut, R. R., Bollen, A., Sun, T. T., Hershey, J. W. B., Sundberg, J. & Pierce, L. R. (1973) Biochemistry 12, 3266-3273 Wold, F. (1967) Methods Enzymol. 11, 617-645 Yamamoto, K. R. & Alberts, B. M. (1972) Proc. Natl. Acad. Sci. US.A. 69, 2105-2109