T H E JOURNAL OF BIDLOCICAL CHEMISTRY VoI. 256, No. 24, Issue of December 25. pp. 12866-12874, 1981 Prrrcted irr U. S A

Differentiation-dependent Expression of Catecholamine-stimulated Adenylate Cyclase ROLES OF THE 0-RECEPTOR AND G/F PROTEIN IN DIFFERENTIATING 3T3-Ll ADIPOCYTES*

(Received for publication, March 20, 1981)

Eseng Lait$, Ora M. RosenflII, and Charles S . Rubin+**+$ From the $Departments of Molecular Pharmacology, qMedicine, and **Neuroscience, Albert Einstein College of Medicine, Bronx, New York I0461

3T3-Ll Preadipocytes possess an adenylate cyclase system that is highly stimulated (-4.5-fold) by 10 PM GTP in the absence of @-adrenergic agonists and other hormones. During the differentiation of preadipocytes into adipocytes, a sharp decline in this activity precedes the manifestation of catecholamine-stimulated (6- to 10-fold) adenylate cyclase activity. Preadipocytes ex- hibit 1900 high affinity @-receptorsJcel1, but the maxi- mal occupation of these receptors by isoproterenol re- sults in only a 20% elevation in the rate of CAMP syn- thesis. During differentiation, P-receptor number in- creases by only 60-70% to 3100 sites/cell. /I-Receptors in preadipocyte membranes have a KO of 0.2 n~ for [12613 iodohydroxybenzylpindolol and decrease their affinity for /I-agonists in the presence of GTP. During prolonged treatment with 5 p~ isoproterenol, preadipocytes down regulate more than 80% of their receptors. No signifi- cant changes in these properties occurred during adi- pocyte development. Moreover, p-receptors in preadi- pocyte and adipocyte membranes exhibited similar af- finities and stereoselectivities for a series of b-agonists and &antagonists. In contrast to the unaltered prop- erties of the 8-receptor, the regulation of adenylate cyclase activity by GTP changes markedly during adi- pocyte development. GTP (10 ELM) stimulates the pre- adipocyte enzyme equally well in the absence or pres- ence of isoproterenol, while the GTP-mediated en- hancement of adipocyte membrane adenylate cyclase activity is highly dependent on the presence of the @- adrenergic agonist. These observations suggested that G/F protein exhibited different sets of regulatory (coupling) properties in the presence of similar 8-recep- tors in undifferentiated and differentiated cells.

In complementary experiments, the cholera toxin- catalyzed [32P]ADP-ribosylation of 3T3-Ll membranes revealed dramatic differences between the G/F sub- units of preadipocytes and adipocytes. In adipocyte membranes, 13-fold more 32P was incorporated into the 42,000-dalton component and 4-fold more labeling was observed in the 49,000- to 50,000-dalton doublet than in the corresponding G/F subunits in preadipocyte mem-

AM21248 and Diabetes Research and Training Grant 5P6OAM-20541. * This work was supported by National Institutes of Health Grant

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

§ Supported by National Institutes of Health Grant 5T32GM7288. This study is in partial fulfillment of the requirements for the Ph.D. degree.

\I Recipient of an Irma T. Hirschl Career Scientist Award. if Recipient of Career Development Award K04AM-00190 from

the National Institutes of Health.

branes. Furthermore, determinations of the relative proportions of the two 32P-labeled components of G/F disclosed a preponderance (4.2:1) of the 49,000- to 50,000-dalton species in preadipocyte membranes and nearly equal amounts of the 42,000-dalton and 49,000- to 50,000-dalton polypeptides in adipocyte membranes. These findings raise the possibility that differentiation- associated changes in the regulation and coupling prop- erties of catecholamine-stimulated adenylate cyclase may be determined principaIly by modulation of the levels, proportions, and/or properties of the constitu- ents of G/F.

The regulation of catecholamine-stimulated adenylate cy- clase by hormones and guanine nucleotides depends upon interactions among the &adrenergic receptor, the guanine nucleotide-binding regulatory protein or G/F,' and the cata- lytic component of the enzyme (reviewed in Refs. 1 and 2). Recent investigations employing a combination of biochemi- cal, genetic, and membrane fusion techniques (1-5) have provided substantial evidence that these three functional com- ponents of the adenylate cyclase system are distinct and physically separable proteins. In addition, biochemical char- acterizations of mutants of S49 lymphoma cells (6-9) and naturally occurring variants of other cells (7,9) have disclosed that G/F plays a central role in mediating the interactions between the &receptor and the catalytic component.

Physiologically responsive target cells for @-adrenergic ag- onists are highly specialized and terminally differentiated. Presumably, these cells contain 8-receptors, G/F and catalytic protein in appropriate stoichiometries, conformations, and membrane environments to optimize ligand recognition and signal transmission. There is very little known about how these components are integrated to become a hormonally responsive adenylate cyclase system during cell and tissue differentiation.

The murine 3T3-Ll cell line appears to be especially suita- ble €or studying the assembly of a hormonally sensitive ade- nylate cyclase system. Differentiation of 3T3-LI preadipo- cytes into adipocytes can be elicited rapidly and uniformly in cell culture (10). This process is accompanied by the acquisi- tion of a catecholamine-sensitive adenylate cyclase and a lipolytic response to catecholamines (11). The physiological development of the catecholamine-responsive adenylate cy-

' The abbreviations used are: G/F, the regulatory component of adenylate cyclase which appears to be a site of action of guanine nucleotides and fluoride; '251-KYP, ['"I]iodohydtoxybenzylpindo~o~ Gpp(NH)p, guanyl-5"yl imidodiphosphate; SDS, sodium dodecyl sul- fate.

12866

3T3-Ll Adenylate Cyclase 12867

clase system in differentiated 3T3-Ll cells enables the study of developmental changes in the P-receptor, G/F protein, and catalytic unit and the degree to which these components interact at various stages during differentiation.

In this communication, we ( a ) report the occurrence of an adenylate cyclase system in 3T3-Ll preadipocytes that is highly stimulated by GTP in the absence of hormones, ( b ) find that a decline in this activity is coordinated with the appearance of catecholamine-responsive adenylate cyclase during adipocyte differentiation, ( c ) characterize the specific binding and coupling properties of ,&adrenergic receptors in preadipocytes and adipocytes and (d) document large, differ- entiation-associated changes in the relative proportions and apparent amounts of the 42,000-dalton and 49,000- to 50,000- dalton species of G/F subunits as determined by labeling with cholera toxin and [32P]NAD.

EXPERIMENTAL PROCEDURES

Cell Culture"ST3-LI and 3T3-C2 cell lines were originally ob- tained from Dr. Howard Green (Harvard Medical School). Cells were grown in 150-mm tissue culture dishes (Lux) or 530-cm2 square tissue culture plates (Nunc) as previously described (10). Preadipocytes were seeded a t a density of -2500 cells/cm2 and grown to confluence in 4-5 days (approximately four cell divisions).

Murine S49 lymphoma cell lines (wild type and cyc- variants) were kindly provided by Dr. Alfred Gilman (Univ. of Texas Health Sciences Center, Dallas) and were propagated in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated horse serum, non- essential amino acids, and 2 m~ glutamine.

