The Hepatic Glucagon Receptor SOLUBILIZATION, CHARACTERIZATION, AND DEVELOPMENT OF AN AFFINITY ADSORPTION ASSAY FOR THE SOLUBLE RECEPTOR*

The hepatic glucagon receptor was covalently labeled with [1261-Tyr'o]monoiodoglucagon (['261]MIG) by use of the heterobifunctional cross-linker hydroxysuccinimidyl p-azidobenzoate. Labeling of the M, = 63,000 peptide was sensitive to glucagon and GTP at concentrations at which they affect ['261]MIG binding to the receptor. The labeled receptor was solubilized with Lubrol-PX, and the hydrodynamic characteristics of the receptor were determined. The molecular parameters of the solubilized receptor are: s~,+, = 4.3 2 0.1, Stokes radius = 6.3 0.1 nm, frictional coefficient f/fo = 1.8, and a calculated M. = 119,000. Incubation of liver membranes at 32 "C for 15 min prior to the addition of [12SI]MIG permitted us to identify the high molecular weight form (M, = -1 13,000) of the receptor by direct sodium dodecyl sulfate-gel electrophoretic analysis. The M, = 63,000 peptide can be adsorbed to wheat germ lectin-Sepharose. The glycoprotein nature of the receptor has been utilized to develop an assay for the detergent-solubilized receptor that uses wheat germ lectin-Sepharose as a solid matrix to adsorb the [12'1] MIG-receptor complex.

3 Established Investigator, American Heart Association.
Glucagon is the major regulator of carbohydrate metabolism in the liver. Glucagon initiates its action by binding to receptors on the plasma membranes. The glucagon-bound receptor activates the stimulatory regulator of adenylyl cyclase (1) which, in turn, associates with and activates the catalyst, resulting in enhanced CAMP synthesis rates (2). A detailed understanding of the primary events in glucagon action requires the isolation, characterization, and reconstitution of the purified components of the system. While the stimulatory regulator has been purified and extensively characterized (3)(4)(5) , relatively little has been known about the glucagon receptor.
Recently, we have used the procedures of Johnson et al. (6) to covalently attach ['251-Tyr'o]monoiodoglucagon to its receptor using the heterobifunctional cross-linker HSAB' (7). This resulted in identification of a M, = 63,000 peptide as the hormone-binding subunit of the glucagon receptor. In the previous studies, we had shown that only a portion of the M, = 63,000 protein was required for hormone binding and interaction with the stimulatory regulator; however, since all analysis of the receptor was carried out by SDS-gel electrophoresis, it was not possible to study the subunit structure of the receptor. In this article, we have analyzed the glucagon receptor in nondenaturing detergent solutions and found the native receptor in the membrane and in detergent solution to be a dimer of the M, = 63,000 peptide. We have found that the M , = 63,000 peptide specifically adsorbs to WGL-Sepharose. We have utilized this glycoprotein nature of the receptor to develop a new assay for the solubilized receptor. We also describe the binding properties of the solubilized receptor. sized and purified by reverse chromatography over a C18 pBondapak column as described in detail elsewhere (11).
WGL-Sepharose was synthesized according to the procedure of Adair and Kornfield (12) with some minor modifications. One hundred mg of WGL were coupled to 40 ml of packed CNBr-activated Sepharose GB-CL in the presence of 10 mM NaHC03, 100 mM NaC1, and 200 mM GlcNAc overnight at 4 "C. After the coupling, the gel was washed with 2 liters of ice-cold 10 mM NaHC03, 100 mM NaCl, and then reacted with 1.5 g of glycine in 10 mM NaHC03, 100 mM NaCl overnight. After the reaction, the gel was washed with 2 liters of 10 mM NaHC03 and 100 mM NaCl and stored at 4 "C as a 1:2 suspension in the same solution. Sepharose was activated with 100 mg of CNBr per ml of packed gel in the presence of 2 M Na2C03 according to the procedure of Parikh et al. (13). The WGL-Sepharose retained its binding characteristics for 4-6 weeks and showed no significant leaching of the WGL.

Preparation of Membranes
Rat liver plasma membranes were prepared according to the procedure of Neville (14) as described by Pohl et al. (15). Prior to use, the membranes were suspended in 20 mM phosphate buffer, pH 7.5 (10 mg of protein/ml) and mixed with 100 mM diisopropyl fluorophosphate dissolved in propylene glycol, such that the final concentration of diisopropyl fluorophosphate was 10 mM. The mixture was held on ice for 10 min. Subsequently, the membranes were washed twice with 10 volumes of 20 mM phosphate buffer, pH 7.5, and diluted to the appropriate concentration for use in experiments. Handling of and treatment with diisopropyl fluorophosphate was carried out in a vented fume hood. Cyc-S49 cell membranes were prepared according to the procedure of Ross et al. (16) with minor modifications described elsewhere (17).