Differentiation of 3T3-Ll Cells to Adipocytes-Confluent preadi- pocytes (day 0 ) were fed with fresh medium containing 0.5 mM 1- methyl-3-isohutylxanthine and 0.25 p~ dexamethasone (IO). After 48 h (day Z), this medium was aspirated and the cells were refed with fresh standard medium. Cells were fed again by the addition of 30% (v/v) fresh medium 96 h after initiating drug treatment (day 4). Day 6 ceUs (adipocytes) were harvested 48 h after the second refeeding. By day 6, 80-90% of the cells have differentiated into rounded cells with large lipid droplets.

Preparation of Cell Membranes-All steps were performed at 4 "C. Dishes of 3T3-LI or C2 cells were rinsed twice in buffer A (10 mM potassium phosphate, pH 7.4, and 0.14 M NaCI). Cells were then scraped off the dish in buffer A with a rubber policeman. Cell number was determined and the cells were harvested by centrifugation at 350 X g for 4 min. The cells were then resuspended (-10-15 X 10' cells/ ml) in buffer B (5 mM Tris buffer, pH 7.4, containing 5 mM NaCl, 0.1 mM ethylene glycol bis(P-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA), and 1 mM dithiothreitol). After 3 min, MgCI? was added to a final Mg2+ concentration of 0.5 mM. Two min later, the cells were homogenized with 20 strokes of a tight fitting Dounce homogenizer. The homogenate was centrifuged at 650 X g for 10 min to remove nuclei and the supernatant fraction was collected. The pellet was washed once and the second supernatant fraction was combined with the first. The combined postnuclear supernatant fraction was centri- fuged at 45,000 X g for 15 min. The membrane pellet was washed once in buffer A without added dithiothreitol and resuspended at a final concentration of 2-3 mg of protein/ml.

Preadipocytes have less membrane protein per cell than do adi- pocytes. Preadipocytes yielded approximately 0.079 -+ 0.020 mg of protein/106 cells, while adipocytes yielded 0.208 -+ 0.028 mg of protein/ lo6 cells (average of seven sets of differentiating cells).

Membranes from S49 lymphoma cells were prepared in a similar manner. Cells were washed twice by centrifugation at 350 X g for 5 min in buffer A. They were then resuspended in buffer B and processed as described above.

Adenylate Cyclase Assay-Adenylate cyclase activity was deter- mined as described by Salomon et al. (12). An ATP preparation (Sigma A-2383) low in contaminating GTP was used as substrate. The reaction mixture contained 0.5 mM [cw3*P]ATP (specific activity

150 cpm/pmol), 5 mM MgCl?, 20 mM phosphocreatine, 20 units/ml of creatine phosphokinase, 1 mM CAMP, and 0.5 mM dithiothreitol in 25 mM Tris buffer, pH 7.4 (37 "C).

I-HYP Binding to Membranes-Membranes were incubated with I2'II-HYP at 37 "C in 50 mM Tris-HC1, pH 7.4, 10 mM MgCI?, 1 mM pyrocatechol, and 1 mM ascorbic acid. Unless otherwise specified, the l2'1-HYP concentration was 300-400 PM. Membranes from

I25

300,000-500,000 cells were used in each assay in a total incubation volume of 100 4. The binding reaction was terminated by filtration on Gelman AE1 glass fiber filters under low vacuum. Filters were washed with 25 ml of 50 mM Tris-HCl, pH 7.4, and 10 mM MgSO, at 37 "C, and radioactivity was determined in a y spectrometer.

Nonspecific binding was defined as the "'1 radioactivity that was not displaced either by 60 p~ isoproterenol or 2 pM propranolol. This was usually -20% of total binding. All binding data were corrected for nonspecific binding.

GTPase Assay-Total GTPase activity was measured by following the degradation of ["HIGTP to ['HIGDP and ["HIGMP under the conditions of the adenylate cyclase assay. Nucleotides were separated and identified on polyethyleneimine cellulose with 0.5 M (NH,)~SOJ as solvent (13).

ADP-Ribosylation of G/F-Membranes (30 pg-200 p g ) were sus- pended in a total volume of 60 p1 of 50 mM potassium phosphate buffer, pH 7.8, containing 20 mM thymidine, 5 mM ATP, 5 mM MgCL, 0.1 mM Gpp(NH)p, and 5-10 PM ["'PINAD (-40-50 cpm/fmol). ADP- ribosylation was initiated by the addition of activated cholera toxin (2.4 or 6 pg) and incubations were carried out at 30 "C for 45 min to obtain maximal labeling of the 42,000-dalton and 49,000- to 50,000- dalton components of G/F. Cholera toxin (1 mg/ml) was preactivated by incubation with 20 mM dithiothreitol in 10 mM potassium phos- phate buffer, pH 7.8, for 15 min at 37 "C. Control samples without cholera toxin received an equal volume of 20 mM dithiothreitol in 10 mM potassium phosphate buffer, pH 7.8. Incubations were terminated by the addition of 2 ml of ice-cold 10 mM potassium phosphate, pH 7.8, and the membranes were isolated by sedimentation at 40.000 X g for 15 min at 4 "C. After carefully aspirating the supernatant fluid, the membrane pellet was solubilized in 25 p1 of Tris-HC1 buffer, pH 6.7, containing 10% glycerol, 2% SDS, and 1% ,8-mercaptoethanol. Each sample was heated at 95 "C for 3 min and the total sample was applied to one lane of a 10% polyacrylamide gel containing 0.1% sodium dodecyl sulfate. SDS-Polyacrylamide gel electrophoresis, staining, destaining, and autoradiography were performed as previ- ously described (14).

Quantitation of '?P Incorporated into G/F-Using the correspond- ing autoradiogram as a guide, slices of dried gels containing labeled components of G/F were cut out and dissolved by heating at 55 "C in 0.4 ml of 30% H,Os (14) for 16 h. After cooling to room temperature, each sample received 3.6 ml of scintillation fluid (700 ml of toluene/ 300 ml of Triton X-100/4 g of Omnifluor (New England Nuclear)) and the radioactivity was determined in a scintillation spectrometer.

Protein Determinations-Protein was measured by the method of Lowry et al. (15) using bovine serum albumin as the standard.

Materials-GTP (dilithium salt) and pyruvate kinase were pur- chased from Boehringer Mannheim. ATP (A-2383), cyclic AMP, Gpp(NH)p, phosphocreatine, creatine phosphokinase, phosphoenol- pyruvate (tricyclohexylamine salt), chromatographic alumina (neu- tral), (-)- and (+)-isoproterenol bitartrate, (-)-epinephrine bitar- trate, (-)-norepinephrine bitartrate, (-)-alprenolo1 tartrate, (&)-pro- pranolol hydrochloride, catechol, ascorbic acid, and thymidine were obtained from Sigma. (-)- And (+)-propranolol hydrochloride were provided by Ayerst Research Laboratories. Phentolamine hydrochlo- ride was from Ciba-Geigy.

RESULTS

Comments on the Normalization of the Data-3T3-Ll Adipocytes are larger and yield more membrane protein per cell than the preadipocytes from which they have developed. Consequently, comparisons between preadipocytes (day 0) and adipocytes (day 6) can be normalized either to the amount of membrane protein or to the number of cells from which the membranes are derived. Because the relative proportion of plasma membranes (which contain the cyclase system) in the crude membrane preparations isolated from adipocytes may be different from that in crude preadipocyte membranes, normalization of the data to protein could be misleading. Thus, all results were expressed in terms of cell number. Conversion of the data to normalization by amount of protein can be readily approximated from the values given under "Experimental Procedures."

Development of P-Agonist-stimulated Adenylate Cyclase- Adenylate cyclase activity was measured in both homogenates

12868 3T3-Ll Adenylate Cyclase

FIG. 1. Activation of adenylate cyclase by isoproterenol and prostaglandin E1 (E&) during adipocyte differentiation. Ade- nylate cyclase activity was determined as described under “Experi- mental Procedures.” Incubations were carried out at 37 “C for 15 min. The bar graphs present the per cent activation of adenylate cyclase activity observed in the presence of hormone (A, 60 p~ isoproterenol + 10 p~ GTP; B, 200 p~ PGEl + 10 PM GTP), where per cent stimulation = 100 X (adenylate cyclase activity in the presence of hormone and GTP - adenylate cyclase activity in the presence of GTP alone) + adenylate cyclase activity in the presence of GTP alone. Membranes (0) or cell homogenates (El) were prepared from 3T3-L1 cells at 48-h intervals during the course of differentiation of a single set of cells (see “Experimental Procedures”). Membranes were also prepared from 3T3-C2 cells (El) subjected to the differentia- tion protocol.

and partially purified membranes prepared from differentiat- ing 3T3-Ll cells. Activation of adenylate cyclase by the p- adrenergic agonist, isoproterenol, was very limited in mem- branes from undifferentiated cells (day 0) but increased mark- edly during adipocyte development (Fig. U), reaching a max- imum value by day 6 of the differentiation protocol (see “Experimental Procedures”). Typically, isoproterenol stimu- lated adenylate cyclase activity 500-900% in membranes from differentiated 3T3-Ll cells but elevated enzyme activity only 20% above values obtained with GTP alone in preadipocyte membranes. In contrast, the degree of stimulation of adenylate cyclase by prostaglandin El remained relatively constant dur- ing the conversion of preadipocytes to adipocytes (Fig. 1B). The acquisition of catecholamine-stimulated adenylate cy- clase activity was correlated with differentiation rather than drug treatment. 3T3-C2 cells,’ which do not differentiate after exposure to dexamethasone and methylisobutylxanthine, did not develop catecholamine-stimulated adenylate cyclase ac- tivity (Fig. 1A).

The data in Fig. 1 show that isoproterenol elicited very similar levels of adenylate cyclase activation in homogenates and partially purified membranes. Since the fractionation procedure used to isolate membranes apparently does not alter the functional organization of the p-agonist-sensitive adenylate cyclase system and the membranes can be washed free of endogenous GTP, the membrane preparation was judged to be suitable for studies on the effects of guanine nucleotides on the adenylate cyclase system and the charac- terization of the &receptor.

Characterization of the /?-Adrenergic Receptor-The p- adrenergic receptor of 3T3-Ll cells was characterized during differentiation using the radiolabeled ,&adrenergic antagonist, [1251]iodohydroxybenzylpindolol (‘251-HYP). In initial studies carried out on 3T3-Ll adipocyte membranes, ‘“1-HYP bind- ing was found to be rapid and reversible(Fig. 2), saturable, and specifically displaceable by a series of p-adrenergic ago- nists (Fig. 3A) and antagonists (Fig. 3B). Under the standard assay conditions, equilibrium binding was attained within 20 min at 37 “C and saturation of the specific p-binding sites occurred at a concentration of 2 nM ‘251-HYP.

The binding of &adrenergic ligands by the membranes was also highly stereoselective (Fig. 3). The physiologically active (-)-isomers of isoproterenol and propranolol were approxi-

The 3T3-CZ cell line, derived from the same precursor as the 3T3- L1 cell line, retains the characteristics of fibroblasts during prolonged maintenance in culture as a confluent monolayer.

- Add 2wM Propronolol I

0 \

1000

‘ 0 20 40 60

MINUTES FIG. 2. Kinetics of association (W) and dissociation

(0- - -0) of ‘261-HYP with adipocyte (day 6) membranes. Mem- branes from 3.8 X 10’ cells were incubated as described under “Ex- perimental Procedures” with 0.4 nM ‘2’I-HYP at 37 ”C. All data are corrected for nonspecific binding. To measure dissociation, 2 p~ (&)- propranolol was added in 50 mM Tris-HCI, pH 7.4, and 10 mM MgS04 to make the final assay volume 1.0 ml.

LOG [ogonisl] (MI

I -3

\ I 60

nn

LOG [ontogonist] (M)

FIG. 3. Competitive displacement of ‘251-HYP binding to adi- pocyte (day 6) membranes by p-adrenergic agonists and an- tagonists. Membranes were incubated with ’251-HYP as described under “Experimental Procedures” for 30 min at 37 “C in the presence of the indicated concentrations of agonist (A) and antagonist ( B ) drugs. The data are presented as percentages of total specific binding determined as B/Bo X 100 where 30 = ’251-HYP bound in the absence of competing ligands and B = ‘“I-HYP bound in the presence of the indicated concentrations of agonists or antagonists. Nonspecific bind- ing was determined in the presence of 60 p~ isoproterenol (A) or 2 PM propranolol (B) .

mately 100-fold more potent than the corresponding (+)-iso- mers in competitively inhibiting lZ5I-HYP binding. The a- adrenergic antagonist, phentolamine, did not displace lZ5I- HYP at concentrations up to 100 pM.

p-Adrenergic receptors in preadipocytes and adipocytes were compared in equilibrium binding studies. Data from a typical experiment are plotted according to the method of Scatchard (16) in Fig. 4. Membranes from preadipocytes and adipocytes both yielded linear plots, consistent with a single set of noninteracting binding sites. Day 0 cells (preadipocytes) exhibited approximately 1850 sites/cell with a dissociation constant of 0.20 -t 0.03 nM, while day 6 cells (adipocytes) possessed 3050 sites/cell with an apparent KO of 0.22 f 0.06 nM (averages of four separate sets of differentiating cells). The

3T3-Ll Adenylate Cyclase 12869

number of sites per cell varied among different batches of cells, but the proportional change in receptor number between day 0 and day 6 in each batch was consistent.

The kinetics, saturability, and reversibility of binding and the stereoselective displacement of '''I-HYP by /3-adrenergic agonists and antagonists observed using preadipocyte mem- branes also paralleled the results obtained with adipocyte membranes (data not shown).

Regulation of @-Adrenergic Receptor Affinity and Num- ber-Two indices of the ability of the /3-receptor to interact productively with G/F are the down regulation of /%receptors upon prolonged exposure to agonists (17) and the decrease in receptor affinity for agonists effected by guanine nucleotides (18, 19).