fZ5I]MIG Binding to Liver Membranes
Liver membranes (0.2-1.0 mg of protein/ml) were incubated in a mixture containing 20 mM phosphate buffer, pH 7.5, 1 mM EDTA, 1% (w/v) BSA, 1 nM ['261]MIG, and appropriate additives. When used, the final concentration of glucagon was 1 p~ and that of GTP 0.1 mM, or as specified. Incubation was carried out at 32 "C for 15 min. Incubation volumes varied from 0.1 ml for routine filtration assays to 10 ml for cross-linking and further analysis of the receptor.
When ["51]MIG binding to membranes was analyzed by filtration, the 100-p1 samples were diluted to 1 ml with ice-cold 20 mM phosphate, 1% (w/v) BSA, pH 7.5. The diluted sample was filtered through 0.45-p cellulose acetate filters (Oxoid) which had been previously soaked in 10% BSA. The tube containing the sample was washed once with 1 ml of 1% BSA. The filters were washed with 10 ml of ice-cold 20 mM phosphate buffer, 0.1% BSA (wash buffer), and the radioactivity bound to the filters was measured with a y counter.
Large volume (2-10 ml) binding assays were carried out in 30-ml ultracentrifuge tubes. After the incubation, the samples (1-ml aliquots) were diluted to 25 ml with ice-cold 20 mM wash buffer. All further handling of the membranes was at 0-4 "C unless specified otherwise. The membranes were sedimented by centrifugation at 100,000 X g for 20 min. The supernatant was removed by aspiration, and the pellet was resuspended in 1 ml of 20 mM phosphate buffer, pH 7.5, diluted to 25 ml with phosphate buffer, and sedimented by centrifugation. The final pellet was resuspended in 20 mM phosphate buffer, pH 7.5, at a concentration of 5-10 mg of protein/ml. This two-step washing procedure resulted in removal of all unbound [lZ5I] MIG, such that the picomoles of ['251]MIG bound per mg of membrane protein were essentially identical to those seen by filtration (see Table  1).
Cross-linking of Bound rZ5I]MIG to Liver Membranes HSAB was weighed in the dark and dissolved in dimethyl sulfoxide in a brown bottle to give a concentration of 20 mM. The membranes in 13-ml ultracentrifuge tubes were mixed with the HSAB solution such that the final concentration was 200 p~. The samples were kept in the dark for 3 min. Subsequently, they were irradiated in a final volume of 1.0 ml for 9 min under a mercury arc lamp (Gates Inc., Arthur H. Thomas 6281-H10) at a distance of 16 cm. The samples were held on ice during the UV irradiation. After the irradiation, the samples were diluted to 10 ml with 25 mM Tris-HC1, pH 7.5, 100 p M GTP, and 1 p~ glucagon. This mixture was incubated for 15 min at 32 "C. The membranes were then sedimented by centrifugation at 100,000 X g for 20 min. This procedure routinely resulted in 1-3% of the bound hormone being covalently attached to the M. = 63,000 peptide.