Both 3T3-Ll preadipocytes and adipocytes exhibit the down regulation of their @-receptors with similar kinetics upon persistent treatment with isoproterenol (Table I). More than one-half of the @-receptors were lost within 1 h and the number of receptors observed after 6 h closely approached a stable plateau value of approximately 16% of the initial recep- tor density seen at 20 h and subsequent times.

Guanine nucleotides negatively modulate the affinity of the &adrenergic receptor for agonists in membranes that contain fully functional G/F protein (18, 19). Fig. 5 shows that GTP

'2sI-HYP BOUND (frnoles/lO6cells)

FIG. 4. Scatchard plots of 1261-HYP binding data for preadi- pocyte ( O " 0 ) and adipocyte (0"io) membranes. Preadi- pocyte (day 0) and adipocyte (day 6) membranes were incubated to equilibrium at 37 "C with "'I-HYP over the concentration range of 20-500 PM. Specific Iz5I-HYP binding was determined as described under "Experimental Procedures."

TABLE I Agonist-induced P-adrenergic receptor loss in preadipocytes and

adipocytes Isoproterenol was added to the culture medium to a final concen-

tration of 5 p ~ . After the indicated incubation times, cells were harvested and membranes prepared as described under "Experimen- tal Procedures." IZ5I-HYP binding was performed as described with 0.7 nM '*'I-HYP. Control specific binding was 1.2 fmol/lOG cells in preadipocyte membranes and 2.0 fmol/lOG cells in adipocyte mem- branes. Addition of 5 PM isoproterenol for 2 min prior to cell harvest did not alter control values for '"I-HYP binding, thereby indicating that prebound and free drug were eliminated during membrane preparation.

% of control specific '*'I-HYP binding Incubation

time Preadipocyte membranes

Adipocyte membranes

h 0 1 6

20

100 48 19 16

100 4 0 23 1:

(3

P z z u\ \ a

\:\

4

LOG fiSOPROTERENOL] (MI FIG. 5. Effects of GTP on the isoproterenol displacement of

'261-HYP binding to &receptors in preadipocyte and adipocyte membranes. Membranes from preadipocytes (A) and adipocytes (B) were incubated with "'1-HYP as described under "Experimental Procedures." The data are presented as percentages of total specific

trations of isoproterenol in the absence (A-A and t".) and presence of 100 KM GTP (A-A and O " - O ) . The data represent the averages of & S. D. of three separate experiments. The '*'I-HYP concentration was between 0.5 IIM and 0.6 nM.

125 I-HYP binding remaining in the presence of the indicated concen-

TABLE 11 Stimulation of adenylate cyclase by guanine nucleotides during the

differentiation of 3T3-Ll adipocytes Adenylate cyclase activities were determined in membrane prepa-

rations isolated during the course of adipocyte differentiation as described under "Experimental Procedures." The data represent du- plicate determinations in a typical experiment using membranes from a single set of cells. GTP and Gpp(NH)p were present at concentra- tions of IO PM. The data are normalized to the number of ce!ls from which the membranes were derived.

Day 2 5.4 9.9 1.8 47.9 Day 4 9.2 15.2 1.7 110.9

8.6 4.8 12.4

Day 6 13.7 20.2 1.5 149.5 10.9 7.4 7.3

decreases the affinity of the receptor for isoproterenol in both preadipocyte and adipocyte membranes. In contrast to the equally avid binding of '251-HYP by both membranes (Fig. 4), adipocyte membranes displayed a slightly higher affinity for isoproterenol and a somewhat larger shift in the presence of GTP. The significance of these small differences is unknown.

Activation of Adenylate Cyclase by Guanine Nucleotides in the Absence of Catecholamines-Adenylate cyclase in pre- adipocyte (day 0 ) membrane preparations was highly sensitive to stimulation by GTP alone. In the typical experiment pre- sented in Table 11, 10 p~ GTP elicited a 4.5-fold increase in enzyme activity. Responsiveness to GTP alone rapidly dimin- ished during adipocyte development after a brief lag period. GTP responsiveness was unaltered 8 h after the initiation of

12870 3T3-Ll Adenylate Cyclase

the differentiation protocol (see “Experimental Procedures”) and decreased by only 20% after 24 h (data not shown). During the next 24 h, however, the stimulation of activity by GTP dropped by approximately 67% (Table 11, day 2). Similarly, low levels of GTP stimulation were noted as the cells pro- gressed through the adipocyte differentiation program (Table 11, days 4 and 6). In contrast, responsiveness of adenylate cyclase to Gpp(NH)p, a nonhydrolyzable analog of GTP, increased. In undifferentiated (day 0) cells, GTP and Gpp(NH)p activated cyclase to a similar degree; by day 6, adenylate cyclase activity in the presence of Gpp(NH)p was more than 7 times higher than that observed with GTP (Table 11).

Since the activation of adenylate cyclase by Gpp(NH)p has a lag phase before reaching steady state levels, the data in Table I1 on GPP(NH)p-stimulated activity represent net cAMP synthesized during the linear and lag components of the 15-min assay. This lag does not change the basic obser- vation. Fig. 6 shows that the lag period of Gpp(NH)p activa- tion is similar in preadipocyte and adipocyte membranes, while GTP activation has no lag phase in either case. The dramatic increase in the ratio of Gpp(NH)p-stimulated activ- ity to GTP-stimulated activity is also reflected in the steady state rates of CAMP synthesis. Values for this ratio (corrected for the contribution of basal activity) were 3 in preadipocytes and 18 in adipocytes.

The loss of activation by GTP alone was correlated with differentiation. When membranes were isolated from 3T3-C2 cells throughout the course of a standard dexamethasone- methylisobutylxanthine treatment protocol, GTP (10 PM) stimulated adenylate cyclase activity 6- to 8-fold at all stages (Table 111). Furthermore, the ratio of cyclase activity in the presence of Gpp(NH)p to that observed in the presence of GTP also remained constant at 1.7 2 0.2. Thus, the properties of guanine nucleotide-activated adenylate cyclase in mem-

MI NU TES FIG. 6. Time course of adenylate cyclase activation by GTP

and Gpp(NH)p. Adenylate cyclase activity was determined as de- scribed under “Experimental Procedures” except that the initial reaction volume was 215 p1 and contained -2 pCi/ml of [’HICAMP. At the indicated times during the incubation, 25-pI aliquots were removed and transferred to tubes containing 100 pl of 2% SDS, 1.4 mM CAMP, and 20 mM ATP. Analyses were performed as described by Salomon et al. (12). The data are from a typical experiment. o ” - o , Adipocyte membranes incubated with 10 ELM Gpp(NH)p; M, adipocyte membranes incubated with 10 ,UM GTP A- - -A, preadipocyte membranes incubated with 10 Cpp(NH)p; A- - -A, preadipocyte membranes incubated with 10 MM GTP.