Detergent Solubilization
Solubilization of Membranes Containing Covahntly Labeled Receptor-Liver membranes (5-10 mg of protein/ml) that had been crosslinked to ['251]MIG were solubilized in the presence of 25 mM Na-Hepes, pH 8.0, 1% Lubrol-PX. The samples were held on ice for 30 min with periodic agitation. Subsequently, the mixture was centrifuged at 100,000 X g for 60 min. This procedure resulted in total solubilization of the M, = 63,000 peptide, since no labeled peptide was observable in the 100,000 X g pellet. The supernatant was removed and used as the source of solubilized glucagon receptor.
in 20 mM CHAPS, 0.5 M NaCl, 0.5 mM EDTA, 1.0 mM MgCL, and Preparation of CHAPS Extract-Liver membranes were suspended 25 mM NaHepes, pH 8.0, at a final concentration of 6-8 mg of protein per ml and stirred for 30 min at 0-4 "C. The detergent suspension was centrifuged at 100,000 X g for 60 min. The clear supernatant was removed carefully and made 16% (w/v) with respect to sucrose. Aliquots were quick frozen in a dry ice-acetone bath and stored at -70 "C until use. Typically, under these conditions, about 50% of the membrane protein is solubilized and 25-30% of specific glucagonbinding sites are recovered in the supernatant.
Binding Assay for the Solubilized Receptor-CHAPS extract was incubated in a final volume of 300 p1 in the presence of 1  (w/v) sucrose, 0.5% Lubrol-PX, 50 mM KCl, 1 mM EDTA, and 2 mM MgClZ (wash buffer) were added to each tube. Prior to use, the WGL-Sepharose was washed with 10 volumes of the wash buffer. The samples were then shaken for 2 h in the cold at a moderate speed so as to impede the gel from settling. At the end of the 2-h period, the samples were diluted to 5 ml with ice-cold wash buffer and centrifuged at 1500 X g for 15 min in a refrigerated IEC centrifuge. The supernatant was aspirated, and the gel was resuspended again in 5 ml of wash buffer. After a second centrifugation, the supernatant was removed and the gel was counted in a Tracor y counter. When [lZ5I] MIG was used in the absence of CHAPS extract protein, 0.08 to 0.3% (600-2500 cpm out of 720,000 cpm) of the total counts adsorbed to the gel in a manner unaffected by the absence or presence of 0.1 M GlcNAc.
Sucrose Density Gradient Ultracentrifugation-The solubilized receptor preparation (150 11) was mixed with 10 p1 of malate dehydrogenase (10 mg/ml; 443 units/mg), 5 pl of catalase (10 mg/ml, 11,000 units/mg), 10 p1 of fumarase (12 mg/ml; 380 units/mg), and 50 p1 of cytochrome c (10 mg/ml). This mixture was loaded on a 5-15% sucrose gradient in either Hz0 or D,O. The gradients also contained 25 mM NaHepes, 1 mM EDTA, and 0.5% Lubrol-PX. When DzO was used as solvent, 9 volumes of 99% DzO were mixed with 1 volume of 250 mM NaHepes, 10 mM EDTA, and 5% Lubrol-PX, pH 8.0. This mixture was then used to dissolve the sucrose. Samples were centrifuged for 16 h at 33,000 rpm in a SW 50.1 rotor at 4 "C. After centrifugation, the samples were fractionatedusing a Hoeffer gradient fractionator. Approximately 30 fractions of 160 pl each were collected. The fractions were then analyzed for the location of the marker enzymes and labeled receptor as described below.
Gel Filtration-400 pl of solubilized receptor were mixed with 10 pl of fumarase, 10 pl of catalase, 10 pl of malate dehydrogenase, 2.8 mg of cytochrome c, and 10 pl (10 mg/ml, 740 units/mg) of Pgalactosidase. The sample was made 30% (v/v) with respect to etbylene glycol. The sample was then introduced to a AcA-34 Ultrogel column (48 x 1 cm) that had been previously equilibrated with 10 mM NaHepes, pH 8.0, 1 mM EDTA, 50 mM NaC1,0.5% Lubrol, and 30% (v/v) ethylene glycol. Approximately 100 fractions of 450 each were collected over a 16 to 20-h period. The void volume of the column was determined by the elution of Dextran Blue in separate runs. Fractions were subsequently analyzed for location of marker enzymes and labeled receptor.

Assay for Marker Enzymes
Malate Dehydrogenase-5-lO pl of sample were added to 1 ml of 50 mM Tris-HC1, pH 7.5, 1.5 mM oxaloacetic acid, and 43 p M NADH. Assay tubes were held on ice for 10 min. Subsequently a decrease in absorbance at 340 nm was recorded after 1 min of incubation in the cuvette a t room temperature.
Catalase-5-10 pI of sample were added to 1 ml of 50 mM Tris-HCI, pH 7.5, and 1.2-7 mM H202. After a 5-min incubation a t room temperature (22-24 "C), the decrease in absorbance a t 240 nm was recorded.
Fumarase-5-10 p1 of sample were added to 1 ml of 50 mM Tris-HCI, pH 7.5, and 40 mM Tris malate. After incubation for 3.5-5 min a t room temperature, the reaction was stopped by the addition of 50 p1 of 1 N HC1. The increase in absorbance a t 250 nm was measured.
e-Galactosida~e-20-50 p l of sample were added to 1 ml of 50 mM Tris-HCI, pH 7.5, and 2 mM o-nitrophenyl-b-D-galactopyranoside. The samples were incubated a t 32 "C until a clearly visible color change occurred (5-10 min). The reaction was stopped by the addition of 0.5 ml of 1 M NazCO3. Increase in absorbance was measured at 500 nm. Details of the marker enzyme assays were kindly provided by Dr. Eva Neer (Department of Cardiology, Harvard Medical School, Boston, MA 02115).