TABLE I11 Stimulation of adenylate cyclase by guanine nucleotides in 3T3-C2

treated with dexamethasone and I-methyl-3-isobutylxanthine Adenylate cyclase activities were determined in membrane prepa-

rations isolated from 3T3-C2 cells treated for 48 h with dexametha- sone and 1-methyl-3-isobutylxanthine as described in the protocol for the differentiation of 3T3-LI cells under “Experimental Procedures.” Unlike 3T3-Ll cells, the 3T3-C2 cells do not differentiate under these culture conditions. GTP and Gpp(NH)p were present at concentra- tions of 10 ,MM. The data are normalized to the number of cells from which the membranes were derived.

(Adenylate cyclase activity (pmol cAMP svnthesized/106 c e l k l 5 mi;

Day 0

seem to be quite simiiar. branes from 3T3-C2 cells and undifferentiated 3T3-LI cells

1.7 13.6 61.2 8.2 37.0 4.5 Day 6

1.5 1.7 9.5 68.5 5.7 41.3 7.2 Day 4

11.7 59.8 1.9

7.6 38.6 5.1 Day 2 12.0 49.4 6.4 26.1 4.1

Dose response curves for the effects of guanine nucleotides on adenylate cyclase activity in membranes derived from preadipocytes (day 0) and adipocytes (day 6) disclosed that the decreased ability of GTP to activate the enzyme in adi- pocytes could not be attributed to a decreased affinity for GTP (Fig. 7A), nor could the increased efficacy of Gpp(NH)p in adipocyte membranes be ascribed to an elevated K , for the analog (Fig. 7B).

Finally, the foregoing results cannot be explained by a higher rate of nonspecific hydrolysis of GTP by adipocyte membranes during the cyclase assay. h e n total GTPase activity was determined under the conditions of the adenylate cyclase assay (see “Experimental Procedures”), only -10% of added C3H]GTP was degraded to GDP and GMP by either preadipocyte or adipocyte membranes.

GTP-dependent Activation of Adenylate Cyclase by Iso. proterenol-When adipocyte membrane adenylate cyclase ac- tivity was assayed using the standard phosphocreatine/crea- tine phosphokinase ATP-regenerating system, enzyme activ- ity was highly and equally stimulated by isoproterenol in the presence (Fig. 1) and absence of GTP. Although unexpected, this result is not unique. Adenylate cyclase activity in highly purified liver plasma membranes is readily activated by glu- cagon in the absence of added GTP (20,211. Such observations have been attributed to the presence of contaminating GTP in the membranes or assay reagents, the generation of GTP during the assay, or the occurrence of non-nucleotide activa- tors in the components of the ATP-regenerating system (22).

Attempts to unmask GTP-dependent, isoproterenol-stimu- lated activity by employing highly purified adipocyte mem- branes isolated by the procedure of Ross et aZ. (23) or by purifying ATP and the components of the ATP-regenerating system, phosphocreatine and creatine phosphokinase (24), were unsuccessful.

GTP-dependent isoproterenol activation of adenylate cy- clase was demonstrated, however, when an alternative ATP- regenerating system, phosphoenolpyruvate/pyruvate kinase, was substituted for phosphocreatine/creatine phosphokinase. Thus, 3T3-Ll adipocyte membrane adenylate cyclase seems to be highly sensitive to a factor(s) inherent in or generated by phosphocreatine/creatine phosphokinase that mimics GTP. While the nature of the factor remains unknown, it is unlikely to be a contaminating guanine nucleotide because the purification procedures used should have eliminated such nucleotides and the addition of either phosphocreatine or creatine phosphokinase alone to assays containing the phos- phoenolpyruvate/pyruvate kinase-regenerating system did

3T3-Ll Adenylate Cyclase 12871

not alter the GTP requirement for isoproterenol stimulation. Typical results of assays performed with the phosphoenol-

pyruvate/pyruvate kinase-regenerating system are provided in Table IV. GTP alone increased adipocyte adenylate cyclase activity very little, but the combination of 6 p~ isoproterenol and 10 ~ L M GTP stimulated the enzyme approximately 10-fold relative to the basal value. In contrast, preadipocyte adenylate cyclase was activated by GTP whether or not hormone was present.

Labeling and Comparison of G / F Subunits in Adipocyte a n d Preadipocyte Membranes-The striking changes in hor- mone-dependent and -independent activation of adenylate cyclase by guanine nucleotides during differentiation sug- gested that qualitative and/or quantitative changes in G/F

5 0 K -

4 2 K ' 4 9 K -

FIG. 7. Activation of adenylate cyclase by guanine nucleo- tides. Adenylate cyclase activities in preadipocyte (t".) and adipocyte (0- - -0) membranes were determined as described under "Experimental Procedures" in the presence of various concentrations of GTP (A) or Gpp(NH)p (B) . The points represent the average of duplicate determinations in a typical experiment.

TABLE IV Effects of GTP and isoproterenol on adenylate cyclase in

preadipocyte and adipocyte membranes Adenylate cyclase assays were carried out as described under

"Experimental Procedures" except that the standard ATP regener- ating system was replaced with 5 mM phosphoenolpyruvate and 25 pg/d of pyruvate kinase. Incubations were carried out for 15 min a t 37 "C.

Adenylate cyclase activity (pmol CAMP synthesized/ 10" cells/l5 min)

Basal +lOpM + 6 p ~ iso- GTP + 6 p ~ GTP proterenol Isoprotere-

+10 pM

nol

Preadipocytes 10.1 35.0 13.2 36.0

Adipocytes 8.6 14.4 21.3 83.6 (day 0 )

(dav 6)

A B Day Day Day Day

WT cyc- 0 6 0 6

I 2 3 4 5 6 7 8 1 2 nnnn

" - C T - + " + - + " k + + FIG. 8. Autoradiogram of [32P]ADP-ribosylat-d G/F cum-

ponents in 549 and 3T3-Ll membranes. A, membranes from 349 WT cells (lanes 1 and 2). S49 cyc cells (lanes 3 and 4 , 3T3-Ll preadipocytes (day 0, lanes 5 and 6). and 3T3-LI adipocytes (day 6, lanes 7 and 8) were incubated with 10 PM [,"PINAD in the absence or presence of 40 pg/ml of preactivated cholera toxin for 45 min a t 30 "C as described under "Experimental Procedures." Membranes were then washed, solubilized, and denatured and subjected to SIX-poly- acrylamide gel electrophoresis and autoradiography as indicated un- der "Experimental Procedures." Solubilized membrane protein from 1.2 X 10' S49 WT cells, 1.2 X 10' S49 cyc- cells, 1.5 X 10'' 3T3-LI preadipocytes, and 5.2 X 10" adipocytes were applied to the indicated lanes. Samples in even numbered lanes were incubated with cholera toxin; samples in odd numbered lanes received [."PINAD but no toxin. B, in a separate experiment, membranes from 1.3 X IO" pread- ipocytes (day 0, lane 1 ) and 1.3 X 10'' adipocytes (day 6, lane 2) were labeled and analyzed as described in A.

may occur. Experiments aimed at characterizing several of the subunits of G/F were based on the observation that activated cholera toxin catalyzes the transfer of the [""PIADP- ribose moiety of ['"PINAD to the subunits of G/F in a highly specific manner in several membrane systems (25-27).

Fig. 8 shows an autoradiogram of a 0.1% SDS and 10% polyacrylamide gel used to resolve ['"PIADP-ribosylated poly- peptides in 3T3-Ll adipocyte and preadipocyte membranes. ['"PIADP-ribosylated membranes isolated from wild type mu- rine S49 lymphoma cells (S49 WT) and S49 cyc- variants (6) were also analyzed on this gel to provide specific, internal markers and controls. A protein with a molecular weight of 42,000 and polypeptide doublet with molecular weight values of 49,000 and 50,000 were specifically labeled by cholera toxin in membranes from undifferentiated and differentiated 3T3- L1 cells (Fig. 8A, Lanes 6 and 8) and S49 WT cells (Fig. 8A, lane 2). None of these polypeptides were labeled in S49 cyc- membranes (Fig. 8A, Lane 4 ) which lack G/F activity (9, 27). All other ADP-ribosylated proteins were labeled equally well in the absence of cholera toxin (Fig. 8A, Lanes I, 3,5, a n d 7 ) . In contrast to the large number of substrates for endogenous ADP-ribosylation in crude S49 membrane preparations, it appears that both types of 3T3-Ll membranes are relatively deficient in ADP-ribose acceptor proteins (except for G/F components) and/or endogenous ADP-ribosyltransferase ac- tivity.

The three proteins specifically labeled via cholera toxin in 3T3-Ll membranes appear to correspond to the ADP-ribo- sylated constituents of the nearly homogeneous rabbit liver

12872 3T3-Ll Adenylate Cyclase n c w

4000 Adlpocyte

4 9 - 5 0 K ~ M e m b r a n e s P 0 0 a

f n n N

it 0

CELL NUMBER x 10-6

FIG. 9. Quantitation of ADP-ribosylation of G/F compo- nents. Membranes prepared from the number of preadipocytes and adipocytes indicated were [32P]ADP-ribosylated and analyzed by SDS-polyacrylamide gel electrophoresis as described under Fig. 8 and "Experimental Procedures." The amounts of 32P incorporated into the 42,000-dalton polypeptide (A- - -A, preadipocyte membrane; w, adipocyte membranes) and the 49,000- plus 50,000-dalton doublet (A- - -A, preadipocyte membranes; o " 0 , adipocyte mem- branes) were determined by cutting out gel slices containing the labeled polypeptides as described under "Experimental Procedures."

G/F protein isolated by Northup et al. (28).3 Examination of the [3ZP]ADP-ribosylation patterns in mem-

branes from 3T3-Ll preadipocytes and adipocytes revealed significant differences. Fig. 8B compares the relative intensi- ties of labeling of the G/F subunits when membranes from equal numbers of preadipocytes and adipocytes were analyzed. The number of ADP-ribosylation sites available in mem- branes from differentiated cells is clearly much higher.

In Fig. 8A, the amount of preadipocyte membranes was increased to facilitate the determination of the relative amounts of "P incorporated into the 42,000-dalton protein and the 49,000- to 50,000-dalton doublet. The relative inten- sities of labeling of the 42,000-dalton polypeptide and the 49,000- to 50,000-dalton doublet were dramatically altered during differentiation (Fig. 8A, lanes 6 and 8). In preadipocyte membranes, incorporation of 32P into the higher molecular weight doublet predominated, whereas in adipocyte mem- branes, the 42,000-dalton polypeptide served as a better ac- ceptor of ADP-ribose than either the 49,000- or 50,000-dalton proteins.

The incorporation of [32P]ADP-ribose into the 42,000-dal- ton protein and the 49,000- to 50,000-dalton doublet (taken as a sum) was a linear function of the amount of membrane protein exposed to cholera toxin and [32P]NAD over a wide range (Fig. 9). Approximately 13 times more "P was incor- porated into the 42,000-dalton component of adipocyte G/F than the corresponding subunit of preadipocyte G/F (Fig. 9). In addition, the amount of 32P in the 49,000- to 50,000-dalton doublet in adipocyte membranes was 4-fold higher than that in membranes from undifferentiated cells. In preadipocyte membranes, the ratio of the labeled doublet to the 42,000- dalton protein was 4.2, while in adipocyte membranes it was 1.3.

The preceding results were independent of the length of the

In the SDS-polyacrylamide gel system used in the present studies, the upper 52,000-dalton band of Northup et al. (28) is resolved into a doublet to which we have assigned molecular weights of 49,000 and 50,000 using pyruvate kinase (57,000), fumarase (48,000), and aldolase (38,000) as markers. Johnson et al. (27) have reported the ADP- ribosylation of a 42,000-dalton polypeptide and 52,000- to 53,000- dalton doublet in S49 WT and rat hepatoma (HTC-4) membranes. By using S49 WT membranes as a standard (e.g. Fig. SA), it is clear that the 49,000- to 50,000-dalton doublet corresponds to the 52,000- dalton 549 polypeptide of Northup et al. (28) and the 52,000-53,000- doublet of Johnson et al. (27); similarly, the 42,000-dalton protein corresponds to a protein of identical moIecular weight in the nomen- clature of Johnson et al. (27) and the 45,000-dalton protein of Northup et al. (28).

u 0

\ 9,

d 2000- Adlpocytt 49-50K

0 2400-

w 2000-

8 1600-

a 0 1200 P

u t

-

; 000- A"" 49-5OU s N n

u

MINUTES FIG. 10. Time course of ADP-ribosylation of G/F compo-

nents in preadipocyte and adipocyte membranes. ADP-ribosy- lation was performed using 10 PM ["'PINAD and 100 pg/ml of cholera toxin as described under "Experimental Procedures" and Fig. 8 for the times indicated at 30 "C. Gel slices containing the labeled proteins were cut out and the amount of 32P incorporated into the protein was determined in a liquid scintillation spectrometer. Slices were obtained from gel lanes that received membrane protein from either 2.0 X IO' preadipocytes or 8.2 X lo5 adipocytes. 42,000-dalton polypeptide (A- - -A, preadipocyte membranes; U, adipocyte membranes). 49,000- and 50,000-dalton doublet (A- - -A, preadipocyte membranes; o " 0 , adipocyte membranes). The addition of 50% more ["*PINAD after 20 min incubation did not alter the amount of "P incorporated into either the 42,000-dalton polypeptide (W, preadipocyte mem- branes; V, adipocyte membranes) or the 49,000- to 50,000-dalton doublet (0, preadipocyte membranes; V, adipocyte membranes) over the succeeding 30 min.

ADP-ribosylation reaction. Fig. 10 shows that the labeling of the 42,000-dalton protein and the 49,000- to 50,000-dalton doublet in both preadipocyte and adipocyte membranes fol- lowed similar time courses and reached saturation.

DISCUSSION

The appearance of catecholamine-sensitive adenylate cy- clase activity during 3T3-Ll adipocyte differentiation (Fig. 1) could reflect the de novo synthesis and accumulation of p- adrenergic receptors, an increase in the level of the G/F protein, or changes in the coupling and regulatory properties of pre-existing components of the cyclase system.