SDS-Polyacrylnmide Gel Electrophoresis
SDS-PAGE was run according to the procedure of Laemmli (18). Routinely, samples were analyzed on a 10% gel with a 3% stacking gel. When membranes were loaded onto the gel, 25-50 pg of protein per sample in a volume of 50 pl were used. When sucrose density gradient fractions were used, 35 p1 of sample were mixed with 35 pl of buffer containing 2% SDS and 5% fi-mercaptoethanol. 200-pl aliquots of the gel filtration column fractions were mixed with 50 pl of 1 mg/ml of BSA and 2 ml of cold (-10 "C) acetone. The precipitated proteins were collected by centrifugation a t 1500 X g for 10 min at -10 "C. The precipitates were dissolved in a buffer containing 1% SDS and 2.5% 0-mercaptoethanol. Prior to application onto the gel, all samples were incubated for 1 h a t 32 "C in the presence of 1% SDS and 2.5% fi-mercaptoethanol. Heating a t higher temperatures was avoided since this led to aggregation and subsequent retention of the samples in the stacking gel during electrophoresis.
After electrophoresis, the gels were stained with Coomassie Blue, destained, and dried. The gels were then exposed to Kodak XAR-5 film, in cassettes equipped with intensifying screens (Cronex Xtra-Lite). Generally, exposure was for 48 h a t -70 "C. Subsequently, the autoradiograms were developed and scanned with a Kontes fiberoptics scanner equipped with a Hewlett-Packard integrator. In general, 48h exposure allowed for scanning within the linear ranges of the instrument.

Assay for the Stimulatory Regulator of Adenylyl Cyclase
The regulatory component was assayed by its capacity to reconstitute the cyc-S49 cell membrane adenylyl cyclase in the presence of 10 pM GTPyS and 20 mM MgC12 as described in detail elsewhere ( 2 ) .
['"PICAMP formed was quantified according to the method of Salomon et al. (19) as modified by Bockaert et al. (20).

Protein Measurement
Proteins were measured by the procedure of Lowry et al. (21). When CHAPS was present during estimation, its concentration was 0.2 mM. This concentration did not interfere with the protein estimation.

Replication of Results
All experiments were performed at least thrice. Experiments with the CHAPS extract were performed 4 to 6 times using three different batches of CHAPS extract protein from different membrane preparations and using two different batches of WGL-Sepharose. While there was variability in the extent of binding obtained as well as the level of nonspecific binding (20-50% of total binding), all the basic characteristics of the binding activity were observed throughout. Representative experiments are shown.

Covalent Labeling of the Glucagon Receptor
When liver membranes were exposed to ['251]MIG and binding was measured by removal of unbound label by filtration or centrifugation, similar levels of specific receptor sites per mg of protein were observed irrespective of the method used to separate bound label from free label (Table I). Since it is relatively easy to recover the membranes after centrif-ugation, we routinely separated the membrane-bound ["'I] MIG from free ["'II]MIG by centrifugation. When the membranes that had [I2'II]MIG specifically bound were treated with HSAB and then analyzed by SDS-PAGE on 10% gels, the label was associated with one macromolecular species which migrated with an apparent M, = 63,000. When excess unlabeled glucagon was included during the binding reaction, the labeling of the M, = 63,000 peptide was abolished. However, insulin, ACTH, and vasopressin had no effect on the labeling of the M, = 63,000 peptide ( Fig. 1). This indicated that the M, = 63,000 peptide specifically recognized glucagon or its radiolabeled analog. The inclusion of unlabeled glucagon during the binding reaction not only abolishes the labeling of the M, = 63,000 peptide but also significantly decreases the

[ '~I I M I G bound (frnollrng protein)
None The assay was carried out in a final volume of 1 ml a t 0.6 mg of 100-pl aliquots were withdrawn at the end of the incubation and 2 0 0 4 sample was diluted to 2.0 ml and centrifuged as described protein/ml.  [8-arginine]vasopressin (8- and glucagon ( 3 p M ) . After the incubation, the membranes were washed free of unbound label and incubated with HSAB in the dark and under UV illumination. The membranes were then washed, and protein and bound counts in the final pellets were estimated. It was found that incubation in the absence of any unlabeled hormone or in the presence of insulin, ACTH, or [8-arginine] vasopressin resulted in 29,000-32,000 cpm bound per 40 pg of membrane protein. When glucagon was present, 1200 cpm were bound to 40 pg of membrane protein. Forty pg of protein for each sample were loaded on the gel. The samples were then subjected to electrophoresis followed by autoradiography. An autoradiogram (48-h exposure) is shown. amount of label that travels with the dye front. This indicates that the label associated with the dye front probably represents the fraction of the receptor-bound [12sI]MIG that is not cross-linked by HSAB, rather than nonspecifically adsorbed [lZsII]MIG.
We then studied the effect of inclusion of various concentrations of unlabeled glucagon during the binding reaction on the extent of the labeling of the M , = 63,000 peptide by 1 nM ["'I]MIG. We found that the concentration range in which glucagon affected labeling of the M , = 63,000 peptide ( Fig. 2) agreed with the range in which it interacts with the glucagon receptor and stimulates adenylyl cyclase (22).
T o establish that the M , = 63,000 peptide labeled is a part of the guanine nucleotide-sensitive glucagon receptor, we compared the effect of various concentrations of GTP on the amount of [lZsII]MIG bound to receptors in the membrane and on the amount of labeling obtained after cross-linking. One such experiment is shown in Fig. 3  Liver membranes (0.25 ml/mg) were incubated with 1 nM ["'IIJMIG, 1 mM ATP, a nucleoside triphosphate-regenerating system consisting of 20 mM creatine phosphate, 0.2 mg/ml of creatine phosphokinase, and 0.02 mg/ml of myokinase and indicated concentrations of G T P in a final volume of 3 ml. After 15 min at 32 "C three 100-pI aliquots were withdrawn and filtered to estimate hormone binding to membranes (bottom). The remaining membranes were washed free of unbound [1251]MIG and incubated with HSAB in the dark and under UV illumination. The membranes were then washed to remove the cross-linker, and proteins were estimated in the final pellet. Thirty-one pg of protein for each sample were applied to the gel. The gel was dried and then exposed to x-ray film. The autoradiogram (48-h exposure) was scanned using a Kontes fiberoptics scanner equipped with a Hewlett-Packard integrator. brol-PX. This detergent was chosen because the partial specific volume of Lubrol-PX is known (0.956 ml/g), and it has been reported that the glucagon receptor maintains guanine nucleotide sensitivity after solubilization with Lubrol-PX (23).

Hydrodynamic Measurements
Gel Filtration ouer AcA 34 Ultrogel-The [1251]MIG-glucagon receptor complex which was cross-linked with HSAB was solubilized with Lubrol-PX and chromatographed over the Calculation of Molecular Parameters-The sedimentation coefficient of the receptor in D20 was not significantly different from that in H20. This observation indicates that the M, = 63,000 peptide in Lubrol-PX solution has a partial specific volume similar to that of standards used (25, 26). Hence, the average partial specific volume of the calibrating proteins (0.738) was used for calculating the molecular weight of the receptor (25). The ~2 0 ,~ value used for these calculations was the experimentally determined value using H20 as solvent (Fig. 5). From the Stokes radius and sedimentation coefficient, an M, = 119,000 was calculated (Table 11

Subunits of the Glucagon Receptor
In order to test if the subunits of the receptor were linked via disulfide bridges, we analyzed the receptor by SDS-PAGE before and after treatment with varying concentrations of pmercaptoethanol (Fig. 6). It was found that addition of varying concentrations of P-mercaptoethanol during solubilization of the sample did not result in any decrease in the size of the band in the M, = 63,000 region. In fact, with increasing concentrations of P-mercaptoethanol, there was a small increase in the apparent molecular weight (-3000). The calculated M, of the labeled peptide in the absence of 8-mercaptoethanol is 60,000, while at 10,25, and 50 mM, B-mercaptoethanol is 63,000. Such increases in apparent molecular weights are attributed to the breakage of intrachain disulfide bridges (27). We then searched for the M, = 120,000 form in the membranes to determine if the high molecular weight form was an artifact of the solubilization procedure. For this pur- Stokes radius was obtained from a standard curve of a (K,, = distribution coefficient) uersu. 9 Stokes radius of the marker protein.
Since there was no significant differences in the experimental S values in H20 and D20, the HZ0 value was used as ~20.~.
Calculated from the following equation, where D (partial specific volume (ml/g)) was assumed to be 0.738, the average for all the marker proteins. Calculated according to the equation, where 6, the solvation factor, was assumed to be 0.2 g of solventlg of protein. protein also recognized glucagon specifically in a GTP-dependent manner as is observed with the M, = 63,000 peptide ( Fig. 7).

94K-
Since many hormone receptors are glycoproteins, we tested if the glucagon receptor was also a glycoprotein. We extracted membranes that contained receptors that had ["'II]MIG covalently attached with 1% Lubrol-PX. The Lubrol-PX extracts were incubated with WGL-Sepharose in the absence and presence of various sugars. Equivalent amounts of the gel supernatants were analyzed by SDS-gel electrophoresis for the presence of the M , = 63,000 peptide. It was found that the M , = 63,000 receptor was not present in solution after exposure to WGL-Sepharose. However, addition of GlcNAc but not glucose or galactose allowed for the hormone-binding subunit to remain in solution (Fig. 8). This experiment indicates that the receptor is a glycoprotein that can specifically interact with WGL. The specific sugar sensitivity of the q-113 .5