Equilibrium binding studies, using the ,&adrenergic antag- onist 1251-HYP, disclosed the presence of 2000 high affinity P-receptors/preadipocyte (Fig. 4). During differentiation, this number increased only 70%, while receptor affinity for the antagonist remained the same (Kn = 0.2 nM). &Receptors in preadipocyte and adipocyte membranes also had very similar affinities for a series of P-adrenergic agonists and other antag- onists (Fig. 3). Since saturation of the pre-existing ,&receptors on preadipocytes with isoproterenol had little effect on the activity of the catalytic unit (Fig. 1 and Table IV), it is difficult to explain the dramatic development of catecholamine-stim- ulated cyclase activity by the addition of a relatively small number of the same type of receptors. We, therefore, consid- ered the possibility that differentiation-dependent alterations in the interactions between G/F and the /3-receptor might control the expression of P-agonist-stimulated activity.

Two aspects of P-receptor regulation that depend upon the communication between the receptor and G/F are the short term regulation of receptor affinity observable in membranes and the long term depletion of &receptors in cells chronically exposed to &adrenergic agonists (17-19). Adipocytes and pre- adipocytes appeared to be equally effective in executing the down regulation of ,&receptors at similar rates in response to prolonged treatment with isoproterenol (Table I) and mem-

3T3-Ll Adenylate Cyclase 12873

branes derived from these cells displayed similar susceptibil- ities to the negative modulation of agonist affinity by GTP (Fig. 5) and Gpp(NH)p. Thus, preadipocyte P-receptors are clearly capable of productive regulatory interactions (direct or indirect) with G/F and the acquisition of catecholamine-sen- sitive adenylate cyclase activity is not correlated with changes in these interactions.

Further studies of adenylate cyclase activity revealed marked differentiation-dependent changes in responsiveness to guanine nucleotides (Tables I1 and IV). GTP (10 phd stimulated preadipocyte adenylate cyclase equally well in the absence or presence of isoproterenol, while the GTP-mediated enhancement of adipocyte membrane adenylate cyclase activ- ity was highly dependent on the presence of the p-adrenergic agonist (Table IV). In contrast, Gpp(NH)p activated the enzyme in the absence of hormone in both preadipocyte and adipocyte membranes. The net stimulatory effect of Gpp(NH)p increased 20-fold (relative to the net activation observed with GTP alone) during differentiation as the result of an increase in catalytic activity and a sharp decline in the effect of GTP alone (Table 11). There are several possible explanations for these results.

The model of adenylate cyclase proposed by Cassel and Selinger (29) postulates that the catalytic component of the enzyme is activated by a complex of G/F and GTP. A specific, G/F-associated GTPase terminates stimulation of the enzyme by converting GTP to GDP. In this model, hormone-receptor complexes activate adenylate cyclase by facilitating the dis- placement of tightly bound GDP by GTP at the regulatory site of G/F. It is possible that the G/F protein in adipocytes requires the presence of hormone to allow GTP to activate the adenylate cyclase (as modeled by Cassel and Selinger (29)), while the G/F protein in preadipocytes may exist in a conformation where GTP can readily displace bound GDP. Alternatively, the specific GTPase activity of the G/F protein may be diminished (or absent) in the preadipocyte adenylate cyclase system. During differentiation, the rate of cleavage of the y-phosphate bond of GTP may increase sharply at the guanine nucleotide regulatory site, thereby lowering the steady state concentration of G/F-GTP complexes available for catalytic unit activation. Thus, one view of the develop- ment of catecholamine responsiveness in 3T3-Ll adipocytes is that a differentiation-dependent alteration in the properties of G/F is responsible for the loss of activity in the presence of GTP and this activity may be transiently restored by inter- actions between G/F and agonist-occupied ,&receptor. Unfor- tunately, it has not been possible to directly measure G/F- associated GTPase activity in 3T3-Ll cell membranes4 and membranes derived from other mammalian cells (1, 30-32) because of the presence of high levels of membrane-associated GTP-degrading enzymes that are unrelated to the adenylate cyclase system. In addition, there are no reports documenting hormone-dependent facilitation of the displacement of GDP by GTP at a regulatory site in mammalian cell membranes. Nevertheless, when adipocyte membranes were incubated with cholera toxin and NAD under conditions known to result in the ADP-ribosylation of components of G/F (see Figs. 8- 10) and the inactivation of G/F-associated GTPase (33, 34), the stimulation of adipocyte membrane adenylate cyclase by GTP alone rose to a level comparable to that observed in preadipocyte membranes5 Moreover, adenylate cyclase activ- ity in the toxin-treated, adipocyte membranes became refrac- tory to further stimulation by isoproterenol after activation by GTP. These preliminary observations, while not definitive,

E. Lai, and C. S. Rubin, unpublished observations. E. Lai, 0. M. Rosen, and C. S. Rubin, manuscript in preparation.

are consistent with differentiation-dependent changes in the rate of hydrolysis of GTP or the rate of entry of GTP at a regulatory site on G/F. Detailed kinetic studies are in progress to further characterize the system and establish whether or not there is a significant increase in the turn-off rate (35) of the adenylate cyclase reaction during adipocyte development.

The kinetic observations indicating that G/F exhibited different sets of regulatory properties in adipocyte and pre- adipocyte membranes were correlated with marked differ- ences in the patterns and levels of cholera toxin-catalyzed [32P]ADP-ribosylation of G/F subunits (Figs. 8-10). In adi- pocyte membranes, 13-fold more 32P was incorporated into the 42,000-dalton component and 4-fold more labeling was observed in the 49,000- to 50,000-dalton doublet than in the corresponding G/F subunits in preadipocyte membranes (Figs. 8 and 9). While the quantitation of the amount of ADP- ribosylation suggests that the concentration of G/F rises sharply during differentiation, we cannot rule out the possi- bility that qualitative rather than quantitative changes occur. For example, the differences observed in the ADP-ribosylation of G/F may reflect differentiation-directed conformational modifications of the G/F subunits which enhance their capac- ity as substrates in the cholera toxin-catalyzed reaction. Al- ternatively, the greater amount of ["'PIADP-ribose incorpo- rated into G/F subunits observed in adipocytes may actually indicate an increase in the number of molecules of G/F protein per cell.6 A large increase in the amount of G/F, coupled with the small increase in /I-adrenergic receptor number described above, could account for the great increase in the responsive- ness of the adenylate cyclase system to isoproterenol during the differentiation of 3T3-Ll cells.

Regardless of whether the changes in ADP-ribosylation patterns reflect changes in the amount of G/F or its ability to be labeled, the results show that there is a significant change in G/F constituents during cell differentiation. There is little information on the roles of the individual subunits of the G/F regulatory component. Northup et al. (28) purified rabbit liver G/F to near homogeneity and showed that G/F was probably composed of three peptides of molecular weights 35,000, 42,000, and 52,0004 with unknown stoichiometries. The latter two components are substrates for ADP-ribosylation. The 49,000- to 50,000-dalton doublet wgs evident at the leading edge of the G/F peak during purification on heptylamine- Sepharose, while the 42,000-dalton protein was prevalent at the trailing portion of the peak and the 35,000-dalton protein was observed throughout (28). Thus, the differential increases in the labeling of the 42,000-dalton polypeptide and the 49,000-50,000 doublet in adipocyte membranes (Figs. 8-10) may reflect the appearance of distinct subspecies of G/F during differentiation. Hudson and Johnson (36) reported that partial proteolytic digests of [3'P]ADP-ribosylated 42,000-dal- ton protein and the 49,000- to 50,000-dalton doublet yielded identical peptide maps except for a single fragment unique to the larger doublet. These observations suggest that these G/ F components are structurally and functionally related and raise the possibility that the 49,000- to 50,000-dalton doublet serves as a precursor of the 42,000-dalton polypeptide.