FIG. 7. Identification of a high molecular weight form of the glucagon receptor by SDS-gel electrophoresis.
Liver membranes (0.6 mg/ml) were incubated a t 32 "C for 20 min in 20 mM phosphate buffer, pH 7.5, 1 mM EDTA, and 0.1% BSA. After this incubation, the membranes were then exposed to 1 nM ['251]MIG under standard binding conditions without any other additives (-) or in the presence of 100 p M GTP, 3 p M glucagon, or 1.8 pM insulin. The membranes were then washed free of unbound ['9]MIG and then treated with HSAB in the dark and under UV illumination. After cross-linking, the membranes were washed and subjected to SDS-gel electrophoresis on 5% gels. The gels were stained, destained, dried, and then subjected to autoradiography.
A 72-h exposure is shown. The 100, OOO X g supernatant was then exposed to Sepharose or WGLsepharose. 100 pl of the Lubrol extract were exposed to 70 pl of packed Sepharose (Con, control) or WGL-Sepharose in a final volume of 600 pl. Exposure to WGL-Sepharose was carried out without any further additions (-) or in the presence of 0.1 M GlcNAc, 0.1 M glucose (Glc) or 0.1 M galactose (Gal). After overnight incubation a t 0 "C, the supernatants were removed and 70-pl aliquots were electrophoresed on SDS-polyacrylamide gels. The gel was then dried and subjected to autoradiography. An autoradiogram (48-h exposure) of M , = 63,000 region is shown. binding reaction is in agreement with the known specificity of WGL (12) and indicates that the glucagon receptor contains GlcNAc and/or sialic acid residues.

Assay for the Solubilized Receptor
Since the receptor is a glycoprotein, we reasoned that if we could solubilize the receptor in a state that still retained hormone-binding capability, we should be able to separate the receptor-bound ['251]MIG from free ['251]MIG by adsorption onto WGL-Sepharose. This method for the separation of bound label from free label would then be based on an affinity chromatography of the receptor and would lend specificity to the separation procedure. Such specificity could be particularly valuable since standard methods based on size such as gel filtration yield unacceptably high backgrounds with labeled glucagon due to aggregation of glucagon. Precipitation with polyethylene glycol also yields high backgrounds. Consequently, no specific binding of the labeled glucagon to solubilized receptor can be detected in a routine manner using these assays.
We made extracts of liver membranes using a variety of detergents (Fig. 9). The extracts were incubated with 0.2 nM ['251]MIG in the absence and presence of 3 p~ unlabeled glucagon for 3 h on ice. After the binding reaction, the mixture was exposed to WGL-Sepharose for 2 h to adsorb the hormone-receptor complex. The gel was then washed and counted. Using this procedure, we found significant differences between the ['251]MIG bound to the gel in the absence versus that bound in the presence of excess unlabeled glucagon. The difference varied with the detergent extract used and was the greatest when CHAPS extracts were used. Even though the difference in ['251]MIG bound to the gel was due to the addition of unlabeled glucagon, this observation alone is not sufficient to conclude that the difference in ['251]MIG bound in the absence and presence of unlabeled glucagon is due to the receptor. In order to establish that the binding observed in the CHAPS extract does indeed represent [lZ5I] MIG interaction with the solubilized glucagon receptor, we tested 1) if the binding could be inhibited by excess unlabeled glucagon but not by other peptide hormones when added during the hormone-binding reaction and 2) if the binding could be abolished by the inclusion of 0.1 M GlcNAc along with WGL-Sepharose. In the experiment shown in Fig. 10, we exposed the CHAPS extract to ['251]MIG without any other hormones and in the presence of glucagon, insulin, ACTH, and vasopressin. After incubation for hormone binding, the reaction mixtures were exposed to WGL-Sepharose in the presence and absence of 0.    IV cyc-S49 cell membrane reconstituting activity by CHAPS extract protein CHAPS extract was prepared as described under "Experimental Procedures" and heated at room temperature (22-24 "C) for 15 min to inactivate intrinsic adenylate cyclase activity. The CHAPS extract was then diluted 3-fold with 25 mM NaHepes, pH 8.0. 200 pl of the diluted extract were mixed with 200 pl of cyc-membranes in 25 mM NaHepes, 1 mM EDTA. The final concentration of cyc-membrane protein was 2.2 mg/ml and that of CHAPS extract protein 0.67 mg/ ml. The reconstituted mixture was held on ice for 20 min.
Subsequently, 20 p1 of the mixture were added to 30 pl of solution containing the other reagents required for the measurement of adenylyl cyclase and incubated for 30 min at 30 "C. Final concentration of MgC12 was 10 mM, GTP and GTPyS, 10 pM, and that of isoproterenol (ISO), 50 p~. Values are mean f S.D. of triplicate determination.

Binding Characteristics of the Solubilized Receptor
We characterized further the ['251]MIG-binding activity with respect to its guanine nucleotide sensitivity, since in the intact membrane, the glucagon receptor binds ['251]MIG in a GTP-sensitive manner (28). We found that the specific binding of ['251]MIG in solution did not display GTP sensitivity (Table 111). We then compared the effect of varying concentrations of unlabeled glucagon on the extent of binding in the absence and presence of GTP. Using the CHAPS extract, it was found that glucagon competed with a n IC50 of 56 nM in the presence and absence of G T P (Fig. 11, bottom). Using liver membranes, the ICs0 of glucagon was 3.3 nM in the absence of GTP and 35 nM in the presence of G T P displaying the characteristic differential affinity induced by guanine nucleotides (Fig. 11, top). In three other experiments, the IC50 with which glucagon inhibited ['251]MIG binding to the solubilized receptor was 33-70 nM while the with which it inhibited ['251]MIG binding in the presence of GTP to the membrane-bound receptor was 35-80 nM. Thus, the solubilized receptor displays the low affinity GTP-insensitive binding of agonist.
We tested whether the lack of guanine nucleotide effect on the solubilized receptor was due to the lack of co-extraction of the stimulatory regulator of adenylyl cyclase. To this end, we tested for the presence of the stimulatory regulator in the CHAPS extract by assaying for reconstitution of cyc-S49 cell membrane adenylyl cyclase. As shown in Table IV, the glucagon receptor containing CHAPS extract also contains the stimulatory regulator of adenylyl cyclase.