Our results suggest that 3T3-Ll adipocytes may acquire catecholamine-sensitive adenylate cyclase activity by altering the properties of G/F to effect the coupling of both pre- existing and newly synthesized p-receptors and catalytic pro- tein. A preponderance of the 49,000-50,000 doublet and a low

A third possibility is that G/F components in preadipocyte mem- branes contain endogenous ADP-ribose moieties that are removed during differentiation. At present, however, there is no evidence to support the possible existence of an endogenous ADP-ribosyltrans- ferase which covalently modifies G/F subunits.

12874 3T3-Ll Adenylate Cyclase

level of 42,000-dalton protein are associated with the G/F of hormone-insensitive, GTP-activated preadipocyte adenylate cyclase. The finding that adipocyte membranes have a very large increase in the labeling of the 42,000-dalton protein relative to the 49,000- to 50,000-dalton doublet, along with reports that avian erythrocyte adenylate cyclase has only the 42,000-dalton G/F subunit and is totally dependent on hor- mone for GTP activation (25, 37, 38) suggest the speculation that the relative concentrations of the 42,000-dalton and 49,000- to 50,000-dalton proteins or some other factor which alters their relative suitabilities as substrates for ADP-ribo- sylation may determine whether the interaction of G/F with GTP requires the presence of agonist occupied P-receptor. Detailed biochemical studies on the nature of the functional and structural changes in G/F subunits during adipocyte differentiation will now be required to explore this working hypothesis.

Acknowledgments-We thank Cathy Fung for tissue culture as- sistance and Donna Putnam for typing and proofreading of the manuscript.

2. 1.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

REFERENCES Rodbell, M. (1980) Nature 284, 17-22 Ross, E. M., and Gilman, A. G. (1980) Annu. Reu. Biochem. 49,

Orly, J., and Schramm, M. (1976) Proc. Natl. Acad. Sei. U. S. A .

Schramm, M. (1979) Proc. Natl. Acad. Sei. U. S. A . 76, 1174-

Neufeld, G., Schramm, M., and Weinberg, N. (1980) J. Biol.

Bourne, H. R., Coffino, P., and Tomkins, G. M. (1975) Science

Insel, P. A,, Maguire, M. E., Gilman, A. G., Bourne, H. R., Coffino, P., and Melmon, K. L. (1976) Mol. Pharmacol. 12,1062-1069

Haga, T., Ross, E. M., Anderson, H. A., and Gilman, A. G. (1977) Proc. Natl. Acad. Sei. U. S. A. 74,2016-2020

Ross, E. M., Howlett, A. C., Ferguson, K. M., and Gilman, A. G. (1978) J. Biol. Chem. 253, 6401-6412

Rubin, C. S., Hirsch, A,, Fung, C., and Rosen, 0. M. (1978) J. Biol. Chem. 253, 7570-7578

Rubin, C. S., Lai, E., and Rosen, 0. M. (1977) J. Biol. Chem. 252, 3554-3557

Salomon, Y., Londos, C., and Rodbell, M. (1974) Anal. Biochem. 58, 541-548

533-564

73,4410-4414

1178

Chem. 255,9268-9274

187, 750-752

13. Bohme, E., and Schultz, G . (1974) Methods Enzymol. 38, 27-38 14. Rubin, C. S., Rangel-Aldao, R., Sarkar, D., Erlichman, J., and

15. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J.

16. Scatchard, G. (1949) Ann. N. Y. Acad. Sci. 51, 660-672 17. Su, Y.-F., Harden, T. K., and Perkins, J. P. (1980) J. Biol. Chem.

18. Maguire, M. E., Van Arsdale, P. M., and Gilman, A. G. (1976)

19. Lefkowitz, R. J., Mullikin, D., and Caron, M. G. (1976) J. Biol.

20. Kimura, N., and Nagata, N. (1977) J. Biol. Chem. 252,3829-3835 21. Yamamura, H., Lad, P. M., and Rodbell, M. (1977) J. Biol. Chem.

22. Johnson, R. A. (1980) J. Biol. Chem. 255,8252-8258 23. Ross, E. M., Maguire, M. E., Sturgill, T. S., Biltonen, R. L., and

24. Iyengar, R., Mintz, P. W . , Swartz, T. L., and Birnbaumer, L.

25. Gill, D. M., and Meren, R. (1978) Proc. Natl. Acad. Sei. U. S. A .

26. Schleifer, L. S., Garrison, J. C., Sternweis, P. C., Northup, J. K., and Gilman, A. G. (1980) J . Biol. Chem. 255,2641-2644

27. Johnson, G. L., Kaslow, H. R., and Bourne, H. R. (1978) J . Biol. Chem. 253,7120-7123

28. Northup, J. K., Sternweis, P. C., Smigel, M. D., Schleifer, L. S., Ross, E. M., and Gilman, A. G. (1980) Proc. Natl. Acad. Sei. U.

29. Cassel, D., and Selinger, Z. (1978) Proc. Natl. Acad. Sci. U. S. A.

30. Cassel, D., and Selinger, Z. (1976) Biochim. Biophys. Acta 452,

31. Salomon, Y., and Rodbell, M. (1975) J. Biol. Chem. 250, 7245-

32. Bitoni, A. J., Moss, J., Tandon, N. N., and Vaughan, M. (1980) J.

33. Cassel, D., and Selinger, 2. (1977) Proc. Natl. Acad. Sei. U. S. A.

34. Birnbaumer, L., Swartz, T. L., Abramowitz, J., Mintz, P. W . , and

35. Cassel, D., Eckstein, F., Lowe, M., and Selinger, Z. (1979) J . Biol.

36. Hudson, T. H., and Johnson, G. L. (1980) J . Bwl. Chem. 255,

37. Cassel, D., and Pfeuffer, T. (1978) Proc. Natl. Acad. Sei. U. S. A.

38. Kaslow, H. R., Farfel, Z., Johnson, G. L., and Bourne, H. R. (1979)

Fleischer, N. (1979) J. Biol. Chem. 254, 3797-3805

(1951) J. Biol. Chem. 193,265-275

255, 7410-7419

Mol. Pharmacol. 12,335-339

Chem. 251,4686-4692

252, 7964-7966

Gilman, A. G. (1977) J. Biol. Chem. 252,5761-5775

(1980) J . Biol. Chem. 255, 11875-11882

75, 3050-3054

S. A . 77,6516-6520

75,4155-4159

538-551

7250

Biol. Chem. 255,2026-2029

74,3307-3311

Iyengar, R. (1980) J. Biol. Chem. 255,3542-3551

Chem. 254,9835-9838

7480-7486

75, 2669-2673

Mol. Pharmacol. 15,472-483