Size and Subunit Structure of the Glucagon Receptor-
Receptors that stimulate adenylyl cyclase bind hormone in a guanine nucleotide-sensitive manner. Current evidence indicates that this guanine nucleotide sensitivity to hormone binding is conferred by the stimulatory regulator of adenylyl cyclase (29). Interaction between the hormone-occupied re- During hydrodynamic measurements, the receptor was monitored by the presence of the M , = 63,000 peptide along the density gradient or gel filtration column fractions. It appears that in the presence of nondenaturing detergents, the M, = 63,000 peptide behaves as part of a larger protein. Since the hydrodynamic measurements were performed in the presence of 0.5% Lubrol-PX, it does not seem likely that the larger size is due to nonspecific aggregation of hydrophobic protein. It also does not appear that the larger size is due to a receptor-detergent complex since no significant differences in the S values were measured using HzO and D20 as solvents during centrifugation. The lack of difference in S values indicates that the partial specific volume of the glucagon receptor is the same as that of the globular proteins used as standards and suggests that there is no significant binding of detergent to the protein; if the protein bound significant amounts of Lubrol-PX, the partial specific column of the protein-detergent complex would be greater than 0.74, since the partial specific volume of Lubrol is 0.956 ml/g (27). This would have resulted in different S values in H,O and DzO. The lack of difference in measured S values using HzO and D20 as solvents was rather unexpected. It is also possible that since the glucagon receptor is a glycoprotein, the partial specific volume of the receptor is lower than that of globular protein and the observed lack of shift in HzO uersw D,O is because the receptor-detergent complex coincidentally has a partial specific volume of 0.74. This appears unlikely since we can also observe the high molecular weight form in the membrane by direct SDS-gel electrophoretic analysis. There appears to be no a priori way of predicting whether a certain membrane protein, after solubilization, will bind significant amounts of detergent. Glycoprotein IIb and 111 of platelets which are known to be integral transmembrane proteins show also no differences in S values between H,O and D,O in the presence of Triton X-100 and their calculated molecular weights from hydrodynamic parameters agree with the observed molecular weights on SDS gels (30).
The data indicate that the M, = 63,000 peptide is a part of a larger protein ( M , = 110,000-120,000) both in detergent solution and in the membranes. The identification of a high molecular weight form in the membranes indicates that the observed size of the receptor in detergent solution is not an artifact of solubilization. Furthermore, the size of the glucagon receptor determined by hydrodynamic measurements is similar to the molecular weight obtained for the functional receptor by target size analysis (31).
To observe the presence of the high molecular weight species of the receptor in the membranes after solubilization with SDS, it was essential that membranes be incubated at 32 "C prior to the addition of ['251]MIG. This would suggest that specific aggregation of the receptor occurs in the vacant state and that addition of ['251]MIG allows dissociation of the aggregates as suggested by Rodbell (32). Alternatively, incubation in the absence of ['251]MIG may permit the receptor to orient itself such that it makes available reactive groups to which HSAB can &tach to cross-link the subunits of the receptor. Our data do not allow us to distinguish between these two alternate explanations.
The identity of the second subunit of the receptor protein merits comment. Since we are measuring the size of the agonist-occupied receptor, it may be suggested that a portion of the stimulatory regulator could account for the higher molecular weight. The experimental procedures we use preclude this possibility. After covalent attachment of [1251]MIG to receptor, we treat the membranes with GTP and excess unlabeled glucagon. Treatment with guanine nucleotide dissociates the receptor-regulator complex. Further, when guanine nucleotides are not included in the gradient, the stimulatory regulator displays a s20,w of 5.2, while the receptor has a szo,w of 4.3. The position of the receptor is unaffected by inclusion of guanine nucleotides during centrifugation. If the second subunit is not part of the stimulatory regulator, then the possibility exists that it is a subunit dissimilar from the hormone-binding subunit. While this possibility cannot be conclusively precluded for the glucagon receptor, it appears unlikely. Recent studies on the pure ,&adrenergic receptors have shown that the hormone-binding subunit is sufficient to confer hormone responsiveness to the system, indicating that there are no additional components that intervene between the receptor and the stimulatory regulator (33,34). At the present time, there is no reason why this situation should not be applicable to the glucagon receptor as well. Another possibility is that the native protein is a dimer of the hormonebinding subunit. In view of the size of the low ( M , = 63,000) and high (Mr = 110,000-120,000) molecular weight form, this appears to be a reasonable explanation. Definitive proof to support this contention must await purification of the glucagon receptor.
Solubilization and Assay for Functional Glucagon Receptor in Solution-The results show that specific ['251]MIG-binding activity has been solubilized by exposure of rat liver membranes to CHAPS. As with other systems, the choice of detergent appears to be crucial for the successful solubilization of the active receptor since significant binding activity could only be obtained when CHAPS was used. Repeated attempts to use Lubrol-PX or cholate during solubilization have not yielded solubilized receptors that retain hormone-binding activity. The binding activity does not sediment when subjected to centrifugation at 220,000 X g for 2 h and freely passes through a 0.22-c( filter, indicating that it is truly soluble. The binding activity in solution displays characteristics of the glucagon receptor. These are: 1) specificity of competition by peptide hormones; 2) similar rates of glucagon binding; and 3) similar affinities for glucagon as observed in the membrane in the presence of GTP.
The method described in this article for the measurement of ['251]MIG binding to the receptor represents a new a n d novel approach. It is based on the observation that the receptor is a glycoprotein, as demonstrated by the GlcNAc-sensitive adsorption of the receptor to WGL-Sepharose. We have utilized this structural feature of the receptor to develop an assay for receptor function (i.e. hormone binding), demonstrating the utility of structural studies on covalently labeled trace proteins. In the assay described here, we use the WGL-Sepharose as an insoluble matrix onto which the hormonereceptor complex is adsorbed under conditions where the adsorption of the free hormone is minimal. This allows for the removal of the free hormone in the supernatant after a low speed centrifugation. A different effect of lectin has been observed in another receptor system. It has been reported that concanavalin A stabilizes the conformation of epidermal growth factor receptor in detergent solution such that hormone binding is facilitated (35). No such effect of WGL was observed on the glucagon receptor. Addition of WGL during the binding reaction did not allow us to observe specific binding of ['251]MIG to the receptor when polyethylene glycol precipitation technique was used to separate the free ligand from the bound. Thus, WGL does not appear to promote ['"I] MIG interaction with the CHAPS-solubilized receptor.
While this study has specifically focused on optimizing conditions for the measurement of the glucagon receptor, the general principle used for the measurement of hormone binding to the solubilized receptor should be useful for other receptors and glycoproteins as well. The procedure will work if the appropriate lectin-Sepharose is used to absorb the glycoprotein of choice. The simplicity of this method compares favorably with that of the polyethylene glycol precipitation technique which indeed is not always usable. In the case of the glucagon receptor, the availability of a simple and reproducible binding assay after solubilization represents a major step toward the purification of the receptor.
Binding Characteristics of the Solubilized Glucagon Receptor-It is noteworthy that the glucagon receptor in solution displays low affinity guanine nucleotide-insensitive binding of ['251]MIG. Studies on the P-adrenergic system had shown that solubilization of the unoccupied receptor results in a form that binds agonists with low affinity unaffected by guanine nucleotides (36). Genetic manipulations on the S49 lymphoma cell system have shown that such low affinity binding represents binding to receptor unaffected by a functional stimulatory regulator of adenylyl cyclase (15,37). This was substantiated by Shorr et al. who showed that the purified frog erythrocyte P-adrenergic receptor binds agonists with low affinity in a guanine nucleotide-insensitive manner (38). Thus, by a number of criteria, low affinity nucleotide-insensitive agonist binding is a characteristic of the free receptor. The studies presented in this article demonstrate that the solubilized glucagon receptor also binds hormone with low affinity in a guanine nucleotide-insensitive manner, indicating that it is a free receptor. If follows, therefore, that when agonist probes such as [1251]MIG are used to study receptor behavior in intact membranes in the absence of guanine nucleotides, the data obtained represent the behavior of the receptor-stimulatory regulator complex and not that of the receptor itself. The functional characterization of the soluble receptor as "free" agrees with the structural studies which also indicate that the ['251]MIG receptor in detergent solution is not the receptor-stimulatory regulator complex.
The lack of guanine nucleotide effect in solution is not due to the functional absence of the stimulatory regulator in the CHAPS solution, since it can reconstitute guanine nucleotide, and, more importantly, hormonal sensitivity to cyc-,549 cell membrane adenylyl cyclase. It appears that reconstitution into a lipid environment may be necessary for such receptorstimulatory regulator interactions to occur, as has been found to be the case for the @-adrenergic receptor.